Country or region and language

    Choose your country or region and your language.

    Suggested for you

      All countries and regions

        Choose your character

        It shows in the top bar and on your profile. Change it any time.

        Interviews in Kubernetes

        Work in the tool while you explain your decisions.

        Choose a role and level. Each interview has a new task, follow-up questions and a report that checks your work and your explanation.

        Speak or type. Sign in to save your results across devices.

        Timed assessment

        • 6 practical questions
        • 25 minutes
        • Mid level

        Questions are drawn from a bank of 18. Your report includes a skill breakdown and lessons to practice next.

        Practice interviews

        Junior tasks cover the fundamentals. Mid and senior tasks add debugging, changing requirements and review decisions.

        6 interviews

        • Site reliability engineer
        • Junior
        • 30 min

        A dispatch deployment capacity change

        Scale the named deployment to the requested replicas and confirm all requested replicas are ready. Preserve the unrelated documentation workload. Scope: namespace incident-120, deployment dispatch-120, capacity task target 3 replicas; available image nginx:1.27.

        Start interview
        Skills and rubricA dispatch deployment capacity change
        • Complete the requested operation, identify the resource and compare its result with one source fact.
        • I change the deployment replica target because the controller owns the desired Pod count. I inspect ready replicas as well as the requested number and preserve the unrelated workload. An accepted scale command records intent, while a scheduling or image failure can still prevent the requested capacity from serving.
        • I change the deployment replica target because the controller owns the desired Pod count. I inspect ready replicas as well as the requested number and preserve the unrelated workload. An accepted scale command records intent, while a scheduling or image failure can still prevent the requested capacity from serving.
        • Platform engineer
        • Junior
        • 30 min

        A missing application setting

        Create the exact ConfigMap and key the existing deployment references. Restore readiness without removing its configuration dependency. Scope: namespace incident-121, deployment dispatch-121, capacity task target 4 replicas; available image nginx:1.27.

        Start interview
        Skills and rubricA missing application setting
        • Complete the requested operation, identify the resource and compare its result with one source fact.
        • I read the Pod event and match the ConfigMap name, key and namespace because a similar map elsewhere cannot satisfy the reference. I create the required value and check readiness without removing the dependency. A saved map alone does not prove the consuming process received its configuration, so I also read the named environment value.
        • I read the Pod event and match the ConfigMap name, key and namespace because a similar map elsewhere cannot satisfy the reference. I create the required value and check readiness without removing the dependency. A saved map alone does not prove the consuming process received its configuration, so I also read the named environment value.
        • DevOps engineer
        • Mid
        • 30 min

        An unavailable release image

        Repair the deployment image using the available release reference. Complete its rollout without reducing the required replica count. Scope: namespace incident-122, deployment dispatch-122, capacity task target 5 replicas; available image nginx:1.27.

        Start interview
        Skills and rubricAn unavailable release image
        • Demonstrate the named failure boundary and repeat the relevant check after the requirement changes.
        • I compare the image reference in the Pod template with the approved available release because restarting a missing image repeats the failure. I change the controller template and inspect the rollout and ready count. Editing one owned Pod or reducing replicas can hide the incident while leaving the intended deployment release incomplete.
        • The approved image starts but fails readiness. Compare probe path, port and application logs before accepting the release because image availability alone cannot establish service readiness. Retain the prior working template and define the recovery decision.
        • Site reliability engineer
        • Mid
        • 30 min

        A Service that selects the wrong workload

        Repair the Service selector so it targets the named dispatch deployment. Keep the Service port and target port intact. Scope: namespace incident-123, deployment dispatch-123, capacity task target 3 replicas; available image nginx:1.27.

        Start interview
        Skills and rubricA Service that selects the wrong workload
        • Demonstrate the named failure boundary and repeat the relevant check after the requirement changes.
        • I compare the Service selector with Pod labels because a matching resource name does not establish endpoint membership. I repair the selector and preserve both port values, then inspect ready backends. A Service can exist and still route to no Pods, and endpoints alone cannot prove that the selected application listens on the target port.
        • A new workload shares the app label. Inspect endpoint membership and narrow the selector with a stable distinguishing label because matching too many ready Pods can route to the wrong application. Test the intended path and an excluded workload.
        • Platform engineer
        • Senior
        • 30 min

        A report job that must not overlap

        Change the supplied CronJob to 04:15 UTC and skip a new run while a prior run is still active. Preserve the supplied report command. Scope: namespace incident-124, deployment dispatch-124, capacity task target 4 replicas; available image nginx:1.27.

        Start interview
        Skills and rubricA report job that must not overlap
        • Defend the namespace, controller ownership, approval and recovery boundaries before another engineer applies this change.
        • The report duration now crosses the next scheduled time. Check Forbid and make report processing idempotent because skipping overlap does not guarantee exactly-once business effects. Retain Job history and distinguish a skipped schedule from a failed execution.
        • I retain the namespace, manifests, controller status, Pod events and relevant logs because another engineer needs to reproduce the diagnosis. I compare readiness and traffic evidence with the requested repair and record the unresolved observation. A successful configuration write alone cannot establish application recovery or permission to change another workload.
        • DevOps engineer
        • Senior
        • 30 min

        One namespace in the incident record

        Count Pods only in the supplied incident namespace and write the count to answer.txt. Exclude table headers and unrelated namespaces. Scope: namespace incident-125, deployment dispatch-125, capacity task target 5 replicas; available image nginx:1.27.

        Start interview
        Skills and rubricOne namespace in the incident record
        • Defend the namespace, controller ownership, approval and recovery boundaries before another engineer applies this change.
        • A second namespace now contains the same workload label. Preserve the incident namespace filter because a label alone does not identify scope. Recount without headers and compare the named namespace with the original incident record.
        • I retain the namespace, manifests, controller status, Pod events and relevant logs because another engineer needs to reproduce the diagnosis. I compare readiness and traffic evidence with the requested repair and record the unresolved observation. A successful configuration write alone cannot establish application recovery or permission to change another workload.

        Interview Hub access

        The Interview Hub is included in All Access. Choose a plan or sign in to your account to continue.

        View plansSign in or join

        Your Work Gym account

        Save your progress and pick up on another device. Your browser's progress comes with you.

        or

        You can keep practicing without an account.

        Change your plan

        Checking the price

        Help and feedback

        Ask a question, tell us what is not working, or share an idea. We reply by email.

        What is it about?

        We use it only to reply to you.