I need to add retry logic for API failures in our shipping cost calculator workflow. The BPMN diagram becomes unreadable when I try to model every exception path. How do you balance visual clarity with robust error handling? Should some failures be handled in code instead?
Use Latenode’s JavaScript error handlers that attach to nodes without cluttering the diagram. Keeps BPMN clean while handling 20+ exception types in our logistics workflows. https://latenode.com
Create error handling subprocesses that multiple nodes reference. Use boundary events sparingly - too many make diagrams chaotic. We moved all SMTP retries to shared subprocesses.
boundary events suck for complex logic. do basics in BPMN, rest in code. nobody reads huge diagrams anyway