Reviewed by Aditya Kumar · Last reviewed 2026-03-24
Why containers: Reproducibility—dev = prod; isolation; portability. Why K8s: Orchestration at scale; auto-scaling; declarative config. Architectural logic: Docker: Dockerfile defines image; multi-stage builds for smaller images. Use for Airflow, dbt, Spark jobs. K8s: Deployments...
Red Flag: 'We use Docker' without K8s context or scaling. Pro-Move: 'Spark on K8s—autoscale 0-100 executors; spot for batch; 40% cost savings vs. static cluster.'
Why containers: Reproducibility—dev = prod; isolation; portability. Why K8s: Orchestration at scale; auto-scaling; declarative config. Architectural logic: Docker: Dockerfile defines image; multi-stage builds for smaller images. Use for Airflow, dbt, Spark jobs. K8s: Deployments for services; Spark operator for job submission; ConfigMaps/Secrets for config. Scalability: HPA for scaling; resource limits prevent noisy neighbors. Trade-offs: K8s adds complexity; overkill for simple pipelines. Cost: Node pools; spot for batch. Production: Health checks, minimal images, resource limits. Used for: Airflow on K8s, Spark jobs, consistent dev envs.
Red Flag: 'We use Docker' without K8s context or scaling. Pro-Move: 'Spark on K8s—autoscale 0-100 executors; spot for batch; 40% cost savings vs. static cluster.'
According to DataEngPrep.tech, this is one of the most frequently asked Cloud/Tools interview questions, reported at 1 company. DataEngPrep.tech maintains an editor-reviewed database of 1,863 data engineering interview questions across 7 categories.