3 Commits

Author SHA1 Message Date
Antoni Nuñez Romeu fa842d3f6f feat(faro): rewrite dashboards to use Tempo instead of Loki
Run Tests on Branches / Detect Changes (push) Successful in 13s
Run Tests on Branches / Backend Tests (push) Has been skipped
Run Tests on Branches / Frontend Mobile Tests (push) Has been skipped
Run Tests on Branches / Parapharmacy API Tests (push) Has been skipped
Run Tests on Branches / Frontend Tests (push) Has been skipped
Run Tests on Branches / PIP Platform Tests (push) Has been skipped
Faro SDK sends traces (spans with events) via OTLP, not OTLP logs or
metrics. Rewrote all 3 Faro dashboards to use Tempo TraceQL queries
instead of Loki LogQL.

- faro-overview: page views, exceptions, console errors, navigation
- faro-js-errors: exception traces, console error traces
- faro-web-vitals: LCP, FCP, CLS, TTFB via web-vitals span events

All panels now query Tempo (uid P214B5B846CF3925F) with TraceQL.
2026-07-21 17:39:19 +02:00
Antoni Nuñez Romeu f32ff701e1 fix(otel): lazy-init metrics and fix Grafana dashboards
Run Tests on Branches / Detect Changes (push) Successful in 14s
Run Tests on Branches / Backend Tests (push) Successful in 4m3s
Run Tests on Branches / Frontend Tests (push) Has been skipped
Run Tests on Branches / Frontend Mobile Tests (push) Has been skipped
Run Tests on Branches / Parapharmacy API Tests (push) Has been skipped
Run Tests on Branches / PIP Platform Tests (push) Has been skipped
- Fix OTel custom metrics not reaching Prometheus: ESM static import
  hoisting caused metrics.getMeter() to run before sdk.start(), getting
  NoopMeterProvider. Changed to lazy-init pattern so instruments are
  created on first use after the SDK configures the real MeterProvider.

- Fix Grafana dashboards job labels: OTel adds service.namespace prefix
  (job=farmafinder/farmafinder-backend), updated queries to use regex
  match (job=~".*farmafinder-backend").

- Fix histogram metric names: OTel appends unit suffix (_milliseconds)
  to histogram names (e.g. http_request_duration_ms_milliseconds_bucket).

Verified: app_heartbeat_total, http_requests_total, db_query_duration_ms,
redis_cmd_duration_ms, and all other custom metrics now flow correctly
through OTLP -> Alloy -> Prometheus remote_write.
2026-07-21 13:18:18 +02:00
Antoni Nuñez Romeu 839c64c12a feat: add end-to-end observability (metrics, health, mobile RUM, dashboards, alerts)
Run Tests on Branches / Detect Changes (push) Successful in 10s
Run Tests on Branches / Backend Tests (push) Failing after 20s
Run Tests on Branches / Frontend Tests (push) Has been skipped
Run Tests on Branches / Frontend Mobile Tests (push) Failing after 20s
- Backend: OTel metrics via OTLP -> Alloy -> Prometheus (OTEL_METRICS_EXPORTER=otlp)
- New business metrics (src/metrics.js): searches, CIMA latency/errors, cache
  hits/misses, logins, rate-limits, pharmacy writes/links, push sent/failed,
  DB + Redis timings/errors, HTTP req count/duration, heartbeat
- Backend health endpoints /healthz and /readyz
- Mobile (Expo): Grafana Faro RUM via @grafana/faro-react-native
- redis/postgres exporters in docker-compose + Prometheus scrape jobs
- Grafana dashboards (backend, datastores, mobile RUM, overview)
- Prometheus alert rules (farmafinder_*) -> existing Alertmanager (Telegram)
- Design/spec saved to docs/superpowers/specs/
2026-07-13 15:57:52 +02:00