kubectl top shows a Deployment's Pods consistently using far less C…

Question

kubectl top shows a Deployment's Pods consistently using far less CPU and memory than their requests. What is the application-developer response to improve packing?

Accepted Answer

Lower the containers' resource requests toward observed usage while keeping enough headroom for peaks. — Requests are reserved seats—if nobody sits there, other apps can't use those seats. After kubectl top shows steady under-use, trim requests closer to real usage with a little peak cushion. That helps the scheduler pack the node more honestly.

Answer

Delete the Metrics Server so under-utilized Pods stop reporting and appear fully utilized.

Answer

Raise requests further so the scheduler spreads Pods onto more nodes by force.

Answer

Convert every request into an equal limit and remove the Deployment's replicas field.

kubectl top shows a Deployment's Pods consistently using far less CPU and memory than their requests. What is the application-developer response to improve packing?

Short Explanation

Full Explanation