SynfraCore
Synfracore
Start Learning
Navigation

Academies

Platform

RoadmapsLabsCertificationsInterviewPYQsAI AssistantCareer
Start Learning Free🗺️ Learning Roadmaps

EKSInterview Q&A

Most asked interview questions with detailed answers

💬
Verified by practitioners with 5+ years production experience· Updated 2025 · SynfraCore EKS Team
Expert Content

AWS EKS Interview Questions

Core Concepts

Q: What is EKS? Node options?

EKS is AWS's managed Kubernetes service. AWS manages the control plane. Cost: $0.10/hr per cluster (needs verification — confirm current EKS pricing page) + EC2 nodes.

Three node options:

Managed Node Groups: AWS manages EC2 lifecycle (patching, replacement on failure). You define instance types.
Fargate: Serverless — no nodes. Pay per pod vCPU+memory. No DaemonSets, limited storage.
Karpenter: Smart node provisioner. Creates right-sized nodes on demand. Spot+On-Demand. Faster than Cluster Autoscaler.

---

Q: IRSA — IAM Roles for Service Accounts.

bash
# Enable OIDC provider
eksctl utils associate-iam-oidc-provider --cluster my-cluster --approve

# Create IAM role bound to K8s service account
eksctl create iamserviceaccount \
  --cluster my-cluster --namespace default \
  --name s3-reader-sa \
  --attach-policy-arn arn:aws:iam::aws:policy/AmazonS3ReadOnlyAccess \
  --approve

# Pods using s3-reader-sa automatically get temporary AWS credentials
# No access keys stored anywhere

---

Q: EKS VPC CNI networking.

AWS VPC CNI: every pod gets a real VPC IP from the node's ENI (no overlay). Pods are directly routable from VPC.

Limit: pods per node = (ENIs × IPs per ENI) - 1. Enable prefix delegation for more pods:

bash
aws eks update-addon --cluster-name my-cluster --addon-name vpc-cni \
  --configuration-values '{"env":{"ENABLE_PREFIX_DELEGATION":"true"}}'
# Each ENI slot = /28 prefix = 16 IPs -> many more pods per node

---

Q: EKS security checklist.

Control plane:  Private API endpoint | Audit logging enabled
Nodes:          Private subnets | IMDSv2 required | No SSH (use SSM)
Workloads:      IRSA for AWS access | Pod Security Standards
                Network policies | External Secrets Operator
                ECR image scanning + Falco runtime

---

Q: Karpenter vs Cluster Autoscaler.

Cluster AutoscalerKarpenter

|---|---|---|

Node typePre-defined node groupsAny instance type on demand
Speed2-10 min scale-up~30 seconds
SpotManual mixed groupsNative, automatic fallback
Right-sizingFixed size groupsPicks optimal size per workload

Revision Notes

EKS: managed K8s, $0.10/hr control plane + EC2 costs
NODE OPTIONS: Managed Groups (EC2 lifecycle) | Fargate (serverless) | Karpenter (smart provisioner)
IRSA: K8s SA + OIDC -> assume IAM role -> no credentials stored
VPC CNI: real VPC IPs per pod | prefix delegation for more pods per node
KARPENTER: right-sized nodes on demand, fast (~30s), native Spot+OD
SECURITY: private endpoint + private nodes + IMDSv2 + IRSA + network policies
Share:
Join our Community
Daily tips, job alerts, interview help — join engineers learning together
Quick Check — EKS
1 / 2

What is the most important concept to understand about EKS for interviews?

Up Next
🔧
EKSTroubleshooting
Debug common issues with root cause analysis
Also Worth Exploring
← Back to all EKS modules
ProjectsTroubleshooting