The only tool you need to know what is happening and how to fix it.
Go alertingexception-trackerissue-trackinglogsmetricsmonitoringobservability-platformopen-telemetryotelsession-replaytelemetrytraces
notifications: absolute deep links when APP_BASE_URL is set (#355)
FrameAutomatae441b86e9bbc
diff --git a/CLAUDE.md b/CLAUDE.mdindex 500ff152..180b37c4 100644--- a/CLAUDE.md+++ b/CLAUDE.md@@ -258,7 +258,7 @@ JWT_SECRET=<min 32 char secret for JWT signing> # the only variable with no deDB_TYPE= # sqlite | postgres. Defaults to sqlite in every build EXCEPT -tags transactional_pg, where an empty value still selects PostgreSQL. Without that tag SQLite is the only supported main DB (the migration runner applies SQLite-dialect migrations unconditionally), so the default is set once in config.defaultDBType rather than at each DBType call site.PORTS=80,8082 # comma-separated HTTP listen ports. Every port is bound up front; a port that fails to bind (typically :80 as a non-root user) logs a warning (and reports via CaptureException when monitoring is configured), and startup only fails if NO port binds. Order is not significant: a listener that later stops serving takes the process down whichever position it holds.SQLITE_PATH= # main DB path, defaults to ./traceway.db in the working directory. The telemetry DB derives from it (_telemetry.db, or _telemetry.duckdb under -tags telemetry_duckdb).-APP_BASE_URL= # public origin of this server (e.g. https://traceway.example.com). Used as the OAuth issuer / device verification URL and SSO redirect base. If unset, the device-auth + well-known endpoints derive it per-request from the Host / X-Forwarded-* headers; set it explicitly behind a proxy that doesn't forward Host.+APP_BASE_URL= # public origin of this server (e.g. https://traceway.example.com). Used as the OAuth issuer / device verification URL and SSO redirect base, and to absolutize the deep links in notifications. If unset, the device-auth + well-known endpoints derive it per-request from the Host / X-Forwarded-* headers; set it explicitly behind a proxy that doesn't forward Host.TRUSTED_PROXIES= # comma-separated IPs/CIDRs whose X-Forwarded-For/X-Real-IP gin trusts for c.ClientIP() (every per-IP rate limiter keys on it, and `/api/report` stores it on every session as `client.ip`). Unset = loopback only, so no peer can spoof its IP. Private ranges are deliberately NOT trusted by default: on a Docker/k8s network every peer has a private address, the clients included when the port is published directly, so trusting RFC1918 would let any of them forge XFF past every per-IP limiter. Set it to the CIDRs of whatever sits in front (a proxy container's compose network, the ingress pod range, the CDN ranges); otherwise every visitor shares one rate-limit bucket keyed on the proxy. The value REPLACES the default, so it must name every hop in the chain, not just the outermost (a CDN-only list leaves a private ingress untrusted and disables XFF entirely); `*` trusts every peer. A malformed entry exits 1 at startup with a FATAL message, before the database is opened (`trustedProxyNets`, cmd/run.go). The first request carrying X-Forwarded-For/X-Real-IP from a peer outside the list logs a one-time warning naming that peer (`warnOnUntrustedForwardedFor`, cmd/run.go), which is how a forgotten proxy hop shows up. Parsing lives in `Cfg.TrustedProxyList` (config.go), covered by config_test.go.TRUSTED_PROXY_HEADER= # optional header taken verbatim as the client IP on every request (gin TrustedPlatform: X-Real-IP behind ingress-nginx, CF-Connecting-IP behind Cloudflare); it wins over TRUSTED_PROXIES and turns the untrusted-forwarder warning off. Only safe when every request provably passes through a proxy