How Event Driven Architectures Go Wrong & How to Fix Them • Matthew Meckes • GOTO 2024

Common anti-patterns in Event-Driven Architecture (EDA):

• YOLO events: Publishing events without proper design or consideration
• Inside-out events: Leaking implementation details across boundaries
• Pinball events: Excessive event bouncing within domains
• God events: Overloaded events containing too much information
• Sparse events: Events with insufficient information

Best Practices for EDA:

• Use RFC process for event schema approval
• Create events from bounded contexts
• Break down events by types (domain events vs state transfer events)
• Implement anti-corruption layers for legacy system integration
• Maintain proper documentation and standards

Design Recommendations:

• Start with smaller, sparse events and grow as needed
• Split metadata from event data
• Use domain-driven design principles
• Create clear boundaries between bounded contexts
• Document architecture decisions and reasoning

Organizational Considerations:

• Enable small teams to work independently
• Balance between team autonomy and cross-organization standards
• Invest in shared tooling and platforms
• Focus on effective communication between teams
• Implement proper observability and monitoring

Implementation Guidelines:

• Don’t be afraid to use orchestration when needed
• Plan for versioning from the start
• Handle idempotency in event processing
• Use proper event naming conventions
• Consider event schema standardization across the organization

Common Challenges:

• Managing semantic coupling between services
• Handling duplicate events and ordering
• Balancing between decoupling and complexity
• Managing breaking changes across consumers
• Maintaining consistency in event schemas