Event types (domain, integration, notification) in Microservices - System Design Exercise

Design: Event Types in Microservices Architecture

Design focuses on event types and their handling in a microservices environment. Out of scope are detailed UI notification designs and specific business logic implementations.

Functional Requirements

FR1: Support domain events to capture business state changes within a service

FR2: Support integration events to communicate between different microservices asynchronously

FR3: Support notification events to inform users or external systems about important occurrences

FR4: Ensure reliable delivery of events with minimal latency

FR5: Allow event consumers to subscribe selectively to event types

FR6: Support event versioning and schema evolution

FR7: Handle failures and retries for event processing

Non-Functional Requirements

NFR1: System must handle up to 10,000 events per second

NFR2: Event delivery latency p99 should be under 200ms

NFR3: Availability target of 99.9% uptime

NFR4: Events must be durable and not lost in case of failures

NFR5: Support eventual consistency between microservices

Think Before You Design

Questions to Ask

❓ Question 1

❓ Question 2

❓ Question 3

❓ Question 4

❓ Question 5

❓ Question 6

Key Components

Event producers within microservices

Message broker or event bus (e.g., Kafka, RabbitMQ)

Event consumers or subscribers

Event schema registry for versioning

Dead letter queues for failed events

Notification service for user or external alerts

Design Patterns

Event-driven architecture

Publish-subscribe pattern

Event sourcing for domain events

CQRS (Command Query Responsibility Segregation)

Retry and dead letter queue patterns

Schema evolution and backward compatibility

Reference Architecture

 +----------------+       +----------------+       +----------------+
 |  Microservice  |       |  Message Broker |       |  Microservice  |
 |  (Event Prod)  | ----> |  (Event Bus)   | ----> |  (Event Cons)  |
 +----------------+       +----------------+       +----------------+
         |                        |                        |
         | Domain Events          | Integration Events     | Notification Events
         |                        |                        |
         v                        v                        v
 +----------------+       +----------------+       +----------------+
 | Domain Event   |       | Integration    |       | Notification   |
 | Handler       |       | Event Handler  |       | Service       |
 +----------------+       +----------------+       +----------------+

Components

Microservice Event Producer

Any microservice framework (e.g., Spring Boot, Node.js)

Generates domain events when business state changes occur

Message Broker / Event Bus

Apache Kafka or RabbitMQ

Routes integration and notification events asynchronously between services

Microservice Event Consumer

Microservice framework

Consumes integration events to update local state or trigger actions

Notification Service

Dedicated service with email/SMS/push capabilities

Sends notifications to users or external systems based on notification events

Event Schema Registry

Confluent Schema Registry or custom solution

Manages event schemas and supports versioning for compatibility

Dead Letter Queue

Message broker feature

Stores events that failed processing for later inspection or retry

Request Flow

1. 1. A microservice performs a business operation and emits a domain event internally.

2. 2. The domain event is published to the message broker as an integration event if other services need to know.

3. 3. Other microservices subscribe to integration events from the broker and update their own state accordingly.

4. 4. When a notification event is generated (either by domain or integration event handlers), it is sent to the notification service.

5. 5. The notification service delivers messages to users or external systems via email, SMS, or push notifications.

6. 6. Failed event deliveries are routed to the dead letter queue for manual or automated retries.

7. 7. Event schemas are validated and managed via the schema registry to ensure compatibility.

Database Schema

Entities: - Event: id (PK), type (domain/integration/notification), payload (JSON), timestamp, version - Subscription: id (PK), microservice_name, event_type, endpoint - Notification: id (PK), event_id (FK), recipient, status, sent_timestamp Relationships: - One Event can trigger multiple Notifications - Subscriptions link microservices to event types they consume

Scaling Discussion

Bottlenecks

Message broker throughput limits when event volume grows beyond 10K events/sec

Event consumers overwhelmed by high event processing rates

Notification service latency when sending large volumes of messages

Schema registry becoming a single point of failure

Dead letter queue growing large with many failed events

Solutions

Partition topics in message broker and scale brokers horizontally

Scale event consumers horizontally with load balancing and backpressure handling

Use bulk sending and rate limiting in notification service; integrate with third-party providers

Deploy schema registry in a highly available cluster with caching

Implement automated retries and monitoring for dead letter queue; archive old events

Interview Tips

Time: Spend 10 minutes clarifying requirements and constraints, 20 minutes designing the architecture and data flow, 10 minutes discussing scaling and trade-offs, and 5 minutes summarizing.

Explain the differences and purposes of domain, integration, and notification events clearly

Show understanding of asynchronous communication and event-driven design

Discuss how to ensure reliability and ordering in event delivery

Mention schema management and versioning importance

Address scaling challenges and practical solutions

Highlight how notifications integrate with event processing