I have enabled a flow that inserts records into my FileMaker server and it works. Most of the time.
Because I need to know when it doesn't work, I have an Catch node set to flag errors when that node fails which pushes me an alert.
I have had it running all morning. So far it has processed 750 or so records. Four times during this period I have had errors flag: a small percentage to be sure (0.5% of requests). The payload of the request is not abnormal, so it wasn't apparently a mapping issue.
Which begs the question: how can I understand exactly what failed so I can better deal with it?
For example Is it because two requests came too close together? If so, is there a good way to handle this?
also - are you using flow variables to store "working message" data as it flows through ? If so then yes a new message may be overwriting one already in progress... if so you could just add a delay in rate limit mode if timing isn't an issue - or actually attach/move the value to another msg property, or.... other options are available - but difficult to guess.
This flow is very simple: data in -> change to javascript object -> add a .layout and then create the record. No working message or place to overwrite a message (assuming each flow is a separate thread, something I am not entirely sure of).
I have added a debug to show msg.error going forward. Hopefully this will provide some insight. I have not had an error pop since the four this morning. over 1300 messages today with 4 errors (all early on).