Abillity to continue workflow execution when API times out

Currently if a workflow has an external API call, and that API call times out, then workflow execution stops completely at that point instead of proceeding to next steps. This happens even if the API has been marked as “continue workflow execution in case of error” option is selected for the API.

In my opinion it is a bug, but Bubble support refuses to agree to that. So filing it as an idea.

Please upvote this idea if you relate with this problem and feel it should be fixed:
https://bubble.io/ideaboard?idea=1702978698811x865871200398344200

It is the worst when dealing with these types of issues where the User has the wrong concept of how the software should work and the Expected behavior is just what an engineer thought it should be, even if that defies logic.

My first thought was that the continue execution in case of error would suffice here, as a timeout would be an error in most people’s logic, however, sounds support/engineering believe they are two separate issues.

Does the workflow error triggers work at all for this as a workaround?

Yeah, for them definition of expected behaviour is this is how it has been coded irrespective of what would a regular person expect the behaviour to be.

Are you referring to the “when an unhandled error occurs” action that is available in frontend?

I am having this in backend workflow and not frontend, so that is not applicable here.

What I am doing for workaround is that I am calling my external API call within another ‘schedule api workflow’ instead of calling it as is or within a custom event now.
So my guess is that now since it becomes a separate workflow now by it being under another scheduled workflow, even if external API fails, only the newly scheduled workflow will abort and not the parent workflow from which it was invoked.

Of course this is just my common-sense based assumption of how it would behave. I am not sure if this is correct. I could not validate it either as it is not straightforward to replicate the API timeout behaviour.

I think your work around should work to at least let remainder of actions to execute.

Yeah, but it is just not practical to do that for every call. There are many API calls that I make and it is not possible everywhere.

Also, in fact sometime back I had moved from schedule API workflow for some of these APIs to trigger custom event as schedule API workflow is more expensive in terms of WUs. And now I am having to undo some of those and also incurring extra cost.

Yeah, it is pretty upsetting that Bubble doesn’t listen to their user base as much as they should.

I and I know others have asked for the backend workflows to have a ‘workflow error’ available in the same way it is available on the front end, but that has not been put in place yet. I also have requested that in the backend an API call timeout would allow the ‘api error’ to be yes, so that we as developers can mitigate the issues, but we really can not because of a lack of features available from Bubble to do so.

Good news is they have a team dedicated to setting up a feature fo AI inspired designs (that is sarcasm)

1 Like

:grinning: