Skip to content

armgit / Pulse

Public Imported from GitHub
0

Real-time monitoring dashboard for Proxmox VE, PBS, Docker, Kubernetes, TrueNAS and vSphere. Self-hosted, with alerts for failed backups, full disks, unhealthy containers and offline nodes.

Merge pending maintenance repairs with current main

rcourtman2fc54a5a1f55
diff --cc .github/workflows/unified-agent-native.yml
index 3b7cf0d281,365e454c9a..a979545d77
--- a/.github/workflows/unified-agent-native.yml
+++ b/.github/workflows/unified-agent-native.yml
@@@ -183,6 -168,69 +168,69 @@@ jobs
set -euo pipefail
timeout --version | grep -q 'GNU coreutils'
+ # Server/LXC fixtures execute Linux-only scripts, not a macOS installer.
+ # Keep their Bash/GNU metadata, archive and rename semantics local to that step;
+ # native macOS agent/runtime tests still use the runner's own utilities.
+ - name: Supply Linux server installer fixture tools on macOS
+ if: runner.os == 'macOS'
+ shell: bash
+ run: |
+ set -euo pipefail
+ brew install bash gnu-tar gawk
+ linux_bash="$(brew --prefix bash)/bin/bash"
+ test -x "$linux_bash"
+ "$linux_bash" -c '(( BASH_VERSINFO[0] >= 5 ))'
+ coreutils_bin="$(brew --prefix coreutils)/bin"
+ fixture_bin="$(mktemp -d "$RUNNER_TEMP/pulse-installer-fixture-tools.XXXXXX")"
+ ln -s "$linux_bash" "$fixture_bin/bash"
- for tool in stat chmod chown mv; do
++ for tool in stat chmod chown mv ln; do
+ executable="$coreutils_bin/g$tool"
+ test -x "$executable"
+ "$executable" --version | grep -q 'GNU coreutils'
+ ln -s "$executable" "$fixture_bin/$tool"
+ done
+ gnu_tar="$(brew --prefix gnu-tar)/bin/gtar"
+ test -x "$gnu_tar"
+ "$gnu_tar" --version | grep -q 'GNU tar'
+ ln -s "$gnu_tar" "$fixture_bin/tar"
+ # The NAS boot filter rejects embedded NUL/control bytes with octal
+ # ranges. Use the same binary-aware awk semantics as its Linux jobs,
+ # not the native macOS parser, only within these server fixtures.
+ gnu_awk="$(brew --prefix gawk)/bin/gawk"
+ test -x "$gnu_awk"
+ "$gnu_awk" --version | grep -q 'GNU Awk'
+ ln -s "$gnu_awk" "$fixture_bin/awk"
+ printf 'PULSE_INSTALLER_FIXTURE_BIN=%s\n' "$fixture_bin" >> "$GITHUB_ENV"
+
+ # Observe the existing generated NAS watchdog/collector control before
+ # Node acquisition and the large installer inventory. The runner's two
+ # phases partition the same discovered checks; no check runs twice or is
+ # dropped, and the production process/cleanup deadlines stay unchanged.
+ - name: Test NAS removal contracts before frontend preparation
+ if: matrix.unix
+ shell: bash
+ run: |
+ set -euo pipefail
+ if [[ -n "${PULSE_INSTALLER_FIXTURE_BIN:-}" ]]; then
+ export PATH="$PULSE_INSTALLER_FIXTURE_BIN:$PATH"
+ fi
+ python3 scripts/installtests/run_unix_contracts.py --phase preflight
+
+ - name: Set up Node.js for native runtime test assets
+ uses: actions/setup-node@949feb2413d6458794dcd2491c4babbbce0c15c1 # v7.1.0
+ with:
+ node-version: '24'
+ cache: 'npm'
+ cache-dependency-path: 'frontend-modern/package-lock.json'
+
+ - name: Install frontend test dependencies
+ working-directory: frontend-modern
+ run: npm ci
+
+ - name: Build real frontend assets for API-linked native tests
+ working-directory: frontend-modern
+ run: npm run build
+
- name: Test native agent runtime
run: >-
go test
diff --cc docs/UPGRADE_v6.md
index ec3222b328,bfef9cf395..482bb3f0df
--- a/docs/UPGRADE_v6.md
+++ b/docs/UPGRADE_v6.md
@@@ -1,11 -1,16 +1,16 @@@
# Upgrade to Pulse v6
- This guide covers practical upgrade steps for existing Pulse installs moving to v6.
+ This guide covers the v5-to-v6 migration and its compatibility changes. For
+ updates within v6, use [Updating Pulse](AUTO_UPDATE.md); the historical packets
+ below are context, not an instruction to install an old version or downgrade.
+ Choose an actually [published software release](https://github.com/rcourtman/Pulse/releases),
+ not a draft, source tag or Helm-chart asset. Keep the existing runtime edition:
+ public Community downloads do not replace a private Pro build.
-For the current v6 beta packet, see:
+For the current v6 release candidate packet, see:
-- `docs/releases/RELEASE_NOTES_v6.6.0-beta.3.md`
-- `docs/releases/V6_CHANGELOG_v6.6.0-beta.3.md`
+- `docs/releases/RELEASE_NOTES_v6.6.0-rc.4.md`
+- `docs/releases/V6_CHANGELOG_v6.6.0-rc.4.md`
For the current stable v6 packet, see:
diff --cc docs/release-control/v6/internal/subsystems/agent-lifecycle.md
index 1a1edcbf43,160e55bce6..56c97ef559
--- a/docs/release-control/v6/internal/subsystems/agent-lifecycle.md
+++ b/docs/release-control/v6/internal/subsystems/agent-lifecycle.md
@@@ -15,38 -15,92 +15,124 @@@
## Purpose
+### Lost container mutation acknowledgements reconcile through reads
+
+Typed container lifecycle execution attempts its mutation once. A transport
+error preserves uncertain execution and, while the original operation context
+remains live, reads the same container and waits for its existing health check.
+Readback never manufactures a mutation acknowledgement or extends a canceled
+operation. Fixed, non-sensitive timeout/cancellation/unknown messages survive
+receipt sanitization. `docker_lifecycle_test.go` covers lost acknowledgement,
+health recovery, cancellation, exactly-once mutation and receipt redaction.
+
+
+### Verified agent actions remain attached to their originating occurrence
+
+An agent action's execution and verification evidence belongs to its original
+action and investigation. Replaying that result after a finding reopens must
+preserve the new occurrence's investigation eligibility and current record.
+The shared findings store validates session identity and the persisted
+regression boundary atomically; current finding reads select current work
+while explicit action history retains earlier receipts. The three-occurrence
+Patrol action regression in
+`internal/api/ai_handlers_patrol_actions_additional_test.go` also reconstructs
+persisted findings and action audit state before the third occurrence.
+
+### Collected backup evidence does not require command-agent access
+
+Guest investigation manifests include the existing read-only storage tool for
+collected backup attempts. The org-scoped recovery reader preserves task error
+text independently of PVE command clients and connected agents. Missing command
+access remains an execution boundary; reading a failure or a successful task
+does not grant a lifecycle action or prove a restorable artifact. Manifest,
+persisted-task and API/tool parity regressions cover this separation.
+
+ ### Native CI tooling refresh (14 October 2026)
+
+ The native agent matrix uses the exact setup-node v7.1.0 and upload-artifact
+ v7.0.2 pins in the [deployment tooling contract](deployment-installability.md#node-setup-and-artifact-action-refresh-14-october-2026).
+ The Node 24 application toolchain, locked frontend cache, real asset build before
+ all-platform runtime tests, generated Windows command execution and retained
+ Linux removal receipt are unchanged. `TestNativeAgentWorkflowBuildsRealFrontendBeforeRuntime`,
+ `TestNativeAgentRejectsBrokenImagesBeforeFrontendPreparation` and
+ `TestNativeWindowsExecutesGeneratedInstallCommand` bind those actual consumers
+ to the refreshed pin without relaxing ordering or platform controls. Local
+ source proof is not native Windows, appliance or installed lifecycle acceptance.
+
+ ### Optional configuration-signing guidance fits narrow installers (13 October 2026)
+
+ The three existing signing environment names wrap within their advice card.
+ Their complete text, values, optional semantics, trust defaults and copied
+ installation/removal commands are unchanged. The source regression in
+ `InfrastructureOperationsModel.test.tsx` preserves those exact names and the
+ wrapping class; desktop/phone light/dark browser proof must inspect the actual
+ installer and preserve separate-token/private-file copy behaviour. This is
+ readability, not an installed agent or QNAP recovery result.
+
+ ### Backend bootstrap preserves silent input on BusyBox hosts
+
+ Setup, Proxmox and container bootstrap commands prefer standalone `stty`,
+ otherwise the BusyBox `stty` applet. They distinguish absent terminals, missing
+ or unsupported tools and failed snapshot/echo/restoration before installation.
+ The privileged Bash child keeps the selected tool and complete saved terminal
+ state through EXIT/HUP/INT/TERM cleanup. Overflow is drained while silent;
+ empty, interrupted or oversized input never reaches the installer. The isolated
+ POSIX root/sudo prefix emits that child once, without moving credentials into
+ shell source, argv, environment or history. Existing complete-download,
+ preflight, private-file, scope, profile, identity and TLS boundaries remain.
+ `portable_bootstrap_test.go` exercises all three producers with real PTYs and
+ BusyBox, full terminal restoration and closed privilege/download fixtures;
+ it is not native appliance installation or recovery evidence.
+
+ ### Unraid/QNAP removal retires exact boot entries and confirms watchdog quiescence
+
+ The Unix uninstall caller selects these platforms from their vendor/installed
+ footprints, not from an ordinary recovered Linux state directory or QNAP state
+ finder alone. Existing exact state-removal authority must select the wrapper's
+ state directory. Regular root-owned files and parents are required; FAT boot
+ volumes cannot be assumed to implement Unix chmod semantics. Unknown wrapper
+ ownership and missing observation tools refuse before mutation.
+
+ One complete bounded boot-file parse admits only the exact generated Unraid
+ launch or canonical QNAP deferred block/legacy direct launch for that wrapper.
+ All duplicates retire; unrelated entries (including co-installed names and
+ legacy Telegraf) remain. Foreign/indirect targets, partial or changed blocks,
+ linked/non-regular files, binary/incomplete content and over-1-MiB files refuse
+ without evaluating or disclosing their text. The original remains in a private
+ `.pulse-removal.*` recovery directory outside the active boot namespace through
+ remote confirmation. A changed source during staging or a concurrently created
+ file, directory or symlink refuses without clobbering it; interrupted staging
+ retains the actual original and never restores automatic activation. New boot
+ content uses no-clobber file creation, compatible with FAT without hard links;
+ this is not an atomic whole-content swap or a lock against external root writers.
+
+ Complete bounded kernel evidence returns only numeric IDs and roles. Only exact
+ selected watchdog/collector processes receive checked TERM, watchdog first,
+ with independent absence observations. Editor sessions, backup wrappers and
+ co-installed names are not targets. Relative/foreign same-name processes or
+ unattributable QNAP deferred autorun or live Unraid go shells refuse instead
+ of signalling another boot owner. A live go shell may already hold later boot
+ commands in memory; editing its file cannot establish quiescence. Failed/diagnostic/duplicate/empty reads and ignored/refused signals
+ retain identity, credentials, trust and recoverable files. QNAP mount and
+ unmount each have checked bounded results, including cleanup after partial mount.
+
+ Whole boot/process and applicable init-manager quiescence closes again before
+ same-instance server removal and before local erasure; late activation refuses,
+ even after remote removal succeeded. NAS paths do not subsequently run the
+ unchecked global pkill or delete arbitrary discovered QNAP state/whole legacy
+ Unraid directories. Successful/absent/repeated removal still requires complete
+ fresh observations. No automatic restart, credential churn or transport-policy
+ change follows a partial operation.
+
+ `TestInstallSHNASBootOwnersQuiesceBeforeRecoveryErasure` requires all eighteen
+ actual-function/full-uninstall caller controls in `agent_nas_teardown.py`. The
+ boot filesystem transformations are real private user operations; flash/kernel
+ faults are closed models. Two generated watchdog/child controls use real private
+ ordinary-user processes, not appliance mounts or root services. Native signed
+ Unraid/QNAP/Linux/PowerShell, TLS/identity/repetition/cleanup and containing
+ publication remain separate obligations; this source proof does not resolve
+ #1099 or claim installed restoration.
+
### Patrol maintenance coverage adds no agent authority
The Patrol operator-state projection in `internal/api/router.go` now reads the
diff --cc docs/release-control/v6/internal/subsystems/api-contracts.md
index 36fbc9aaf5,116d5a271f..5efd96bc41
--- a/docs/release-control/v6/internal/subsystems/api-contracts.md
+++ b/docs/release-control/v6/internal/subsystems/api-contracts.md
@@@ -21,66 -20,19 +21,79 @@@
## Purpose
+### Reporting telemetry observes canonical generation and delivery boundaries
+
+`internal/api/reporting_activity.go` records best-effort content-free events to
+the root config persistence wired through the existing settings-store boundary.
+Organization context contributes only to a local opaque deduplication key. Direct
+resource and VM inventory downloads count only a full successful response write,
+excluding HEAD, cancellation and short writes. Distinct generation days cover
+returning manual use without retaining download identities. Scheduled report generation, file
+saving and existing SMTP transport outcomes each contribute their own counter.
+A successful reporting period is recorded only after final schedule state saves.
+Interrupted email recovery records uncertainty after persisting that outcome.
+Retries still use the same scheduling/delivery system and frozen reporting window.
+
+No endpoint, report payload, delivery state, retry policy, pricing or entitlement
+changes. The activity adapter cannot block a report on tally failure. Scheduler
+restart/retry, distinct-period reuse, complete download and mock-suppression proof
+lives in `internal/api/reporting_activity_test.go`. Notifications retains SMTP
+acceptance authority, and the activity recorder never claims inbox delivery.
+The shared adapter is declared by exact path in the paid-surface allowlist.
+The real boundary regression rejects sibling reporting implementation files and
+private licensing imports added to this observer, preserving the private renderer
+and licensing bridge boundaries.
+
+
+### Container verification distinguishes transitional health
+
+Both agent and independent lifecycle readbacks classify a running container's
+`starting` health as inconclusive, not a final postcondition contradiction.
+Execution and verification remain separate axes: a confirmed readback cannot
+turn a lost mutation acknowledgement into acknowledged execution. The action
+summary retains the agent's sanitized error. The Docker action-result tests
+cover this projection without changing receipt protocol versions or authority.
+
+
+### Finding-scoped investigation reads belong to the current occurrence
+
+The finding investigation and transcript endpoints return only the session
+owned by the finding's current occurrence. After a resolved finding is
+re-detected, a retained earlier session returns `404` through these current
+views until a new session exists. Explicit session and action history retain
+their original evidence. Outcome and record reconciliation check session
+identity and the persisted regression boundary under the findings lock before
+changing current state. `TestPatrolHistoricalActionCannotResolveRecurringFinding`
+in `internal/api/ai_handlers_patrol_actions_additional_test.go` covers three
+occurrences, restart, historical reads and duplicate action callbacks.
+
+### Recovery point reads stay coherent between HTTP and investigations
+
+The points endpoint and investigation adapter use the same org-scoped manager
+reader. Demo composition reads the canonical fixture graph through the recovery
+query helpers; ordinary tenants read their persisted store. Moving the existing
+in-memory filters into recovery preserves point, series and facet query semantics
+and pagination bounds. The API/tool parity regression checks that a collected
+backup failure remains readable without a PVE client or command agent. The
+investigation backup inventory uses the same `KindBackup` and canonical subject
+query as the points endpoint, retaining point identity, repository references,
+tri-state verification and evidence provenance in real and demo modes. Tool
+pagination translates global offsets into the shared source's bounded pages.
+No route, entitlement, recovery mutation or action approval is added.
+
+ ### Copied backend bootstrap commands have one portable private reader
+
+ Existing setup artifact aliases and Proxmox/container command fields retain
+ separate credential response fields. Their command text now selects standalone
+ `stty` or, only when absent, BusyBox `stty`, with distinct fail-closed terminal,
+ tool and mode diagnostics. One isolated root/sudo prefix emits the privileged
+ reader once, preserving complete fetch/preflight and 0700/0600 path handoff.
+ No route, payload key, issuance scope, TLS policy or install authority changes.
+ `TestContract_PortablePrivateBootstrapPayloads` checks the serialized producer
+ boundary; full configapi PTY/tool/privacy/caller tests exercise the commands.
+ Browser acceptance must use the actual producer fields through their existing
+ setup/agent consumers, not a frontend-file guard skip or fabricated commands.
+
### Patrol setup panel reuses existing AI endpoints
The Patrol page setup panel adds no endpoint or payload field. It calls
@@@ -12470,51 -12217,22 +12496,71 @@@ reference route checks cover caller fil
Final-source checks must retain this upstream behavior beside the reviewed
reporting/help repairs; source proof is not installed credential recovery.
