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.
Go aialertsdashboarddockerdocker-monitoringgohomelabinfrastructure-monitoringkubernetesmonitoringmspproxmoxproxmox-backup-serverproxmox-mail-gatewayproxmox-veself-hostedtruenastypescriptvspherewebhooks
Keep guest memory report evidence distinct from Pulse performance
pulse-triage[bot]3ce8f5450dad
diff --git a/.github/ISSUE_TEMPLATE/bug_report.yml b/.github/ISSUE_TEMPLATE/bug_report.ymlindex 7154f3de42..12442bf1cf 100644--- a/.github/ISSUE_TEMPLATE/bug_report.yml+++ b/.github/ISSUE_TEMPLATE/bug_report.yml@@ -138,8 +138,12 @@ body:For notification reports, distinguish an already-observed **Test** result from ordinary alert delivery: single, grouped/digest or resolved. Give the destination type (such as email or Telegram), built-in or custom template, original time and redacted error or missing host/resource context where known. A successful Test does not establish ordinary delivery or correct identity.Use existing messages and queue details only; blank or "unknown" is valid. Do not send a Test, trigger an alert, retry/replay a queued message, clear the queue or change notification settings just to complete this report. Keep destination addresses, webhook URLs, chat IDs, tokens and full notification settings or payloads private.- For CPU, memory or disk-write reports, use existing readings or safe, passive observations. Say whether the reading covers the Pulse process, its container or the whole host, and give its units, measurement window and uptime. Include allocated CPUs or memory limit, fleet size, polling interval and open dashboards where relevant and known. A process-start CPU average is not a recent sampling window; database size is not a write rate.+ For CPU, memory or disk-write reports about Pulse itself, use existing readings or safe, passive observations. Say whether the reading covers the Pulse process, its container or the whole host, and give its units, measurement window and uptime. Include allocated CPUs or memory limit, fleet size, polling interval and open dashboards where relevant and known. A process-start CPU average is not a recent sampling window; database size is not a write rate.Existing screenshots are useful; write "unavailable" when safe collection is not possible. Do not restart, create load or change polling or retention just to measure. Do not attach raw profiles, heap dumps, databases or full process command lines; summarise only relevant counters after local review.++ For VM or LXC memory disagreements, say where each existing reading appears (overview, History, alert, Patrol or another tool) and whether the comparison was measured inside the affected guest or on its hypervisor node. Give original times, units and memory source or sample age only if already known or displayed; use consistent private aliases and write "unknown" for missing details.+ Available memory, buff/cache and a process's RSS measure different things: RSS alone is not total guest usage, and a cache-inclusive hypervisor footprint does not establish guest pressure. A Proxmox VM API connection may still obtain QEMU guest-agent (QGA) readings without a Pulse agent.+ Use existing observations only. Do not run commands, Diagnostics or guest-agent probes, install an agent, change memory settings or workloads, or repeat a backup just to answer.validations:required: falsediff --git a/.github/ISSUE_TEMPLATE/v6_rc_feedback.yml b/.github/ISSUE_TEMPLATE/v6_rc_feedback.ymlindex 183feb8575..fee1d2bfb3 100644--- a/.github/ISSUE_TEMPLATE/v6_rc_feedback.yml+++ b/.github/ISSUE_TEMPLATE/v6_rc_feedback.yml@@ -180,9 +180,13 @@ body:For notification reports, distinguish an already-observed **Test** result from ordinary alert delivery: single, grouped/digest or resolved. Give the destination type (such as email or Telegram), built-in or custom template, original time and redacted error or missing host/resource context where known. A successful Test does not establish ordinary delivery or correct identity.Use existing messages and queue details only; blank or "unknown" is valid. Do not send a Test, trigger an alert, retry/replay a queued message, clear the queue or change notification settings just to complete this report. Keep destination addresses, webhook URLs, chat IDs, tokens and full notification settings or payloads private.- For CPU, memory or disk-write reports, use existing readings or safe, passive observations. Say whether the reading covers the Pulse process, its container or the whole host, and give its units, measurement window and uptime. Include allocated CPUs or memory limit, fleet size, polling interval and open dashboards where relevant and known. A process-start CPU average is not a recent sampling window; database size is not a write rate.