Slice 2 of observability-service-registry. The monitoring router resolves
observability components from the service registry instead of env vars.
- routers/monitoring.py: removed _alertmanager_client/_webhook_client env
readers + the get_settings import. Added _resolve_service_record(store,
service_type, service_id?) -> ServiceRecord|None (requested instance with
type+enabled checks, else first enabled instance), plus _base_url/_timeout/
_auth_headers (Bearer from api_key)/_status_response helpers.
- /alerts + /alertmanager-status now take service_id? + Depends(store),
resolve an alertmanager service, return graceful not-configured/
unreachable payloads including service_id/name; status down-branches now
include peers:[] + error (fixes prior type drift).
- NEW /grafana-status (probes /api/health) and /prometheus-status (probes
/-/healthy then /api/v1/status/buildinfo) returning
{up,version,service_id,name,error}.
- Webhook receiver is now log-only (dropped the outbound
ALERTMANAGER_WEBHOOK_URL forward).
- tests: rewrote TestAlertmanager + TestAlertmanagerWebhook to mock
_resolve_service_record/requests.get (not-configured via empty registry);
added TestGrafanaStatus/TestPrometheusStatus and a TestResolveServiceRecord
unit class covering service_id match/type-mismatch/disabled and first-
enabled/none-enabled paths.
Orphaned config fields alertmanager_url/alertmanager_webhook_url and the
env-var removal land in Slice 5. ruff clean; 240 backend tests pass.
Reviewed fresh-context (read-only): no blockers.
Rename grafana-prometheus-polish -> observability-service-registry and
rewrite proposal/design/tasks for the approved vision: all observability
integration (alertmanager, grafana, prometheus) configured as service-
registry instances in the UI, surfaced on a dedicated page, with widgets
per service definition -- nothing in the env.
Key scope decisions captured:
- Add alertmanager as a 6th service type + active_alerts widget.
- Rewire /alerts + /alertmanager-status to resolve from service records
(first-enabled-instance default; no primary flag in v1).
- Add /grafana-status + /prometheus-status health endpoints.
- Observability page discovers services; kill VITE_GRAFANA_URL /
VITE_PROMETHEUS_URL deep-links.
- Webhook receiver stays log-only (drop the outbound forward).
- Remove PROMETHEUS_FILE_SD_DIR + the file-writer; external Prometheus
uses http_sd_configs against GET /api/monitoring/prometheus-targets.
build_node_exporter_targets + that endpoint stay.
- PROMETHEUS_ENABLED stays (Manage's own /metrics toggle).
- End state: zero observability *service* env vars.
Plan = 5 slices, each <=400 changed lines, green tests/lint/build,
commit per slice.
Manage now connects to existing Grafana/Prometheus/Alertmanager instances
and never deploys its own stack.
- docker-compose.yml / docker-compose.dev.yml: removed prometheus, loki,
alloy, grafana, alertmanager, node-exporter services, the monitoring
network, and observability named volumes; they now ship only backend +
frontend. Dev frontend now joins the web network so the Vite dev proxy
can reach the backend.
- backend: alertmanager_url default is now empty; /api/monitoring/alerts
and /alertmanager-status return graceful "not configured" responses
when ALERTMANAGER_URL is unset. Added not-configured tests.
- docker-compose.observability.yml: kept as the optional standalone
example; header clarifies Manage does not deploy it.
- Removed orphaned combined monitoring/prometheus/prometheus.yml
(standalone stack uses prometheus.standalone.yml).
- Docs (README, REQUIREMENTS decision log, monitoring-logging-design,
observability-runbooks, context.md, MIGRATION_PLAN, frontend/README,
CHANGELOG) updated to the connect-to-existing model.
VITE_GRAFANA_URL / VITE_PROMETHEUS_URL remain as optional frontend
deep-link overrides. .env.example still needs a manual update (safety
policy blocks assistant edits): set ALERTMANAGER_URL empty/optional and
move standalone-only vars out of the root file.
- jellyfin-service-registry: proposal, design, and tasks for completing
the Jellyfin migration off machine-level config.
- grafana-prometheus-polish: proposal, design, and tasks for improving
the Grafana/Prometheus observability integration.
Both are planning-only artifacts; implementation not started.
Move finished change directories to openspec/changes/archive/:
- configurable-dashboard-widgets
- decommission-monitoring-poller
- service-registry
- unify-tasks-on-services
All associated implementation has been merged to main.