+### Successful extended history HTTP reads
+
+`chartapi.Service.WriteHistoryJSON` and `WriteHistoryPayload` share one
+post-write observation boundary for metric history and the main, workload,
+infrastructure and storage chart handlers, including cached bytes. Existing
+response bodies, range/entitlement checks and failure behavior remain the API
+contract. Only GET responses requested beyond seven days that contain a real
+returned numeric metric point strictly older than seven days qualify. Long
+empty windows, metadata timestamps, failed encoding/writes, cancellation,
+HEAD and mock responses cannot increment the tally. Mock and live cache keys
+are separated. The optional observer receives no request or content.
+`history_activity_test.go` and `telemetry_history_reads_test.go` exercise this
+boundary. Schema v19 count semantics and retention belong to security-privacy.
+
+### Saved-node connection tests use draft TLS settings (16 October 2026)
+
+`POST /api/config/nodes/{id}/test` accepts an optional JSON body containing
+`verifySSL` and `fingerprint` for PVE, PBS and PMG. Omission retains the saved
+value, `false` disables CA verification, and an empty fingerprint clears the
+saved pin for that probe. An empty body remains compatible with existing
+callers. The bounded decoder rejects malformed bodies and unknown fields.
+The endpoint retains its existing administrator and settings-write authority.
+Credentials are read server-side, and overrides affect only the temporary
+client. A test must not save configuration or alter runtime connection health.
+This is neutral to storage-recovery because no persistence or recovery path changes.
+
+`config_handlers_connection_test.go` exercises actual TLS probes for all three
+platforms, saved credentials, omitted/cleared settings and unchanged configuration
+and health. `frontend-modern/src/api/__tests__/nodes.test.ts` binds the optional
+body to the saved-node route, including explicit false and empty values.
+
+### Router chart service initializes once for concurrent requests (16 October 2026)
+
+Each router constructs and publishes its chart service once, including routers
+created without the normal constructor. Eager construction uses the same guarded
+accessor as lazy history and chart handlers. Concurrent requests share that
+router's service, while different routers retain separate services and resolve
+the tenant monitor from each request context. Payload and route authority are
+unchanged. `TestChartServiceInitializationIsConcurrentAndRouterScoped` in
+`internal/api/contract_test.go` and `TestMultiTenant_TenantIsolation` are the
+direct race-detector proofs. This is a separate CI stabilization slice.
+
+The additive reporting-catalog `backupHealthAvailable` flag describes runtime
+renderer presence independently of entitlement. New backup schedules are offered
+only when it is present, while existing schedules remain editable after a runtime
+change. Legacy catalogs default to false. API and catalog-model tests cover
+absence and registration so staggered source upgrades do not offer unavailable
+rendering. No price or entitlement changes.
++
+ ### API-rejection guidance in diagnostics (14 October 2026)
+
+ `GET /api/diagnostics` distinguishes `proxmox_permission_denied` (403) from
+ `proxmox_auth_rejected` (401), with `proxmox_api_error` for other observed HTTP
+ failures. Typed PVE/PBS statuses outrank response text. Retained scheduler text
+ uses the first explicit producer status marker, not numbers in endpoint or
+ token identities. Authentication guidance requires confirmed invalidity before
+ replacement; permission guidance preserves both scoped ACLs and privilege
+ separation. The existing `errors` array now carries these bounded diagnoses and
+ recovery guidance into Errors Detected, using positional connection references
+ rather than adding private identities. It does not claim to list every fault.
+ PVE live-monitor reconciliation and separate PBS monitor/probe evidence remain.
+ Routes, caller scope/token filtering, cache policy and probe count are unchanged.
+ `TestDiagnosticsAPIStatusOutranksResponseText`,
+ `TestDiagnosticsSerializedAPIStatusMarkers`,
+ `TestComputeDiagnosticsAPIRejectionsPreserveCredentials` and
+ `TestContract_DiagnosticsAPIRejectionGuidanceSurvivesSerialization` cover real
+ HTTP statuses, private-body controls, same-credential fixture recovery and wire
+ attribution. They are not installed permission or guest-recovery acceptance.
diff --cc docs/release-control/v6/internal/subsystems/cloud-paid.md
index e084399e37,25675415de..b09e11eb29
--- a/docs/release-control/v6/internal/subsystems/cloud-paid.md
+++ b/docs/release-control/v6/internal/subsystems/cloud-paid.md
@@@ -15,27 -15,17 +15,38 @@@
## Purpose
+### Reporting measurement precedes commercial feature decisions
+
+Schema-v20 reporting usage exports only aggregate generation by closed report
+kind, distinct successful-generation days for manual and scheduled use, file-save
+and SMTP outcomes, backup source-data states, distinct successful
+saved-schedule periods and repeat periods, plus measurement availability and
+elapsed observation days. The receiver admits these fields only for schema v20
+or later with an available tally and positive observation days. It bounds counts
+and repeat/success relationships without changing billing, pricing, grants,
+customer entitlements or existing opt-out behavior.
+
+The public adoption report uses each recent active published non-mock rotating
+ID's latest rolling counts. Old senders, missing receiver columns and unavailable
+tallies are unmeasured, not zero usage. Partial elapsed observation is explicit.
+Neither rotating installation IDs nor the coarse nonfree flag establish paying
+customer identity, demand or willingness to pay. No join to license keys,
+accounts, purchases or support is introduced. `pulse-pro:license-server/telemetry_test.go`
+verifies receiver round-trip, schema/availability gates and bounds. Public
+snapshot, privacy and adoption-query tests verify the shared semantics. Receiver
+deployment and a client release are necessary before collecting fleet evidence.
+
+ ### Retired mobile settings keep information separate from entitlement
+
+ Pulse Mobile retirement guidance is a full-text labelled note, not a paid
+ feature offer with a clamped sales description. Instances without the existing
+ Relay capability can read the date and phone-alert alternatives but cannot
+ pair devices or configure connections. Entitled phones keep their existing
+ configuration and pairing controls until 31 March 2027. The change introduces
+ no commercial offer, billing read, entitlement substitution or service change.
+
+ `RelaySettingsPanel.runtime.test.tsx` and the current first-session/paid-settings
+ browser consumers verify the readable notice and unchanged access boundaries.
### Opt-in confirmed-operation transport keeps paid contexts unchanged
@@@ -3715,29 -3705,12 +3726,39 @@@ Mobile settings gate and Alerts push pa
provided remote access to the web UI. Pulse Mobile retires on 31 March 2027;
until then licensed instances keep pairing and push unchanged.
+ The backend feature catalog follows the same boundary: `relay`, `mobile_app`
+ and `push_notifications` have no generic upgrade reason. Both full entitlements
+ and non-billing commercial posture omit those retired offers. Missing extended
+ history advertises only Pro's 90 days, never Relay or remote web access. Catalog
+ roles, capability grants, subscriber terms and the legacy `relay` runtime's
+ 14-day compatibility limit remain unchanged. The retirement HTTP regression
+ checks both commercial projections and runtime capabilities for Community,
+ legacy Relay and Pro; continuing Relay subscribers' Pro grants remain owned by
+ the license server, not inferred from a legacy tier payload.
++
+### Extended history measurement does not establish customer identity
+
+The license server accepts schema-v19 `successful_history_reads_over_7d_30d`
+as a nonnegative aggregate integer, migrates storage without rewriting older
+rows, and resets pre-v19 supplied values to zero. The public reporter treats
+older schemas or missing receiver columns as unmeasured. Current coarse
+`paid_license` posture is not a paying-subscription census and must not be
+joined to billing accounts. This adds no price, entitlement, refund or billing
+behavior. Receiver intake and migration tests pin the additive schema boundary.
+Deploying the compatible receiver and release adoption remain separate from
+source verification. Count and disclosure semantics belong to security-privacy.
+
+### Business backup health report renderer
+
+New backup health PDF/CSV rendering lives in pulse-enterprise's reportingadmin
+package behind `GenerateBackupHealthReportFunc`. BusinessHooks installs that
+private callback without forking the canonical reporting handlers, scheduler or
+delivery system. Public code owns task collection, retained evidence, tenant
+scope, effective branding and the shared rendering contract. The public adapter
+fails explicitly when the private renderer is absent.
+
+Existing advanced-reporting and branding entitlements, pricing, grants and
+metric/recovery retention remain unchanged. This source change does not deploy
+a runtime or change customer subscriptions. Private renderer tests cover formats,
+missing/observed-zero distinction, formula-safe CSV and PDF generation. Runtime
+hook tests and final-source Pro browser delivery qualify registration/composition.
diff --cc docs/release-control/v6/internal/subsystems/deployment-installability.md
index a65840abb7,49c91022a8..2ae2676443
--- a/docs/release-control/v6/internal/subsystems/deployment-installability.md
+++ b/docs/release-control/v6/internal/subsystems/deployment-installability.md
@@@ -15,23 -15,78 +15,94 @@@
```
## Purpose
+### Telemetry schema release disclosures follow receiver compatibility
+
+`docs/releases/TELEMETRY_SCHEMA_V19.md` carries the extended-history telemetry notice for
+inclusion in the first release containing schema v19. Source landing neither
+deploys the compatible license-server receiver nor releases upgraded clients.
+The reporter exposes missing receiver columns and pre-v19 clients as unknown
+coverage, never as evidence of zero use. Existing telemetry opt-out, ID rotation
+and retention are unchanged. Payload details belong to security-privacy.
+`docs/releases/TELEMETRY_SCHEMA_V20.md` likewise carries the reporting outcome
+and repeat-use notice for the first schema-v20 release. Schema and tally
+availability gate measurement, including unknown old receivers and clients.
+The telemetry privacy tests pin both notices. These disclosure files use the
+telemetry release-disclosure proof route, rather than changing release promotion,
+artifact dispatch, opt-out, pricing or customer entitlements.
+
+
+ ### Issue metadata checks before publication (14 October 2026)
+
+ Build and Test runs the locked issue parser and silent-triage policy tests on
+ proposed source, not only after publication when an issue event obtains its
+ write credential. Parser/test/manifest changes, issue templates, triage guidance
+ and every workflow read by the policy tests admit the check. Missing bases admit
+ it too; complete receipt-only ranges and unrelated inputs do not. Guidance-only
+ changes still run it even when application runtime checks are inapplicable.
+ Mixed source or dependency ranges retain every existing runtime/audit check.
+
+ The bounded hosted job uses Node 24, read-only repository permission, checkout
+ without persisted credentials, no dependency cache and a locked install without
+ lifecycle scripts. Both install and test failures propagate; no App token,
+ issue mutation, tolerant exit or production operation belongs to this job.
+ The separate issue-event pre-token test remains mandatory and unchanged.
+
+ `TestCIIssueMetadataUsesSourceOnlyTriageChecks` executes the actual workflow's
+ classification and recipes through `test_ci_browser_proof_only.py`, including
+ real Git whole-range/lookalike controls and install/test failure exits. Actual
+ locked parser tests remain independently required. These source controls do not
+ establish a hosted pass, live issue classification or reporter resolution.
+
+ ### Unraid/QNAP removal retires exact boot entries and confirms watchdog quiescence
+
+ The Unix uninstall caller selects these platforms from their vendor/installed
+ footprints, not from an ordinary recovered Linux state directory or QNAP state
+ finder alone. Existing exact state-removal authority must select the wrapper's
+ state directory. Regular root-owned files and parents are required; FAT boot
+ volumes cannot be assumed to implement Unix chmod semantics. Unknown wrapper
+ ownership and missing observation tools refuse before mutation.
+
+ One complete bounded boot-file parse admits only the exact generated Unraid
+ launch or canonical QNAP deferred block/legacy direct launch for that wrapper.
+ All duplicates retire; unrelated entries (including co-installed names and
+ legacy Telegraf) remain. Foreign/indirect targets, partial or changed blocks,
+ linked/non-regular files, binary/incomplete content and over-1-MiB files refuse
+ without evaluating or disclosing their text. The original remains in a private
+ `.pulse-removal.*` recovery directory outside the active boot namespace through
+ remote confirmation. A changed source during staging or a concurrently created
+ file, directory or symlink refuses without clobbering it; interrupted staging
+ retains the actual original and never restores automatic activation. New boot
+ content uses no-clobber file creation, compatible with FAT without hard links;
+ this is not an atomic whole-content swap or a lock against external root writers.
+
+ Complete bounded kernel evidence returns only numeric IDs and roles. Only exact
+ selected watchdog/collector processes receive checked TERM, watchdog first,
+ with independent absence observations. Editor sessions, backup wrappers and
+ co-installed names are not targets. Relative/foreign same-name processes or
+ unattributable QNAP deferred autorun or live Unraid go shells refuse instead
+ of signalling another boot owner. A live go shell may already hold later boot
+ commands in memory; editing its file cannot establish quiescence. Failed/diagnostic/duplicate/empty reads and ignored/refused signals
+ retain identity, credentials, trust and recoverable files. QNAP mount and
+ unmount each have checked bounded results, including cleanup after partial mount.
+
+ Whole boot/process and applicable init-manager quiescence closes again before
+ same-instance server removal and before local erasure; late activation refuses,
+ even after remote removal succeeded. NAS paths do not subsequently run the
+ unchecked global pkill or delete arbitrary discovered QNAP state/whole legacy
+ Unraid directories. Successful/absent/repeated removal still requires complete
+ fresh observations. No automatic restart, credential churn or transport-policy
+ change follows a partial operation.
+
+ `TestInstallSHNASBootOwnersQuiesceBeforeRecoveryErasure` requires all eighteen
+ actual-function/full-uninstall caller controls in `agent_nas_teardown.py`. The
+ boot filesystem transformations are real private user operations; flash/kernel
+ faults are closed models. Two generated watchdog/child controls use real private
+ ordinary-user processes, not appliance mounts or root services. Native signed
+ Unraid/QNAP/Linux/PowerShell, TLS/identity/repetition/cleanup and containing
+ publication remain separate obligations; this source proof does not resolve
+ #1099 or claim installed restoration.
+
### TrueNAS SCALE removal retires every owned middleware boot task before recovery erasure
The [agent lifecycle removal contract](agent-lifecycle.md#truenas-scale-removal-retires-every-owned-middleware-boot-task-before-recovery-erasure)
diff --cc docs/release-control/v6/internal/subsystems/frontend-primitives.md
index 7caba44d04,624db03e4d..710512d884
--- a/docs/release-control/v6/internal/subsystems/frontend-primitives.md
+++ b/docs/release-control/v6/internal/subsystems/frontend-primitives.md
@@@ -242,10 -332,9 +332,15 @@@ the chart's native hover title and assi
An independent qualified memory reading is not paused by a disk deferral.
Neither Last seen, backup completion nor running power state renews memory.
Unannotated unrelated platforms and actual zero readings retain their meaning.
++Selected agent memory freshness is independent of VM/LXC classification. A stale
++agent retains its original observation as last known, while an independent
++current Proxmox observation remains current. Guest kind still determines the
++provider's cache interpretation.
++
A current reading from a cache-inclusive Proxmox source (`status-mem`,
-`status-freemem`, `cluster-resources`, `derived-total-minus-used`, the same set
-the backend marks `MayIncludeCache`) keeps its numbers, colour and thresholds;
+`status-freemem`, `cluster-resources`, `derived-total-minus-used`, the same
+sources and guest kinds the backend marks `MayIncludeCache`) keeps its numbers,
+colour and thresholds;
the shared presentation adds `mayIncludeCache` and a `cacheNote` saying a high
percentage alone does not show memory pressure and which fix can supply the
guest's own figure. A VM row shows `May include cache` (`Cache?` in compact
@@@ -9771,37 -9722,60 +9880,95 @@@ existing table evidence. No route, gues
permission changes. This presentation result does not establish thaw, covered
filesystem writes, native restart recovery or containing released software.
+### Client backup report scheduling and data-quality states
+
+Reports adds Client backup health to the existing schedule editor. PDF/CSV,
+weekly/monthly cadence, current workload/tag scope and existing email/disk
+delivery controls reuse the schedule workflow. Empty selection and tags mean
+all current VMs and system containers. The picker restricts this kind to those
+workload types and has no performance-report selection cap. Its existing default
+cap remains unchanged for other consumers. Switching kinds retains only valid
+workload selections and preserves the existing digest/inventory constraints.
+
+The schedule card presents last report data quality separately from delivery
+acceptance. Missing/incomplete history never reads as zero failures or completed
+delivery failure. The browser payload cannot supply server-owned quality fields.
+Copy states the PVE task scope, PBS destinations and unsupported-provider state.
+Pricing and current advanced-reporting/branding entitlement behavior are intact.
+
+Proofs: `ReportingPanel.test.tsx`, `ResourcePicker.test.tsx` and
+`reportingSchedulesModel.backupHealth.test.ts` under Settings tests cover report
+kinds, quality/delivery coexistence, unrestricted workload selection, payloads
+and kind switching. Final-source browser proof covers the actual editor, client
+scope, reload, delivery/recovery states and narrow/desktop layouts.
+
+The shared picker uses semantic theme colours for workload names, selection,
+tags and chips. Large selections remain in a bounded scrollable summary so
+client filters and delivery controls stay reachable. Toggle rows expose pressed
+state, and keyboard removal scrolls selected chips into view. The browser matrix
+covers light and dark themes at all three widths.
+
+The additive reporting-catalog `backupHealthAvailable` flag describes runtime
+renderer presence independently of entitlement. New backup schedules are offered
+only when it is present, while existing schedules remain editable after a runtime
+change. Legacy catalogs default to false. API and catalog-model tests cover
+absence and registration so staggered source upgrades do not offer unavailable
+rendering. No price or entitlement changes.
