Senior DevOps Interview Questions
516 questions across 39 categories
Deeper trade-off and design questions — why a specific approach was chosen, what breaks at scale, and how to reason about a system you didn't build.
What's the tradeoff between alpine and slim/distroless base images for a Node.js service?
IntermediateDockerNodejs6 min
How does Docker layer caching behave differently on CI runners without a persistent Docker daemon between builds, and how would you mitigate the resulting slowdown?
IntermediateDocker6 min
Your CI pipeline's Docker build takes 12 minutes and produces a 1.8GB image for a small Node.js service. How would you diagnose and fix it?
IntermediateDockerNodejsCI/CD10 min
How would you further reduce a Docker image's size if the runtime still needs native binary dependencies, beyond what a multi-stage build alone achieves?
AdvancedDocker7 min
How does Docker's DNS-based service discovery actually work under the hood for containers on a user-defined network?
IntermediateDocker6 min
How would container-to-container networking troubleshooting change once these containers move to Kubernetes, where networking works differently from plain Docker?
IntermediateDockerKubernetes7 min
A Docker host's disk usage keeps growing even though containers are regularly removed — how do anonymous volumes cause this, and how do you clean them up?
IntermediateDocker6 min
How do you actually back up and restore a Docker named volume, given that the data lives inside Docker-managed storage you can't directly browse from the host?
IntermediateDocker6 min
A container fails to write to a bind-mounted directory with permission denied, even though the host directory has permissive permissions — why?
AdvancedDocker7 min
A container was recreated to deploy a new image version, and all its data disappeared — what went wrong with the volume configuration?
IntermediateDocker6 min
How would you design volume mounts for a containerized production database, considering both data persistence and backup requirements?
AdvancedDocker7 min
How would you harden a container to run with a read-only root filesystem, given that some part of the application still needs to write to disk somewhere?
AdvancedDocker7 min
Two containers need to share the same data — one writes, another reads. How would you design this safely with Docker volumes?
IntermediateDocker6 min
When would you use a tmpfs mount instead of a regular volume, and why does it matter for handling sensitive data like decrypted secrets?
IntermediateDocker6 min
Standard Docker volumes are tied to a single host — how would you give containers access to network-backed storage that survives even the host itself being replaced?
AdvancedDocker7 min
What's actually different between Cloud Functions Gen 1 and Gen 2, beyond just a version number?
IntermediateGCPCloud Functions6 min
A Cloud Function's p99 latency is dominated by cold starts under bursty traffic — how would you mitigate this, and what's the actual cost trade-off?
IntermediateGCPCloud Functions7 min
A Pub/Sub-triggered Cloud Function occasionally processes the same message twice — is this a bug, and how do you actually fix it?
IntermediateGCPCloud FunctionsPub/Sub7 min
A Cloud Function has been silently failing for days with no alert firing — what's typically missing in the observability setup, and how do you fix it?
IntermediateGCPCloud Functions6 min
A Cloud Function times out mid-execution after partially completing a multi-step operation — what state does it leave behind, and how do you design around this?
AdvancedGCPCloud Functions7 min
After adding a Serverless VPC Access connector to a Cloud Function, cold starts got noticeably slower — why, and is this avoidable?
AdvancedGCPCloud Functions6 min
How would you decide between Cloud Functions, Cloud Run, and GKE for a new workload, beyond just 'serverless is simpler'?
IntermediateGCPCloud FunctionsCloud Run7 min
A Cloud Function works perfectly with the Functions Framework locally but fails immediately in production — how do you systematically narrow down why?
IntermediateGCPCloud Functions7 min
How would you audit an entire GCP organization to find every principal holding Owner or Editor at the project level, before a security review?
AdvancedGCP7 min