Hi @osseonews ,
It appears on your workflow that you shared with the HTTP trigger was last triggered 2 days ago, and there’s been an update since then.
Do you have another example that is currently having the issue? We’d just like to validate the problem is still ongoing as well as give an example workflow that consistently shows this error.
Hi,
I believe we opened up 2 workflows that had that error. I’m not sure of other ones offhand. We haven’t really used these workflows yet, as they were set up to go live once our new website went live. But, then we got these errors during testing, so didn’t proceed yet. If you want, you can test them again. Let me know.
Hi @osseonews ,
Of both of the workflows I’ve see shared, they both had this error 2 days ago but have not been triggered since.
We don’t have access to the HTTP source itself, so you’re free to test at any time but I believe the newest change made yesterday might have already fixed the issue.
We are getting this again on other workflows. I’ve opened a new one up for support to check out. Do you have any idea what is causing this? Why do workflows that haven’t run in awhile throw internal errors for no reason?
Hi @osseonews it appears that the AWS Lambas powering your workflows have gone into a very “cold” state since they haven’t been triggered for so long.
My suggestion is to either use a timer workflow to keep them warm, or to have a retry mechanism on the other side of the workflow that initiates it.
But this is a bug we’re tracking, it’s not very common and it’s only happening on very infrequently used workflows.
Thanks. is your recommendation about “cold” state still accurate? Cold Starts - was going to set this up, but wanted to be sure it’s still the correct approach.
@osseonews that’s right, but this state your workflows are entered are beyond a cold state, we’re referring to them as a “frozen” state internally. We’re trying to understand the timing of these Lambdas going into that state.
You won’t need to re-trigger them every 5 minutes, it seems that it’s only happening to workflows that haven’t been triggered over a few weeks.