All of the catalog
Timed challenge
Restore the broken deploy
go-api has been deployed with a nonexistent image tag — pods are stuck in ImagePullBackOff. Your job: identify the problem, update the deployment to a valid image, and get the rollout healthy before the clock runs out.
WorkloadPar 10m−5 per hint
Definition on GitHubHow it is scored
- Par time is 10m. Beating it scores higher; the clock starts when you do.
- Each hint costs 5 points. Taking one is often still the right call.
- Grading is a real check against the live cluster, so a fix that only looks right scores nothing.
What gets set up
Starts from the bad-deploy-rollout incident, injected for you before the timer begins.
Related challenges
- Find the memory leakecho-server pods keep getting OOMKilled at startup — a memory limit was recently changed. Find the resource limit that was misconfigured and restore the pod to a healthy Running state.
- Make the SLO greengo-api's SLO is breaching — the service selector is broken and no traffic is reaching pods. The SRE dashboard shows the error rate above the threshold. Fix the root cause and confirm the SLO recovers.