Basic Event-Driven Architecture Interview Questions
1. What is Event-Driven Architecture (EDA)?
Event-Driven Architecture (EDA) is a software design pattern where the flow of the program is determined by events, which are signals from the environment or other programs. Components in EDA communicate asynchronously by producing and consuming events, enabling decoupled interactions between services.[1][5]
2. How does EDA differ from traditional request-response architectures?
In traditional request-response architectures, components communicate synchronously through direct calls, creating tight coupling. EDA uses asynchronous event communication where producers don’t wait for consumers, promoting loose coupling and better scalability.[1][3]
3. What are the core components of an Event-Driven Architecture?
The main components are event producers (generate events), event consumers (process events), and event brokers (route events between producers and consumers). Additional components include dispatchers, aggregators, and listeners.[1][4][5]
4. What is the role of an event broker in EDA?
An event broker acts as a central hub that receives events from producers, filters them, and routes them to subscribed consumers. It ensures reliable delivery and decoupling between producers and consumers.[1][5]
5. What is an event producer in EDA?
An event producer is a service or application that detects significant state changes and generates events containing relevant data. For example, a Zoho CRM service might produce a “UserRegistered” event when a new user signs up.[4][5]
6. What is an event consumer in EDA?
An event consumer subscribes to specific events from the event broker and processes them when received. Multiple consumers can handle the same event independently, like notification and analytics services consuming a “OrderPlaced” event.[1][4]
7. What are the primary benefits of Event-Driven Architecture?
Key benefits include decoupling of components, independent scalability of producers and consumers, improved resilience through fault isolation, and extensibility by adding new consumers without changing producers.[3][5]
8. Name three common use cases for EDA.
EDA excels in real-time applications (live notifications), scalable systems (e-commerce order processing), and systems requiring flexibility (microservices communication in Paytm’s payment workflows).[4][5]
Intermediate Event-Driven Architecture Interview Questions
9. What is the difference between an event and a command in EDA?
An event represents something that has already happened (past tense, like “OrderPlaced”), while a command represents a request to perform an action (future tense, like “PlaceOrder”). Events are declarative; commands are imperative.[3]
10. Explain the publish-subscribe pattern in EDA.
In publish-subscribe, producers publish events to a topic without knowing consumers. Consumers subscribe to topics of interest and receive relevant events through the broker, enabling fan-out scenarios where one event triggers multiple consumers.[1][4]
11. What is event sourcing?
Event sourcing is a pattern where state changes are stored as a sequence of immutable events rather than updating current state directly. The current state is derived by replaying events from the log.[3][1]
12. What are the advantages of event sourcing?
Event sourcing provides a complete audit trail, enables temporal queries (state at any point in time), prevents data loss through event replay, and supports building projections for different views.[3]
13. How do you ensure exactly-once processing in EDA?
Exactly-once processing is achieved through idempotency (making operations repeatable without side effects), transactional outbox patterns, and consumer acknowledgments. Design handlers to safely process duplicate events.[3][6]
14. What is the Competing Consumers pattern?
The Competing Consumers pattern load balances messages across multiple consumer instances. Consumers compete to claim messages from a queue, ensuring high throughput by parallel processing.[6]
15. How would you implement retry logic in an event-driven system at Flipkart?
Implement retries with exponential backoff for transient failures, distinguish between retryable and permanent errors, set maximum retry limits, and use dead letter queues (DLQ) for failed events after retries.[6]
16. What is a Dead Letter Queue (DLQ) in EDA?
A DLQ stores events that fail processing after maximum retries. It helps debug issues, prevents event loss, and allows manual intervention or separate processing workflows.[6]
17. Explain fanout messaging in Event-Driven Architecture.
Fanout messaging occurs when a single event is delivered to multiple consumers or topics. This pattern is useful for broadcasting updates, like sending an “OrderShipped” event to notification, analytics, and inventory services.[4]
18. What challenges arise with eventual consistency in EDA?
Eventual consistency means changes propagate asynchronously, creating temporary inconsistencies. Systems must embrace this trade-off for scalability, using techniques like event versioning and compensation transactions.[3]
Advanced Event-Driven Architecture Interview Questions
19. How do you handle event ordering in distributed EDA systems?
Event ordering is maintained within partitions using message keys (like customer ID), timestamps in payloads, or designing commutative operations. Global ordering across partitions is often sacrificed for scalability.[3]
20. Compare orchestration and choreography in event-driven systems.
Orchestration uses a central coordinator to direct workflow via commands. Choreography uses events where services react autonomously, promoting better decoupling but requiring careful event design.[3]
21. What is the Retry Messages pattern and when to use it?
The Retry Messages pattern requeues failed messages for later processing. Use it for transient failures like network issues, with backoff delays and limits to prevent infinite loops.[6]
22. How would you design an EDA system for real-time inventory updates at Swiggy?
Use event producers from order services publishing “OrderPlaced” events to a broker. Inventory consumers subscribe, reserve stock atomically, and publish “StockReserved” events. Use partitioning by product ID for ordering.[3][5]
23. What is an event dispatcher in EDA?
An event dispatcher routes events to appropriate handlers based on type or routing keys. It manages processing flow and can implement load balancing across multiple handlers.[5]
24. Explain the role of an aggregator in Event-Driven Architecture.
An aggregator collects and combines multiple related events into a single meaningful event, reducing complexity. For example, combining multiple “ItemAddedToCart” events into a “CartReady” event.[5]
25. How do you test event-driven systems?
Test with contract testing for event schemas, unit tests for handlers, integration tests simulating event flows through brokers, and end-to-end tests verifying complete workflows with mocked components.[4]
26. What is the Async Request-Response pattern in EDA?
This pattern combines EDA with request-response by publishing a request event and correlating the response via unique IDs. Useful for scenarios needing acknowledgment without tight coupling.[6]
27. How do you handle schema evolution in event-driven systems?
Use backward/forward compatible schema changes, versioned events, consumer-side transformations, or dual-writing to old/new formats during migration. Avoid breaking changes to existing consumers.[3]
28. Design an EDA notification system for Salesforce using event patterns.
Order service publishes “OrderConfirmed” events. Notification consumer subscribes, enriches with user data, and publishes “NotificationReady” events. Multiple channel consumers (email, SMS, push) compete for processing.[4][6]
29. What are the trade-offs of using EDA versus synchronous communication?
EDA offers decoupling and scalability but introduces complexity in debugging, consistency guarantees, and ordering. Synchronous communication is simpler but creates coupling and scaling bottlenecks.[1][3]
30. How would you implement idempotency in an EDA payment processor at Atlassian?
Include unique event IDs and processing timestamps. Consumers track processed IDs in a database or cache. Payment operations use idempotency keys to safely retry without duplicate charges.[6][3]
