Amazon EKS — Introduction Practice Questions
Test your understanding of Amazon EKS fundamentals with these 10 multiple-choice questions.
Question 1
What does AWS manage in an EKS cluster?
- A. Worker nodes and pod scheduling
- B. The Kubernetes control plane (API server, etcd, scheduler)
- C. Application deployments and scaling
- D. VPC networking and security groups
Show Answer
Correct Answer: B
| Option | Explanation |
|---|---|
| A | Worker nodes are managed by you (or AWS manages node lifecycle if you use managed node groups, but not pod scheduling decisions) |
| B | AWS manages the control plane — API server, etcd, scheduler, and controller manager — across multiple AZs |
| C | Application deployments are your responsibility |
| D | You create and manage your own VPC and security groups |
Question 2
How much does an EKS cluster (control plane) cost per month approximately?
- A. Free
- B. ~$25/month
- C. ~$73/month
- D. ~$150/month
Show Answer
Correct Answer: C
| Option | Explanation |
|---|---|
| A | EKS has no free tier for the control plane |
| B | This is below the actual cost |
| C | $0.10/hour × 730 hours/month ≈ $73/month per cluster |
| D | This would be the cost if extended support is enabled ($0.10 + $0.60/hr) |
Question 3
What is Fargate in the context of EKS?
- A. A separate container orchestrator
- B. A serverless compute engine that runs pods without managing nodes
- C. A managed node group type
- D. An EKS add-on for networking
Show Answer
Correct Answer: B
| Option | Explanation |
|---|---|
| A | Fargate is not an orchestrator — EKS (Kubernetes) is the orchestrator |
| B | Fargate is a serverless compute engine. Each pod runs in its own isolated micro-VM. No EC2 instances to manage. |
| C | Fargate is not a node group — it eliminates nodes entirely |
| D | Fargate is a compute option, not a networking add-on |
Question 4
Which API server endpoint mode is recommended for production EKS clusters?
- A. Public only
- B. Private only or Public + Private with CIDR restrictions
- C. Public with no restrictions
- D. It doesn’t matter — all modes are equally secure
Show Answer
Correct Answer: B
| Option | Explanation |
|---|---|
| A | Public-only exposes the API server to the internet — not ideal for production |
| B | Private endpoint limits access to within the VPC. Public + Private with CIDR restrictions allows admin access from specific IPs while keeping internal traffic private. |
| C | Unrestricted public access is the least secure option |
| D | Endpoint mode directly affects who can reach your API server |
Question 5
Which EKS data plane option does NOT support DaemonSets?
- A. Managed Node Groups
- B. Self-Managed Nodes
- C. Fargate
- D. All options support DaemonSets
Show Answer
Correct Answer: C
| Option | Explanation |
|---|---|
| A | Managed Node Groups run on EC2 instances and fully support DaemonSets |
| B | Self-Managed Nodes also run on EC2 and support DaemonSets |
| C | Fargate does not support DaemonSets. Each Fargate pod runs in its own micro-VM with no concept of a shared node. |
| D | Incorrect — Fargate is the exception |
Question 6
What is the SLA for the EKS Kubernetes API server?
- A. 99.9%
- B. 99.95%
- C. 99.99%
- D. 100%
Show Answer
Correct Answer: B
| Option | Explanation |
|---|---|
| A | This is lower than the actual SLA |
| B | AWS guarantees 99.95% availability for the EKS API server endpoint |
| C | This would be an exceptionally high SLA — EKS does not guarantee this |
| D | No AWS service guarantees 100% availability |
Question 7
What tool is the official CLI for creating and managing EKS clusters?
- A. kubectl
- B. aws eks
- C. eksctl
- D. helm
Show Answer
Correct Answer: C
| Option | Explanation |
|---|---|
| A | kubectl interacts with Kubernetes clusters but doesn’t create EKS clusters |
| B | aws eks is the AWS CLI subcommand for EKS — it can create clusters but is lower-level |
| C | eksctl is the official, purpose-built CLI tool for creating and managing EKS clusters. It abstracts away the complexity of VPC creation, IAM roles, and node groups. |
| D | Helm is a package manager for Kubernetes — it installs applications, not clusters |
Question 8
How does EKS ensure high availability of the control plane?
- A. By running the control plane in a single AZ with a standby
- B. By running the API server and etcd across at least 3 Availability Zones
- C. By replicating the control plane to another region
- D. By using Fargate for the control plane components
Show Answer
Correct Answer: B
| Option | Explanation |
|---|---|
| A | A single AZ with standby would not provide the same level of availability |
| B | The EKS control plane (API server + etcd) runs across at least 3 AZs within a region, with automatic failover |
| C | EKS does not replicate the control plane to another region |
| D | The control plane components are not run on Fargate |
Question 9
What is the primary difference between EKS and ECS?
- A. EKS is free and ECS is paid
- B. EKS uses Kubernetes as the orchestrator while ECS uses AWS-proprietary orchestration
- C. ECS supports more container types than EKS
- D. EKS only runs on Fargate
Show Answer
Correct Answer: B
| Option | Explanation |
|---|---|
| A | EKS has a control plane fee ($73/mo); ECS has no control plane fee |
| B | EKS runs upstream CNCF-conformant Kubernetes. ECS uses AWS’s own task/service model. This is the fundamental difference — EKS workloads are portable to any K8s cluster. |
| C | Both EKS and ECS run Docker/OCI containers |
| D | EKS runs on EC2, Fargate, or both |
Question 10
What happens when an EKS cluster’s Kubernetes version exits standard support?
- A. The cluster is immediately deleted
- B. The cluster enters extended support at an additional $0.60/hour
- C. The cluster continues running with no additional cost
- D. AWS automatically upgrades the cluster to the latest version
Show Answer
Correct Answer: B
| Option | Explanation |
|---|---|
| A | AWS does not delete clusters when support ends |
| B | Clusters on unsupported versions enter extended support at an extra $0.60/hr (~$438/month). Eventually, if extended support also ends, AWS will auto-upgrade the cluster. |
| C | There is an additional cost for extended support |
| D | Auto-upgrade happens only after extended support ends, not immediately |
