# Observability 002: Alerts And Dashboard Hardening

## Goal

Make the preproduction observability stack actionable by adding alert rules, better operator dashboards, pinned image versions, and operational counters for services that commonly fail silently.

## Feature Spec

- `docs/FEATURES/observability.md`

## Scope

- Pin Grafana, Prometheus, Loki, Tempo, and Alloy image tags in the observability compose overlay.
- Add Prometheus alert rules for API health, error rate, latency, usage silence, feedback bugs, email failures, blob failures, and background job failures.
- Expand the Grafana dashboard with health, usage, operational failure, alert, log, and trace-oriented panels.
- Add backend counters for email delivery, blob storage operations, and background job runs.
- Document alerting and safe Grafana exposure expectations.

## Out Of Scope

- Notification delivery integration for alerts.
- Client-facing status page.
- Cloud observability backends.
- Full product analytics or session tracking.

## Validation

```bash
dotnet build backend/Socialize.slnx
dotnet test backend/Socialize.slnx
docker compose -f deploy/compose.yml -f deploy/observability/compose.observability.yml config
```