Prometheus & Grafana labs - 10 hands-on challenges

Solve 10 real Prometheus & Grafana challenges in an isolated container in your browser. No setup, no fake shells.

• Install Prometheus From Scratch - beginner · Prometheus is staged as a tarball, not installed. Extract it, install the binary, and get the real server healthy on :9090.
• Install Grafana From Scratch - beginner · Grafana's tarball is staged too. Extract it, install it, start the real server, and reach its login page on :3000.
• Wire Up Your First Scrape Target - beginner · Prometheus and node_exporter are both running, but Prometheus was never told to scrape it.
• Connect Grafana to Prometheus - intermediate · Grafana has zero datasources configured. Provision a Prometheus datasource via YAML so real queries work.
• Import a Real Dashboard - intermediate · A real community dashboard JSON is sitting unimported. Provision it so it renders real live data.
• Stop the Alert Fatigue - intermediate · An alerting rule with no sustained-duration check pages on every tiny fluctuation instead of real problems.
• The Silent Outage - intermediate · A target has been down for hours and nobody noticed, because nothing alerts on up == 0.
• The Cardinality Bomb - advanced · A metric labeled with a raw per-request ID is generating a new time series on almost every scrape.
• Disk Full Before Morning - advanced · Prometheus has no retention.size flag against a fixed disk budget and is heading for a crash-loop.
• The Migration That Broke Every Dashboard - advanced · A relabel_configs change during a migration silently renamed a label, and a saved dashboard query now returns nothing.