++
+ ### One explicit guest backup completion time (12 October 2026)
+
+ The Workloads backup badge and name-cell indicator disclose the same UTC
+ completion date/time as the guest drawer, including on the fine-pointer hover
+ shortcut. `formatBackupCompletionTime` normalises the original instant once per
+ disclosed readout; the semantic `time` preserves its ISO timestamp. Local browser
+ timezone and locale cannot silently change the visible calendar day or time.
+ The completion text wraps at narrow widths. Unknown/future/absent readings still
+ have no completion element; running or ended activity and clock ticks do not
+ renew it. Only replacement evidence changes the supplied completed time.
+
+ `BackupCompletionTime.test.tsx`, `BackupStatusAccess.test.tsx` and the formatter
+ controls connect both disclosure paths to the drawer, offset inputs, updates,
+ uncertainty and disposal. The `backup-completion-time` browser fixture uses real
+ VM/LXC rows, policy producer, hover, keyboard/touch dialogs and the guest drawer
+ in non-UTC desktop and phone contexts. This is unambiguous presentation, not
+ PBS verification ingestion, independent thaw, filesystem liveness, native
+ guest restart recovery or containing software availability.
+
+ ### Host drawers distinguish unknown filesystem usage from empty filesystems
+
+ The existing Proxmox node and Docker host per-filesystem rows use the shared
+ host disk presentation. An explicit negative usage sentinel, negative or
+ non-finite used bytes, or absent/non-positive/non-finite capacity produces
+ **Usage unavailable**, with any finite positive capacity retained separately.
+ Unavailable usage has neutral text and no utilization bar; a retained capacity
+ alone is not a healthy zero. Finite byte ratios preserve their existing
+ threshold presentation. Real zero-byte values, including zeros omitted by the
+ canonical DiskInfo JSON representation, remain valid empty filesystems when
+ capacity is known and no unknown sentinel is present.
+
+ The shared usage text wraps in both existing drawers, rather than relying on
+ a tooltip or rendering an old percentage after a new unavailable read. The
+ existing mount filtering, row disclosure and single-filesystem summary stay
+ unchanged. The connected `HostDrawer.filesystem-readings.test.tsx` and shared
+ `DrawerDiskListCard.readings.test.tsx` controls cover unknown/capacity/zero,
+ invalid bytes and same-mounted-drawer replacement. The desktop Chromium and
+ narrow touch WebKit `browser-tests/host-disk-usage.cjs` fixture consumes synthetic
+ Resources responses through the actual platform tables and drawers. No new
+ route, monitoring policy, provider operation or native recovery claim is added.
+
+ The shared `filesystemByteUsage.ts` admission also applies to filesystem inputs
+ of `stackedDiskBarModel.ts`: unusable byte rows remain neutral in tooltips and
+ do not contribute used bytes, capacity, bars, combined percentages or highest
+ usage. All-unmeasured filesystem lists with no independent accepted aggregate
+ render N/A, not a fabricated zero. Independently supplied aggregate metrics and
+ retained/unavailable source classifications preserve their existing semantics.
+ The connected stacked-model controls retain explicit aggregate fallback tests;
+ the browser fixture checks mixed and all-unknown lists in the actual summaries
+ as well as both drawers, including live replacement at narrow widths.
+
+ The aggregate bar admits byte sublabels only with finite non-negative measured
+ used bytes and finite positive capacity. An independently valid percentage
+ remains visible beside invalid capacity without formatting NaN/Infinity bytes.
+ Model/component controls cover non-finite, absent and non-positive capacity;
+ `workload-disk-metrics.cjs` checks both Bars (including the tooltip) and Sparklines
+ through the same REST reading at desktop and phone widths.
diff --cc docs/release-control/v6/internal/subsystems/monitoring.md
index 58df29d96e,82d8df5a0e..dea75ac8a1
--- a/docs/release-control/v6/internal/subsystems/monitoring.md
+++ b/docs/release-control/v6/internal/subsystems/monitoring.md
@@@ -17,16 -17,322 +17,332 @@@
## Purpose
+### Storage investigation history shares the chart source boundary
+
+`ReadStorageMetricHistory` and `GetStorageMetricsForChart` use the same monitor
+selector for memory, retained and generated storage history. Analytical reads
+retain the requested window and propagate retained-store errors. Chart-only
+lookback remains presentation behavior and cannot supply older observations as
+current investigation history. Generated demo samples carry `mock_synthetic`;
+they are never evidence of a real storage growth rate. The requested-window and
+store-error regression is `internal/monitoring/storage_history_test.go`.
+
+ ### Conflicting saved endpoint pins cannot select certificate trust
+
+ For equivalent Proxmox endpoint URL spellings, all nonempty saved member pins
+ must agree under the TLS verifier's existing case/colon normalization. Exact
+ spelling and map iteration can no longer select one of contradictory pins.
+ Resolution returns a fixed, credential-free trust-conflict error before a TLS
+ probe, password login, token read, first-use capture or polling-client creation.
+ The member stays unreachable to Pulse; an independently trusted sibling can
+ continue ordinary monitoring. No saved pin is removed or overwritten, and the
+ error cannot become an empty pin/CA or opt-out fallback. Agreeing pin formats,
+ nonempty invalid pins, primary-pin fallback, declared-member first-use and the
+ existing scheme/port/base-path/IPv4-mapped trust boundaries remain unchanged.
+
+ `TestClusterConflictingSavedPinsNeverAuthenticate` covers startup, recovery and
+ lazy selection in token/JSON/form modes, three URL spellings and both VerifySSL
+ settings with real TLS sockets. It checks zero TLS/login/read work for conflicts,
+ no installed client or trust mutation, sibling use, and accepted monitoring
+ following a normal reload with reconciled fixture pins. Existing established-pin,
+ first-use, cancellation, ownership and guest-admission controls still apply.
+ `TestClusterSavedPinResolutionRejectsConflicts` and
+ `TestClusterFirstUseCaptureDoesNotReconcileConflictingPins` retain the conflict
+ through normalized authorities, invalid nonempty pins and a capture race.
+ `TestClusterConflictingPinsEndpointProjection` uses the production endpoint
+ builder, real clients and monitor projection in all three auth modes: native
+ Online, identity and saved trust are not changed by failed Pulse reachability.
+ The producer also records contradictory exact duplicate saved rows before
+ its map/deduplication can overwrite them, and carries those endpoint conflicts
+ through initialisation, missing-client reconnect and cluster refresh. The
+ cluster owns a copy of that conflict set; aliases cannot escape it, and a
+ later repeated row cannot clear it. Normal configuration reload with agreeing
+ pins is required to resume that endpoint. A non-conflicting sibling is not
+ refused. The public constructor's optional conflict inputs preserve existing
+ callers and do not introduce a new trust mode. Producer tests cover both
+ routes, pin formatting, repeated contradictions and configuration ownership.
+ The projection cases include exact duplicates as well as URL aliases.
+ A credential-free agreement check also holds automatic topology refresh while
+ saved trust is contradictory: discovery cannot collapse the rows, move an
+ alias or substitute a pin to clear that refusal. Current trusted sibling
+ clients keep monitoring; topology changes wait for independently reconciled
+ configuration. `TestClusterPinConflictsSurviveMonitorLifecycle` exercises the
+ actual initialisation, missing-client retry and refresh/rebuild paths in all
+ three auth modes and both layouts, including normal refresh after reconciliation.
+ Saved-view consolidation also refuses a view with internal pin contradictions,
+ so map indexing cannot hide an earlier contradictory row before monitoring
+ starts. `TestConsolidatePVEInstancesRetainsInternallyConflictingPins` covers
+ both merge orders, repeated rows, same-node/address evidence and agreeing formats.
+ Internal agreement uses the TLS verifier's case/colon comparison even when
+ legacy identity formatting would erase or simplify a malformed nonempty pin;
+ colon-only, whitespace and prefixed pins cannot hide a contradiction.
+ The endpoint-ownership routing control uses a five-second fixture client and
+ a ten-second enclosing caller lifetime, rather than a one-second latency
+ criterion. It records elapsed time and both targets' request counts; unchanged
+ pin, admission and routing assertions remain required. Runtime timeouts and
+ separate timeout/cancellation controls are unchanged. This is not a diagnosis
+ or dismissal of the retained earlier one-second timeout.
+ Lock-evidence fixture cases use independent base-path namespaces while each
+ case's clients share the same guard key. A recycled loopback port cannot inherit
+ an earlier synthetic guest's cooldown. `TestGuestAgentLockEvidenceFixtureIsolation`
+ shares one listener and retains an earlier uncertainty fence unchanged while an
+ independent fixture completes its lock/command checks. No runtime fence is
+ cleared, expired or weakened; native guest recovery is not established.
+ This establishes source/consumer behaviour, not native recovery or attribution
+ of a guest, PBS or certificate report.
+
+
+ ### Cluster recovery requires success of the operation actually probed
+
+ Startup and unhealthy-member recovery read `/nodes`; lazy construction uses the
+ same probe. A failed nodes or login request cannot count as successful recovery
+ because its response text mentions a guest agent, QMP or `guest-get-*`. The old
+ per-guest substring exception is removed from these non-guest operations.
+ Failed selection retains its actual sanitised reason and observation time, and
+ installs no polling client. Successful probe plus accepted full-timeout session
+ is still required. The existing temporary-rate-limit path may admit an ordinary
+ request, but does not report the failed probe as passed or clear a known failure.
+
+ The single-endpoint last-resort path remains available with its cached client;
+ selecting it no longer marks that endpoint healthy before a request succeeds.
+ A later accepted ordinary read still clears the failure. Actual per-guest
+ errors remain request-level errors, without marking an otherwise working node
+ unhealthy. Certificate pins, first-use trust, credential encodings, request and
+ recovery cadence, rate-limit backoff and guest/backup fences are unchanged.
+
+ `TestClusterNodeProbeRequiresSuccess` covers startup/recovery/selection in token,
+ JSON password and form password modes, three guest-like failure bodies at the
+ nodes or login stage, no extra login/full client and distinct later recovery.
+ `TestClusterSingleEndpointFallbackKeepsFailedHealth` checks cached selection,
+ failed ordinary read and later success. `TestClusterNodeProbeRateLimitKeepsOrdinaryRead`
+ preserves genuine rate-limit handling. `TestClusterNodeProbeEndpointProjection`
+ carries the actual failed and later recovered member through the production
+ monitor projection while independent native Online, saved trust and identity
+ remain unchanged. These are source/consumer controls, not attribution or
+ installed acceptance for a reported PBS/QGA/backup incident.
+
+
+ ### Node inventory admission consumes the complete response
+
+ The PVE `/nodes` reader consumes and closes the entire existing 8 MiB-bounded
+ body before returning inventory or allowing startup/recovery/lazy probes to
+ succeed. A valid JSON prefix with a missing transport tail, extra JSON values,
+ oversized suffix, malformed UTF-8, missing/null data, conflicting errors or
+ ambiguous/case-overwritten fields is not an authoritative inventory. Each row
+ must have a canonical nonempty, unique node name; null/unnamed/duplicate rows
+ and ambiguous supported node fields reject the whole response rather than
+ publish partial or invented empty inventory. Validation errors contain no body
+ material. Unique provider metadata and explicit empty arrays remain accepted,
+ including scoped tokens with no visible nodes. Existing node fields and offline
+ status are preserved without changing provider permissions.
+
+ `node_inventory_response_test.go` exercises actual HTTP reads and all supported
+ fields, the explicit-empty controls, byte/transport bounds and no replay.
+ `cluster_node_inventory_test.go` checks pinned HTTPS in token, JSON-password and
+ form-password modes at startup, recovery and lazy selection: failed success
+ installs no client and never advances to a full session; distinct complete
+ recovery remains usable. The cached fallback keeps its earlier failure until a
+ complete ordinary read succeeds. `TestClusterNodeInventoryEndpointProjection`
+ drives missing/duplicate/incomplete responses and subsequent ordinary recovery
+ through the production monitor projection while native Online, identity and
+ saved pins remain intact. Recovery cadence, auth encodings, TLS admission,
+ rate-limit policy, other API readers and guest/backup fences are unchanged.
+ These are source and consuming-surface proofs, not attribution or installed
+ acceptance for any reported PBS, guest restart or workload-safety incident.
+
+
+ ### First-use cluster trust includes password construction
+
+ Declared, previously untrusted Proxmox members use the existing credential-free
+ TLS capture and pinned retry whether the inherited primary pin fails during
+ password login or during the token-authenticated nodes probe. Startup, ordinary
+ unhealthy recovery and lazy first selection share this path. Capture still
+ requires typed local pin-mismatch evidence and cannot replace a primary, saved
+ member or concurrent winner pin; it never sends an HTTP request or credentials.
+ The pinned retry is not repeated to obtain a nodes count. Failed or cancelled
+ authentication installs no polling session; a separately accepted full-timeout
+ client is required before recovery becomes healthy. Refusal retains the
+ established runtime pin for a distinct later ordinary recovery. Saved trust,
+ CA/redirect policy, login encodings, cadence and guest/backup fences do not change.
+
+ `TestClusterFirstUseSessionAdmission` covers JSON password, existing form
+ fallback and token modes in all three phases, then ordinary member reads.
+ The adjacent refusal, established-pin and cancellation controls exercise actual
+ TLS sockets. Concurrent lazy selection retains serialized construction while
+ health/trust inspection stays usable during capture. `TestClusterFirstUsePasswordEndpointProjection` carries first-use
+ probe/full-session refusal and later public-read recovery into the monitor's
+ cluster-member projection without changing native Online state, identity or
+ saved fingerprints. This is source/consumer qualification, not installed
+ cluster recovery or attribution of an existing user report.
+
+
+ ### Proxmox password success admits a complete session, not just HTTP 200
+
+ PVE and PBS consume the entire existing byte-bounded login body before
+ publishing ticket, CSRF token and the normal two-hour expiry together. The
+ shared `internal/securityutil/proxmox_session.go` parser requires canonical,
+ unambiguous envelope/session keys, complete UTF-8 JSON, a nonempty cookie-safe
+ ticket and a nonempty header-safe CSRF token. Unknown unique metadata such as
+ username and capabilities remains supported. Partial/null, duplicate or
+ case-conflicting credentials, extra JSON values, incomplete reads and oversized
+ tails fail without replacing the retained tuple or advancing PVE generation.
+ Validation errors contain no response credentials. A malformed success is not
+ a provider credential refusal and does not trigger JSON-to-form fallback.
+
+ `session_response_admission_test.go` in both client packages exercises these
+ controls and eight joined expired readers, zero subsequent API/form requests
+ on invalid success, retained state and distinct later ordinary recovery with
+ coherent cookie/CSRF pairs. Cluster recovery additionally checks empty and
+ partial success for the retained polling client, with and without a previous
+ client. `TestPBSInvalidSessionKeepsFirstPollBoundary` exercises the real pinned
+ constructor, normal startup retry, poll executor and production broadcast:
+ invalid success installs no client, valid construction is not healthy telemetry,
+ and only the accepted first poll publishes recovered node/datastore readings.
+
+ TLS trust, credentials, auth encodings and 400/415 fallback, API-token paths,
+ renewal ownership, retries/cadence, guest/backup fences and History are unchanged.
+ The parser establishes response integrity, not cryptographic ticket validation
+ or native/installed acceptance. It neither attributes nor resolves a reported
+ PBS, Windows guest, restart or workload-safety incident.
+
+
+ The invalid-QGA-success matrix uses a distinct synthetic base path for each
+ case/reader, so OS reuse of a closed ephemeral listener cannot borrow a previous
+ fixture's live cooldown. `TestGuestAgentInvalidSuccessFixtureIsolation` shares
+ one listener deliberately and checks two independent first admissions and both
+ retained second-client fences. No production key, alias, TTL or guard state is
+ changed or cleared to qualify these tests.
+
+ ### Proxmox recovery sessions honour the owning caller
+
+ PVE/PBS context-aware constructors bind initial password login, including the
+ existing JSON/form fallback, to the initiating context. Cluster selection,
+ first-use pinned retry, synchronous construction and unhealthy-member recovery
+ carry that context through both the short probe and the retained polling
+ session. The six-second background recovery budget also bounds construction;
+ ending a caller is neither a credential refusal nor successful recovery.
+ Cancelled selection/recovery preserves prior errors and does not install a
+ partial or nil client. Startup retry exits without publishing a new health
+ outcome or advancing another provider after its monitor stops. Discarded probe
+ transports close their idle connections.
+
+ The compatibility constructors retain their background-context behaviour.
+ Accepted clients retain their configured full timeout, and subsequent ordinary
+ reads use their own contexts, not the ended construction context. TLS pins,
+ CA verification, redirects, first-use admission, auth encodings, retry cadence,
+ provider permissions, guest/backup fences and History are unchanged.
+
+ `TestClientConstructionHonoursContext` in both client packages covers ended
+ password/token construction, in-flight JSON/form cancellation/deadlines,
+ independent later reads and wrong-pin refusal. Cluster caller-budget controls
+ cover selection/recovery at first login, probe and full login, then trusted
+ ordinary recovery. Startup cancellation covers PVE, PBS and multi-endpoint
+ cluster construction. The connected PBS first-poll control emits the actual
+ broadcast for desktop/phone consumption: cancellation cannot turn an offline
+ server healthy, and accepted construction alone still cannot do so. These are
+ secret-free local fixtures, not installed recovery, native guest safety or an
+ attribution of any reporter's failure.
