DevOps Foundations · Module 7: Operations, observability and backup · Lab
Find the Error Source from Logs and Simple Metrics
45 min hands-on · Core
A sick fixture service emitting structured logs plus a simple metrics endpoint; all fixtures under /tmp/dlab-m07-01.
Two ways to do this lab: in your browser on Killercoda (free, no install), or on your own machine as a local guide. Killercoda runs one free scenario at a time: if you see a waiting queue, close other Killercoda tabs and wait a minute.
Objectives
- Start from the metric graph and name the scope of the spike
- Join graph to log lines with a shared request id
- Name the downstream cause with quoted evidence
Step 1
Read the graph first
Open the error-rate and latency graphs, name which endpoint spikes, when it starts, and its share of traffic. Write the scope before touching any log line.
Step 2
Join to the lines
Take exemplars or spike-window request ids from the graph, pull those exact lines, and find the shared downstream error across them. Quote the lines that name it.
Step 3
Name the cause, not the symptom
Separate the caller symptom (checkout 500s) from the downstream cause with evidence for each. State the fix direction the evidence supports.
How to confirm it worked
- Spike scope written from the graph before any log reading
- Join key quoted linking graph window to log lines
- Shared downstream error quoted from the lines
- Cause and symptom separated with the fix direction stated