ReadMeFeed / Community knowledge
Real questions. Useful conversations. Find the people who know your stack.
DevOps & observability · #prometheus
Learn how to implement a lightweight, secure /health endpoint in Dropwizard that uses the built‑in HealthCheck framework, avoids data leakage, and supports asynchronous checks and Prometheus metrics.
Dashboard variables from Prometheus labels are powerful, but high-cardinality labels and panel repetition can silently multiply queries. Learn how to build a curated variable chain, check cardinality before saving, and why variables are not alert boundaries.
Learn how to drive Kubernetes Horizontal Pod Autoscaler with application‑specific metrics (e.g., requests‑per‑second) using Prometheus Adapter, complete with a reproducible Go service example, common pitfalls, and a checklist for production rollout.
Learn how to diagnose why Grafana alert rules show 'No data' or fail to send notifications, with a step‑by‑step checklist for datasource health, query validity, timing, contact points, and silencing.
Learn how to eliminate Prometheus dashboard timeouts by using Recording Rules to pre-compute expensive PromQL expressions and reduce query-time latency.
Learn how to push Prometheus samples to a dedicated long‑term store like Thanos or Cortex, keep query performance fast, and avoid storage bloat. The post walks through architecture, config, a worked example, trade‑offs, and a quick sanity‑check checklist.
Prometheus remote write forwards samples to external storage by tailing the local WAL. A working setup needs only a YAML block, but a reliable one needs queue tuning and monitoring, because a full queue means silent data loss.
A complete walkthrough of Grafana unified alerting: a Prometheus CPU rule using node_exporter idle-CPU rate, Slack contact point, label-based routing, and how to test it safely.
Stop duplicating Grafana dashboards per environment. Learn how query and custom variables, chained filtering, and the =~ operator give you one dynamic dashboard — plus the multi-select mistake that causes 'No Data'.
dropwizard
grafana
kubernetes