+
+
+ ### Cluster recovery requires a usable polling session
+
+ A short endpoint probe cannot establish successful recovery until the client
+ retained for ordinary polling has been constructed successfully. A refused or
+ malformed password login for that second client preserves the existing client
+ (or its absence), records the failure and does not count or publish recovery.
+ An endpoint must never become healthy with a nil polling client. A later
+ ordinary eligible recovery can install an accepted full-timeout session and
+ clear its own error; usable siblings continue to serve polling throughout.
+
+ `TestClusterRecoveryRequiresUsableSession` covers pinned HTTPS 401/403,
+ malformed success and JSON-to-form refusal, with and without a retained client,
+ then accepted recovery and ordinary reads. `TestClusterRecoverySessionEndpointProjection`
+ drives public cluster reads and background recovery through the monitor's
+ per-endpoint projection: native Online state, identity and saved pins survive,
+ while PulseReachable and PulseError reflect the actual polling session. The
+ existing recovery interval, authentication encodings, TLS trust/first-use
+ policy, failover order and guest-command safety state are unchanged. These
+ synthetic controls establish neither installed trust recovery nor a reported
+ incident's cause.
+
+ ### Rejected infrastructure URLs stay out of diagnostics
+
+ PVE/PBS constructors validate a host before logging it. URL syntax errors omit
+ `url.Parse`'s copy of the supplied input; ordinary validation reasons remain.
+ Cluster logs, health snapshots, recovery summaries and aggregate errors use
+ display-only labels:
+ a valid base URL or `[invalid endpoint]`. Health snapshots preserve each entry
+ even when rejected endpoints share that label, and are built only for enabled
+ log events. These labels never supply routing,
+ alias membership, saved configuration or TLS-pin identity. HTTP/HTTPS defaults,
+ base paths, authentication, CA/pin/opt-out choices and guest fences are unchanged.
+
+ `TestClientRejectedHostDiagnosticsContainNoCredentials` and
+ `TestClusterRejectedEndpointDiagnosticsContainNoCredentials` exercise actual
+ constructors, single/multiple endpoint routing and recovery with synthetic
+ userinfo/query/fragment and malformed URLs. Accepted-host and
+ `TestInfrastructureURLValidationPolicyUnchanged` controls preserve ordinary
+ routing and diagnostics. This bounds client diagnostics, not arbitrary provider
+ error text, previously saved logs, native trust recovery or an installed service.
+
+ ### PBS password renewal publishes one complete shared session
+
+ Concurrent reads of an expired PBS password session share one in-flight
+ authentication attempt. Ticket, CSRF token and expiry are published together;
+ each ordinary request uses one complete tuple. Only renewal is shared: the
+ subsequent API reads remain independent, and a cancelled waiter cannot cancel
+ the owner. Caller cancellation and deadlines stay context errors, not credential
+ refusals, including token reads that never attempt a login. Joined callers receive the owner's failure without reissuing it;
+ a later ordinary call may try again. JSON authentication and the existing
+ 400/415 form fallback retain their semantics. Tokens do not acquire session
+ authority, and API 401/403 replies remain terminal without adding a replay.
+
+ `pkg/pbs/auth_session_concurrency_test.go` exercises pinned local HTTPS expiry,
+ encoding fallback/refusal, malformed responses, independent reads, cookie/CSRF
+ pairing, owner/waiter cancellation and ordinary later recovery. Its accepted
+ node/datastore responses also consume through the canonical PBS projection.
+ TLS pins/CA/opt-out and redirect policy, credentials, ticket lifetime, provider
+ permissions, wire shapes, polling schedules, guest/backup fences and stored
+ History are unchanged. These synthetic controls do not establish an installed
+ connection, native recovery or the cause of a reported PBS/guest failure.
+
+ ### Connection retirement preserves unrelated guest safety state
+
+ Retiring a consolidated PVE connection cancels and clears only its own guest
+ metadata lifecycle. Connection labels may contain `|`; matching a label prefix
+ cannot establish ownership. The metadata key's two fixed node/VMID suffixes
+ are read from the right, while diagnostic memory snapshots use their recorded
+ `Instance`. A separate `site|other` connection keeps its outstanding owner,
+ OS-command suppression, original observations and refresh/backoff deadlines
+ when `site` is retired. The retired connection's cancelled fetch still owns
+ its key until completion, and shared transport uncertainty is never cleared.
+
+ `TestGuestMetadataRetirementOwnershipContract` covers exact cleanup, nested and
+ Unicode labels, diagnostic origins and in-flight HTTP bodies at all three
+ metadata stages. `TestGuestMetadataRetirementPollingKeepsSuppression` exercises
+ both clients and ordinary collectors: no reissued suppressed OS command, no
+ unrelated identity loss, and continued independent filesystem/CPU/memory reads.
+ Wire formats, TTLs, guest-command/backup-lock/TLS fences and stored History are
+ unchanged. These are synthetic source controls, not native thaw/restart recovery,
+ a reporter's failure cause or installed/release acceptance.
+
+ ### Rate refusals cannot permanently disable guest OS discovery
+
+ An observed HTTP 429 from guest OS-info preserves the existing OS-command
+ failure count and suppression state, like access refusal or stopped QGA. It
+ neither exercised the OS command nor supplies a fresh OS observation. Ordinary
+ metadata backoff remains in force; a subsequent admitted poll can learn the OS
+ without a restart or clearing command-safety state. Only the typed wire status
+ admits this distinction: provider text quoting 429 is not a rate refusal.
+
+ `TestGuestMetadataRateLimitRecoveryContract` covers both HTTP client types and
+ ordinary collectors with cold/stale metadata, OS-like 429 bodies, unchanged
+ backoff and accepted filesystem/memory readings, subsequent Windows discovery
+ before Linux-only memory admission, and genuine repeated-failure suppression
+ with a quoted 429. Existing unsupported-command, backup-lock and uncertain-reply
+ fences are unchanged. Its fixtures use local sockets and synthetic credentials,
+ not native guest recovery, History storage or a reported failure's cause.
+
### Cluster endpoint membership belongs to its client
`NewClusterClient` snapshots the supplied endpoint slice before registering
@@@ -7404,41 -7752,64 +7855,103 @@@ through `publishPVEGuests` to the canon
updates its former direct-call assertions without changing Docker ingestion or
the runtime ownership gate; the behavioral withdrawal controls remain required.
+### Backup history collection carries completeness
+
+`GetBackupTaskHistory` in `pkg/proxmox/backup_history.go` enumerates a bounded
+ninety-day PVE task window with pagination and a fixed collection cutoff. Offline
+nodes, failed reads, repeated/capped pages and unconfirmed `Sys.Audit` authority
+remain explicit partial evidence. A successful HTTP response alone cannot prove
+visibility of other users' backup jobs. The cluster client retains its canonical
+transport failover wrapper. Older adapters can still supply live tasks but
+cannot qualify report observation coverage.
+
+`pollBackupTasks` persists mapped task evidence synchronously through the
+organization's recovery manager before recording collection quality. The
+continuity allowance derives from the configured backup interval/cycle cadence.
+An unavailable manager or failed write cannot certify complete observation.
+Multi-guest attribution requires a nontruncated, contiguous completed task log
+with a terminal marker and parsed guests. Incomplete logs remain unqualified
+and eligible for later collection. No backup artifact or restore assurance is
+created by an execution outcome.
+
+Proofs: `pkg/proxmox/backup_history_test.go` covers pagination, permissions,
+partial reads and offline nodes. `internal/monitoring/backup_task_history_test.go`
+covers durable failures surviving an empty poll and log qualification.
+
+Backup collection starts with a bounded retained-history backfill, then uses a
+durable collection checkpoint with a day of overlap. The oldest known pending
+execution remains in range until completion, including after restart. Failed
+reads do not advance the checkpoint. A capped backfill may advance a later
+incremental read without backdating observation coverage. Another node's
+failure cannot be hidden by a truncation reason. The live API/state keeps its
+existing 200 provider tasks per node plus their guest details, while the recovery
+store retains executions beyond that live window. Checkpoint, long-running-job,
+failed-read, capped-backfill and bounded-live-payload regression tests exercise
+these boundaries.
+
+Recovery rollup and alert readers take a locked snapshot of the manager pointer,
+matching the existing setter during live monitor dependency rebinding. The
+`TestBackupHealthRecoveryRebindingCanOverlapActiveReaders` race proof exercises
+rebinding while both consumers remain active.
++
+ ### Complete guest enumeration before publication (14 October 2026)
+
+ PVE `cluster/resources`, node `qemu` and node `lxc` enumeration consume and
+ close the whole existing 8 MiB-bounded HTTP response before admission. Only
+ HTTP 200 containing a complete UTF-8 JSON object with an explicit canonical
+ `data` array is authoritative inventory. Missing/null data, duplicate or
+ case-variant supported fields, conflicting errors, trailing values, incomplete
+ bodies, invalid rows and duplicate guest identities fail the whole enumeration;
+ no partial inventory or provider body is returned as a successful reading.
+ Cluster guest identities require a node and positive cluster-wide unique VMID.
+ Node lists require positive unique VMIDs; existing LXC string integers remain.
+ Scoped empty arrays, optional names/status/metrics, null counter presence,
+ templates, unique provider metadata and unfiltered non-guest resources remain
+ supported. Observation timestamps are assigned only to an admitted inventory.
+
+ A refused efficient enumeration can use the existing ordinary node fallback.
+ Failed node enumerations retain their earlier guest observations, History and
+ backup-template eligibility; a complete fallback can still remove deleted
+ guests and restore ordinary sampling. Complete cluster recovery returns to
+ its efficient route. This changes no TLS/pin, credentials, permissions,
+ polling cadence, guest-command/backup-lock checks, completion uncertainty,
+ publication ownership or grace-period rules.
+
+ `guest_inventory_response_test.go` covers complete-body and all-or-nothing
+ admission on all three real HTTP readers; `io_counters_test.go` covers
+ scoped-empty, extended fields, counter presence and receipt-time controls.
+ `diagnostic_snapshots_test.go` covers actual efficient/fallback collection,
+ owned publication, retained observation times, breach/History/template
+ continuity, independent-connection preservation and normal recovery/removal.
+ These controlled source proofs do not establish native thaw/restart recovery,
+ installed trust, report attribution or release availability.
+
+ Failed-node retention selects canonical guest membership but takes the unchanged
+ complete tuple from that connection's last accepted producer state. Enrichment's
+ partial previous-state projection is not an observation: it must not turn an
+ unavailable/retained memory reading into zero, lose CPU/disk usage or change the
+ original in-guest collection state. A node-scoped detached snapshot copies only
+ selected guests, not unrelated host disks or Kubernetes state. The existing
+ canonical/enrichment path remains the fallback when no accepted source tuple
+ exists; linked-agent evidence is never promoted to Proxmox ownership. The
+ same publication owner still admits or withdraws the eventual generation.
+ `proxmox_guest_observations_test.go` checks scope, original memory provenance,
+ CPU/disk/read time and nested detachment, alongside the end-to-end admission
+ controls above. The first complete admission proof exposed the former partial
+ retention and is retained as failed evidence, not weakened into acceptance.
+
+ ### Permission-denial diagnostics retain scoped access (14 October 2026)
+
+ A PVE token's 403 error explains that privilege-separated access is the
+ intersection of user and token ACLs on the intended resource. It no longer
+ says to inspect only the user or includes the configured username/provider
+ body. The actual HTTP status, authentication-error wrapping and guest permission
+ classification remain; there is no retry, token rotation, access widening or
+ TLS change. Diagnostics distinguishes rejected authentication from permission
+ and other wire failures without promoting a denied resource to a reading.
+ `TestClientRequest_403TokenPermissionHint` and
+ `TestGetNodeStatusPermissionRefusalDoesNotInventMemory` cover the ordinary
+ HTTP boundary, unknown-not-zero memory and resumed CPU/memory with the same
+ credential. `TestComputeDiagnosticsAPIRejectionsPreserveCredentials` covers
+ pinned PVE/PBS fixture use and preserves the separate PBS scheduler state.
+ Native scoped-fleet recovery, QGA/thaw and installed trust remain separate.
diff --cc docs/release-control/v6/internal/subsystems/security-privacy.md
index f836c1048b,89c7fcea83..8277a028e6
--- a/docs/release-control/v6/internal/subsystems/security-privacy.md
+++ b/docs/release-control/v6/internal/subsystems/security-privacy.md
@@@ -18,91 -18,146 +18,231 @@@
## Purpose
+### Restricted outbound connections keep approved destinations pinned
+
+`pkg/securityutil/outbound_http.go` owns restricted HTTP destination validation,
+redirect policy and DNS pinning. `pkg/securityutil/pinned_dial.go` owns shared
+numeric-address connection scheduling, including notification callers. The
+dialer never resolves a hostname or expands destination permissions. Callers
+must validate every candidate before handing it over.
+
+The primitive validates and normalises numeric IPs, rejects malformed input,
+deduplicates equivalent IPv4 forms, respects the requested TCP address family,
+and races at most two staggered serial workers under one positive connection
+timeout. Remaining time is shared among each worker's remaining candidates.
+Caller cancellation cancels in-flight work. A successful connection cancels
+competitors and closes any late successful loser without writing an HTTP
+payload to it. Transport TLS and destination-policy checks remain in place.
+
+`pkg/securityutil/pinned_dial_test.go` verifies family fallback, later-address
+budget, concurrency, cancellation, late-connection closure, numeric-only input
+and error preservation. Notification loopback controls exercise both clients
+against a real stalled TCP address and a healthy receiver. The registry routes
+both shared runtime files to these controls and the existing outbound policy
+tests, keeping this primitive owned without moving notification allowlisting.
+
+### Reporting usage remains content-free and distinguishable from missing measurements
+
+`internal/config/reporting_activity.go` is the canonical durable count store.
+It retains 30 UTC-day buckets of closed report kinds and outcomes. Local opaque
+schedule/period SHA-256 markers expire after 90 days and use existing config
+storage encryption when configured. No raw report, recipient, schedule, scope,
+organization identifier, range or error content enters the tally. Only schema
+v20 scalar counters, measurement availability and elapsed observation days leave
+the installation through the existing opt-out-controlled heartbeat. Local tally
+errors never fail reporting. Mock mode never records real reporting activity.
+
+Generation counts include separate attempts and complete direct response writes.
+Distinct successful-generation days measure activity, including manual downloads
+and scheduled retries, with same-day attempts counted once. Multiple days alone
+cannot establish a fresh successful period or returning manual use. The adoption
+summary keeps multi-day generation separate from successful repeat periods.
+Saved files, SMTP acceptance, definite failures and uncertain acceptance remain
+separate. Persisted successful saved-schedule windows count once. A successful
+repeat period requires a different window on that schedule within 90 days,
+including monthly reuse, so retries and same-period reruns do not inflate that
+period metric. The aggregate
+query uses only each eligible rotating ID's latest counts and marks old-schema,
+missing-column and unavailable-tally data unmeasured. Elapsed observation days do
+not establish continuous uptime, and SMTP acceptance does not prove inbox receipt.
+No customer/account joins or paying-customer census are permitted.
+
+`internal/config/reporting_activity_test.go`,
+`internal/api/reporting_activity_test.go`, the telemetry payload/disclosure tests,
+the server snapshot tests and the adoption-report tests verify this boundary.
+Both privacy document copies and `docs/releases/TELEMETRY_SCHEMA_V20.md` disclose
+the payload. Client source landing and compatible receiver activation remain
+separate from release and installed fleet measurement.
+
+
+### Investigation evidence retains tenant and runtime-source boundaries
+
+Recovery HTTP and tool reads share `Manager.ListPointsForOrg`: a nonblank tenant
+is required, cancellation and read failures remain errors, and ordinary reads
+use that tenant's cached store. The optional demo source follows the current
+fixture graph and never merges persisted tenant rows into demo evidence. The
+real-to-demo-to-real regression checks both consumers and retained store data;
+the persisted task regression excludes another tenant and another guest with
+the same provider VM ID. Historical storage reads preserve source labels and
+requested time bounds. These read paths add no credential, execution or approval
+authority.
+
+### Reports and metric summaries read only the requesting organization
+
+The server's report engine (`pkg/server/server.go`) resolves the metrics store
+of each request's organization through the multi-tenant monitor
+(`EngineConfig.MetricsStoreForOrg`) and refuses a request that names no
+organization, so no report path falls back to the default organization's
+store. Report handlers also reject a resource that is not in the requesting
+organization's inventory, scheduled runs resolve their scope against their
+owning organization's inventory, and Pulse Assistant's `pulse_summarize`
+reads its executor's organization. A resource ID that exists in two
+organizations yields each organization only its own history.
+`internal/api/reporting_org_scope_test.go`,
+`TestReportingScheduledRunReadsOnlyItsOwningOrganizationsMetrics`,
+`pkg/reporting/engine_org_scope_test.go` and
+`TestSummarizeToolReadsOnlyItsOrganizationsMetrics` hold this boundary.
+
+ ### Conflicting saved endpoint pins cannot select certificate trust
+
+ For equivalent Proxmox endpoint URL spellings, all nonempty saved member pins
+ must agree under the TLS verifier's existing case/colon normalization. Exact
+ spelling and map iteration can no longer select one of contradictory pins.
+ Resolution returns a fixed, credential-free trust-conflict error before a TLS
+ probe, password login, token read, first-use capture or polling-client creation.
+ The member stays unreachable to Pulse; an independently trusted sibling can
+ continue ordinary monitoring. No saved pin is removed or overwritten, and the
+ error cannot become an empty pin/CA or opt-out fallback. Agreeing pin formats,
+ nonempty invalid pins, primary-pin fallback, declared-member first-use and the
+ existing scheme/port/base-path/IPv4-mapped trust boundaries remain unchanged.
+
+ `TestClusterConflictingSavedPinsNeverAuthenticate` covers startup, recovery and
+ lazy selection in token/JSON/form modes, three URL spellings and both VerifySSL
+ settings with real TLS sockets. It checks zero TLS/login/read work for conflicts,
+ no installed client or trust mutation, sibling use, and accepted monitoring
+ following a normal reload with reconciled fixture pins. Existing established-pin,
+ first-use, cancellation, ownership and guest-admission controls still apply.
+ `TestClusterSavedPinResolutionRejectsConflicts` and
+ `TestClusterFirstUseCaptureDoesNotReconcileConflictingPins` retain the conflict
+ through normalized authorities, invalid nonempty pins and a capture race.
+ `TestClusterConflictingPinsEndpointProjection` uses the production endpoint
+ builder, real clients and monitor projection in all three auth modes: native
+ Online, identity and saved trust are not changed by failed Pulse reachability.
+ The producer also records contradictory exact duplicate saved rows before
+ its map/deduplication can overwrite them, and carries those endpoint conflicts
+ through initialisation, missing-client reconnect and cluster refresh. The
+ cluster owns a copy of that conflict set; aliases cannot escape it, and a
+ later repeated row cannot clear it. Normal configuration reload with agreeing
+ pins is required to resume that endpoint. A non-conflicting sibling is not
+ refused. The public constructor's optional conflict inputs preserve existing
+ callers and do not introduce a new trust mode. Producer tests cover both
+ routes, pin formatting, repeated contradictions and configuration ownership.
+ The projection cases include exact duplicates as well as URL aliases.
+ A credential-free agreement check also holds automatic topology refresh while
+ saved trust is contradictory: discovery cannot collapse the rows, move an
+ alias or substitute a pin to clear that refusal. Current trusted sibling
+ clients keep monitoring; topology changes wait for independently reconciled
+ configuration. `TestClusterPinConflictsSurviveMonitorLifecycle` exercises the
+ actual initialisation, missing-client retry and refresh/rebuild paths in all
+ three auth modes and both layouts, including normal refresh after reconciliation.
+ Saved-view consolidation also refuses a view with internal pin contradictions,
+ so map indexing cannot hide an earlier contradictory row before monitoring
+ starts. `TestConsolidatePVEInstancesRetainsInternallyConflictingPins` covers
+ both merge orders, repeated rows, same-node/address evidence and agreeing formats.
+ Internal agreement uses the TLS verifier's case/colon comparison even when
+ legacy identity formatting would erase or simplify a malformed nonempty pin;
+ colon-only, whitespace and prefixed pins cannot hide a contradiction.
+ The endpoint-ownership routing control uses a five-second fixture client and
+ a ten-second enclosing caller lifetime, rather than a one-second latency
+ criterion. It records elapsed time and both targets' request counts; unchanged
+ pin, admission and routing assertions remain required. Runtime timeouts and
+ separate timeout/cancellation controls are unchanged. This is not a diagnosis
+ or dismissal of the retained earlier one-second timeout.
+ Lock-evidence fixture cases use independent base-path namespaces while each
+ case's clients share the same guard key. A recycled loopback port cannot inherit
+ an earlier synthetic guest's cooldown. `TestGuestAgentLockEvidenceFixtureIsolation`
+ shares one listener and retains an earlier uncertainty fence unchanged while an
+ independent fixture completes its lock/command checks. No runtime fence is
+ cleared, expired or weakened; native guest recovery is not established.
+ This establishes source/consumer behaviour, not native recovery or attribution
+ of a guest, PBS or certificate report.
+
+
+ ### First-use cluster trust includes password construction
+
+ Declared, previously untrusted Proxmox members use the existing credential-free
+ TLS capture and pinned retry whether the inherited primary pin fails during
+ password login or during the token-authenticated nodes probe. Startup, ordinary
+ unhealthy recovery and lazy first selection share this path. Capture still
+ requires typed local pin-mismatch evidence and cannot replace a primary, saved
+ member or concurrent winner pin; it never sends an HTTP request or credentials.
+ The pinned retry is not repeated to obtain a nodes count. Failed or cancelled
+ authentication installs no polling session; a separately accepted full-timeout
+ client is required before recovery becomes healthy. Refusal retains the
+ established runtime pin for a distinct later ordinary recovery. Saved trust,
+ CA/redirect policy, login encodings, cadence and guest/backup fences do not change.
+
+ `TestClusterFirstUseSessionAdmission` covers JSON password, existing form
+ fallback and token modes in all three phases, then ordinary member reads.
+ The adjacent refusal, established-pin and cancellation controls exercise actual
+ TLS sockets. Concurrent lazy selection retains serialized construction while
+ health/trust inspection stays usable during capture. `TestClusterFirstUsePasswordEndpointProjection` carries first-use
+ probe/full-session refusal and later public-read recovery into the monitor's
+ cluster-member projection without changing native Online state, identity or
+ saved fingerprints. This is source/consumer qualification, not installed
+ cluster recovery or attribution of an existing user report.
+
+
+ ### Proxmox password success admits a complete session, not just HTTP 200
+
+ PVE and PBS consume the entire existing byte-bounded login body before
+ publishing ticket, CSRF token and the normal two-hour expiry together. The
+ shared `internal/securityutil/proxmox_session.go` parser requires canonical,
+ unambiguous envelope/session keys, complete UTF-8 JSON, a nonempty cookie-safe
+ ticket and a nonempty header-safe CSRF token. Unknown unique metadata such as
+ username and capabilities remains supported. Partial/null, duplicate or
+ case-conflicting credentials, extra JSON values, incomplete reads and oversized
+ tails fail without replacing the retained tuple or advancing PVE generation.
+ Validation errors contain no response credentials. A malformed success is not
+ a provider credential refusal and does not trigger JSON-to-form fallback.
+
+ `session_response_admission_test.go` in both client packages exercises these
+ controls and eight joined expired readers, zero subsequent API/form requests
+ on invalid success, retained state and distinct later ordinary recovery with
+ coherent cookie/CSRF pairs. Cluster recovery additionally checks empty and
+ partial success for the retained polling client, with and without a previous
+ client. `TestPBSInvalidSessionKeepsFirstPollBoundary` exercises the real pinned
+ constructor, normal startup retry, poll executor and production broadcast:
+ invalid success installs no client, valid construction is not healthy telemetry,
+ and only the accepted first poll publishes recovered node/datastore readings.
+
+ TLS trust, credentials, auth encodings and 400/415 fallback, API-token paths,
+ renewal ownership, retries/cadence, guest/backup fences and History are unchanged.
+ The parser establishes response integrity, not cryptographic ticket validation
+ or native/installed acceptance. It neither attributes nor resolves a reported
+ PBS, Windows guest, restart or workload-safety incident.
+
+
+ ### Rejected infrastructure URLs stay out of diagnostics
+
+ PVE/PBS constructors validate a host before logging it. URL syntax errors omit
+ `url.Parse`'s copy of the supplied input; ordinary validation reasons remain.
+ Cluster logs, health snapshots, recovery summaries and aggregate errors use
+ display-only labels:
+ a valid base URL or `[invalid endpoint]`. Health snapshots preserve each entry
+ even when rejected endpoints share that label, and are built only for enabled
+ log events. These labels never supply routing,
+ alias membership, saved configuration or TLS-pin identity. HTTP/HTTPS defaults,
+ base paths, authentication, CA/pin/opt-out choices and guest fences are unchanged.
+
+ `TestClientRejectedHostDiagnosticsContainNoCredentials` and
+ `TestClusterRejectedEndpointDiagnosticsContainNoCredentials` exercise actual
+ constructors, single/multiple endpoint routing and recovery with synthetic
+ userinfo/query/fragment and malformed URLs. Accepted-host and
+ `TestInfrastructureURLValidationPolicyUnchanged` controls preserve ordinary
+ routing and diagnostics. This bounds client diagnostics, not arbitrary provider
+ error text, previously saved logs, native trust recovery or an installed service.
+
### Patrol maintenance coverage widens no authority
Patrol inherits only maintenance from an ancestor window scoped
@@@ -3426,46 -3430,16 +3578,60 @@@ controls in `send_report_test.go` and `
status/endpoint, non-disclosure and resumed ordinary reporting with the same
credential. They are not a credential inventory probe or native recovery proof.
+### Concurrent chart service publication retains tenant boundaries (16 October 2026)
+
+`internal/api/router.go` initializes each router's chart service through the
+same guarded accessor used by concurrent history and chart requests. The
+service remains scoped to its router and resolves the monitor from the current
+request context. This synchronizes publication without changing authentication,
+authorization, tenant selection or disclosure. The concurrent initialization
+regression in `internal/api/contract_test.go` and the existing
+`TestMultiTenant_TenantIsolation` race test verify that boundary.
+
+### Instance administration refuses organization owners (17 October 2026)
+
+`ensureAdminSession`, the session check behind `RequireAdmin` and the settings
+scope helpers, admits only instance administrators: the configured admin or an
+instance RBAC admin grant, whichever organization the request names. It used
+to admit any manager of the request's organization, which on a multi-tenant
+install let an organization owner rotate the instance administrator's API
+token and receive its secret, widen token scopes, change SSO providers and
+system settings, apply updates and restart the service. API tokens follow the
+same boundary: `ensureAdminSession` and `ensureSettingsScope` refuse a token
+bound to any organization other than the default one (`tokenActsForInstance`)
+whatever its scopes. Before that, an org-bound `settings:write` token resolved
+to its own organization without naming it, passed the tenant middleware and
+reached system settings, SSO providers, token management, relay, log level,
+updates and diagnostics. `internal/api/org_bound_token_authority_test.go` holds
+the token matrix. Routes that act only on
+the request organization opt in through `RequireOrgAdmin`,
+`ensureOrgAdminSession` or `ensureOrgSettingsWriteScope`; every other route fails
+closed. `internal/api/org_admin_session_authority_test.go` holds the session
+matrix and `TestSettingsWriteScopeSplitsInstanceAndOrgManagerSession` the
+settings scope split.
+
+### Backup report renderer hook preserves authority
+
+The server supplies `GenerateBackupHealthReportFunc` from BusinessHooks before
+router and scheduler construction, and clears it for runtimes without the hook.
+The callback receives already-scoped recovery evidence and effective branding.
+It adds no route, organization selection, credential, action or licensing
+authority. Existing report administrator/settings scopes and advanced-reporting
+checks remain before generation and delivery. Missing private rendering fails
+explicitly. `pkg/server/server_test.go` retains runtime identity separation.
+Backup reporting tests cover missing renderer and unchanged entitlement refusal.
+The live disposable Pro runtime proof exercises actual registered rendering.
++
+ ### Diagnostics does not prescribe credential or ACL churn (14 October 2026)
+
+ Proxmox 403 is an access denial, not evidence that a saved token must be
+ rotated or privileges broadened. Diagnostic recovery keeps intentional target
+ boundaries, both user/token ACLs and privilege separation; 401 replacement is
+ conditional on confirmed invalidity. Typed HTTP status precedes untrusted
+ provider text. The existing displayed error summary uses connection positions
+ and fixed diagnoses/guidance, not host, token, username or provider-body text.
+ No credential, TLS, authorization, caller-filtering or endpoint contract is
+ weakened. The diagnostic wire/identity and ordinary same-credential recovery
+ controls in `diagnostics_access_test.go`, `diagnostics_additional_test.go`,
+ `contract_test.go` and `client_request_test.go` cover these properties; they do
+ not establish provider or installed acceptance.
diff --cc docs/release-control/v6/internal/subsystems/storage-recovery.md
index 7ae4b3701f,980a11295e..79b077b9fc
--- a/docs/release-control/v6/internal/subsystems/storage-recovery.md
+++ b/docs/release-control/v6/internal/subsystems/storage-recovery.md
@@@ -21,40 -21,30 +21,64 @@@
## Purpose
+### Patrol timer checkpoints preserve run and recovery history
+
+The existing Patrol history envelope retains the next scheduled deadline and
+its effective interval. History and timer writes share a serialized
+read-modify-write boundary, preserving each other's records, daily tallies and
+cursor. A failed history read must fail the timer write instead of replacing
+unreadable evidence. The configuration and Patrol history tests cover concurrent
+writes, file reconstruction and failed-write retry. This cursor schedules a
+check only; it grants no additional recovery execution authority.
+
+### Recurring findings preserve the ownership of recovery evidence
+
+Storage and recovery findings use the same occurrence-bound investigation
+projection as other Patrol findings. A previous verified action remains
+historical evidence when the finding is re-detected; it cannot resolve the new
+occurrence or replace its investigation record. Finding-scoped investigation
+and transcript reads select the current session, including an explicit absent
+state before a new session starts. Explicit session and action history retain
+the earlier result. The shared three-occurrence and persisted-restart controls
+live in `internal/api/ai_handlers_patrol_actions_additional_test.go`.
+
+### Recovery APIs and investigations share the org-scoped point reader
+
+`Manager.ListPointsForOrg` selects the composed recovery evidence source for
+both the HTTP points API and investigation tools. Real installations retain
+tenant-scoped SQLite reads, including errors; demo composition projects the
+current canonical fixture graph through shared recovery filters and pagination.
+It does not seed a second store or grant recovery execution. Source selection is
+bound before consumers share the manager and follows runtime demo-mode changes
+without exposing persisted rows through demo reads. Manager boundary tests preserve tenant
+validation, cancellation and read failures; the API/tool parity regression in
+`internal/api/investigation_evidence_parity_test.go` reads actual graph-owned
+failure text without a PVE command connection.
+
+ ### Known PBS health survives unreadable backup inventories (13 October 2026)
+
+ Backups keeps the separately supplied PBS server/datastore projection mounted
+ while PVE and PBS inventories load, fail or recover. Inventory notices and
+ source-local retries remain explicit; both failures still suppress the inventory
+ summary rather than claim zero restore points or no backups. PBS counts are
+ Loading/Unavailable until its own valid response arrives; resource health,
+ capacity, identity and detail access remain independent. No backup execution,
+ agent requirement, credential/access change or reporter recovery is implied.
+ `ProxmoxBackupsTable.test.tsx` covers both pending reads, each source failure,
+ both failures and ordinary recovery without replacing the known server table.
+
+ ### Portable installation input does not change recovery authority
+
+ The backend setup/Proxmox/container copied-command reader can use BusyBox `stty`
+ when standalone `stty` is absent. Tool or terminal-mode failures stop before
+ installation; full same-tool restoration and silent overflow draining retain
+ the private credential boundary. The shorter root/sudo dispatch adds no root
+ profile, state/identity replacement, backup operation or incident-replay right.
+ Reported QNAP installed boot/watchdog/removal/reinstall recovery remains a
+ separate native outcome; source PTYs and publishing an incoming updater do
+ not establish it. `portable_bootstrap_test.go` and the serialized API contract
+ cover these producer boundaries without touching stored monitoring data.
+
### Patrol maintenance coverage adds no recovery execution
Patrol's operator-state projection now projects the active recurring
diff --cc docs/release-control/v6/internal/subsystems/unified-resources.md
index a7fb62876c,ae1c7ca7eb..ceb9d09e21
--- a/docs/release-control/v6/internal/subsystems/unified-resources.md
+++ b/docs/release-control/v6/internal/subsystems/unified-resources.md
@@@ -23,22 -23,63 +23,79 @@@ and sort the complete canonical change
## Purpose
+### Docker object aliases do not identify their reporting host
+
+Containers and other Docker child objects own their native IDs, names and
+subject-typed discovery/metrics targets. Reporting-agent IDs, Docker host IDs,
+hostnames and task placement nodes remain routing metadata, not child aliases.
+Canonical hostname/platform presentation must not reintroduce those parent
+identities. Missing child targets cannot promote the parent to primary identity.
+Host aliases and same-entity host facets retain their existing semantics.
+
+`TestDockerChildIdentityDoesNotClaimItsCommandHost` covers all nine child kinds,
+owned aliases and missing-target boundaries. The registered Docker log-read
+control in `internal/ai/tools/internal_routing_identity_test.go` consumes actual
+registry ingestion with collected containers, images and networks. Genuine
+duplicate names must still fail closed. Source controls alone do not establish
+installed command transport or autonomous recovery.
