The shop worker gets 'bad address' for the web backend on every job, so nothing it processes completes.
The shop's background worker calls the web backend at http://web/ and has failed every job for fifteen minutes.
The web container itself is healthy: it came up after the morning's host patching and nginx logged a clean start.
After the patch reboot, someone recreated web by hand from a docker run command in their shell history.
Sam suspects the host's upstream DNS.
The fix is yours to find: hints, the recovery checks and the debrief unlock inside the incident.
The alert fires and the clock starts. Read the page and the briefing.
Work in a simulated shell with realistic output: logs, configs, services.
Change the system the way you would in production. Hints are there if you get stuck.
Automated checks verify the recovery, then the debrief explains what happened.
Free, instant, and it works on your phone. No signup: start as a guest and save your progress later.