Question

A web application takes about 30 seconds after container start to load its cache before it can serve requests correctly. During this warm-up period, users report errors because the Service routes traffic to Pods that are not yet ready. You want the Service to stop sending traffic to a Pod until its warm-up completes, without restarting the container. Which addition to the Pod spec best solves this?

Accepted Answer

A readiness probe that only reports success once the application can serve requests

Answer

A liveness probe that checks the application health endpoint every 30 seconds

Answer

A higher CPU request so the cache loads faster during startup

Answer

An increased Deployment replica count to spread the load across more Pods