+
+ ### Workload aggregate disk presence survives projection (14 October 2026)
+
+ `useWorkloads` does not default an absent aggregate disk usage/used byte field
+ to zero merely because capacity was supplied. Both REST and owning snapshots
+ preserve the difference between capacity-only inventory and finite measured
+ usage. Explicit zero percentages, percentage-only metrics and valid byte-only
+ metrics remain available. This aggregate metric contract is distinct from
+ DiskInfo filesystem records whose zero byte fields may be omitted. Identity,
+ producer selection, collection, retention and History transport are unchanged.
+
+ Raw byte metrics always serialize a generic `value`, even its default zero.
+ That scalar cannot confer a current percentage: an explicit `percent` or a
+ valid used/total byte pair is required when `unit` names bytes. Legacy untyped
+ value-only percentages and explicit percent-unit values remain usable.
+
+ Verification: `useWorkloads.test.ts` exercises mounted REST and owning-snapshot
+ refreshes and `workloadDiskUsage.test.ts` covers numeric admission. The
+ `workload-disk-metrics` browser fixture consumes the REST mapper into production
+ rows/drawer with synthetic evidence; browser proof is retained separately and
+ establishes no installed/native recovery.
+
+
+ ### Workloads preserve platform guest interfaces — issue #2757
+
+ Both the `/api/resources` and owning canonical-snapshot Workloads projections
+ retain `proxmox.networkInterfaces` when the linked agent reports no interface
+ list. A present agent list still wins, including an explicit empty list;
+ lists from different producers are not merged. Names, addresses, MACs and
+ reported traffic retain their association. The selected identity/IP order,
+ canonical row and History coordinates do not change.
+
+ `networkInterfacesFromAgent` is internal frontend provenance. Only the selected
+ agent list inherits `agentStale` in the drawer's Network section. A silent linked
+ agent still marks its other readings last known, but cannot mark independently
+ supplied Proxmox interfaces as its last report. Legacy drawer callers keep their
+ existing agent-freshness semantics. This adds no wire field, guest query,
+ collection command, permission or native freshness/recovery guarantee.
+
+ `useWorkloads.test.ts` checks REST/snapshot VM and LXC source transitions,
+ agent precedence, empty/absent lists, IP order and stable canonical identity;
+ `GuestDrawer.test.tsx` checks reactive source-specific last-known presentation
+ without clearing agent RAID/filesystem warnings. Desktop and narrow browser
+ proof must consume producing JSON, not manually populated WorkloadGuest rows.
+ The row's existing Network tooltip uses semantic foreground/muted tokens for
+ names, addresses, MACs, traffic and the address-only fallback. Fixed dark-palette
+ text must not make the newly retained names unreadable on a light surface.
+
+ ### Independently known PBS remains available in Backups (13 October 2026)
+
+ The resource-owned PBS table is outside backup-inventory loading/error gates.
+ Its canonical identity, datastore health/capacity and details remain usable
+ when either or both inventory endpoints fail. Error/loading counts are not zero
+ backup evidence. The inventory-only surface retains incomplete notices and
+ ordinary retries. `ProxmoxBackupsTable.test.tsx` checks stable server-table
+ ownership through loading, single/both failures and valid response recovery.
+ This consumes existing resources; it adds no producer or native recovery claim.
+
### Shared interpretation of selected guest memory
`GuestMemoryEvidenceForResource` and the VM/container `MemoryEvidence` accessors
diff --cc frontend-modern/browser-tests/resource-stopped-filesystem.cjs
index 826564e1f1,826564e1f1..03364df79b
--- a/frontend-modern/browser-tests/resource-stopped-filesystem.cjs
+++ b/frontend-modern/browser-tests/resource-stopped-filesystem.cjs
@@@ -3,7 -3,7 +3,8 @@@ const fs = require('node:fs')
const path = require('node:path');
const crypto = require('node:crypto');
const { chromium, webkit } = require('playwright');
--const root = '/workspace/frontend-modern';
++const workspace = process.env.PULSE_BROWSER_PROOF_ROOT || '/workspace';
++const root = path.join(workspace, 'frontend-modern');
const origin = 'http://127.0.0.1:5343';
const hash = (file) => crypto.createHash('sha256').update(fs.readFileSync(file)).digest('hex');
const runtime = [
@@@ -12,6 -12,6 +13,7 @@@
'frontend-modern/src/components/Infrastructure/ResourceDetailDrawer.tsx',
'frontend-modern/src/components/Workloads/GuestDrawerHistory.tsx',
'frontend-modern/src/utils/workloadGuestPresentation.ts',
++ 'frontend-modern/src/utils/memoryObservation.ts',
];
const settle = (page) =>
page.evaluate(async () => {
@@@ -20,13 -20,13 +22,18 @@@
});
async function verify(original = false) {
-- const output = `/workspace/tmp/resource-stopped-filesystem/browser-${original ? 'original' : 'final'}`;
++ const output = path.join(
++ workspace,
++ `tmp/resource-stopped-filesystem/browser-${original ? 'original' : 'final'}`,
++ );
fs.mkdirSync(output, { recursive: true });
const result = {
result: 'incomplete',
phase: original ? 'original-adverse-observation' : 'candidate-verification',
playwright: require('playwright/package.json').version,
-- runtime_hashes: Object.fromEntries(runtime.map((file) => [file, hash('/workspace/' + file)])),
++ runtime_hashes: Object.fromEntries(
++ runtime.map((file) => [file, hash(path.join(workspace, file))]),
++ ),
cases: [],
screenshots: [],
cleanup: { browsers: false, server: false },
@@@ -35,9 -35,9 +42,8 @@@
};
assert.equal(
result.playwright,
-- JSON.parse(fs.readFileSync('/workspace/tests/integration/package-lock.json')).packages[
-- 'node_modules/@playwright/test'
-- ].version,
++ JSON.parse(fs.readFileSync(path.join(workspace, 'tests/integration/package-lock.json')))
++ .packages['node_modules/@playwright/test'].version,
);
process.chdir(root);
const { createServer } = await import(path.join(root, 'node_modules/vite/dist/node/index.js'));
@@@ -53,6 -53,6 +59,7 @@@
await server.listen();
for (const [name, engine, width, height, dark] of [
['chromium-desktop-light', chromium, 1280, 900, false],
++ ['chromium-tablet-light', chromium, 1024, 768, false],
['webkit-phone-dark', webkit, 390, 844, true],
]) {
browser = await engine.launch(
@@@ -308,6 -308,6 +315,68 @@@
);
await apply({ type: 'vm', power: 'stopped', status: 'online', reason: '', usage: 90 });
await unavailable();
++ // Selected memory freshness belongs to its source, independently of guest type.
++ for (const type of ['vm', 'system-container']) {
++ await apply({
++ type,
++ power: 'running',
++ status: 'online',
++ memorySource: 'agent',
++ memoryState: 'current',
++ memoryUsage: 0,
++ agentStale: false,
++ });
++ assert.match(
++ await chart.locator('[data-history-current="memory"]').innerText(),
++ /0\.0%\s*current/,
++ );
++ const memoryReads = reads();
++ await apply({ agentStale: true });
++ assert.equal(await chart.locator('[data-history-current="memory"]').count(), 0);
++ assert.match(
++ await chart.locator('[data-history-last-known="memory"]').innerText(),
++ /0\.0%/,
++ );
++ assert.match(
++ await chart.locator('[data-history-deferred="memory"]').innerText(),
++ /Last known/,
++ );
++ assert.match(
++ await chart.locator('[data-history-deferred="memory"]').innerText(),
++ /2026-10-10 00:00:00 UTC/,
++ );
++ record.states.push({ state: `${type}-agent-stale`, text: await chart.innerText() });
++ await capture(`${type}-agent-stale`);
++ await apply({ memorySource: 'cluster-resources', memoryUsage: 95 });
++ assert.match(
++ await chart.locator('[data-history-current="memory"]').innerText(),
++ /95\.0%\s*current/,
++ );
++ await capture(`${type}-provider-current`);
++ await apply({ memoryState: 'last-known' });
++ const memoryText = await chart.locator('[data-history-deferred="memory"]').innerText();
++ if (type === 'vm') assert.match(memoryText, /may include cache/i);
++ else assert.doesNotMatch(memoryText, /may include cache/i);
++ await capture(`${type}-provider-retained`);
++ await apply({ memorySource: 'agent', memoryState: 'unknown', agentStale: false });
++ assert.equal(await chart.locator('[data-history-current="memory"]').count(), 0);
++ assert.match(
++ await chart.locator('[data-history-deferred="memory"]').innerText(),
++ /Freshness unknown/,
++ );
++ await apply({ memoryState: 'current' });
++ assert.equal(reads(), memoryReads);
++ }
++ record.checks.push(
++ 'A stale selected agent retains its original memory observation for VM and LXC; independent Proxmox memory stays current, with the cache caveat only for VMs.',
++ );
++ await apply({
++ type: 'vm',
++ power: 'stopped',
++ memorySource: 'agent',
++ memoryUsage: 0,
++ agentStale: false,
++ });
mode = 'stored';
const read = page.waitForResponse((r) => r.url().includes('/api/metrics-store/history'));
await page.clock.runFor(30_000);
diff --cc frontend-modern/browser-tests/resource-stopped-filesystem.tsx
index dc7c35b3b4,dc7c35b3b4..73a2bf24e1
--- a/frontend-modern/browser-tests/resource-stopped-filesystem.tsx
+++ b/frontend-modern/browser-tests/resource-stopped-filesystem.tsx
@@@ -25,6 -25,6 +25,10 @@@ type Snapshot =
platform: Resource['platformType'];
scopes?: Resource['platformScopes'];
seen: number;
++ memorySource?: string;
++ memoryState?: string;
++ memoryUsage?: number;
++ agentStale?: boolean;
};
function Fixture() {
@@@ -54,9 -54,9 +58,14 @@@
},
cpu: { current: 10 },
memory: {
-- current: 0,
-- observation: { state: 'current', source: 'agent', observedAt: '2026-10-10T00:00:00Z' },
++ current: snapshot().memoryUsage ?? 0,
++ observation: {
++ state: snapshot().memoryState ?? 'current',
++ source: snapshot().memorySource ?? 'agent',
++ observedAt: '2026-10-10T00:00:00Z',
++ },
},
++ agent: { stale: snapshot().agentStale === true },
disk: { current: snapshot().usage },
network: { rxBytes: 0, txBytes: 200 },
proxmox: {
diff --cc frontend-modern/browser-tests/unix-private-install.cjs
index 96e520158d,62f3566191..1d4c744fbb
--- a/frontend-modern/browser-tests/unix-private-install.cjs
+++ b/frontend-modern/browser-tests/unix-private-install.cjs
@@@ -2,24 -2,45 +2,46 @@@
const assert = require('node:assert/strict');
const fs = require('node:fs');
const path = require('node:path');
+ const http = require('node:http');
+ const crypto = require('node:crypto');
const { chromium, webkit } = require('playwright');
+ const hash = (bytes) => crypto.createHash('sha256').update(bytes).digest('hex');
(async () => {
-- const root = '/workspace/frontend-modern';
- process.chdir(root);
- console.log(
- JSON.stringify({
- fixtureHTML: fs.existsSync(path.join(root, 'browser-tests/unix-private-install.html')),
- fixtureTSX: fs.existsSync(path.join(root, 'browser-tests/unix-private-install.tsx')),
- }),
- );
- const artifacts = path.join(root, 'node_modules', 'unix-private-browser');
- fs.mkdirSync(artifacts, { recursive: true });
- const { createServer } = await import(path.join(root, 'node_modules/vite/dist/node/index.js'));
- const server = await createServer({
- root,
- configFile: path.join(root, 'vite.config.ts'),
- server: { host: '127.0.0.1', port: 5241, strictPort: true },
- const artifacts = '/workspace/tmp/unix-private-install';
++ const workspace = process.env.PULSE_BROWSER_PROOF_ROOT || '/workspace';
++ const root = path.join(workspace, 'frontend-modern');
++ const artifacts = path.join(workspace, 'tmp/unix-private-install');
+ const preview = path.join(artifacts, 'static');
+ const bindingBytes = fs.readFileSync(path.join(preview, 'binding.json'));
+ const binding = JSON.parse(bindingBytes);
+ for (const [file, expected] of Object.entries(binding.sources))
+ assert.equal(hash(fs.readFileSync(path.join(root, file))), expected, file);
+ for (const [file, expected] of Object.entries(binding.artifacts))
+ assert.equal(hash(fs.readFileSync(path.join(preview, file))), expected, file);
+ const served = [];
+ const server = http.createServer((req, res) => {
+ const file = new URL(req.url, 'http://fixture.invalid').pathname.slice(1);
+ if (!['GET', 'HEAD'].includes(req.method) || !Object.hasOwn(binding.artifacts, file)) {
+ res.writeHead(404);
+ res.end();
+ return;
+ }
+ const bytes = fs.readFileSync(path.join(preview, file));
+ assert.equal(hash(bytes), binding.artifacts[file]);
+ served.push({ file, sha256: hash(bytes) });
+ res.writeHead(200, {
+ 'Content-Type': file.endsWith('.js')
+ ? 'text/javascript'
+ : file.endsWith('.css')
+ ? 'text/css'
+ : 'text/html',
+ 'Cache-Control': 'no-store',
+ });
+ res.end(bytes);
});
+ const captures = [],
+ consumerCommands = [],
+ requests = [];
+ let cleanup = { browser_closed: false, server_closed: false };
const results = [];
let browser;
const secret = 'browser-synthetic-install-token';
@@@ -38,7 -63,37 +64,38 @@@
viewport: { width, height: 900 },
...(engine === 'webkit' ? { isMobile: true, hasTouch: true } : {}),
});
++ await page.clock.install();
const errors = [];
+ const capture = async (name, locator) => {
+ if (locator)
+ await locator.evaluate((el) => window.scrollBy(0, el.getBoundingClientRect().top - 48));
+ await page.evaluate(
+ () =>
+ new Promise((resolve) => requestAnimationFrame(() => requestAnimationFrame(resolve))),
+ );
+ assert.equal(
+ await page.evaluate(() => document.documentElement.scrollWidth <= innerWidth),
+ true,
+ name,
+ );
+ const file = `${engine}-${tone}-${name}.png`;
+ await page.screenshot({ path: path.join(artifacts, file) });
+ captures.push({
+ file,
+ sha256: hash(fs.readFileSync(path.join(artifacts, file))),
+ engine,
+ tone,
+ width,
+ height: 900,
+ });
+ };
+ const activate = async (locator) => {
+ if (engine === 'webkit') await locator.tap();
+ else {
+ await locator.focus();
+ await page.keyboard.press('Enter');
+ }
+ };
page.on('pageerror', (e) => errors.push(e.message));
page.on('console', (m) => {
if (m.type() === 'error') {
@@@ -48,6 -103,6 +105,8 @@@
});
page.on('requestfailed', (r) => console.log('REQUEST_FAILED', r.url(), r.failure()));
let minted = 0;
++ let lookupReads = 0;
++ const seen = new Date(Date.now() - 5000).toISOString();
let optionalAuth = false;
await page.addInitScript(() => {
window.__copies = [];
@@@ -84,6 -142,6 +146,21 @@@
},
});
}
++ if (url.pathname === '/api/agents/agent/lookup') {
++ lookupReads++;
++ return route.fulfill({
++ json: {
++ agent: {
++ id: 'fixture-connected',
++ hostname: 'fixture-host',
++ displayName: 'Fixture host',
++ connected: true,
++ status: 'online',
++ lastSeen: seen,
++ },
++ },
++ });
++ }
if (url.pathname === '/api/state')
return route.fulfill({ json: { connectedInfrastructure: [] } });
return route.fulfill({ json: { data: [] } });
@@@ -162,6 -248,6 +267,22 @@@
await page
.getByRole('button', { name: 'Show advanced connection and install options', exact: true })
.click();
++ const signing = page
++ .locator('div')
++ .filter({ has: page.getByText('Config signing (optional)', { exact: true }) })
++ .last();
++ for (const key of [
++ 'PULSE_AGENT_CONFIG_SIGNATURE_REQUIRED=true',
++ 'PULSE_AGENT_CONFIG_SIGNING_KEY',
++ 'PULSE_AGENT_CONFIG_PUBLIC_KEYS',
++ ]) {
++ const code = signing.getByText(key, { exact: true });
++ await code.scrollIntoViewIfNeeded();
++ const bounds = await code.boundingBox();
++ assert.ok(bounds && bounds.x >= 0 && bounds.x + bounds.width <= width, key);
++ assert.match(await code.getAttribute('class'), /wrap-anywhere/);
++ }
++ await capture('config-signing', signing);
await page
.getByLabel('Connection URL (Agent → Pulse)', { exact: true })
.fill('https://pulse.example/base/');
@@@ -203,12 -289,13 +324,23 @@@
'Never insert the token into the command',
),
);
- await page.screenshot({
- path: path.join(artifacts, `${engine}-commands.png`),
- fullPage: true,
- });
+ await capture(
+ 'trust-options',
+ page.getByLabel('Custom CA certificate path (optional)', { exact: true }),
+ );
+ consumerCommands.push({ engine, tone, kind: 'linux-install', command: unixCommands[0] });
++ await page.getByPlaceholder('Hostname or agent ID').fill('fixture-host');
++ await page.getByRole('button', { name: 'Check status', exact: true }).click();
++ await page.getByText('First host connected', { exact: true }).waitFor();
++ const age = page.getByText(/^Last seen /);
++ const initialAge = await age.innerText();
++ const readsBeforeTick = lookupReads;
++ await page.clock.runFor(65_000);
++ assert.notEqual(await age.innerText(), initialAge);
++ assert.equal(lookupReads, readsBeforeTick);
++ await capture('connected-age', age);
optionalAuth = true;
- await page.goto('http://127.0.0.1:5241/browser-tests/unix-private-install.html?optional=1');
+ await page.goto(`${origin}/browser-tests/unix-private-install.html?optional=1`);
await page.getByRole('button', { name: 'Confirm without token', exact: true }).click();
const optionalCode = page.locator('code').filter({ hasText: 'bootstrap_dir=$(mktemp' });
const optionalCommands = await optionalCode.allTextContents();
@@@ -242,15 -329,38 +374,38 @@@
await browser.close();
browser = undefined;
}
- assert.equal(captures.length, 16);
++ assert.equal(captures.length, 24);
+ } finally {
+ if (browser) await browser.close();
+ cleanup.browser_closed = true;
+ await new Promise((resolve) => server.close(resolve));
+ cleanup.server_closed = true;
+ fs.writeFileSync(
+ path.join(artifacts, 'consumer-commands.json'),
+ JSON.stringify({ source: binding.source_sha, consumerCommands }, null, 2) + '\n',
+ );
fs.writeFileSync(
path.join(artifacts, 'result.json'),
- JSON.stringify({ playwright: require('playwright/package.json').version, results }, null, 2),
+ JSON.stringify(
+ {
+ source: binding.source_sha,
+ binding_sha256: hash(bindingBytes),
+ source_inputs: Object.keys(binding.sources).length,
+ playwright: require('playwright/package.json').version,
+ results,
+ captures,
+ requests,
+ served,
+ cleanup,
+ },
+ null,
+ 2,
+ ) + '\n',
);
- console.log(JSON.stringify({ result: 'passed', artifacts, results }));
- } finally {
- if (browser) await browser.close();
- await server.close();
}
+ console.log(
+ JSON.stringify({ result: 'passed', artifacts, results, captures: captures.length, cleanup }),
+ );
})().catch((e) => {
console.error(e);
process.exitCode = 1;
diff --cc frontend-modern/public/docs/UPGRADE_v6.md
index ec3222b328,bfef9cf395..482bb3f0df
--- a/frontend-modern/public/docs/UPGRADE_v6.md
+++ b/frontend-modern/public/docs/UPGRADE_v6.md
@@@ -1,11 -1,16 +1,16 @@@
# Upgrade to Pulse v6
- This guide covers practical upgrade steps for existing Pulse installs moving to v6.
