Q&A
ive got a workflow that calls an external API and sometimes that API returns 500 errors or just times out. in zapier the whole zap would just silently fail and id find out days later when someone complained
how does error handling work here? specifically:
- can i set up automatic retries for transient failures?
- can i route to a different branch if a step fails (like send me a slack notification)?
- is there a log somewhere that shows me which executions failed and why?
this is the number one reason we left make.com, their error handling was terrible. you couldnt even see the actual error response half the time
error handling works way better than zapier/make. heres the deal:
1. retries - you can configure automatic retries on any node. set how many times to retry and the delay between attempts. great for flaky APIs
2. error branch - every node has an "on error" output you can connect to a separate branch. so you can catch the failure and route it to a slack notification, add a row to an error log table, send yourself an email, whatever
3. execution log - theres a full log of every workflow run. you can filter by status (success/failed/running) and click into any failed execution to see exactly which node failed and the full error response including headers and body
this was a huge selling point for us too. the visibility into failures is night and day compared to zapier
this is exactly what i was hoping to hear. the error branch concept is perfect. going to set up error > slack notification on all our critical workflows
the execution log is a lifesaver. we had a workflow fail at 2am last week and i could see the full API response the next morning without having to reproduce the issue. in make i would have had zero context
+1 the error handling here is solid. one tip - set up a dedicated error notification workflow that all your other workflows route to on failure. saves you from configuring slack/email notifications in every single workflow
Community