Kubernetes Advanced Quiz

← Back to Kubernetes Quiz

Welcome! ☸️
Test your expertise on Security, Helm, Scheduling, and Troubleshooting.

Instructions:

• Select the best answer for each question.
• Your score will be shown at the end.
• Aim for 100% to earn the Kubernetes Expert status!

#

What is Helm?

A package manager for Kubernetes

A monitoring tool

A container runtime

A CI/CD pipeline

Helm is the package manager for Kubernetes meant for managing complex applications via Charts.

#

What is a Taint and Toleration used for?

To ensure that pods are not scheduled onto inappropriate nodes

To group nodes together

To encrypt node communication

To monitor node health

Taints repel pods. Tolerations allow pods to schedule on tainted nodes.

#

What is a NetworkPolicy?

A specification of how groups of Pods are allowed to communicate with each other

A router configuration

A load balancer rule

A DNS entry

NetworkPolicy acts as a firewall for pods, controlling traffic flow at the IP address or port level (Layer ¾).

#

What is RBAC in Kubernetes?

Role-Based Access Control

Rule-Based Access Control

Remote By-pass Access Control

Routing Based Access Control

RBAC regulates access to computer or network resources based on the roles of individual users.

#

Which component is the single source of truth for the cluster?

etcd

Controller Manager

Scheduler

API Server

etcd is the consistent, highly-available key-value store for all cluster data.

#

What is a Sidecar container?

A helper container that runs alongside the main container in the same Pod

A container running on a different node

A backup container

An init container

A Sidecar runs in the same Pod, sharing the network and storage, assisting the main app (e.g., logging, proxy).

#

What is the purpose of a Horizontal Pod Autoscaler (HPA)?

Automatically scales the number of Pods based on observed CPU utilization

Scales the number of nodes

Scales the size of the pod

Restarts failed pods

HPA scales the number of Pod replicas (horizontal scaling).

#

What is a CRD (Custom Resource Definition)?

An extension of the Kubernetes API that allows you to define custom resources

A standard resource type

A container runtime definition

A cloud resource definition

CRDs extend the K8s API with your own object types.

#

What command would you use to drain a node for maintenance?

kubectl drain

kubectl delete node

kubectl remove

kubectl maintenance

kubectl drain safely evicts all pods from a node, respecting PDBs.

#

What is a PodDisruptionBudget (PDB)?

Limits the number of Pods of a replicated application that can be down simultaneously

Limits the budget for cloud costs

Limits the CPU usage of a Pod

Limits the network bandwidth

PDBs ensure high availability during voluntary disruptions (like node maintenance) by ensuring a minimum number of pods remain running.

#

What happens if you have both HPA and VPA (Vertical Pod Autoscaler) on the same metric?

They can conflict and cause "thrashing" (rapid scaling up and down)

They work perfectly together

VPA takes precedence

HPA takes precedence

Running HPA and VPA on the same metric (e.g., CPU) is not recommended as they will fight over the resource size vs replica count.

#

In Kubernetes Networking, what is the role of CNI (Container Network Interface)?

To configure network interfaces in containers and handle IPAM (IP Address Management)

To load balance traffic

To manage DNS

To encrypt traffic

CNI plugins (Calico, Flannel) are responsible for inserting a network interface into the container namespace and assigning it an IP address.

#

What is the effect of automountServiceAccountToken: false in a Pod spec?

The Pod will not have the service account token mounted, increasing security

The Pod cannot communicate with the network

The Pod cannot start

The Pod runs as root

Disabling automounting prevents the Pod from accessing the API server unless explicitly configured, which reduces the attack surface.

#

What is OOMKilled (Exit Code 137)?

The container used more memory than its limit and was killed by the kernel

The container was killed by a user

The container crashed due to a code error

The node ran out of disk space

OOMKilled means "Out Of Memory". The process was terminated by the Linux OOM Killer because it exceeded its cgroup memory limit.

#

What is the purpose of an Admission Controller?

To intercept requests to the Kubernetes API server and validate or mutate them before persistence

To control who can log in

To manage ingress traffic

To schedule pods

Admission Controllers (Validating/Mutating) intercept requests after authentication/authorization but before the object is saved to etcd. They can reject requests or modify objects (e.g., injecting sidecars).

#

What is the readOnlyRootFilesystem security context used for?

Forces the container's root filesystem to be read-only, preventing attackers from modifying binaries

Prevents the container from writing logs

Makes the volume read-only

Prevents the pod from starting

This security setting hardens the container by preventing writes to the root filesystem, thwarting many persistent attacks.

#

How does etcd maintain consistency?

Using the Raft consensus algorithm

Using Paxos

Using simple replication

Using a master-slave model

etcd uses the Raft consensus algorithm to ensure data consistency across the cluster quorum.

#

What is an Operator in Kubernetes?

A method of packaging, deploying, and managing a Kubernetes application using custom controllers and CRDs

A human administrator

A maintenance script

A type of Service

An Operator is a software extension that uses custom resources to manage applications and their components (encoding operational knowledge into software).

#

If a Node goes into NotReady state, how long does Kubernetes wait before evicting pods (default)?

5 minutes

Immediately

1 minute

1 hour

The pod-eviction-timeout defaults to 5 minutes. After this, the controller manager will taint the node to evict pods.

#

What is kube-proxy's "IPVS" mode?

Uses the Linux IPVS kernel module for load balancing, which is faster and more scalable than iptables

Uses simple user-space proxying

Uses a physical load balancer

Uses DNS round robin

IPVS mode scales much better than iptables for clusters with thousands of Services because it uses hash tables instead of linear lists.

Quiz Progress

0 / 0 questions answered (0%)

0 correct

Quiz Complete!

0%

Reset quiz

---

📚 Study Guides

• Kubernetes Tutorials
• Kubernetes Interview Questions

---

📬 Weekly DevOps, Cloud & Gen AI quizzes & guides