Random "identity required" error

[Raphaelle_J]

For a few days, i have been having a very puzzling error with some of my flows :

This "identity required" errors happens once in while, and if i relaunch the flow manually with the circular arrow button in the top menu, the flows works perfectly, which means that the input parameters are not the problem.

Plus, the users triggering the flow (with a Form) and myself (my account bears the connections for these flows) are logged-in, so really, i can't figure out what this random "identity required" error is.

This is of the outmost importance since we use this flow in a healthcare setting to inform providers of importants lab tests results.

Thus, any delay in the execution of this flow car endanger patients lives.

Of course, i don't have a Power Automate Premium Account, so no way to contact the support directly to know what is going on.

The only thing to do i could think about is to delete the "List Channels" step and replace it with another "List Channels" action, thinking that maybe a more up-to-date portion of code would work better. But i still got 2 flow failures this morning.

Any help in fixing these errors would be welcomed.

[sf240337]

Hi,

I have the same problem since 15 days.

[Raphaelle_J]

I found a "temporary" workaround in order to avoid flow failure :
 in the parameters of the step which fails unexpectedly, i configured a fixed interval of 20 seconds to relaunch the step, with a maximum of 90 retries.

I also added a parallel branch which send me an email if this steps fails nevertheless.

Nevertheless, this does not explain why this step fails randomly with an "identity required error"...

[leverson]

I have the same issue last 8 days. I added the retry but seems not to work for me 

[Raphaelle_J]

I have slightly less flow failures than previously, but still have a lot.

This is baffling.

Anyone from Microsoft to help us ?

[leverson]

I reduce the concurrency to 1 just in case there were conflicts but just had another failure

getting very frustrating to re-run flows due to this

[Raphaelle_J]

So i read in another thread about "random" errors that some users had gotten rid of the problem by.... *drum rolls intensifies*... duplicating the flow by using "Save As" (and deactivating the old one of course).

I have done this 2 hours ago, and i have not had a single failure since then (on a grand total of 19 automatically triggered flows).

Which is even MORE baffling, because what would i have problem with 1 flow, but not with its exact copy ?

Still, I will be monitoring the problem during the week-end, and if no other failure is observed, I will accept my own post as solution.

[Raphaelle_J]

Unfortunately i still have errors this morning 😣

So duplicating the flow is not a solution.

[leverson]

Based on the fact the failing step failed once and then ran ok the 2nd time, I added in a duplicate step "list channels" and then set the 2nd channel listing to run after failed/skipped etc

I've not had any failures over the weekend or as yet today.

Not sure if i mentioned but I also reduced concurrency to only all 1 instance of the flow at any one time 

[Raphaelle_J]

Thank you for your input, i immediately add this duplicated step in my flow.

Since i do not have an "apply to each" step in my flow, i do not have the concurrency control setting.

[leverson]

I have this set on the trigger action settings

Looks to be available despite an Apply to each loop

[Bearenstone]

Unfortunately this did not work for me either.
The "sacrificial" listing actually passed and the old one still failed, despite both using the same connection.

[Raphaelle_J]

I had no error between 2:43 PM (when i added the second step) and 4:19 PM (when i added the concurrency = 1 parameter).

So i changed the concurrency to the default 25 at a time (you can't disable the concurrency setting once it has been enabled).

I am watching the flow hoping it won't fail again.

[Bearenstone]

In my (failed) case, concurrency was enabled, but there were no concurrent runs. The closest runs where minutes away for a flow that lasts a few seconds.
I set it to concurrency =1 now to see if that will make any difference.

[Bearenstone]

No dice, it still fails. Perhaps less than before, hard to tell.

[Raphaelle_J]

Same for me, still fails but less than before.

I noticed that sometimes it was the first "List channel" action that failed, sometimes the second one. So what i will do is i will try to implement another step, this way :

* if step 1 fails, take the variables from step 2

* if step 2 fails, take the variables from step 1

* if both fails (unlikely), send an email to people who can retrigger the flow

* if none fails, take the variables from step 1

I'll post updates as usual.

Someone just told me that this recent flow failures may have happened since i increased the number of channels in my team, which this plausible. Did you do the same recently ?

[Bearenstone]

Regarding the addition of channels, yes.
I couldn't pinpoint exactly when the errors started but now that you say it, it is very likely around the same time we added additional channels. We are at 15 channels now (including "General"), 4 of which were added very recently.

[Raphaelle_J]

I have 57 channels + "General" 😭

[Bearenstone]

I removed 3. Let's hope for your sake that's a dud 😄

[Bearenstone]

60-odd runs in and so far, so good. No errors yet.