The legacy SSH-scraping MonitoringPoller and its endpoints were
decommissioned earlier; update the backend README endpoint list and
Monitoring description to match the current Alertmanager + Prometheus
targets + Grafana observability model.
- Add shared task_runner.run_saved_task helper used by routers/tasks.py and
widgets/sources.py SshTaskWidgetSource.
- Saved tasks now target ssh_tasks service instances via default_service_id;
the legacy default_machine_id and saved_task_runs are removed.
- Actions page lists ssh_tasks services for default and run-time selection.
- Update types, API client, hooks, tests, docs, and changelog.
Backend tests: 222 passed. Frontend lint/build/test: clean (71 passed).
Follow-up #1 to the service-registry change. Jellyfin/Jellyseerr now resolve
from the service registry, so the machine-level app fields are dead config.
- dependencies.py: drop dead _jellyseerr_client_for; simplify _resolve_machine
to SSH-only.
- settings_store.py + routers/settings.py: remove jellyfin_*/jellyseerr_* from
machine default config, get_machine_config, normalization, row mappers, and
MachineInput.
- frontend types + Settings.tsx: drop the fields and the Jellyfin/Jellyseerr
form sections + service options.
- Update frontend test fixtures.
Existing DB rows may still carry these keys in config_json; they are inert and
drop on the next machine save. Verification: backend ruff clean, pytest 222;
frontend lint 0 errors, build success, 70 tests.
Slice 4b frontend half. Jellyfin-touching pages now select a Jellyfin service
instance instead of a machine.
- api/client.ts: Jellyfin-backed calls (counts/libraries/activity/users, media
status/build/stop/force-stop, queryMedia) send jellyfin_service_id.
- hooks/useDashboard, useUsers, useMedia: selector param renamed to
jellyfinServiceId.
- pages/Media + Applications: list jellyfin service instances and persist
jellyfin_service_id in the URL.
- Dashboard (widgets) and Users (default instance) need no selector change.
- Update Applications + Media tests for the new hook/param.
Files/SSH transport keeps machine_id. Verification: frontend lint 0 errors,
build success, 70 tests; backend ruff clean, 222 tests.
Slice 4b backend half. Jellyfin and Jellyseerr clients are now resolved from
service instances instead of machine-level app config.
- Add jellyseerr service definition (6 service types total); add user_id to
the Jellyfin service config.
- dependencies.py: jellyfin_service_id query param + _service_record
(decrypt-on-read); get_jellyfin_client / get_jellyseerr_client / get_user_id
resolve against the service registry (first enabled instance as fallback).
- SSH/Files transport (get_ssh_client) unchanged; still uses machine_id.
- Update service-registry tests for 6 types.
Selection model: split params — ?jellyfin_service_id= for Jellyfin/Jellyseerr,
?machine_id= for SSH/Files. Frontend threading follows in the next PR.
Verification: backend ruff clean, pytest 222 passed; frontend green (unchanged).
PR 4a of the runtime service registry change.
- Remove addon pages (/addons/:addonId, AddonPage, addons/*) superseded by
service pages.
- Remove grafana_url/prometheus_url from backend config, compose, .env.example,
and README (URLs now live on service records; VITE_ frontend deep-link vars
retained).
- Add Services page (/services) with create/list/delete + sidebar nav, so
services are configurable in the tool itself and service pages are reachable.
- Update docs/REQUIREMENTS.md service-registry section; add CHANGELOG.md with
the breaking-upgrade note (MANAGE_ENCRYPTION_KEY required; grafana/prometheus
env vars removed; default widget seeding removed).
Verification: backend ruff clean, pytest 222 passed; frontend lint 0 errors,
build success, 70 tests passed.
PR 3 of 4 for the runtime service registry change.
- Add service + new-shape widget TypeScript types; widgets carry service_id
+ widget_kind (service-bound) or null (built-in).
- Add services API client + TanStack Query hooks; reconcile the widget API
client/hooks to the new endpoints (remove sources/types; add builtin kinds).
- Add closed frontend service registry (integrations/registry.ts) mirroring the
backend, with resolveWidget(widget, services) mapping a widget to its
component + refresh interval.
- Add ServicePage at /services/:serviceType/:serviceId with config view,
empty-on-edit secret inputs + 'set' badges, enable toggle, delete, and the
service's widget-kind list.
- Register /services/:serviceType/:serviceId in App.tsx.
- Reconcile the six widget components to refreshIntervalMs + description props;
rewrite WidgetConfigDialog around a service -> widget-kind picker.
- Update Dashboard test; add integrations/registry.test.ts.
