Observability stack — Prometheus / Grafana / Loki — POC
Why. You can't run what you don't measure. A platform without observability fails silently.
What I deploy
- Prometheus: metric collection, alert rules, service discovery.
- Grafana: dashboards (latency, errors, saturation — RED/USE method).
- Loki: log aggregation correlated with metrics.
- Alertmanager: alert routing, SLO/SLA, on-call.
Stack
Prometheus · Grafana · Loki · Alertmanager · Kubernetes
POC — SRE test bed, not deployed to production.