Three tickets for wrecks-imgkyl: "the API keeps restarting", "the worker never starts", "the website is down". Triage them without fixing anything yet:

kubectl get pods -n wrecks-imgkyl -o wide
kubectl describe pod market-worker -n wrecks-imgkyl
kubectl get endpoints market-web -n wrecks-imgkyl
kubectl describe svc market-web -n wrecks-imgkyl

For each ticket, write down the STATUS or number you saw and the layer it points at: Pod, controller, Service, config or cluster.