Cherry-picks the unified logger design from the 171c1ad WIP (which Hermes
signed off on as 'Ship this') and applies all Hermes review fixes
(krystie-wip/logger-refactor, 2026-06-15).
src/utils/logging.js is now the single entry point for:
- log.info / log.warn / log.error / log.debug (with level filtering,
color-coded dev output, JSON prod output)
- log.audit() / log.auditMiddleware() (audit-log.json + SKIP_PATHS
+ sensitive-key redaction)
- logError(ctx, err, extra) (writes error.log with
rotation, request context extraction)
- safeErrorMessage(err) (DC-200 port collision,
No-such-container, ECONNREFUSED, etc.)
Existing src/security/audit-logger.js kept untouched — routes/errorlogs.js
still uses auditLogger.query/clear, no callers migrated.
Hermes' must-fixes (all addressed):
[1] Syntax error on logger.js:401 — old logger.js at repo root is gone;
refactored src/utils/logging.js is the new home, no Chinese IME bug.
[2] /health/live and /health/ready endpoints — untouched in src/app.js.
[3] Tests — added __tests__/logging.test.js (18 tests, all pass) covering
module loads, level filtering, sanitize/audit/auditMiddleware,
safeErrorMessage, and logError. Full suite: 897/897 pass across 31
suites (was 879 + 18 new).
Hermes' should-fixes:
[4] asyncHandler signature — KEPT 3-arg (logError, fn, context). 49 route
files still call it this way; src/app.js's boundAsyncHandler unchanged.
[5] platformPaths.pkiRootCert — UNTOUCHED, still used in src/app.js.
[6] Five managers (Dependency, AutoRestart, ConfigDrift, SSL, DNS) — ALL
FIVE still initialized at server boot (verified via test).
[7] ok(res, ...) helper — UNTOUCHED, all routes still use it.
[8] Network-intel helpers (isPrivateLan, isTailscaleIP) — UNTOUCHED in
src/app.js, no duplicate inline logic added.
- setLevel() now updates both GLOBAL_LEVEL and the singleton log._level,
so level-filter tests don't pollute later tests.
- Logger.audit() and Logger.error() now return promises so await works.
- Logger._log() awaits writeErrorLog so callers using await can rely on
the error.log being flushed.
- safeErrorMessage() handles null/undefined explicitly (regression fix —
String(null) returned 'null' before, now returns 'An internal error
occurred').
- src/app.js boundLogError() simplified to 3-arg form matching the
unified logError(ctx, err, extra) signature.
- createLogger(level) alias exported so existing src/app.js callers work.
- logError, safeErrorMessage, LOG_LEVELS still exported.
- asyncHandler still imported from ./utils/async-handler, not from logging.
- No changes to routes/* (audit-logger.js still consumed unchanged).
- jest: 897/897 tests pass across 31 suites
- node -e "require('./src/app.js')" loads cleanly
- node server.js boots through full init (all 5 managers start)
- Color-coded logger output visible in dev mode (no NODE_ENV)
- JSON output in production mode (NODE_ENV=production)
Three coordinated fixes for the System Overview widget:
1. routes/monitoring.js — flatten getAllStats() shape from
{current:{cpu:{percent},memory:{percent}}} to {cpu,memory,memoryUsage}
so the widget's Number() coercion actually produces numbers, not NaN.
Skill reference: references/totp-and-system-overview-pitfalls.md §3.
2. routes/health.js — add summary block to /health-checks/status response.
Widget looks for {healthy, unhealthy, total} but only per-service objects
existed. Permissive on healthy side (up|healthy|online), strict on
unhealthy (down|unhealthy|offline|error); anything else counted as
unknown. Same skill §3 reference.
3. middleware.js — add /api/v1/monitoring/stats to PUBLIC_ROUTES and the
rate-limit skip list. The widget polls it every 5s from the dashboard;
cookie-auth works but listing it explicitly makes it future-proof
against auth-cookie expiry and prevents per-second 429s.
