Unseen or hidden time in apply to each loop

[MonkMonk]

I'm currently working on a subflow which contains a loop which has just under 1000 iterations.  Runtimes for this loop have taken between 2hrs and 13hrs, but after browsing the iterations, the average duration of each iteration was 8 or so seconds.

This would suggest the loop should run for in and around 2 hours, give or take, but nowhere near the 6hrs or 13 hours being shown in the runtime for the flow.

I'm wondering if there is a way to find which iterations could be taking minutes or hours, or where this extra time is coming from, or simply whats going on and how to fix it

Note: the flow is updating a json array with 3000 objects

[PieterVeenstra]

First important thing to do is remove the nested apply to each steps.

I have a number of posts that will help you with this.

in the inner apply to each you are only massaging the data hence use a select instead of the apply to each. You could probably even do the same for the outer one and turn everything into one select that changes the data into what you want it to be. A single select would run within seconds.

this is quite a good example of using select instead of apply to each:

https://sharepains.com/2021/03/01/sum-sharepoint-column-power-automate/

When you find yourself having to use an apply to each then the Pieter's method will help. This method helps you reference an action inside an apply to each , outside the apply to each. Collecting an array of all the values.

https://sharepains.com/2020/03/11/pieters-method-for-advanced-in-flows/

[MonkMonk]

I will definitely try and get rid of the nested apply to each, it's a smart idea for the future.

But in a new development, I found that the reason for excessively long runtimes is due to some iterations taking hours eg

you can see that the condition is taking two hours, but the actions within the condition add up to a few seconds. The condition is just testing whether an integer is equal to 1 or 0 so I'm kind of baffled by what's going on here.

[HelloWorldCovid]

Have you found the answer? I'm running to the same issue. Accumulative loop iterations is far shorter than

shown on loop!

[IPC_ahaas]

Did you find the reason why? I have a similar situation, running through 6000 rows of data (migration from SharePoint List to Dataverse) and it takes several hours.

[takolota]

@IPC_ahaas 

This batch Dataverse template may help you completely avoid some loops. It should take about 6 minutes for 6000 rows.

https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/Dataverse-Batch-Update-Create-and-Upsert...

[IPC_ahaas]

Woah, that is a whole lot. No explanation as to what it's doing and why though.