that overwrites the header.UPLOAD_MAX_CONCURRENT=4 # concurrent /api/sourcemaps/upload + /api/symbols/upload requests; excess waits 30s then gets 503. Separate from INGEST_MAX_CONCURRENT so a CI upload burst cannot starve telemetry ingest. Each upload holds a whole file in memory while parsing, so this bounds peak upload memory at roughly concurrency x largest file. It is a count-based gate, not a memory budget: on a memory-capped container size it from the limit rather than trusting the default. Same gate and bounded waiting room as ingest, its own instance. The source-map warm-up that follows an upload (`services.GenerateTWArtifacts`) runs outside this gate on its own 2-worker pool with a 64-job queue: the stale `.tw` artifact is deleted and the cache invalidated inline, the rebuild is queued, identical pending jobs are coalesced, and a full queue drops the warm-up (rate-limited `CaptureException`) because lookups build a missing artifact on demand anyway.diff --git a/backend/app/config/config.go b/backend/app/config/config.goindex 2734cb51..9dfefdb1 100644--- a/backend/app/config/config.go+++ b/backend/app/config/config.go@@ -178,6 +178,29 @@ func (c *Cfg) TwilioEnabled() bool {(c.TwilioFromNumber != "" || c.TwilioMessagingServiceSID != "")}+// devBaseURL is the origin assumed when APP_BASE_URL is unset. Links that+// must be clickable regardless (emails, on-call pages) fall back to it; the+// Vite dev server is where a dashboard without a configured origin lives.+const devBaseURL = "http://localhost:5173"++// PublicBaseURL is the dashboard origin (APP_BASE_URL without surrounding+// whitespace or a trailing slash), or empty when the operator has not set one.+func (c *Cfg) PublicBaseURL() string {+ if c == nil {+ return ""+ }+ return strings.TrimRight(strings.TrimSpace(c.AppBaseURL), "/")+}++// PublicBaseURLOrDev is PublicBaseURL for callers that need some origin to+// build a link with, even on an instance that never configured one.+func (c *Cfg) PublicBaseURLOrDev() string {+ if base := c.PublicBaseURL(); base != "" {+ return base+ }+ return devBaseURL+}+func LoadFromEnv() *Cfg {return &Cfg{JWTSecret: os.Getenv("JWT_SECRET"),diff --git a/backend/app/notifications/dispatch.go b/backend/app/notifications/dispatch.goindex 09d43436..cc875def 100644--- a/backend/app/notifications/dispatch.go+++ b/backend/app/notifications/dispatch.go@@ -45,6 +45,7 @@ func dispatch(rule *models.NotificationRuleWithChannel, msg Message) bool {if rule.Severity != "" {msg.Severity = Severity(rule.Severity)}+ msg.URL = dashboardURL(msg.URL)// Escalation channels do not send anything themselves: they open (or// dedup into) an on-call page, and the escalator + outbox do alldiff --git a/backend/app/notifications/dispatch_outbox_test.go b/backend/app/notifications/dispatch_outbox_test.goindex f58306d2..37954f03 100644--- a/backend/app/notifications/dispatch_outbox_test.go+++ b/backend/app/notifications/dispatch_outbox_test.go@@ -4,9 +4,12 @@ package notificationsimport ("database/sql"+ "encoding/json"+ "strings""testing""time"+ "github.com/tracewayapp/traceway/backend/app/config""github.com/tracewayapp/traceway/backend/app/db""github.com/tracewayapp/traceway/backend/app/dbtest""github.com/tracewayapp/traceway/backend/app/models"@@ -150,3 +153,55 @@ func TestSeedCooldownsIncludesOutbox(t *testing.T) {t.Error("outbox-backed seeding should keep the rule in cooldown after restart")}}++func setAppBaseURL(t *testing.T, base string) {+ t.Helper()+ previous := config.Config.AppBaseURL+ config.Config.AppBaseURL = base+ t.Cleanup(func() { config.Config.AppBaseURL = previous })+}++func queuedMessage(t *testing.T) Message {+ t.Helper()+ rows := outboxRows(t)+ if len(rows) != 1 {+ t.Fatalf("expected 1 outbox row, got %d", len(rows))+ }+ var msg Message+ if err := json.Unmarshal(rows[0].Message, &msg); err != nil {+ t.Fatalf("decode queued message: %v", err)+ }+ return msg+}++// The link reaches an adapter two ways — the URL field, which webhook,+// telegram, pushover and sms render, and the "View details" body line, which+// is the only one Slack and GitHub ever show — so both are asserted.