End-to-end test (unauthenticated):
GET /api/v1/monitoring/stats -> {cpu: 8.71, memory: 0.37, ...}
GET /api/v1/health-checks/status -> {summary: {healthy:11, unhealthy:4, total:15}}
Three cascading bugs in server.js's workflow engine init block:
1. fetchT was referenced but never imported from ./src/utils/http
2. notification-manager was called as factory function but the module
now exports a class (NotificationManager) - need 'new'
3. servicesStateManager was referenced in workflowCtx but only created
later inside an async IIFE (out of scope at workflow init time)
Result: every container start logged
Workflow engine failed to initialize - fetchT is not defined
and the workflow engine never actually wired to resourceMonitor/
updateManager event sources. The 'app' context workflow engine
still ran but didn't get those connections.
Fix:
- Import fetchT at top of file
- Use 'new' for NotificationManager instantiation
- Hoist servicesStateManager creation before workflow init and
remove the duplicate inside the health-checker async IIFE
Verified: container restart shows
[server] Workflow engine initialized
[ResourceMonitor] Workflow engine configured
[UpdateManager] Workflow engine configured
in the log, no more errors at startup.
Also bumps VERSION to current SHA (bump from c64bbe2).
- status/js/totp-recovery.js: NEW. Wires up recovery panel on the TOTP
gate. Pastes Base32 -> /api/v1/totp/setup -> /verify-setup -> session.
Exposes window._refreshRecoveryLink() called by totp-auth.js.
- status/js/totp-auth.js: showTotpOverlay() now calls
_refreshRecoveryLink() so the recovery link hides when TOTP is healthy
and appears when it's broken.
- status/js/totp-settings.js: removed setupSection.style.display='none'
so 'Import existing secret' is always visible; added 'Download backup
file' button after setup that exports the Base32 + recovery
instructions as JSON.
- status/index.html: added 'Lost access? Recover with saved Base32
key ->' link to the TOTP overlay plus the recovery panel itself;
added title tooltip to the auth card reminding users to save the
Base32 on first setup.
- status/build.js: include JS('totp-recovery.js') in the core bundle
after totp-auth.js (since recovery registers a hook auth calls).
- credential-manager.js: add diagnose(key) method that distinguishes
ok | missing | unreadable | corrupt instead of silently returning null
- crypto-utils.js: silent fallback to .encryption-key.bak when primary
can't decrypt existing credentials; first-run bootstrap writes .bak;
rotateKey() backs up old key before swap
- routes/auth/totp.js: new public /api/v1/totp/recovery-info endpoint
returns {status, isSetUp, hint} so UI can show meaningful errors
- middleware.js: add /totp/recovery-info to PUBLIC_ROUTES so the
locked-out user can read the diagnostic without being logged in
Three logical changes grouped:
1. Widget bundle rebuild + sami-files logo (from previous session)
- status/dist/{init,core,features,onboarding}.js rebuilt from latest source
- status/sw.js cache bumped to dashcaddy-shell-594ec75648 to force SW refresh
- status/assets/sami-files.png added (Sami Files service card logo)
2. status/build.js: include monitoring-widgets.js in bundle
- The original build.js was missing monitoring-widgets.js from its JS()
bundle list — that's why the System Overview widget never showed up
in the live init.js until we ran the live /var/www/dashcaddy-status/
build.js. Now consistent.
3. dashcaddy-api/scripts/dashcaddy-update.sh restart_container(): preserve
TOTP secret across container recreates
- Was only setting SERVICES_FILE; container fell back to image-local
/app/credentials.json + /app/.encryption-key (auto-generated fresh
every recreate), which broke TOTP for the bind-mounted secret at
/app/data/credentials.json
- Added CREDENTIALS_FILE + ENCRYPTION_KEY_FILE env vars pointing at
/app/data/ so the container reads from the bind-mounted host data dir
- See skill: software-development/dashcaddy/references/totp-and-system-overview-pitfalls.md §9
4. Auto-updater integration (pulled from upstream release):
- dashcaddy-api/VERSION: dev → c64bbe2
- dashcaddy-api/health-checker.js, middleware.js, package.json,
routes/backups.js, src/app.js: new release code (bundled workflows,
/api/auth/ → /api/v1/ back-compat rewrite, backup storage limits)
Working tree accumulated 172 untracked/modified files from the auto-updater:
- 19 secret/runtime files in dashcaddy-api/data/ that should never be tracked
- 199 byte-identical duplicates of tracked files dumped at root by an
outdated rsync/cp step
- 6 scratch debug scripts (cm_check.js, login_test.js, full_test.js, ...)
