Skip to content
Code by Pawpu

Engineering Lab

Debug This

Inspect simulated logs, errors, metrics, and symptoms, then name what actually failed.

Loading incident…

How the challenge works

  • Read symptoms, logs, and metrics. Mark the evidence that actually matters.
  • Pick one root cause. The score rewards the cause and the relevant evidence — not every highlighted line.

Key concepts

Change + signal

Incidents usually start with a deploy, a flag, growth, or a dependency. Tie the timestamp to a signal.

Disconfirm

Green health checks and unused systems are useful. They stop you from boiling the ocean.

Common mistakes

  • Blaming the database because it is on the dashboard, even when it is idle.
  • Selecting every log line so the score looks thorough.

Interview tips

  • Say what you would check first and what would falsify it.
  • Separate user symptom, system signal, and recent change.

Related challenges

Built by Al Beltran