+func TestDispatchDashboardLinks(t *testing.T) {+ cases := []struct {+ name string+ base string+ want string+ }{+ {"absolute when APP_BASE_URL is set", "https://traceway.example.com/", "https://traceway.example.com/issues/0123456789abcdef"},+ {"relative when it is not", "", "/issues/0123456789abcdef"},+ }+ for _, tc := range cases {+ t.Run(tc.name, func(t *testing.T) {+ fixture := setupDispatchDB(t)+ setAppBaseURL(t, tc.base)++ details := ExceptionDetails{Hash: "0123456789abcdef", ErrorType: "RuntimeError"}+ if !dispatch(fixture.Rule, buildNewErrorMessage(details, "api")) {+ t.Fatal("dispatch failed")+ }+ msg := queuedMessage(t)+ if msg.URL != tc.want {+ t.Errorf("URL = %q, want %q", msg.URL, tc.want)+ }+ if !strings.Contains(msg.Body, "View details: "+tc.want) {+ t.Errorf("body lacks %q:\n%s", tc.want, msg.Body)+ }+ })+ }+}diff --git a/backend/app/notifications/messages.go b/backend/app/notifications/messages.goindex 68b84976..e703398a 100644--- a/backend/app/notifications/messages.go+++ b/backend/app/notifications/messages.go@@ -6,9 +6,22 @@ import ("strings""time"+ "github.com/tracewayapp/traceway/backend/app/config""github.com/tracewayapp/traceway/backend/app/models")+// dashboardURL makes a dashboard path absolute when APP_BASE_URL is set.+// Without it the path stays relative: a guessed localhost origin would be+// wrong for everyone reading a Slack post or a GitHub issue, whereas the bare+// path still identifies the page.+func dashboardURL(path string) string {+ base := config.Config.PublicBaseURL()+ if base == "" || !strings.HasPrefix(path, "/") {+ return path+ }+ return base + path+}+func alertEmail(alert models.EmailAlert) *models.NotificationEmail {return &models.NotificationEmail{Template: models.EmailTemplateAlert, Alert: &alert}}@@ -476,7 +489,7 @@ func buildExceptionBody(headline string, d ExceptionDetails) string {}if d.Hash != "" {- fmt.Fprintf(&b, "\n\nView details: /issues/%s", d.Hash)+ fmt.Fprintf(&b, "\n\nView details: %s", dashboardURL("/issues/"+d.Hash))}return b.String()diff --git a/backend/app/oncall/escalator.go b/backend/app/oncall/escalator.goindex 068c9050..f226e9c7 100644--- a/backend/app/oncall/escalator.go+++ b/backend/app/oncall/escalator.go@@ -578,14 +578,8 @@ func newAckToken() string {}// appBaseURL is the dashboard origin for outgoing links. On-call deployments-// should set APP_BASE_URL; the localhost fallback mirrors the email service.-func appBaseURL() string {- base := config.Config.AppBaseURL- if base == "" {- base = "http://localhost:5173"- }- return strings.TrimRight(base, "/")-}+// should set APP_BASE_URL; without it an ack link still has to be clickable.+func appBaseURL() string { return config.Config.PublicBaseURLOrDev() }// ackURLFor builds the public no-login acknowledge link for a delivery token.func ackURLFor(token string) string {diff --git a/backend/app/services/email.service.go b/backend/app/services/email.service.goindex da9a4c30..a91d191f 100644--- a/backend/app/services/email.service.go+++ b/backend/app/services/email.service.go@@ -67,12 +67,7 @@ func (e *emailService) BaseURL() string { return emailBaseURL() }func emailEnabled() bool { return config.Config != nil && config.Config.SMTPEnabled == "true" }-func emailBaseURL() string {- if config.Config != nil && config.Config.AppBaseURL != "" {- return strings.TrimRight(config.Config.AppBaseURL, "/")- }- return "http://localhost:5173"-}+func emailBaseURL() string { return config.Config.PublicBaseURLOrDev() }func RenderEmail(email Email) (string, error) {if email.LogoURL == "" {diff --git a/cli/pkg/mcpserver/knowledge/notifications.md b/cli/pkg/mcpserver/knowledge/notifications.mdindex 322bbb19..049bdc6e 100644--- a/cli/pkg/mcpserver/knowledge/notifications.md+++ b/cli/pkg/mcpserver/knowledge/notifications.md@@ -9,7 +9,7 @@ The body embeds everything for a direct, fast lookup. It contains:- `Hash: <16-hex>` — the exception group → `traceway exceptions show <hash>`.- `Exception ID: <uuid>` — the specific occurrence.- `Occurred at: 2006-01-02 15:04:05 UTC` — the occurrence timestamp. **Convert to RFC3339**: replace the space with `T` and ` UTC` with `Z` (→ `2006-01-02T15:04:05Z`).-- `View details: /issues/<hash>` — the deep link.+- `View details: <url>` — the deep link, `https://<instance>/issues/<hash>` when the server sets `APP_BASE_URL`, otherwise the bare path `/issues/<hash>`.So from a notification, go straight to the occurrence (fast), then pivot reusing the same timestamp:diff --git a/docs/pages/learn/alerts.mdx b/docs/pages/learn/alerts.mdxindex 02f0831d..c37ae5c5 100644--- a/docs/pages/learn/alerts.mdx+++ b/docs/pages/learn/alerts.mdx@@ -104,7 +104,7 @@ When a webhook channel fires, Traceway sends a JSON POST request with the follow| `severity` | `info`, `warning`, or `critical` || `ruleType` | The rule type that triggered the alert || `ruleName` | The name you gave the rule |-| `url` | Deep link to the relevant dashboard page |+| `url` | Deep link to the relevant dashboard page. Absolute when the server sets `APP_BASE_URL`, otherwise a path such as `/issues/<hash>` relative to your dashboard origin || `timestamp` | When the alert fired (ISO 8601) |If you configure an HMAC secret on the channel, each request includes an `X-Traceway-Signature` header containing `sha256=<hex>` of the request body, which you can use to verify that the request came from Traceway.diff --git a/docs/pages/server/all-in-one.mdx b/docs/pages/server/all-in-one.mdxindex b1d907a2..3a424872 100644--- a/docs/pages/server/all-in-one.mdx+++ b/docs/pages/server/all-in-one.mdx@@ -65,7 +65,7 @@ All environment variables have working defaults, no configuration required. Over| `POSTGRES_SSLMODE` | `disable` | PostgreSQL SSL mode || `JWT_SECRET` | *(built-in default, **override in production**)* | Signs all authentication tokens: dashboard sessions and CLI/MCP device logins. The built-in default is public; set a strong unique secret (min 32 chars, `openssl rand -hex 32`) and keep it stable. See [CLI Authentication](/learn/cli-auth#set-a-strong-jwt_secret). || `GIN_MODE` | `release` | Gin framework mode |-| `APP_BASE_URL` | *(none)* | Public URL of your Traceway instance (e.g. `https://traceway.example.com`). Used for SDK setup instructions, email links, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, the device-auth and `/.well-known` endpoints derive it per-request from the `Host` / `X-Forwarded-*` headers; set it explicitly behind a proxy that rewrites `Host`. See [CLI Authentication](/learn/cli-auth). |+| `APP_BASE_URL` | *(none)* | Public URL of your Traceway instance (e.g. `https://traceway.example.com`). Used for SDK setup instructions, email links, deep links in notifications, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, the device-auth and `/.well-known` endpoints derive it per-request from the `Host` / `X-Forwarded-*` headers; set it explicitly behind a proxy that rewrites `Host`. See [CLI Authentication](/learn/cli-auth). || `TRUSTED_PROXIES` | *(loopback)* | Comma-separated CIDRs of the proxies in front of Traceway whose `X-Forwarded-For` / `X-Real-IP` is