- 7 .bak-* files from start.sh and dashcaddy-update.sh rollback branches
- Root-level routes/ directory: dead code, container COPYs dashcaddy-api/routes/
.gitignore now ignores:
- dashcaddy-api/data/ (runtime: credentials, secrets, history)
- start.sh.bak*, scripts/*.bak* (auto-updater rollback backups)
- updates/ (auto-updater runtime state)
- cm_check*.js, *_test.js (scratch debug scripts)
Removed dead code:
- routes/openclaw.js (replaced by dashcaddy-api/routes/openclaw.js)
Recreated runtime scripts that were deleted with their duplicates:
- start.sh (canonical container-start, 47-line full config)
- scripts/dashcaddy-update.sh was already untracked; fixed the tracked
dashcaddy-api/scripts/dashcaddy-update.sh instead (see next commit)
Net change: 172 → 17 files in working tree.
- Recreate status/js/monitoring-widgets.js with robust services count
(reads from window.APPS, #cards DOM, then live fetch as fallback)
- Add sami-files service to data/services.json (Sami Files card)
- Add sami-files template to app-templates.js under 'Files' category
with full systemd deployment docs and Caddy snippet
- Bundle monitoring-widgets.js into init.js
- Add backup_data_dir() and restore_data_dir() using rsync
- Data backed up to backups/{version}/data-backup/ alongside code
- restore_data_dir() called in all three rollback paths (build fail, restart fail, health check fail)
- Add restart_container() that does rm + run to apply new env vars
- Handle action=rollback explicitly (no new version deployment)
- Uses standalone docker build instead of compose for reliability
- Add start.sh at /opt/dashcaddy/start.sh for reboot survival
The data/ directory (services.json, config.json, credentials,
TOTP config, notifications) was never included in the update
backup. Every update wiped user data — services, licenses,
credentials — requiring manual restore.
Now the host-side updater:
- Backs up data/ alongside code files before any update
- Restores data/ on rollback (build failure, restart failure,
or health-check failure)
- openClawRoutes was mounted at root causing /status vs /openclaw/status mismatch
- ctx.docker is a typed wrapper {client,pull,...} — all calls now use docker.client.*
- templates/deploy/removal/restore sub-routers had /apps/ hardcoded in inner routes
causing double-stacking when mounted under /apps (→ /apps/apps/templates etc)
- openclaw.js: GET /status, POST /deploy, GET/POST /proxy/*, DELETE /
1. WebSocket exec auth bypass (exec.js): Require valid JWT or API key
before accepting WebSocket upgrade. Reject unauthenticated requests
with 401 before the upgrade completes.
2. Shell injection in router auto-login (session-handlers.js): Validate
baseUrl against safe hostname pattern before embedding in wget shell
command. Reject with null session if invalid.
3. Path traversal in credentials routes (services.js): Add explicit
serviceId validation (alphanumeric + dash/underscore/dot, max 100
chars) to all three credential endpoints. Removed redundant
try/catch wrapper.
4. execSync injection in CA CSR generation (ca.js): Add sanitize step
replacing any non-alphanumeric domain chars with underscore before
interpolation into shell subj argument. Redundant with existing
validation but provides defense-in-depth.
5. Auth bypass when TOTP disabled (middleware.js): Split the logic
cleanly — disabled TOTP means no auth (initial setup state), enabled
TOTP means all auth methods checked (session/JWT/API key). Removed
the sessionDuration:never conflating shortcut.
- Set Domain=.sami on session + CSRF cookies so browsers send them to all subdomains
- This fixes Caddy forward_auth returning 401 for radarr/sonarr/prowlarr
- Fix login URL concatenation bug (radarr.samilogin -> radarr.sami/login)
- Fix getSetCookie() missing from _httpsFetch/_httpFetch response objects
- Fix array/string handling for set-cookie header in session-handlers fallback
- Refactor csrf-protection to createCSRFMiddleware() factory with cookieDomain support
- Pass renewCSRFToken through middleware deps chain to TOTP route
The /api/v1/services/status endpoint (dashboard card ON/OFF) uses an
HTTPS agent to probe each service. When /app/pki/root.crt is missing
inside the container, it fell back to new https.Agent() which rejects
self-signed certificates. This caused all .sami domain probes to fail
with UNABLE_TO_GET_ISSUER_CERT_LOCALLY, making dashboard cards randomly
flip between ON and OFF depending on whether the Pylon relay responded
before the 10s deadline.
