Kubernetes Pod Scheduling: Resource Requests

Resource requests tell the scheduler how much CPU and memory a pod needs. The scheduler only places a pod on a node that has enough unallocated capacity to satisfy the request. A pod with requests that no node can fit stays Pending indefinitely.

Resource Management for Pods and Containers - Kubernetes Docs

Verify the cluster and check available capacity:

kubectl get nodes
kubectl describe nodes | grep -A5 "Allocated resources"

Task 1 - Schedule With Requests

Deploy a workload pod with explicit resource requests. The scheduler uses these to find a node with enough free capacity.

Steps:

• Create pod named workload, image nginx:alpine
• Resource requests: cpu: 100m, memory: 64Mi

kubectl get pod workload -o wide

containers:
- name: <name>
  image: <image>
  resources:
    requests:
      cpu: <cpu>
      memory: <memory>

Task 2 - Trigger Pending

Deploy an oversized pod that requests more CPU than any node can provide. Observe why it stays Pending.

Steps:

• Create pod named oversized, image nginx:alpine
• Resource requests: cpu: 128 (128 cores - no node has this)

kubectl get pod oversized
kubectl describe pod oversized | grep -A5 Events

The events section explains why the scheduler could not place it.

containers:
- name: <name>
  image: <image>
  resources:
    requests:
      cpu: <cpu>

Task 3 - Fix the Requests

Deploy a fixed pod with the same image but a realistic CPU request that fits on a node.

Steps:

• Create pod named fixed, image nginx:alpine
• Resource requests: cpu: 100m, memory: 64Mi

kubectl get pod fixed -o wide

containers:
- name: <name>
  image: <image>
  resources:
    requests:
      cpu: <cpu>
      memory: <memory>

This challenge is part of the Kubernetes Pod Scheduling skill path.