+ For CPU, memory or disk-write reports about Pulse itself, use existing readings or safe, passive observations. Say whether the reading covers the Pulse process, its container or the whole host, and give its units, measurement window and uptime. Include allocated CPUs or memory limit, fleet size, polling interval and open dashboards where relevant and known. A process-start CPU average is not a recent sampling window; database size is not a write rate.Existing screenshots are useful; write "unavailable" when safe collection is not possible. Do not restart, create load or change polling or retention just to measure. Do not attach raw profiles, heap dumps, databases or full process command lines; summarise only relevant counters after local review.+ For VM or LXC memory disagreements, say where each existing reading appears (overview, History, alert, Patrol or another tool) and whether the comparison was measured inside the affected guest or on its hypervisor node. Give original times, units and memory source or sample age only if already known or displayed; use consistent private aliases and write "unknown" for missing details.+ Available memory, buff/cache and a process's RSS measure different things: RSS alone is not total guest usage, and a cache-inclusive hypervisor footprint does not establish guest pressure. A Proxmox VM API connection may still obtain QEMU guest-agent (QGA) readings without a Pulse agent.+ Use existing observations only. Do not run commands, Diagnostics or guest-agent probes, install an agent, change memory settings or workloads, or repeat a backup just to answer.+- type: checkboxesid: confirmationsattributes:diff --git a/.github/scripts/issue-version-triage.test.cjs b/.github/scripts/issue-version-triage.test.cjsindex dc2225fdba..ba88964d41 100644--- a/.github/scripts/issue-version-triage.test.cjs+++ b/.github/scripts/issue-version-triage.test.cjs@@ -391,8 +391,46 @@ for (const [name, field] of [assert.match(attachment, /Do not attach raw profiles, heap dumps, databases or full process command lines/);assert.match(attachment, /summarise only relevant counters after local review/);});++ test(`guest memory comparisons retain optional target and sample context: ${name}`, () => {+ const attachment = evidenceField();+ assert.match(attachment, /For CPU, memory or disk-write reports about Pulse itself/);+ const guest = attachment.split("For VM or LXC memory disagreements,")[1];+ assert.ok(guest, "guest readings must not inherit Pulse process-only guidance");+ assert.match(guest, /overview, History, alert, Patrol or another tool/);+ assert.match(guest, /inside the affected guest or on its hypervisor node/);+ assert.match(guest, /original times, units and memory source or sample age only if already known or displayed/);+ assert.match(guest, /consistent private aliases.*"unknown" for missing details/);+ assert.match(guest, /Available memory, buff\/cache and a process's RSS measure different things/);+ assert.match(guest, /RSS alone is not total guest usage/);+ assert.match(guest, /cache-inclusive hypervisor footprint does not establish guest pressure/);+ assert.match(guest, /A Proxmox VM API connection may still obtain QEMU guest-agent \(QGA\) readings without a Pulse agent/);+ assert.doesNotMatch(attachment, /required: true|render:/);+ });++ test(`guest memory context does not ask for new collection or workload changes: ${name}`, () => {+ const guest = evidenceField().split("For VM or LXC memory disagreements,")[1];+ assert.ok(guest, "a guest comparison needs its own collection boundary");+ assert.match(guest, /Use existing observations only/);+ assert.match(guest, /Do not run commands, Diagnostics or guest-agent probes, install an agent, change memory settings or workloads, or repeat a backup just to answer/);+ });}+test("guest memory triage preserves original evidence rather than inferring past provenance", () => {+ const guide = fs.readFileSync(path.resolve(__dirname, "../../docs/ISSUE_TRIAGE.md"), "utf8").replace(/\s+/g, " ");+ const guest = guide.split("For VM or LXC memory disagreements,")[1].split("For notification reports,")[0];+ assert.match(guest, /same target and episode/);+ assert.match(guest, /inside the affected guest or on its hypervisor node/);+ assert.match(guest, /original times, units and selected memory source or sample age if already known/);+ assert.match(guest, /unknown is valid, and existing reports need no refile/);+ assert.match(guest, /RSS alone is not total guest usage/);+ assert.match(guest, /A Proxmox VM API connection