Observability & Reliability

Metrics, dashboards, logs, traces, and load tests — knowing what the system is doing before a user tells you.

Prometheus

Scraping numbers off your services and querying them over time — the default metrics database, and PromQL.

Grafana

Dashboards over whatever you already store — and the discipline that keeps them useful after the first week.

Loki

Searchable logs without a full-text index — cheap to run, and queried by the same labels as your metrics.

OpenTelemetry

One vendor-neutral way to emit traces, metrics, and logs — instrument once, and change backends without touching the code.

Load testing written in JavaScript and run from a CLI — finding the breaking point before your users do.

Chaos Engineering

Breaking things on purpose, in daylight, with Chaos Mesh or LitmusChaos — so the first test of your failover isn't the real one.

OpenCost

Telling which team, namespace, and workload spent the money — Kubernetes cost allocation you can self-host.