1. Scenario: A microservice is failing frequently. How will you detect and recover it?
Answer
Detect with health checks, monitoring (Prometheus, App Insights).
Recover with Kubernetes auto-restart, circuit breakers, and retries.
Use dead-letter queues for failed async messages.
2. Scenario: Two services (Order & Payment) must complete together or rollback. How will you ensure consistency?
Answer
Use the Saga Pattern.
Choreography → Payment emits event, Order reacts.
Orchestration → Orchestrator coordinates both.
Compensating transaction cancels payment if the order fails.
3. Scenario: Your product catalog service is very read-heavy. How do you optimize it?
Answer
Apply CQRS: separate read/write models.
Add caching (Redis, CDN).
Use read replicas in the database.
4. Scenario: A user-facing API takes too long because it calls five microservices. How do you optimize latency?
Answer
Use API Gateway aggregation to reduce multiple calls.
Use GraphQL for flexible data fetching.
Implement parallel calls + async messaging.
Add caching at API Gateway.
5. Scenario: One microservice is getting overloaded with requests. How do you scale it?
Answer
Enable horizontal auto-scaling (Kubernetes HPA).
Use load balancer.
Queue requests with message broker.
6. Scenario: Your monolith e-commerce app must migrate to microservices. What’s your approach?
Answer
Identify bounded contexts (Cart, Order, Payment).
Apply Strangler Fig Pattern → build new services around monolith.
Slowly cut dependencies and route traffic to microservices.
7. Scenario: You deployed a new version, and some users are facing issues. How do you roll back?
Answer
Use Blue-Green Deployment → switch back to Blue.
Use Canary Deployment to limit blast radius.
Keep database backward compatibility for rollback.
8. Scenario: Your Payment service is critical and must not fail. How do you design it?
Answer
Multi-region deployment with active-active setup.
Retries + Circuit breakers.
Idempotent APIs (so double charge doesn’t happen).
Audit logs + Event sourcing for transactions.
9. Scenario: How will you debug a request that flows across 10 microservices?
Answer
Use distributed tracing (Jaeger, Zipkin, OpenTelemetry).
Correlate requests with Correlation ID / Trace ID in logs.
Centralized logging with ELK/Grafana Loki.
10. Scenario: One microservice (Search) needs data from many services. How do you avoid tight coupling?
Answer
Use event-driven architecture → each service publishes events.
Search builds its own read model (materialized view).
Avoid direct DB joins across services.
11. Scenario: A service is consuming too much memory and crashing. How do you fix it?
Answer
Profile memory usage with APM tools.
Apply bulkhead pattern to isolate failures.
Increase memory limits in Kubernetes resource quota.
Optimize code, cache heavy queries.
12. Scenario: Two microservices use different technologies (Java & .NET). How do you integrate them?
Answer
Use REST or gRPC for sync calls.
Use Kafka/RabbitMQ for async messaging.
Standardize with OpenAPI/Swagger contracts.
13. Scenario: You want to deploy 50 microservices. How do you manage configuration and secrets?
Answer
Use Kubernetes ConfigMaps/Secrets.
Cloud-native secret stores: Azure Key Vault, AWS Secrets Manager.
Centralized configuration: Spring Cloud Config, Consul.
14. Scenario: How do you ensure zero downtime during microservice upgrades?
Answer
Use rolling updates (Kubernetes).
Blue-Green Deployment.
Keep backward compatibility in APIs and DB schema.
15. Scenario: How do you monitor business KPIs in microservices (e.g., failed orders)?
Answer
Expose custom metrics (e.g., failed orders, success rate).
Collect with Prometheus.
Create Grafana dashboards + alerts in PagerDuty/Slack.
Join the conversation! Your thoughts help the community grow.