I have some users submitting forms and Power Automate is not detecting them. This is quite the issue.
This only happens to some people. Most users aren't having this issue. The user in question is using an up-to-date app and survey. They are using the desktop app (MS Store).
I can't recreate the webhook as that requires every single user update their survey or it wont work. I also cant have multiple webhooks as that would result in duplicates.
Hey abureaux,
Since you're not able to recreate the webhook, is there any errors that you're getting from the flow? I've seen this happen a few times before and I have personally fixed it myself with removing and readding the "when a survey response is submitted" part of the flow.
If that doesn't work, let me know and I can find some more things to try!
That's the thing. No errors anywhere. The survey sent successfully, and I was able to find it on the Portal, exactly where it should be. It's just that Power Automate ignored it. Also, other users, myself included, are still working normally.
I called up the user and we uninstalled and reinstalled S123, copied the sent survey, and then tried again. Power Automate detected it and everything worked as intended.
The only part of my flows I am hesitant of changing is "When a survey response it submitted" because that typically means a new webhook. New webhooks are a bad thing for our workflow* and that would cause more problems than it solves.
*It wouldn't be an issue if Esri fixed surveys to force an update when launched via an external URL rather than ignore the update.
Ah alright I understand where you're coming from, I'm curious if it may need to be redownloaded? Possibly a loading issue or issue with the survey itself that just needs refreshed like an uninstall. Has the reinstallation worked for all users you have attempted it with?
I have observed two scenarios: 1) Re-sending will work as normal, or 2) deleting and re-downloading the survey will fix the error.
I really wish logging wasn't optional. It would be nice if every "send" would log a file of what it did, so then we could go back and find a specific sent item and see exactly what it did.
For more critical/time sensitive items, the complete inability to know if something worked beyond "wait and see" is frustrating to users and erodes their confidence in the system. So I'd love to have more tools at my disposal to reduce the chance of this happening again.