Power Platform Community Forum Thread Details

We have a number of remote machines which are running power automate desktop where the machines are registered as part of machine groups and we are kicking off the flow from power automate cloud.

We are finding that on some machines the flows are working as expected whereas on others it is unable to recognize the initial screen and we cant tell why.

We have verified that the screen resolution and scaling on the machines are the exact same but we cant determine why the flow works on one machine over another.

As we are running the test over a large range of machines approx 400 we are getting a very high failure rate in our tests. We would like to bring this failure rate down but are at a loss to understand what the difference between the flow on one machine is versus running it on another where there is no apparent difference in terms of screen res.

We have tried looking in the power automate logs but they dont really make much sense to us and there are no obvious errors being logged.

Looking to see if anyone has encountered similar problems or if there is a way to proactively build resilience into our scripting to either prevent it or deal with this problem if it occurs again.

Categories:

Power Automate Desktop

Is it possible to use non-image selectors so that this isn't an issue. I know it's a pain, but I only use image selection as a last resort, because they really do need to be practically the same size monitor with the same size resolution and the smallest variances can be detrimental. You may be able to build an error loop into each search for an image.

Something like this:

Move Mouse to Image (with Tolerance 10)

On Error - Move Mouse to Image (with Tolerance 15)

On Error - Move Mouse to Image (with Tolerance 20)

I don't know if this will help. Out of the handful of times I have tried to use tolerance to increase accuracy on other machines, it has only worked once or twice. This also gets very cumbersome if it is a big flow.

Best of luck!