trusted for the real client IP. Every per-IP rate limit keys on it, and `/api/report` stores it as each session's `client.ip`. Unset trusts loopback only, so a proxy container on a Docker network or an in-cluster ingress has to be listed before its header counts. A value replaces the default, so list every hop between the client and this server, not just the outermost one: naming only a CDN leaves a private ingress untrusted and turns `X-Forwarded-For` off entirely. `*` trusts every peer and lets clients spoof their address. Until the proxy is listed, every visitor shares one rate-limit bucket keyed on the proxy, and the server logs a warning the first time it sees a forwarding header from a peer it does not trust. || `TRUSTED_PROXY_HEADER` | *(none)* | A header taken as the client IP on every request, ahead of `TRUSTED_PROXIES`: `X-Real-IP` behind ingress-nginx, `CF-Connecting-IP` behind Cloudflare. Only safe when every request provably passes through a proxy that overwrites the header; a directly reachable backend lets clients set it freely. || `REPORT_MAX_BODY_MB` | `64` | Cap on the decompressed `/api/report` and `/api/profiles/ingest` body in megabytes. Gzip does not raise it. A body over the cap answers `413`. |diff --git a/docs/pages/server/docker-compose.mdx b/docs/pages/server/docker-compose.mdxindex 23f05a8f..22588cc9 100644--- a/docs/pages/server/docker-compose.mdx+++ b/docs/pages/server/docker-compose.mdx@@ -99,7 +99,7 @@ Edit the environment variables in your compose file as needed:| `POSTGRES_PASSWORD` | `traceway` | PostgreSQL password || `POSTGRES_SSLMODE` | `disable` | PostgreSQL SSL mode || `JWT_SECRET` | `change-this-to-a-secure-secret-at-least-32-chars` | Signs all authentication tokens (dashboard sessions and CLI/MCP device logins). Replace the placeholder with a strong, unique secret of at least 32 chars (`openssl rand -hex 32`) and keep it stable. See [CLI Authentication](/learn/cli-auth#set-a-strong-jwt_secret). |-| `APP_BASE_URL` | *(none)* | Public URL of your Traceway instance (e.g. `https://traceway.example.com`). Used for SDK setup instructions, email links, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, the device-auth and `/.well-known` endpoints derive it per-request from the `Host` / `X-Forwarded-*` headers; set it explicitly behind a proxy that rewrites `Host`. See [CLI Authentication](/learn/cli-auth). |+| `APP_BASE_URL` | *(none)* | Public URL of your Traceway instance (e.g. `https://traceway.example.com`). Used for SDK setup instructions, email links, deep links in notifications, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, the device-auth and `/.well-known` endpoints derive it per-request from the `Host` / `X-Forwarded-*` headers; set it explicitly behind a proxy that rewrites `Host`. See [CLI Authentication](/learn/cli-auth). || `TRUSTED_PROXIES` | *(loopback)* | Comma-separated CIDRs of the proxies in front of Traceway whose `X-Forwarded-For` / `X-Real-IP` is trusted for the real client IP. Every per-IP rate limit keys on it, and `/api/report` stores it as each session's `client.ip`. Unset trusts loopback only, so a proxy container on a Docker network or an in-cluster ingress has to be listed before its header counts. A value replaces the default, so list every hop between the client and this server, not just the outermost one: naming only a CDN leaves a private ingress untrusted and turns `X-Forwarded-For` off entirely. `*` trusts every peer and lets clients spoof their address. Until the proxy is listed, every visitor shares one rate-limit bucket keyed on the proxy, and the server logs a warning the first time it sees a forwarding header from a peer it does not trust. || `TRUSTED_PROXY_HEADER` | *(none)* | A header