may still obtain QEMU guest-agent \(QGA\) readings without a Pulse agent/);+ assert.match(guest, /not measured zero or verified recovery/);+ assert.match(guest, /live memory source cannot establish the source of a historical sample/);+ assert.match(guest, /Reconcile the full thread before asking only for a consequential remaining distinction/);+ assert.match(guest, /Do not request commands, Diagnostics or guest-agent probes, an agent install, memory or workload changes, or another backup/);+});+test("performance report triage retains measurement boundaries without demanding unsafe evidence", () => {const guide = fs.readFileSync(path.resolve(__dirname, "../../docs/ISSUE_TRIAGE.md"), "utf8");assert.match(guide, /Pulse process, container or whole host/);diff --git a/docs/ISSUE_TRIAGE.md b/docs/ISSUE_TRIAGE.mdindex cae83c2586..624d089eff 100644--- a/docs/ISSUE_TRIAGE.md+++ b/docs/ISSUE_TRIAGE.md@@ -114,6 +114,25 @@ profiles, heap dumps, databases or full process command lines in a publicthread; use locally reviewed counter summaries, without restarting, creatingload or changing polling or retention just to measure.+For VM or LXC memory disagreements, retain where each existing reading appears+(overview, History, alert, Patrol or another tool), whether the comparison was+measured inside the affected guest or on its hypervisor node, and the original+times, units and selected memory source or sample age if already known. Use+consistent private aliases; unknown is valid, and existing reports need no+refile. Compare the same target and episode before inferring a discrepancy.+Available memory, buff/cache and a process's RSS measure different things:+RSS alone is not total guest usage, and a cache-inclusive hypervisor footprint+does not establish guest pressure. A Proxmox VM API connection may still obtain QEMU guest-agent (QGA) readings+without a Pulse agent.++Keep the displayed number, its qualification and the resulting finding separate.+An unexplained or retained reading is not measured zero or verified recovery.+Today's live memory source cannot establish the source of a historical sample;+do not reattribute History or a past warning from the current view. Reconcile+the full thread before asking only for a consequential remaining distinction.+Do not request commands, Diagnostics or guest-agent probes, an agent install,+memory or workload changes, or another backup just to fill this context.+For notification reports, distinguish an already-observed **Test** result fromordinary alert delivery: single, grouped/digest or resolved. A successful Testdoes not establish ordinary delivery or correct identity. Retain the destinationdiff --git a/frontend-modern/public/docs/ISSUE_TRIAGE.md b/frontend-modern/public/docs/ISSUE_TRIAGE.mdindex cae83c2586..624d089eff 100644--- a/frontend-modern/public/docs/ISSUE_TRIAGE.md+++ b/frontend-modern/public/docs/ISSUE_TRIAGE.md@@ -114,6 +114,25 @@ profiles, heap dumps, databases or full process command lines in a publicthread; use locally reviewed counter summaries, without restarting, creatingload or changing polling or retention just to measure.+For VM or LXC memory disagreements, retain where each existing reading appears+(overview, History, alert, Patrol or another tool), whether the comparison was+measured inside the affected guest or on its hypervisor node, and the original+times, units and selected memory source or sample age if already known. Use+consistent private aliases; unknown is valid, and existing reports need no+refile. Compare the same target and episode before inferring a discrepancy.+Available memory, buff/cache and a process's RSS measure different things:+RSS alone is not total guest usage, and a cache-inclusive hypervisor footprint+does not establish guest pressure. A Proxmox VM API connection may still obtain QEMU guest-agent (QGA) readings+without a Pulse agent.++Keep the displayed number, its qualification and the resulting finding separate.+An unexplained or retained reading is not measured zero or verified recovery.+Today's live memory source cannot establish the source of a historical sample;+do not reattribute History or a past warning from the current view. Reconcile+the full thread before asking only for a consequential remaining distinction.+Do not request commands, Diagnostics or guest-agent probes, an agent install,+memory or workload changes, or another backup just to fill this context.+For notification reports, distinguish an already-observed **Test** result fromordinary alert delivery: single, grouped/digest or resolved. A successful Testdoes not establish ordinary delivery or correct identity. Retain the destination