Flows randomly running hours instead of seconds/minutes

[Anonymous]

Some of the flows I have have started to experience "Slow performance" a couple weeks ago.

They absolutely randomly run for hours instead of regular duration of usually dozens of seconds, and the system threatens to turn off the flows because it thinks they run too many actions (which is absolutely untrue).

Your flow's performance may be slow because it's been running more actions than expected since <datetime>. Your flow will be turned off if it doesn't use fewer actions.

Here's the most recent example. All regular runs are 20 sec max, but this one is 1.5 hrs. I need to have a concurrency control = 1 at a time, so the info that was submitted to Sharepoint in the time frame while it was thinking for 1.5 hrs is basically lost because of this.

If I inspect it, here's the part the system claims ran for 2hrs:

The whole flow is just this:

I don't understand two things: how can such a simple flow run for 1.5 hrs, and how can it claim it has too many actions if there are so few.

If anyone has any tips on how to get rid of this glitch I would greatly appreciate it. Thanks 

[braydenlarson · Accepted solution]

Hello StepanOpalev,

Based off the details provided below it sounds like the flow is being throttled, there are usually 2 major forms of throttling that could be affecting your flow runs.

1. The first is that there are too many API requests being made that you are hitting Daily API Limits

https://docs.microsoft.com/en-us/power-automate/limits-and-config#daily-limits

So if you have a flow that triggers every 10 minutes and completes 10 API requests that would be 60 for that one flow in an hour. If this was suddenly multiplied to 10 different flows each making 10 requests then you suddenly got 600 in an hour. We usually recommend looking at redesigning your flow to reduce the number of loops or if not possible, to reduce the amount of information pulled in to it that is looped through(I.E. if you grab your entire SharePoint list, adding a query filter to reduce the total number of items it collects).

2. The second limit that is normally hit is connector limit. This limit is set on each individual connector and deals with service protection limits set by the services themselves. This is usually seen with a 429 error on the flow run due to it timing out.

Based off the error posted you are likely erroring on the first flow. I would recommend seeing if there are any other flows that are seeing a similar issue and if you have any way to redesign the flow to reduce total calls made in a loop.
Otherwise, since the flow screenshotted on this is using a recurrence trigger, are you able to increase the time between triggers to help reduce the total number of calls it makes?

Should neither of these options be viable, you may need to look at obtaining a license that increases your total Daily API limits. The Document linked above gives details on which each license has set as their daily limit.

Have a good day,