Part 4 · Lesson 20 · Expert

Logs, Metrics & Traces

Find the slow hop with logs, metrics, and traces.

In this interactive Infrastructure lesson from Stencel Academy you don't just read — you type real commands like kubectl logs deploy/promptd --tail=5, curl -s lb-01:9090/metrics | grep latency, trace 4f9d21c7 and watch the machine respond on a live, animated map. It is part of the free How Infrastructure Works course.

Key takeaway: Logs say what happened, metrics say how much, traces say where — you can't fix what you can't see.

Before this lesson

Builds on: CI/CD Pipelines.

What you'll practise

kubectl logs deploy/promptd | grep ERROR — Filter the event stream for the bad news.
curl -s lb-01:9090/metrics | grep requests_total — Counters only go up — rates come from deltas.
psql -h db-01 -c 'EXPLAIN SELECT * FROM prompts WHERE user_id = 42' — Ask the database how it plans to answer.
psql -h db-01 -c 'CREATE INDEX ON prompts (user_id)' — The fix the trace pointed at.

Terms you'll learn

logsmetricsp99 latencydistributed tracespanobservability

Check your understanding

Which signal pinpointed the slow hop inside a single request?

The distributed trace — per-span timings across services
The log lines — they name the slow service
The p50 latency metric — the median revealed it

Open the interactive lesson →