Pod priority and preemption in Kubernetes - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Pod Priority in Kubernetes

What happens when a new pod with a higher priority is scheduled but the cluster lacks free resources?

A The new pod waits indefinitely until resources free up without affecting other pods.

B The new pod shares resources with existing pods causing resource contention.

C The new pod is rejected and never scheduled if resources are insufficient.

D The new pod preempts lower priority pods to free resources and gets scheduled immediately.

💡 Hint

Think about how Kubernetes manages resources when priorities differ.

Kubernetes uses pod priority and preemption to ensure higher priority pods can run by evicting lower priority pods if needed.

💻 Command Output

intermediate

Output of Pod Preemption Event

What is the output of the following command when a pod is preempted due to priority?

kubectl describe pod low-priority-pod

A

Events:
  Type    Reason     Age   From               Message
  ----    ------     ----  ----               -------
  Normal  Scheduled  5m    default-scheduler  Successfully assigned default/low-priority-pod
  Warning Preempted  1m    default-scheduler  Preempted by pod high-priority-pod

B

Events:
  Type    Reason     Age   From               Message
  ----    ------     ----  ----               -------
  Normal  Scheduled  5m    default-scheduler  Successfully assigned default/low-priority-pod

C Error from server (NotFound): pods "low-priority-pod" not found

D

Events:
  Type    Reason     Age   From               Message
  ----    ------     ----  ----               -------
  Warning Failed  1m    kubelet  Pod failed to start

💡 Hint

Look for event messages related to preemption in pod description.

The pod description shows an event with reason 'Preempted' indicating it was evicted to make room for a higher priority pod.

❓ Configuration

advanced

Correct Pod Priority Class YAML

Which YAML snippet correctly defines a PodPriorityClass named high-priority with value 1000 and global default set to false?

A

apiVersion: scheduling.k8s.io/v1
kind: PriorityClass
metadata:
  name: high-priority
value: "1000"
globalDefault: false
description: "High priority pods"

B

apiVersion: v1
kind: PriorityClass
metadata:
  name: high-priority
value: 1000
globalDefault: false
description: "High priority pods"

C

apiVersion: scheduling.k8s.io/v1
kind: PriorityClass
metadata:
  name: high-priority
value: 1000
globalDefault: false
description: "High priority pods"

D

apiVersion: scheduling.k8s.io/v1
kind: PriorityClass
metadata:
  name: high-priority
value: 1000
globalDefault: true
description: "High priority pods"

💡 Hint

Check the correct apiVersion and data types for fields.

The correct apiVersion is scheduling.k8s.io/v1. The value must be an integer, not a string. globalDefault is a boolean. Option C matches all requirements.

❓ Troubleshoot

advanced

Troubleshooting Pod Not Preempting

A high priority pod is stuck in Pending state even though lower priority pods exist. What could be a reason?

A The lower priority pods have <code>podDisruptionBudget</code> preventing eviction.

B The high priority pod's priority class is not defined or missing in the cluster.

C The cluster has no nodes available at all, so preemption cannot occur.

D The high priority pod has <code>tolerations</code> that do not match any node taints.

💡 Hint

Think about what can block eviction of pods even if they have lower priority.

PodDisruptionBudgets can prevent eviction of pods even if preemption is needed, causing higher priority pods to remain pending.

🔀 Workflow

expert

Pod Scheduling and Preemption Order

Given these events, what is the correct order of actions Kubernetes takes when scheduling a high priority pod that requires preemption?

A 1,2,4,3

B 1,2,3,4

C 2,1,3,4

D 1,3,2,4

💡 Hint

Think about the logical sequence from detecting resource shortage to scheduling.

The scheduler first detects resource shortage, then selects pods to preempt, evicts them, and finally schedules the new pod.