Metrics and Observability: Prometheus + Grafana for Backend Engineers

After the quality pipeline course, the student can verify a project before merge: make qa, tests, static analysis, build, healthchecks, and architecture guards. But a green pipeline does not answer how the system behaves under real load. An endpoint can pass tests and still be slow. A queue can build backlog. An Octane worker can slowly grow memory usage. Users can suffer from p95 latency while the average still looks harmless. This course moves the student from intuition to data. Using the existing Laravel/Filament project, the student adds Prometheus and Grafana, exposes /metrics, instruments HTTP routes, queues, and runtime signals, builds a dashboard, defines SLI/SLO, and learns to connect graphs to action. Metrics are not decoration here. Every signal must answer an operational question: how fast, how many errors, what is happening to the queue, what is happening to the worker, and when someone needs to react. The main outcome is a new engineering habit. An architecture decision can no longer be defended with "it feels faster". The student needs a baseline, a metric, a graph, a causal explanation, a tradeoff, and the remaining risk.