Fix: use rejectUnauthorized: false as fallback when CA cert is absent.
The 1.5.0 API cutover removed the legacy `/api` mount, but the
`dashcaddy_auth` Caddy snippet generated by the installer still called
`uri /api/auth/gate/{args[0]}`. forward_auth then hit the notFoundHandler
and propagated a DC-404 to the browser instead of a 401-redirect to the
login flow (seen on requests.sami).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Remove legacy /api/ mount; all routes now under /api/v1/ only
- Update path matchers (CSRF excludes, public routes, audit log, rate limits)
- Move standalone routes (/api/network/ips, /api/docs, /api/docs/spec) to v1
- Update openapi.yaml (110 paths), CA pages, and 4 lingering frontend files
- Add LICENSE (proprietary EULA), CHANGELOG.md (Keep a Changelog format)
- Add .gitea/workflows/ci.yml (test+lint and security audit jobs)
- Fix 9 pre-existing no-empty lint errors so CI starts green
- Drop ad-hoc scratch reports and *.bak files from repo root
All 739 jest tests pass. Lint is clean.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
build.js rewrites three things during `node build.js`:
- status/dist/*.js (bundle output)
- status/index.html (CSP hash for inline bootstrap)
- status/sw.js (cache name derived from bundle content)
release.sh was only staging status/dist/. Result: when a release didn't
touch index.html or sw.js source, the post-build modifications to those
two files were left unstaged, the commit included only dist/, and the
tarball shipped the stale sw.js. Clients then kept the previous SW
cache name -> activate handler never wiped the cache -> precached old
bundles served forever even after they were "updated" on disk.
Now stage index.html and sw.js too. They're tracked (not gitignored)
so a plain `git add` is enough; the commit is a no-op when nothing
actually changed.
container-logs.js called `wireModal(modal, null, closeModal)` — passing
the local `closeModal()` function as a third arg where wireModal expects
button elements. wireModal then did `closeModal.addEventListener('click',...)`,
threw TypeError, and because each module's IIFE is a top-level statement
in the concatenated features.js bundle, every IIFE *after* container-logs
silently skipped: snapshot, smart-arr-connect, notification-settings,
panel-tabs, backup-restore, resource-monitor, health-check, update-
management, docker-resources, compose-import, container-exec, audit-log,
weather, clock, card-badges, theme-builder, and license. Symptoms:
"Customize Theme" did nothing on click, license badge stuck at "FREE TIER"
(because license.js never ran), no weather, etc.
- container-logs.js: drop the wireModal call, wire backdrop click directly
to the local closeModal so the SSE log stream actually stops on close.
- globals.js: harden wireModal — skip any closeBtn that isn't a real
EventTarget. One typo upstream shouldn't take down the rest of features.js
init silently.
When the build runs on a Windows checkout, fs.readFileSync returns the
file with CRLF intact, and the hash of the inline bootstrap script's
body reflects those CRLFs. The release tarball / git transport / Linux
file system strip CRLF on the publishing host, so the browser sees the
LF-only version and computes a different sha256. CSP then blocks the
script — disabling the version widget, theme switcher, and any other
DOM bindings set up in that inline block.
Normalize CRLF -> LF before computing the hash (the on-disk file keeps
its native line endings; only the hash input is normalized). The CSP
allowlist now matches whatever Caddy actually serves.
Three merge-fallout bugs that combined to leave the services grid empty
and most UI inert:
1. error-handler.js was bundled into onboarding.js (loaded 3rd), but
globals.js in core.js (loaded 1st) does `const errorHandler = new
ErrorHandler()` at top level. ErrorHandler was undefined when core.js
ran -> ReferenceError -> globals.js stopped, so window.APPS,
_showTotpOverlay, loadServices, etc. were never set, and init.js
blew up on every call into core's exports.
Moved error-handler.js to the start of the core.js bundle so the
class is on window before any other script touches it.
2. setup-wizard.js also declared `const errorHandler = new ErrorHandler()`
at top level. Classic scripts share the document's top-level lexical
environment, so this collided with globals.js's declaration ->
redeclaration SyntaxError in features.js. Removed setup-wizard.js's
copy; it picks up the global one.