+ This guide covers the v5-to-v6 migration and its compatibility changes. For
+ updates within v6, use [Updating Pulse](AUTO_UPDATE.md); the historical packets
+ below are context, not an instruction to install an old version or downgrade.
+ Choose an actually [published software release](https://github.com/rcourtman/Pulse/releases),
+ not a draft, source tag or Helm-chart asset. Keep the existing runtime edition:
+ public Community downloads do not replace a private Pro build.
-For the current v6 beta packet, see:
+For the current v6 release candidate packet, see:
-- `docs/releases/RELEASE_NOTES_v6.6.0-beta.3.md`
-- `docs/releases/V6_CHANGELOG_v6.6.0-beta.3.md`
+- `docs/releases/RELEASE_NOTES_v6.6.0-rc.4.md`
+- `docs/releases/V6_CHANGELOG_v6.6.0-rc.4.md`
For the current stable v6 packet, see:
diff --cc frontend-modern/src/components/Infrastructure/__tests__/resourceDetailDrawerMetricsHistoryModel.branchcov0712.test.ts
index d9a15c7c44,d90fe3717d..119efcf262
--- a/frontend-modern/src/components/Infrastructure/__tests__/resourceDetailDrawerMetricsHistoryModel.branchcov0712.test.ts
+++ b/frontend-modern/src/components/Infrastructure/__tests__/resourceDetailDrawerMetricsHistoryModel.branchcov0712.test.ts
@@@ -67,18 -67,6 +67,48 @@@ describe('canonical drawer History hono
},
);
+ it('qualifies a retained Proxmox listing reading by the guest kind', () => {
+ const listing = (type: 'vm' | 'system-container') => {
+ const value = { ...resource('last-known'), type };
+ value.memory!.observation!.source = 'cluster-resources';
+ return getResourceMetricsHistoryDeferredMetrics(value).memory.message;
+ };
+ // A QEMU listing counts the guest's cache; a container's has it removed.
+ expect(listing('vm')).toContain('cached memory the VM can reuse');
+ expect(listing('system-container')).not.toContain('cached memory');
+ expect(listing('system-container')).toContain('Source: Proxmox.');
+ });
+
++ it.each(['vm', 'system-container'] as const)(
++ 'keeps %s agent freshness independent of provider memory interpretation',
++ (type) => {
++ const value = { ...resource('current', 0), type, agent: { stale: false } };
++ value.memory!.observation!.source = 'agent';
++ expect(getResourceMetricsHistoryCurrentMetrics(value).memory).toBe(0);
++ value.agent.stale = true;
++ expect(getResourceMetricsHistoryCurrentMetrics(value).memory).toBeUndefined();
++ expect(getResourceMetricsHistoryDeferredMetrics(value).memory).toMatchObject({
++ lastKnownValue: 0,
++ message: expect.stringContaining('stopped reporting'),
++ });
++ expect(getResourceMetricsHistoryDeferredMetrics(value).memory.message).toContain(
++ '2026-10-04 12:00:00 UTC',
++ );
++ expect(getResourceMetricsHistoryDeferredMetrics(value).memory.message).toContain(
++ 'Last known.',
++ );
++
++ // A stale supplemental agent cannot invalidate Proxmox's independent
++ // reading or change the VM/LXC distinction of a listing measurement.
++ value.memory!.observation!.source = 'cluster-resources';
++ expect(getResourceMetricsHistoryCurrentMetrics(value).memory).toBe(0);
++ value.memory!.observation!.state = 'last-known';
++ const message = getResourceMetricsHistoryDeferredMetrics(value).memory.message;
++ expect(message).not.toContain('stopped reporting');
++ expect(message.includes('cached memory the VM can reuse')).toBe(type === 'vm');
++ },
++ );
++
it.each([-1, 101, NaN, Infinity])('withdraws invalid retained memory %s', (value) => {
expect(
getResourceMetricsHistoryDeferredMetrics(resource('last-known', value)).memory.lastKnownValue,
diff --cc frontend-modern/src/components/Infrastructure/resourceDetailDrawerMetricsHistoryModel.ts
index 775ac3940b,719e5cfbb4..a7f9db1ac9
--- a/frontend-modern/src/components/Infrastructure/resourceDetailDrawerMetricsHistoryModel.ts
+++ b/frontend-modern/src/components/Infrastructure/resourceDetailDrawerMetricsHistoryModel.ts
@@@ -173,14 -171,7 +173,15 @@@ const memoryReading = (resource: Resour
resource.memory ? finiteMetric(getMemoryPercent(resource)) : undefined,
resource.memory?.observation,
isProxmoxGuest(resource),
- { agentStale: resource.agent?.stale === true },
+ {
++ agentStale: resource.agent?.stale === true,
+ guestKind:
+ resource.type === 'vm'
+ ? 'vm'
+ : resource.type === 'system-container'
+ ? 'container'
+ : undefined,
+ },
);
const guestDiskReading = (resource: Resource) => {
diff --cc frontend-modern/src/features/docs/__tests__/troubleshootingEntryGuide.test.ts
index 5eadaa9493,7d5d321694..7f6edd4c21
--- a/frontend-modern/src/features/docs/__tests__/troubleshootingEntryGuide.test.ts
+++ b/frontend-modern/src/features/docs/__tests__/troubleshootingEntryGuide.test.ts
@@@ -11,7 -11,7 +11,7 @@@ const renderedGuides = new Map<string,
function renderGuide(name: string): HTMLElement {
// Several destinations share a long guide. Render each immutable document
- // once, rather than making sixteen navigation checks parse it repeatedly.
- // once, rather than making seventeen navigation checks parse it repeatedly.
++ // once, rather than making navigation checks parse it repeatedly.
// entry() clones its section, so the cached article is never modified.
const cached = renderedGuides.get(name);
if (cached) return cached;
@@@ -52,64 -44,7 +52,88 @@@ describe('troubleshooting workflow entr
const section = entry();
const link = section.querySelector('a');
expect(link?.getAttribute('href')).toBe('/docs/VM_DISK_MONITORING#backup-safety');
- expect(section.querySelectorAll('li')).toHaveLength(12);
+ // New, independently verified workflows must not break unrelated safety
+ // checks. Require every existing job once, without relying on list offsets.
+ const requiredWorkflows: [string, string[]][] = [
++ [
++ 'Not sure which version is affected:',
++ ['/docs/TROUBLESHOOTING#find-the-running-server-and-agent-versions'],
++ ],
++ [
++ 'Proxmox or PBS reports a certificate error:',
++ ['/docs/TROUBLESHOOTING#proxmox-or-pbs-certificate-validation-failed'],
++ ],
++ [
++ 'Copied agent install or removal command failed:',
++ ['/docs/UNIFIED_AGENT#copied-agent-command-fails-before-installation'],
++ ],
++ [
++ 'PBS is Active but missing from monitoring:',
++ ['/docs/PBS#pbs-is-active-but-missing-from-monitoring'],
++ ],
++ [
++ 'RAID is reported degraded despite a spare:',
++ ['/docs/TROUBLESHOOTING#raid-array-is-degraded-with-a-spare'],
++ ],
++ [
++ 'Unraid reports missing disks in empty slots:',
++ ['/docs/TROUBLESHOOTING#unraid-reports-missing-disks-in-empty-slots'],
++ ],
+ [
+ 'Cannot sign in:',
+ [
+ '/docs/TROUBLESHOOTING#i-forgot-my-password',
+ '/docs/TROUBLESHOOTING#cannot-login--401-unauthorized',
+ ],
+ ],
+ [
+ 'Update failed or its result is uncertain:',
+ [
+ '/docs/AUTO_UPDATE#update-failed',
+ '/docs/DOCKER#check-a-failed-or-pending-workload-update',
+ ],
+ ],
+ [
+ 'Agent host is missing or reporting is rejected:',
+ ['/docs/UNIFIED_AGENT#agent-lookup-returns-http-401-or-403'],
+ ],
+ [
+ 'Guest readings are missing or disagree:',
+ [
+ '/docs/VM_DISK_MONITORING#current-retained-and-unavailable-disk-readings',
+ '/docs/TROUBLESHOOTING#guest-memory-is-high-but-available-memory-is-plentiful',
+ ],
+ ],
+ [
+ 'Proxmox data belongs to the wrong host, or healthy jobs look Pending:',
+ [
+ '/docs/TROUBLESHOOTING#monitoring-is-mixed-between-proxmox-installations',
+ '/docs/TROUBLESHOOTING#replication-jobs-are-pending-stale-or-missing',
+ ],
+ ],
+ ['TrueNAS readings are stale or absent:', ['/docs/TRUENAS#stale-truenas-data']],
+ [
+ 'Pulse itself uses more memory, CPU or disk writes:',
+ [
+ '/docs/TROUBLESHOOTING#memory-use-keeps-growing',
+ '/docs/TROUBLESHOOTING#excessive-cpu-writes-or-database-growth',
+ ],
+ ],
+ [
+ 'Notification Test works but alerts do not arrive:',
+ ['/docs/TROUBLESHOOTING#test-succeeds-but-real-alerts-are-missing'],
+ ],
+ [
+ 'Private information was posted:',
+ ['/docs/TROUBLESHOOTING#if-you-already-posted-private-information'],
+ ],
+ ];
+ for (const [label, expectedLinks] of requiredWorkflows) {
+ const actual = [...workflow(section, label).querySelectorAll('a[href]')].map((link) =>
+ link.getAttribute('href'),
+ );
+ expect(actual, label).toEqual(expect.arrayContaining(expectedLinks));
+ }
expect(section.querySelector('pre')).toBeNull();
const beforeList = [...section.children]
.slice(
@@@ -127,7 -62,24 +151,17 @@@
it('keeps server updates, workload updates and reporting acceptance distinct', () => {
const section = entry();
- const row = (label: string) => {
- const rows = [...section.querySelectorAll('li')].filter((element) =>
- prose(element.querySelector('strong') as HTMLElement).startsWith(label),
- );
- expect(rows, label).toHaveLength(1);
- return rows[0] as HTMLElement;
- };
- const update = row('Update failed or its result is uncertain');
- const install = row('Agent host is missing or reporting is rejected');
- const command = row('Copied agent install or removal command failed');
- const reporting = row('Guest readings are missing or disagree');
- const memory = row('Pulse itself uses more memory, CPU or disk writes');
- const raid = row('RAID is reported degraded despite a spare');
+ const update = workflow(section, 'Update failed or its result is uncertain:');
++ const install = workflow(section, 'Agent host is missing or reporting is rejected:');
++ const command = workflow(section, 'Copied agent install or removal command failed:');
++ const reporting = workflow(section, 'Guest readings are missing or disagree:');
++ const memory = workflow(section, 'Pulse itself uses more memory, CPU or disk writes:');
++ const raid = workflow(section, 'RAID is reported degraded despite a spare:');
+ expect(raid.querySelector('a')?.getAttribute('href')).toBe(
+ '/docs/TROUBLESHOOTING#raid-array-is-degraded-with-a-spare',
+ );
+ expect(prose(raid)).toContain('A spare-inclusive total is not the required member count');
+ expect(prose(raid)).toContain('do not change the array just to clear a Pulse alert');
expect(prose(update)).toContain('An update of a monitored container is a different job');
const updateLinks = [...update.querySelectorAll('a')].map((link) => link.getAttribute('href'));
expect(updateLinks).toEqual([
diff --cc frontend-modern/src/utils/__tests__/agentInstallCommand.test.ts
index ba725b0946,c0c7be1b52..77565c4616
--- a/frontend-modern/src/utils/__tests__/agentInstallCommand.test.ts
+++ b/frontend-modern/src/utils/__tests__/agentInstallCommand.test.ts
@@@ -389,13 -405,10 +405,14 @@@ history = p.get("history"
env = os.environ.copy()
if history:
env.update(HISTFILE=history, HISTSIZE="100", HISTFILESIZE="100", HISTCONTROL="", PS1="PULSE_TEST$ ")
+# Phases share a fixture directory. A marker left by an earlier run would let
+# the shell's echo of the command line pass for the real prompt.
+try: os.remove(env["READER_PID"])
+except FileNotFoundError: pass
program = ["bash", "--noprofile", "--norc", "-i"] if history else ["sh", "-c", p["command"]]
child = subprocess.Popen(program, stdin=slave, stdout=slave, stderr=slave, preexec_fn=session, env=env)
- if history: os.write(master, (p["command"] + "\\n").encode())
+ command_sent = not bool(history)
+ exit_sent = False
os.close(slave)
output = b""
sent = False
diff --cc internal/api/contract_test.go
index 001ff0ed25,13341e4ac3..25fd6c0606
--- a/internal/api/contract_test.go
+++ b/internal/api/contract_test.go
@@@ -27275,150 -27263,41 +27304,189 @@@ func TestContract_DiagnosticsTLSGuidanc
}
}
+func TestChartServiceInitializationIsConcurrentAndRouterScoped(t *testing.T) {
+ const requests = 64
+ router := &Router{}
+ start := make(chan struct{})
+ services := make(chan *chartapi.Service, requests)
+ for range requests {
+ go func() {
+ <-start
+ services <- router.ensureChartService()
+ }()
+ }
+ close(start)
+
+ var shared *chartapi.Service
+ for range requests {
+ service := <-services
+ if service == nil {
+ t.Fatal("chart service was not initialized")
+ }
+ if shared == nil {
+ shared = service
+ } else if service != shared {
+ t.Fatal("concurrent requests received different chart services")
+ }
+ }
+ if router.ensureChartService() != shared {
+ t.Fatal("later requests replaced the chart service")
+ }
+ if (&Router{}).ensureChartService() == shared {
+ t.Fatal("different routers shared a chart service")
+ }
+}
+
+// Login, status and middleware reads of the SSO configuration run on every
+// request goroutine while an administrator edits providers, and the legacy
+// OIDC_* provider is re-applied on every read. Run under -race.
+func TestSSOConfigConcurrentReadsAndAdminWrites(t *testing.T) {
+ t.Setenv("OIDC_ENABLED", "true")
+ t.Setenv("OIDC_ISSUER_URL", "https://issuer.example.com")
+ t.Setenv("OIDC_CLIENT_ID", "pulse-env")
+
+ router, _ := setupTestRouter(t)
+ router.ensureSSOConfig()
+
+ const (
+ readers = 4
+ writers = 4
+ rounds = 6
+ )
+
+ stop := make(chan struct{})
+ var readersWG sync.WaitGroup
+ for range readers {
+ readersWG.Add(1)
+ go func() {
+ defer readersWG.Done()
+ for {
+ select {
+ case <-stop:
+ return
+ default:
+ }
+ // /api/security/status and the CSRF middleware.
+ _ = router.ensureSSOConfig().GetEnabledProviders()
+ // OIDC and SAML login handlers.