taken as the client IP on every request, ahead of `TRUSTED_PROXIES`: `X-Real-IP` behind ingress-nginx, `CF-Connecting-IP` behind Cloudflare. Only safe when every request provably passes through a proxy that overwrites the header; a directly reachable backend lets clients set it freely. || `REPORT_MAX_BODY_MB` | `64` | Cap on the decompressed `/api/report` and `/api/profiles/ingest` body in megabytes. Gzip does not raise it. A body over the cap answers `413`. |diff --git a/docs/pages/server/haloy.mdx b/docs/pages/server/haloy.mdxindex f19d464d..a27d885e 100644--- a/docs/pages/server/haloy.mdx+++ b/docs/pages/server/haloy.mdx@@ -163,7 +163,7 @@ Point your SDK's connection string at your Traceway domain:| Variable | Default | Description ||----------|---------|-------------|-| `APP_BASE_URL` | `https://traceway.yourdomain.com` | Public URL of the Traceway instance. Also the CLI/MCP OAuth issuer and device-login verification URL. See [CLI Authentication](/learn/cli-auth). |+| `APP_BASE_URL` | `https://traceway.yourdomain.com` | Public URL of the Traceway instance. Also what the deep links in notifications are built from, and the CLI/MCP OAuth issuer and device-login verification URL. See [CLI Authentication](/learn/cli-auth). || `TRUSTED_PROXIES` | *(loopback)* | Comma-separated CIDRs of the proxies in front of Traceway whose `X-Forwarded-For` / `X-Real-IP` is trusted for the real client IP. Every per-IP rate limit keys on it, and `/api/report` stores it as each session's `client.ip`. Unset trusts loopback only, so a proxy container on a Docker network or an in-cluster ingress has to be listed before its header counts. A value replaces the default, so list every hop between the client and this server, not just the outermost one: naming only a CDN leaves a private ingress untrusted and turns `X-Forwarded-For` off entirely. `*` trusts every peer and lets clients spoof their address. Until the proxy is listed, every visitor shares one rate-limit bucket keyed on the proxy, and the server logs a warning the first time it sees a forwarding header from a peer it does not trust. Haloy fronts the app with its own proxy, so list the range that proxy connects from (and your CDN's, if any). || `TRUSTED_PROXY_HEADER` | *(none)* | A header taken as the client IP on every request, ahead of `TRUSTED_PROXIES`: `X-Real-IP` behind ingress-nginx, `CF-Connecting-IP` behind Cloudflare. Only safe when every request provably passes through a proxy that overwrites the header; a directly reachable backend lets clients set it freely. || `REPORT_MAX_BODY_MB` | `64` | Cap on the decompressed `/api/report` and `/api/profiles/ingest` body in megabytes. Gzip does not raise it. A body over the cap answers `413`. |diff --git a/docs/pages/server/local-setup.mdx b/docs/pages/server/local-setup.mdxindex efd7bba6..0d8831b8 100644--- a/docs/pages/server/local-setup.mdx+++ b/docs/pages/server/local-setup.mdx@@ -73,7 +73,7 @@ POSTGRES_SSLMODE=disable| `POSTGRES_USERNAME` | PostgreSQL username || `POSTGRES_PASSWORD` | PostgreSQL password || `POSTGRES_SSLMODE` | PostgreSQL SSL mode (`disable`, `require`, etc.) |-| `APP_BASE_URL` | Public URL of your Traceway instance. Used for SDK connection instructions, email links, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. For local dev set this to `http://localhost:5173` so OAuth callbacks land on the Vite dev server (which proxies `/api` back to the Go backend). |+| `APP_BASE_URL` | Public URL of your Traceway instance. Used for SDK connection instructions, email links, deep links in notifications, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. For local dev set this to `http://localhost:5173` so OAuth callbacks land on the Vite dev server (which proxies `/api` back