3. tooltip-definitions.js closed its `(function(window){...})(window);`
IIFE at line ~171 ("Validation module loaded"), then the TOOLTIP_
DEFINITIONS array, getter helpers, window.TooltipDefinitions export,
and final `debug(...)` log all sat at top level — outside the IIFE,
where `debug` was no longer in scope. Removed the early close and
added one at EOF so the whole file is in one IIFE.
The service worker uses staleWhileRevalidate on /dist/*, so after a
release it would serve old bundles from cache indefinitely (cache is
only wiped when the cache *name* changes, which was hardcoded to
'dashcaddy-shell-v10'). Result: dashboard appears unchanged after a
self-update until the user manually unregisters the SW.
build.js now hashes the concatenated dist bundles and writes
`dashcaddy-shell-<10-hex-chars>` into sw.js. Any change in dist/
produces a fresh cache name; on the next page load the SW's activate
handler deletes all older caches and the new bundles are fetched.
When DashCaddy is installed without `${DASHBOARD_DIR}:/app/dashboard` bind
mounted into the container (e.g. legacy DNS2 setup where Caddy serves from
/var/www/dashcaddy-status/), the self-updater's in-container copy to
/app/dashboard was a silent no-op — leaving the dashboard stale across
self-updates, which led to CSP-hash mismatches and a broken UI.
- self-updater: new hostFrontendDir option (default `/var/www/dashcaddy-status`
on Linux, overridable via DASHCADDY_HOST_FRONTEND_DIR). When set, defer the
frontend copy to the host-side updater by passing frontendStagingDir +
frontendTargetDir in trigger.json. Now also includes `js/` in the copy list.
- dashcaddy-update.sh: read those new trigger fields and sync the dashboard
files on the host. Auto-detect fallback for older self-updaters (no fields
in trigger.json) so a single release upgrade self-heals.
- csrf-protection: skip CSRF validation on /api/system/update-notify. The
endpoint has its own X-DashCaddy-Notify-Secret auth and is only ever called
machine-to-machine; browsers never reach it. Without this, the CSRF cookie
check rejects the notify POST before the secret comparison runs.
- release.sh: the verify step piped curl into `node -p ".../dev/stdin"` which
works on Linux but blows up on Windows/git-bash. Replaced with portable
grep+sed extraction so the same script works on both publisher OSes.
- self-updater: per-instance notify secret (auto-generated), notifyAndApply()
triggers an immediate check+apply for the publishing host
- routes: POST /api/system/update-notify (X-DashCaddy-Notify-Secret gated,
added to public-routes allowlist so TOTP doesn't block machine-to-machine)
- dashcaddy-update.sh: include VERSION in backup/deploy/rollback copy lists;
belt-and-suspenders write trigger.json commit to VERSION post-deploy.
Fixes drift where /app/VERSION stayed at the old commit after self-update.
- release.sh: mirror failures are non-fatal+loud; HTTP-verify get2 after
rsync; auto-notify co-located instance via /opt/dashcaddy/updates/notify-secret
(or honour DASHCADDY_NOTIFY_TARGETS for multi-instance setups).
- Service Filter Bar: search by name, filter by status (online/offline)
- Batch Operations: multi-select containers for start/stop/restart
- Container Snapshots: create and manage Docker checkpoints
- Added filter bar and batch action bar to index.html
- Added snapshot button to Admin tools section
- New JS modules: service-filter.js, batch-operations.js, snapshot.js
- Updated build.js to include new modules in bundle
- Added window.openContainerLogsModal(containerId, containerName) function
- Service cards (grid.js) already call this when clicking the 📋 logs button
- Modal now pre-selects the correct container when called from a card
- Rebuilt dist files
- New container-logs.js module for viewing Docker container logs
- Integrated with existing API endpoints (/logs/containers, /logs/container/:id, /logs/stream/:id)
- Features:
- Select container from dropdown
- View logs with stdout/stderr color coding
- Real-time log streaming via SSE
- Search/filter within logs
- Download logs as text file
- Line count and filter indicators
- Added '📜 Container Logs' button to Tools section in index.html
- Added to features.js bundle via build.js
- Rebuilt dist files
Hermes