Verification: frontend lint 0 errors, build success, 70 tests passed; backend
ruff clean, 222 tests passed.
PR 2 of 4 for the runtime service registry change.
- dashboard_widgets gains service_id + widget_kind columns (legacy
addon_id/widget_type kept but unused).
- Source adapters take (service: ServiceRecord | None, widget_kind, config).
SERVICE_ADAPTERS keyed by service_type; BUILTIN_ADAPTERS for backups/static.
- Backups and static stay as service-less built-ins (service_id nullable),
exposed via GET /api/widgets/builtin.
- SSH task adapter resolves the task + instance, runs over SSH, and appends a
service_task_runs history row on success/failure/timeout/error.
- Retire widgets/registry.py; widget metadata now comes from the integrations
registry + widgets/builtin. Remove /api/widgets/types and /api/widgets/sources.
- Stop default widget seeding (fresh install = empty dashboard).
- Rewrite widget tests around the service-bound + built-in model (26 tests).
Backend-only breaking change; frontend is reconciled in Slice 3. Build/lint
stay green; pytest 222 passed.
- Refresh README feature list and remove references to the legacy
in-app monitoring charts / backend poller.
- Document configurable dashboard widgets, addon pages, and widget env vars.
- Add VITE_PROMETHEUS_URL support to frontend Dockerfile and both compose files.
- Add header comment to .env.example explaining shell-export workflow.
- Update remote server requirements to match current capabilities.
PR 4 of 4 for configurable dashboard widgets.
- Replace hard-coded Jellyfin/Backups dashboard sections with a loop that
renders enabled widget instances by sort_order.
- Add WidgetInstance renderer and WidgetConfigDialog for adding, editing,
enabling/disabling, deleting, and reordering widgets.
- Add addon pages for grafana, prometheus, and ssh-tasks at /addons/:addonId.
- Register /addons/:addonId route in App.tsx.
- Update docs/REQUIREMENTS.md with the widget system design and API.
Verification:
- backend ruff clean; pytest 200 passed
- frontend npm run lint: 0 errors
- frontend npm run build: success
- frontend npm run test -- src/widgets/registry.test.ts: 3 passed
- Remove GRAFANA_URL from backend environment (backend never consumed it).
- Add VITE_GRAFANA_URL to frontend build-args (prod), dev environment, and
frontend/Dockerfile ARG/ENV so Grafana deep-links resolve correctly.
- Add ALERTMANAGER_WEBHOOK_URL to backend environment so the documented
alert-forwarding feature is reachable from compose.
- Document VITE_GRAFANA_URL in .env.example.
Update docs to reflect that Manage no longer scrapes its own system
metrics (slices 1-2). AGENTS.md, REQUIREMENTS.md (decision log +
observability section), monitoring-logging-design.md, MIGRATION_PLAN.md.
Gate: docs only; backend pytest (173) + frontend build/lint/test (22/63)
remain green from slices 1-2.
The 2026-06-16/17 observability update externalised metrics to
Prometheus + node_exporter + Grafana, but the legacy Manage-side
SSH-scraping monitor was never removed. It duplicated the new stack,
ran SSH df on every machine every 300s, and fed nothing (its UI was
deleted in e2ad731). This slice decommissions the duplication.
Removed (backend):
- services/monitoring_poller.py (MonitoringPoller) — entire file
- services/monitoring_actions.py (disk_space, run_machine_operation,
poll_machine_snapshot, build_machine_client) — entire file;
run_machine_operation had only 2 callers (the poller + /disk), both gone
- tests/test_monitoring_actions.py
- endpoints: POST /api/monitoring/poller, GET /machines/{id}/actions,
GET /disk (and the now-dead _resolve_machine helper)
- lifespan wiring (main.py), dependency wrapper (dependencies.py),
poller.start()/kick() from machine save (routers/settings.py)
- SettingsStore: monitoring_machine_actions table CREATE + 2 indexes +
record/list/prune_machine_actions methods; DROP TABLE IF EXISTS on
startup cleans existing DBs (user-approved)
- config knobs: monitoring_poll_interval_seconds,
monitoring_poll_initial_delay_seconds, monitoring_action_retention_days
- test_api.py: TestMonitoring._ensure_machine + test_disk
Kept (fits the new model): /machines, /prometheus-targets, /alerts,
/alertmanager-status, /alertmanager-webhook; the disk_usage JOB template
(manual on-demand, not monitoring); node_exporter_* machine fields
(they point Prometheus at the right host).
Gate: backend pytest 173 passed; ruff clean.
Developer
alex