DebugMidBackendAuto-graded
Find and fix an orders service throwing 500s
Shows how someone behaves during an outage, reading logs and metrics to find the cause and confirming errors stop.The orders service started returning 500s right after a config change went out, and an alert has fired. You have the incident note, the access and error logs, and a metrics dump captured during the bad window. Work from the evidence to the root cause, patch the service, and drive the error rate to zero — without breaking the requests that were working fine before.30 min
python
You're done when
- The service serves the traffic that was failing, and its 5xx rate drops to zero.
- Requests that already worked still return correct totals — no collateral damage.
- The fix addresses the cause visible in the logs and metrics, not just the symptom.
- A guard is in place so the same regression cannot silently return.
What it measures
ObservabilityReading logs and a metrics dump together to locate a fault, rather than guessing from the code alone.
Debugging and profilingMoving from an alert to a root cause under the constraint that the working traffic must keep working.
Incident disciplineVerifying the error rate actually reached zero and leaving behind protection against recurrence.
Your environment
Terminal + editorA full shell in the browser
Claude CodePre-authenticated, no key setup
Isolated microVMYours alone, wiped afterwards
Start this lab
30 min
Start labYou'll sign in to Anthropos to start. The lab opens in a private cloud workspace with Claude Code ready to use.AI Labs is part of Anthropos Workforce. Not on your plan yet? Book a demo