+ if provider := router.getSSOProvider(config.LegacyOIDCProviderID); provider == nil || provider.OIDC == nil || provider.OIDC.ClientID != "pulse-env" {
+ t.Error("environment OIDC provider missing during concurrent edits")
+ return
+ }
+ w := httptest.NewRecorder()
+ router.handleListSSOProviders(w, httptest.NewRequest(http.MethodGet, "/api/security/sso/providers", nil))
+ if w.Code != http.StatusOK {
+ t.Errorf("list providers: status %d", w.Code)
+ return
+ }
+ }
+ }()
+ }
+
+ create := func(id string) int {
+ body, _ := json.Marshal(config.SSOProvider{
+ ID: id,
+ Name: "Provider " + id,
+ Type: config.SSOProviderTypeOIDC,
+ OIDC: &config.OIDCProviderConfig{
+ IssuerURL: "https://issuer.example.com",
+ ClientID: id,
+ },
+ })
+ w := httptest.NewRecorder()
+ router.handleCreateSSOProvider(w, httptest.NewRequest(http.MethodPost, "/api/security/sso/providers", bytes.NewReader(body)))
+ return w.Code
+ }
+ remove := func(id string) int {
+ w := httptest.NewRecorder()
+ router.handleSSOProvider(w, httptest.NewRequest(http.MethodDelete, "/api/security/sso/providers/"+id, nil))
+ return w.Code
+ }
+
+ var writersWG sync.WaitGroup
+ for i := range writers {
+ writersWG.Add(1)
+ go func() {
+ defer writersWG.Done()
+ for j := range rounds {
+ id := fmt.Sprintf("race-%d-%d", i, j)
+ if code := create(id); code != http.StatusCreated {
+ t.Errorf("create %s: status %d", id, code)
+ return
+ }
+ if code := remove(id); code != http.StatusNoContent {
+ t.Errorf("delete %s: status %d", id, code)
+ return
+ }
+ }
+ if id := fmt.Sprintf("keep-%d", i); create(id) != http.StatusCreated {
+ t.Errorf("create %s failed", id)
+ }
+ }()
+ }
+ writersWG.Wait()
+ close(stop)
+ readersWG.Wait()
+
+ providerIDs := func(cfg *config.SSOConfig) []string {
+ ids := make([]string, 0, len(cfg.Providers))
+ for _, provider := range cfg.Providers {
+ ids = append(ids, provider.ID)
+ }
+ sort.Strings(ids)
+ return ids
+ }
+
+ want := []string{"keep-0", "keep-1", "keep-2", "keep-3", config.LegacyOIDCProviderID}
+ sort.Strings(want)
+ if got := providerIDs(router.ensureSSOConfig()); fmt.Sprint(got) != fmt.Sprint(want) {
+ t.Fatalf("published providers = %v, want %v", got, want)
+ }
+
+ persisted, err := router.persistence.LoadSSOConfig()
+ if err != nil {
+ t.Fatalf("load persisted SSO config: %v", err)
+ }
+ if got := providerIDs(persisted); fmt.Sprint(got) != fmt.Sprint(want[:writers]) {
+ t.Fatalf("persisted providers = %v, want %v", got, want[:writers])
+ }
+}
++
+ func TestContract_PortablePrivateBootstrapPayloads(t *testing.T) {
+ const token = "synthetic-portable-install-token"
+ artifact := configapi.BuildSetupScriptInstallArtifact("https://pulse.example", "pbs", "https://pbs.example:8007", "https://pulse.example", false, token, 1900000000)
+ payload := map[string]any{
+ "setup": artifact,
+ "proxmox": AgentInstallCommandResponse{Command: configapi.BuildProxmoxAgentInstallCommand(configapi.AgentInstallCommandOptions{BaseURL: "https://pulse.example", Token: token, InstallType: "pbs", IncludeInstallType: true}), Token: token},
+ "container": map[string]string{"command": buildContainerRuntimeAgentInstallCommand("https://pulse.example", token, false), "token": token},
+ }
+ encoded, err := json.Marshal(payload)
+ if err != nil {
+ t.Fatal(err)
+ }
+ var wire map[string]map[string]any
+ if err := json.Unmarshal(encoded, &wire); err != nil {
+ t.Fatal(err)
+ }
+ for kind, fields := range wire {
+ command, ok := fields["command"].(string)
+ if !ok || strings.Contains(command, token) || strings.ContainsAny(command, "\r\n") || strings.Count(command, "bash -c ") != 1 {
+ t.Fatalf("%s wire command lost its single private body", kind)
+ }
+ for _, required := range []string{`"$@" bash -c`, "tty_control=(stty)", "tty_control=(busybox stty)", "Cannot restore terminal settings.", ` -o "$install_script"`} {
+ if !strings.Contains(command, required) {
+ t.Fatalf("%s wire command lost %s", kind, required)
+ }
+ }
+ credentialKey := "token"
+ if kind == "setup" {
+ credentialKey = "setupToken"
+ if fields["commandWithEnv"] != command || fields["commandWithoutEnv"] != command || fields["downloadURL"] != fields["url"] {
+ t.Fatal("setup artifact aliases or credential-free download changed")
+ }
+ }
+ if fields[credentialKey] != token {
+ t.Fatalf("%s lost its separate credential field", kind)
+ }
+ }
+ }
diff --cc internal/monitoring/guest_agent_transport_integration_test.go
index 6ae60c8084,8a459a7c72..c4483ad7f5
--- a/internal/monitoring/guest_agent_transport_integration_test.go
+++ b/internal/monitoring/guest_agent_transport_integration_test.go
@@@ -90,7 -109,13 +109,12 @@@ func testGuestAgentTransportDeferralKee
}
}))
defer server.Close()
- client, err := proxmox.NewClient(proxmox.ClientConfig{Host: server.URL, TokenName: "fixture@pve!pulse", TokenValue: "fixture", Timeout: guestPollFixtureHangGuard})
+ host := server.URL
- timeout := time.Second
++ timeout := guestPollFixtureHangGuard
+ if strings.HasPrefix(kind, "foreign ") {
- timeout = 15 * time.Second
+ host += "/command-reply-fixture/" + strings.ReplaceAll(kind, " ", "-")
+ }
+ client, err := proxmox.NewClient(proxmox.ClientConfig{Host: host, TokenName: "fixture@pve!pulse", TokenValue: "fixture", Timeout: timeout})
if err != nil {
t.Fatal(err)
}
@@@ -106,8 -132,31 +133,31 @@@
t.Fatal("VM disappeared")
}
m.recordGuestSnapshot("transport", "qemu", "node", 105, GuestMemorySnapshot{Name: vm.Name, Status: vm.Status, RetrievedAt: at, MemorySource: source, Memory: vm.Memory, Raw: raw, Notes: notes})
- m.recordGuestMetrics([]models.VM{vm}, nil, time.Now().Add(-time.Second))
+ m.recordGuestMetrics([]models.VM{vm}, nil, cycleStart)
previous = &vm
+ if preserveOrigins {
+ // The safe fixture's actual builder/broadcast output is also
+ // available for consuming browser proof. Do not hand-author a
+ // fresh/last-known carrier or replace its observation time.
+ m.state.UpdateVMs([]models.VM{vm})
+ front := m.buildBroadcastFrontendStateFromSnapshot(m.state.GetSnapshot(), m.mockModeFence.begin())
+ if len(front.Resources) != 1 {
+ t.Fatal("typed-reply producer lost the canonical guest")
+ }
+ wire, err := json.Marshal(map[string]any{
+ "kind": kind, "phase": phase.Load(), "guestCommands": calls.Load(),
+ "vm": vm, "resources": front.Resources,
+ "history": m.metricsHistory.GetAllGuestMetrics(identity, time.Hour),
+ })
+ if err != nil {
+ t.Fatal(err)
+ }
+ marker := "TYPED_REPLY_PRODUCER"
+ if strings.HasPrefix(kind, "foreign ") {
+ marker = "COMMAND_REPLY_PRODUCER"
+ }
+ t.Logf("%s %s", marker, wire)
+ }
return vm, source
}
initial, source := build()
diff --cc scripts/installtests/agent_upstart_teardown.py
index f8a15970dd,0d42c09a9c..1025de3661
--- a/scripts/installtests/agent_upstart_teardown.py
+++ b/scripts/installtests/agent_upstart_teardown.py
@@@ -123,11 -150,14 +150,17 @@@ class AgentUpstartTeardown(unittest.Tes
tools = root / "tools"; tools.mkdir()
for name in ("ps", "initctl"):
p = tools / name; p.write_text(MODEL); p.chmod(0o700)
+ link = tools / "fixture-link.py"; link.write_text(LINK_MODEL)
+ # Native macOS ln lacks -T. No host ln implementation may determine
+ # whether these closed DSM semantics or restoration controls pass.
+ external_link = tools / "ln"
+ external_link.write_text("#!/bin/sh\nprintf 'unexpected external ln\\n' >&2\nexit 96\n")
+ external_link.chmod(0o700)
if fault == "missing-initctl": (tools / "initctl").unlink()
- init = root / "init"; init.mkdir()
+ # The installer refuses a group- or world-writable job directory.
+ # Give it an explicit mode so a caller umask of 002 cannot make
+ # the trusted fixture look untrusted.
+ init = root / "init"; init.mkdir(mode=0o755)
config = init / "pulse-agent.conf"
unrelated = root / "unrelated"; unrelated.write_text("retain unrelated recovery\n")
if conf == "linked": config.symlink_to(unrelated)
diff --cc scripts/installtests/native_installer_fixture_tools_test.go
index 0161a758e1,a1dd4ebf48..14990a7c48
--- a/scripts/installtests/native_installer_fixture_tools_test.go
+++ b/scripts/installtests/native_installer_fixture_tools_test.go
@@@ -50,9 -159,8 +159,8 @@@ func TestNativeInstallerPhasesKeepNASBe
}
func TestNativeInstallerFixtureToolsAreRealAndStepLocal(t *testing.T) {
- supply, installer := nativeInstallerFixtureSteps(t)
- for _, scenario := range []string{"real tools", "install failure", "missing bash", "legacy bash", "missing stat", "non-GNU chmod", "missing mv", "non-GNU mv", "missing ln", "non-GNU ln", "missing tar", "non-GNU tar"} {
+ supply, preflight, installer := nativeInstallerFixtureSteps(t)
- for _, scenario := range []string{"real tools", "install failure", "missing bash", "legacy bash", "missing stat", "non-GNU chmod", "missing mv", "non-GNU mv", "missing tar", "non-GNU tar", "missing awk", "non-GNU awk"} {
++ for _, scenario := range []string{"real tools", "install failure", "missing bash", "legacy bash", "missing stat", "non-GNU chmod", "missing mv", "non-GNU mv", "missing ln", "non-GNU ln", "missing tar", "non-GNU tar", "missing awk", "non-GNU awk"} {
t.Run(scenario, func(t *testing.T) {
root := t.TempDir()
bin, prefix, tarPrefix, runnerTemp := filepath.Join(root, "bin"), filepath.Join(root, "brew with spaces"), filepath.Join(root, "gnu tar with spaces"), filepath.Join(root, "runner temp")
@@@ -73,7 -182,7 +182,7 @@@ esa
if err := os.WriteFile(filepath.Join(bin, "brew"), []byte(brew), 0700); err != nil {
t.Fatal(err)
}
- tools := []string{"bash", "stat", "chmod", "chown", "mv", "ln", "tar"}
- tools := []string{"bash", "stat", "chmod", "chown", "mv", "tar", "awk"}
++ tools := []string{"bash", "stat", "chmod", "chown", "mv", "ln", "tar", "awk"}
installed := make(map[string]string)
for _, name := range tools {
leaf, toolPrefix := name, prefix
@@@ -156,7 -269,7 +269,7 @@@
}
entries, err := os.ReadDir(fixtureBin)
if err != nil || len(entries) != len(tools) {
- t.Fatal("fixture supply must select only Bash/stat/chmod/chown/mv/ln/tar")
- t.Fatal("fixture supply must select only Bash/stat/chmod/chown/mv/tar/awk")
++ t.Fatal("fixture supply must select only Bash/stat/chmod/chown/mv/ln/tar/awk")
}
for _, name := range tools {
if target, err := os.Readlink(filepath.Join(fixtureBin, name)); err != nil || target != installed[name] {
@@@ -170,9 -283,9 +283,9 @@@
// only Python's containing suite to avoid recursively starting Go.
probe := `#!/bin/sh
set -eu
- test "$*" = 'scripts/installtests/run_unix_contracts.py' || exit 91
+ test "$*" = "scripts/installtests/run_unix_contracts.py --phase $EXPECTED_PHASE" || exit 91
test "$(command -v bash)" = "$PULSE_INSTALLER_FIXTURE_BIN/bash" || exit 92
- for tool in stat chmod chown mv ln tar; do
-for tool in stat chmod chown mv tar awk; do
++for tool in stat chmod chown mv ln tar awk; do
test "$(command -v "$tool")" = "$PULSE_INSTALLER_FIXTURE_BIN/$tool" || exit 93
done
# Exercise the actual Linux-only flags with real executables and file bytes.
@@@ -186,37 -299,34 +299,50 @@@ cp "$PROBE_ROOT/extracted/source" "$PRO
mv -T "$PROBE_ROOT/staged" "$PROBE_ROOT/published"
cmp "$PROBE_ROOT/source" "$PROBE_ROOT/published"
test ! -e "$PROBE_ROOT/staged"
+# DSM Upstart removal links its manual override and restores retired config
+# with GNU ln -T: a hard link that never replaces or descends into a target.
+links="$PROBE_ROOT/links"
+rm -rf "$links"
+mkdir "$links" "$links/occupied-directory"
+ln -T -- "$PROBE_ROOT/published" "$links/linked"
+cmp "$PROBE_ROOT/source" "$links/linked"
+printf 'concurrent\n' > "$links/occupied"
+if ln -T -- "$PROBE_ROOT/published" "$links/occupied" 2>/dev/null; then exit 94; fi
+test "$(cat "$links/occupied")" = concurrent
+if ln -T -- "$PROBE_ROOT/published" "$links/occupied-directory" 2>/dev/null; then exit 95; fi
+test ! -e "$links/occupied-directory/published"
+ # Exercise the NAS filter's actual octal byte range, not a --version double.
+ # Safe text must pass; NUL and other disallowed controls must fail. Neither
+ # false-positive empty matching nor dropping binary bytes is acceptable.
+ printf 'safe boot text\n' > "$PROBE_ROOT/boot-safe"
+ printf 'unsafe\000boot\n' > "$PROBE_ROOT/boot-nul"
+ printf 'unsafe\001boot\n' > "$PROBE_ROOT/boot-control"
+ awk 'BEGIN { bad = 0 } /[\000-\010\013-\037\177]/ { bad = 1 } END { exit bad }' "$PROBE_ROOT/boot-safe"
+ for file in boot-nul boot-control; do
+ if awk 'BEGIN { bad = 0 } /[\000-\010\013-\037\177]/ { bad = 1 } END { exit bad }' "$PROBE_ROOT/$file"; then
+ exit 94
+ fi
+ done
exit "${SUITE_EXIT:-0}"
`
if err := os.WriteFile(filepath.Join(bin, "python3"), []byte(probe), 0700); err != nil {
t.Fatal(err)
}
- // The supply step itself links with the native ln; only the installer
- // step must resolve the supplied GNU ln ahead of a native one.
++ // Supply uses native ln, while installer phases must use the verified GNU ln.
+ if err := os.WriteFile(filepath.Join(bin, "ln"), []byte("#!/bin/sh\nprintf 'native BSD utility reached\\n' >&2\nexit 64\n"), 0700); err != nil {
+ t.Fatal(err)
+ }
- for _, code := range []string{"0", "23"} {
- cmd := exec.CommandContext(ctx, "bash", "-c", installer.Run)
- cmd.Env = append(env, "PULSE_INSTALLER_FIXTURE_BIN="+fixtureBin, "PROBE_ROOT="+root, "SUITE_EXIT="+code)
- output, err := cmd.CombinedOutput()
- if code == "0" && err != nil {
- t.Fatalf("fixture-local path did not reach the suite: %v: %s", err, output)
- } else if code == "23" {
- if exit, ok := err.(*exec.ExitError); !ok || exit.ExitCode() != 23 {
- t.Fatalf("containing suite failure hidden: %v: %s", err, output)
+ for phase, step := range map[string]nativeGuidanceStep{"preflight": preflight, "remaining": installer} {
+ for _, code := range []string{"0", "23"} {
+ cmd := exec.CommandContext(ctx, "bash", "-c", step.Run)
+ cmd.Env = append(env, "PULSE_INSTALLER_FIXTURE_BIN="+fixtureBin, "PROBE_ROOT="+root, "SUITE_EXIT="+code, "EXPECTED_PHASE="+phase)
+ output, err := cmd.CombinedOutput()
+ if code == "0" && err != nil {
+ t.Fatalf("fixture-local path did not reach the suite: %v: %s", err, output)
+ } else if code == "23" {
+ if exit, ok := err.(*exec.ExitError); !ok || exit.ExitCode() != 23 {
+ t.Fatalf("containing suite failure hidden: %v: %s", err, output)
+ }
}
}
}