Event-Driven Architectures at Texas Dreamin’: Part 4

Moving away from Point-to-Point Architectures

If you’re convinced that an event-driven architecture suits your organization’s needs, starting with numerous point-to-point integrations can present challenges. Securing funding for a comprehensive replacement project can be daunting, especially when some legacy systems may not directly support event-driven models. In such cases, an incremental migration approach might be most effective. Begin by converting key applications first, then gradually transition other systems during future updates or replacements. This method facilitates the seamless addition of new applications to the event bus, ensuring scalability and resilience as your IT landscape evolves over time.

Event-Driven Architecture Best Practices

These practices help in effectively leveraging Salesforce Platform Events while mitigating common anti-patterns and optimizing system performance and stability.

• Monitor Your Event Usage: It’s crucial to monitor event usage against allocated limits to prevent service disruptions, manage costs, optimize performance, ensure compliance, plan capacity, identify unusual activities, and maximize ROI.
• Internal Communication Workaround: Avoid using platform events as a workaround for synchronous internal communications within your Salesforce org. This can lead to inefficiencies and misuse of event allocation limits. Platform Events are meant to facilitate integration with external systems.
• Correct Use of Real-time Events: Use real-time platform events judiciously, understanding their immediate publishing nature and potential impact on transactional integrity, especially in scenarios like logging.
• Selective Event Handling: Ensure event handlers execute selectively when multiple components listen for an application event, avoiding unnecessary processing.
• Payload Optimization: Minimize payload size when sending and receiving data to improve transport efficiency and processing times.
• Flow Orchestration: Use subflows or Flow Orchestrator instead of platform events for executing logic across multiple automations to avoid event allocation limits.
• Infinite Trigger Loops: Publishing events from Apex triggers associated with the same event object can trigger infinite loops, impacting system stability.
• Transactional Considerations: Be cautious when publishing events from Apex before DML transactions complete, as event publishing may not roll back with transaction failures.