vitals-vyidnj runs the whole Monster Market and none of it is healthy. The manifests are in ~/vitals-vyidnj/. Start with the overview:

kubectl get deploy,pods,hpa -n vitals-vyidnj
ls ~/vitals-vyidnj

market-web     restarts every few seconds
market-api     Running, never Ready
market-worker  Pending
market-db      CrashLoopBackOff
hpa market-web TARGETS <unknown>/50%

Each one is a different lesson. For each, read before you change: describe pod, Events, Last State, the node's Allocated resources.

Work through them

• web: the liveness probe asks nginx for a path it does not serve. Fix the path in web.yaml and apply it. The HPA also needs something from this Deployment: a CPU request. Add resources.requests.cpu: 100m to the container while you are in the file.
• api: the readiness probe asks the wrong port. Fix it in api.yaml.
• worker: it asks for more CPU than the node has left. describe pod names the resource. Lower the request to something the API and web tiers leave room for; kubectl set resources is quickest.
• db: its memory limit is below what postgres needs. Last State tells you how it died. Raise the limit to 256Mi or more, with a request to match.

Edit a file with sed -i or vim, then kubectl apply -f ~/vitals-vyidnj/<file>. A Deployment rolls out the change on its own.

When you are done

kubectl get deploy,hpa -n vitals-vyidnj
kubectl describe hpa market-web -n vitals-vyidnj

Every Deployment at its replica count, READY equal to it, and the HPA showing a real percentage. The last objective is a question about the difference between two probes.