to the Go backend). || `GOOGLE_CLIENT_ID` | Optional. Google OAuth client ID. Setting both Google variables enables the **Continue with Google** button. See the [SSO guide](/learn/sso). || `GOOGLE_CLIENT_SECRET` | Optional. Google OAuth client secret. || `GITHUB_CLIENT_ID` | Optional. GitHub OAuth App client ID. Setting both GitHub variables enables the **Continue with GitHub** button. See the [SSO guide](/learn/sso). |diff --git a/docs/pages/server/minimal.mdx b/docs/pages/server/minimal.mdxindex 04b7472b..9059a434 100644--- a/docs/pages/server/minimal.mdx+++ b/docs/pages/server/minimal.mdx@@ -57,7 +57,7 @@ Then use `traceway:minimal` instead of `ghcr.io/tracewayapp/traceway:minimal` in| `POSTGRES_PASSWORD` | PostgreSQL password | `password` || `POSTGRES_SSLMODE` | PostgreSQL SSL mode | `disable` || `JWT_SECRET` | Signs all authentication tokens (dashboard sessions and CLI/MCP device logins). Use a strong, unique secret of at least 32 chars; keep it stable. See [CLI Authentication](/learn/cli-auth#set-a-strong-jwt_secret). | `your-secret-here` |-| `APP_BASE_URL` | Public URL of your instance. Used for SDK setup instructions, email links, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, derived per-request from `Host` / `X-Forwarded-*`. | `https://traceway.example.com` |+| `APP_BASE_URL` | Public URL of your instance. Used for SDK setup instructions, email links, deep links in notifications, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, derived per-request from `Host` / `X-Forwarded-*`. | `https://traceway.example.com` || `TRUSTED_PROXIES` | Comma-separated CIDRs of the proxies in front of Traceway whose `X-Forwarded-For` / `X-Real-IP` is trusted for the real client IP. Every per-IP rate limit keys on it, and `/api/report` stores it as each session's `client.ip`. Unset trusts loopback only, so a proxy container on a Docker network or an in-cluster ingress has to be listed before its header counts. A value replaces the default, so list every hop between the client and this server, not just the outermost one: naming only a CDN leaves a private ingress untrusted and turns `X-Forwarded-For` off entirely. `*` trusts every peer and lets clients spoof their address. Until the proxy is listed, every visitor shares one rate-limit bucket keyed on the proxy, and the server logs a warning the first time it sees a forwarding header from a peer it does not trust. | `10.0.0.0/8,192.168.0.0/16` || `TRUSTED_PROXY_HEADER` | A header taken as the client IP on every request, ahead of `TRUSTED_PROXIES`: `X-Real-IP` behind ingress-nginx, `CF-Connecting-IP` behind Cloudflare. Only safe when every request provably passes through a proxy that overwrites the header; a directly reachable backend lets clients set it freely. | `X-Real-IP` || `REPORT_MAX_BODY_MB` | Cap on the decompressed `/api/report` and `/api/profiles/ingest` body in megabytes. Gzip does not raise it. A body over the cap answers `413`. | `64` |diff --git a/docs/pages/server/sqlite.mdx b/docs/pages/server/sqlite.mdxindex 6bab04ec..b7da7d3e 100644--- a/docs/pages/server/sqlite.mdx+++ b/docs/pages/server/sqlite.mdx@@ -67,7 +67,7 @@ If `STORAGE_TYPE` is unset (or set to `local`), blobs are written to `/data/stor| Variable | Default | Description ||----------|---------|-------------|| `JWT_SECRET` | *(required)* | Signs all authentication tokens (dashboard sessions and CLI/MCP device logins). Use a strong, unique secret of at least 32 characters (`openssl rand -hex 32`) and keep it stable. See [CLI Authentication](/learn/cli-auth#set-a-strong-jwt_secret). |-| `APP_BASE_URL` | *(unset)* | Public URL of your instance. Used for SDK setup snippets, password-reset emails, OAuth callback URLs, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, the