Design and operate a monolith on Kubernetes
Overview
This question evaluates a candidate's competency in designing and operating a monolithic application on Kubernetes, covering platform engineering topics such as deployment topology, networking and traffic management, CI/CD and rollout strategies, resource management and autoscaling, stateful dependencies, observability, incident response, and disaster recovery. Commonly asked in System Design interviews to assess practical application of distributed systems and SRE principles—specifically the ability to reason about operational trade-offs, reliability, and incremental migration paths—this prompt sits in the System Design domain and emphasizes practical application with architectural reasoning rather than purely conceptual understanding.
Design and Operate a Monolith on Kubernetes
Context
You are joining an infrastructure team. The backend is a single monolithic service. The company is not pursuing microservices or aggressive horizontal scaling right now. Assume:
- Moderate scale (e.g., peak 2–5k RPS), high availability target across multiple zones, zero-downtime deploys.
- Single cloud provider with managed Kubernetes, managed databases/caches/object storage where possible.
- Mix of web (HTTP), background workers, and scheduled jobs from the same monolith codebase.
Task
Design how you would run this monolith on Kubernetes. Cover:
- Deployment topology
- Pods, replicas, separation of web vs worker vs cron jobs
- Node pools, placement, disruption policies
- Networking and traffic management
- Services, Ingress, TLS, rate limiting, canary traffic splitting
- Configuration and secret management
- Configs, secrets, rotation, feature flags
- CI/CD and rollout strategies
- Pipelines, blue/green vs canary, automated rollback
- Resource management and autoscaling
- Requests/limits, HPA/VPA, cluster autoscaling, capacity buffers
- Handling stateful dependencies
- Databases, caches, object storage, migrations, connections
- Observability
- Logs, metrics, traces, SLOs, alerting
- Incident response
- Runbooks, on-call, kill switches, rollback
- Disaster recovery and backups
- RTO/RPO targets, backups, restore drills
- Architecture trade-offs
- Monolith vs microservices, concrete signals that justify a change, and a low-risk, incremental migration path (service boundaries, data ownership, API contracts, team workflows)
Solution
Login required