Sample interview

Platform / Site Reliability Engineer

7 questions · 60 minutes

Kubernetes platform, delivery pipelines and incident response.

Illustrative: the job post, candidate and results are examples.

Why this interview

Requirements

The job requires

Kubernetes
Linux
Observability and SLOs
Incident response
CI/CD

Claims

The résumé claims

Kubernetes in production
Terraform
Prometheus and Grafana

Focus

So the interview tests

Health checks and rollouts
Process and resource behavior
Incident reasoning
Supply chain

The plan we propose

Every question comes with the reason it was chosen. The interviewer can swap any of them before sending.

1
Choose liveness and readiness behavior
Select all · Advanced · ~5 min
Why this question: Kubernetes in production is on the résumé. Probe design is where restart storms come from.
2
Container ignores termination signals
Multiple choice · Advanced · ~4 min
Why this question: Signals and PID 1 decide whether deploys drain cleanly.
Try it below
3
Two infrastructure applies collide
Multiple choice · Advanced · ~4 min
Why this question: The résumé claims Terraform. State locking is the first thing that breaks with a team.
4
Prioritize a fast error-budget burn
Ordering · Senior · ~5 min
Why this question: The job requires SLOs and incident response. Tests priorities during a fast error-budget burn.
5
Order the failure chain after a migration
Ordering · Advanced · ~5 min
Why this question: Reconstructing a failure chain shows systems thinking under pressure.
Try it below
6
Diagnose too many open files
Multiple choice · Advanced · ~4 min
Why this question: Linux depth: finding a leak instead of raising limits.
7
Verify a deployment artifact
Ordering · Advanced · ~5 min
Why this question: CI/CD ownership includes knowing what production is about to run.

Try two of the questions

Answer as the candidate would. The interviewer guide beside each question is what the interviewer sees.

Order the failure chain after a migration

Candidate view

Exactly what the candidate sees. Try answering it.

An API went fully down about 10 minutes after a routine deploy that included a database migration. The postmortem lists the events below, but they are shuffled.

Drag them into causal order, from the first cause to the final outage.

Drag the steps into order, or use the arrows.

1Request handlers block waiting for a free connection, including the health check endpoint.
2That query switches to a sequential scan and takes seconds instead of milliseconds.
3Requests hold database connections much longer, and the app's connection pool runs out.
4A migration drops an index that a hot query depends on.
5Health checks time out and the load balancer removes instances from rotation.
6Remaining instances receive more traffic and fail faster, until the whole service is down.

Ready when you are.

Interviewer guide

Objective

Test whether the candidate can reconstruct a causal failure chain and spot the design choices that amplified it.

Hints

Ask which event is a cause and which ones are consequences.

Follow-ups

deepenWhat is the difference between liveness and readiness checks, and which should touch the database?
if-quickWhich migration practice would have caught this before deploy?
clarifyWhich single metric would have shown the problem first?

Strong signals

Notices that a database-dependent health check turned a slowdown into a full outage.
Sees that removing instances concentrated load and accelerated the collapse.

Weak signals

Orders the events but treats the outage as caused only by the dropped index.

Rubric

Shuffled. Places consequences before causes in several places.
Correct chain. Orders the chain correctly but cannot name the amplifiers.
Systems view. Explains the amplifying feedback loops and proposes prevention at more than one link.

Container ignores termination signals

Candidate view

Exactly what the candidate sees. Try answering it.

A container starts the app through a shell-form command. During deploys the shell is PID 1, the app does not receive SIGTERM, and shutdown always reaches SIGKILL. What is the most direct correction?

Choose an answer to check it.

Interviewer guide

Objective

Test whether the candidate understands signal delivery to PID 1 in containers.

Hints

Ask them to draw the process tree inside the container with the shell-form command.

Follow-ups

deepenWhen would you add a tiny init such as tini, and what does it do besides forwarding signals?
if-quickOnce SIGTERM arrives, what should the app do in its shutdown window?

Strong signals

Explains that the shell is PID 1 and does not forward SIGTERM to the app.
Uses exec form or exec, and knows when an init process is needed for reaping.

Weak signals

Treats longer kill timeouts as the fix.

Rubric

Timeout tweak. Chooses longer timeouts or SIGKILL first.
Correct fix. Chooses exec form without explaining the PID 1 behavior.
Process model. Explains PID 1 signal handling, zombie reaping and graceful shutdown behavior.

See what the interview actually established

Built only from the interviewer's judgment on each question. The job and résumé decide what gets tested; they never count as evidence.

What the interview established

Strongest evidence was in Containers & Docker, Linux & operating systems and Reliability & incidents. Delivery & cloud showed mixed evidence.

4 of 4 required areas evaluated · 7 questions judged by the interviewer

Required competencies

From the job. Open a row for the questions and notes behind it.

CompetencyRésuméQuestionsEvidence

Delivery & cloudClaimed3Mixed

Choose liveness and readiness behavior
StrongInterviewer
Kept dependency checks out of liveness and explained the restart storm risk.
Two infrastructure applies collide
StrongInterviewer
Remote state with locking, serialized applies per state in CI.
Verify a deployment artifact
MixedInterviewer
Placed verification correctly but could not say what provenance does not protect against.

Containers & DockerClaimed1Strong

Container ignores termination signals
StrongInterviewer
Explained PID 1 signal handling and when an init process is needed.

Linux & operating systemsDemonstrated beyond résumé claims–2Strong

Container ignores termination signals
StrongInterviewer
Explained PID 1 signal handling and when an init process is needed.
Diagnose too many open files
StrongInterviewer
Measured descriptors by type with lsof and traced the missing close.

Reliability & incidentsDemonstrated beyond résumé claims–6Strong

Choose liveness and readiness behavior
StrongInterviewer
Kept dependency checks out of liveness and explained the restart storm risk.
Container ignores termination signals
StrongInterviewer
Explained PID 1 signal handling and when an init process is needed.
Two infrastructure applies collide
StrongInterviewer
Remote state with locking, serialized applies per state in CI.
Prioritize a fast error-budget burn
ExceptionalInterviewer
Rolled back first, verified the burn rate, then investigated. Explained multi-window alerts unprompted.
Order the failure chain after a migration
ExceptionalInterviewer
Ordered the chain and named the database-backed health check as the amplifier before being asked.
Diagnose too many open files
StrongInterviewer
Measured descriptors by type with lsof and traced the missing close.

Also observed

Not required for this role, but the questions produced evidence.

CompetencyRésuméQuestionsEvidence

Git & software delivery–1Mixed

Verify a deployment artifact
MixedInterviewer
Placed verification correctly but could not say what provenance does not protect against.

Secrets & supply chain–1Mixed

Verify a deployment artifact
MixedInterviewer
Placed verification correctly but could not say what provenance does not protect against.

SQL & data storage–1Exceptional

Order the failure chain after a migration
ExceptionalInterviewer
Ordered the chain and named the database-backed health check as the amplifier before being asked.

How they worked

The kind of work the questions asked for, across competencies.

Way of workingJudgedExercisedEvidence

Root-cause analysis44Strong

Debugging33Strong

Design reasoning33Mixed

Incident analysis22Exceptional

Build interviews like this for your roles.

Post a job, and we propose an interview for every applicant from the job and their résumé.

Start interviewing More samples