device-auth and `/.well-known` endpoints derive it per-request from the `Host` / `X-Forwarded-*` headers. See [CLI Authentication](/learn/cli-auth) |+| `APP_BASE_URL` | *(unset)* | Public URL of your instance. Used for SDK setup snippets, password-reset emails, OAuth callback URLs, deep links in notifications, and the CLI/MCP OAuth issuer plus device-login verification URL. If unset, the device-auth and `/.well-known` endpoints derive it per-request from the `Host` / `X-Forwarded-*` headers. See [CLI Authentication](/learn/cli-auth) || `TRUSTED_PROXIES` | *(loopback)* | Comma-separated CIDRs of the proxies in front of Traceway whose `X-Forwarded-For` / `X-Real-IP` is trusted for the real client IP. Every per-IP rate limit keys on it, and `/api/report` stores it as each session's `client.ip`. Unset trusts loopback only, so a proxy container on a Docker network or an in-cluster ingress has to be listed before its header counts. A value replaces the default, so list every hop between the client and this server, not just the outermost one: naming only a CDN leaves a private ingress untrusted and turns `X-Forwarded-For` off entirely. `*` trusts every peer and lets clients spoof their address. Until the proxy is listed, every visitor shares one rate-limit bucket keyed on the proxy, and the server logs a warning the first time it sees a forwarding header from a peer it does not trust. || `TRUSTED_PROXY_HEADER` | *(none)* | A header taken as the client IP on every request, ahead of `TRUSTED_PROXIES`: `X-Real-IP` behind ingress-nginx, `CF-Connecting-IP` behind Cloudflare. Only safe when every request provably passes through a proxy that overwrites the header; a directly reachable backend lets clients set it freely. || `REPORT_MAX_BODY_MB` | `64` | Cap on the decompressed `/api/report` and `/api/profiles/ingest` body in megabytes. Gzip does not raise it. A body over the cap answers `413`. |@@ -166,7 +166,7 @@ docker stop traceway && docker rm traceway|----------|-------|-------|| `JWT_SECRET` | a random string ≥ 32 chars | Required. Generate with `openssl rand -hex 32` || `PORTS` | `${{PORT}}` | Railway exposes a single port via `PORT`. The app's `PORTS` setting accepts a single value, so this binds the server where Railway expects it |-| `APP_BASE_URL` | `https://<your-service>.up.railway.app` | Optional but recommended once you generate a domain. Makes SDK snippets and reset emails point at the right URL |+| `APP_BASE_URL` | `https://<your-service>.up.railway.app` | Optional but recommended once you generate a domain. Makes SDK snippets, reset emails and the deep links in notifications point at the right URL |To use external object storage, also set `STORAGE_TYPE=s3`, `S3_BUCKET`, `S3_REGION`, `S3_ACCESS_KEY`, `S3_SECRET_KEY`, and optionally `S3_ENDPOINT`. Otherwise blobs are written to `/data/storage` on the volume you added above.diff --git a/skills/traceway/SKILL.md b/skills/traceway/SKILL.mdindex 7dfeae19..ca0eaae3 100644--- a/skills/traceway/SKILL.md+++ b/skills/traceway/SKILL.md@@ -86,7 +86,7 @@ The body embeds everything for a direct, fast lookup. It contains:- `Hash: <16-hex>` — the exception group → `traceway exceptions show <hash>`.- `Exception ID: <uuid>` — the specific occurrence.- `Occurred at: 2006-01-02 15:04:05 UTC` — the occurrence timestamp. **Convert to RFC3339**: replace the space with `T` and ` UTC` with `Z` (→ `2006-01-02T15:04:05Z`).-- `View details: /issues/<hash>` — the deep link.+- `View details: <url>` — the deep link, `https://<instance>/issues/<hash>` when the server sets `APP_BASE_URL`, otherwise the bare path `/issues/<hash>`.So from a notification, go straight to the occurrence (fast), then pivot reusing the same timestamp: