Webhooks Failing - Workflow not Found Error Message

[boston85719]

So I have a series of around 12 webhooks setup for a single API provider. All 12 were initialized properly and working properly, 11 of them still do. However, one began to fail sometime between December 28, 2024 at 5PM (GMT+7) and December 28, 2024 at 8PM (GMT+7).

As far as my memory serves me, I made no changes to the backend workflow in question. The name is still the same, no extra characters at the end or beginning as far as I can see in either Bubble or the API provider.

Since there are no differences in the URLs, it seems quite strange that the webhooks would just suddenly begin to fail due to ‘workflow not found’ when the workflow has never been changed since initialization (although some actions within have).

This is the error message reported by the API provider to me… Response: {“statusCode”:404,“body”:{“status”:“NOT_FOUND”,“message”:“Workflow not found container_created”}}

Moreover, the logs show in Bubble the request for workflow as empty instead of showing the name of the workflow as it should and the HTTP request is PATCH instead of POST as it should be.

Has anybody else ever experienced something like this? Anybody have any idea what might be the reason for it, even an idea of me maybe having done something to break the functionality.

I hope this is not a new type of bug to be expecting to arise from time to time, a fully functional webhook just failing due to a bug indicating an existing backend workflow doesn’t exist.

[Jici]

Do some test:
a) Send a request to your endpoint using postman
b) Make sure the url entered in webhooks settings is correct (and not live one, as you show the test version)
c) If you added your own domain, try to switch to your own domain url (It’s been a while, but I got this issue in the past)
d) Patch request is not allowed in Backend WF. The error sound strange as it should return a method not allowed instead. Should be 405 error.
e) try to add another webhook to a tool like requestbin to check if the request from webhook have been updated

I’ve actually are issue with API Connector where API name are not updated. I wonder if this could be related…

[georgecollier]

Make sure you don’t have a trigger of the same name

[boston85719]

There is one! I just changed the name of the trigger. If this works, then there should be a fix by Bubble to automatically append for their purposes ‘database_trigger’ so this wouldn’t cause an issue for others.

[georgecollier]

@boston85719

There is one! I just changed the name of the trigger.

That’ll resolve it. This is a known but undocumented issue that I banged my head over a while back.

Today I learned database triggers can have name conflicts with backend endpoints

Setting up Stripe syncing with Bubble database today and learned for the first time that trigger names can conflict with backend endpoint names. That is to say, if you have a trigger named invoice and a backend workflow named invoice, Bubble will return a workflow not found error when you try to call that endpoint. The trigger will still work. I’d always assumed the name of triggers wouldn’t conflict with endpoints but today shows I was wrong so putting this here for anyone that may bang their …

[Jici]

Completely forget that. Got the same issue too in the past. I agree this should be better documented!

Also, Bubble should consider “type” in naming endpoint conflict. GET, POST, Trigger should be able to have same name if they don’t have the same type. This would help with some API validation that use GET but send the payload after in POST.