Symptoms
You regularly receive the following errors while running your desktop flows:
-
AgentNotReachable - Requested agent is not reachable
-
UnresponsiveAgent - The rpa agent is not responsive
Cause
These errors are caused when the power automate service running on your machine can no longer communicate with the agent in the user session executing your flow. They are typically caused by sub-optimal machine performance such as high CPU usage and may temporarily block the communication between the service and the agent.
Solution
If you are regularly having these errors, please ensure your machine meets the minimum requirements.
If you are still having this problem, you can try increasing the the timeout the power automate service uses to talk to the agent. By default, this value is set to 10 seconds. To do this:
-
Open the registry to: Computer\HKEY_LOCAL_MACHINE\SOFTWARE\WOW6432Node\Microsoft\Power Automate Desktop\Service
-
Add a new DWORD (32-bit) and name it the following: AgentPingTimeoutSeconds
-
Double click on the value to change its setting. Before setting the value, please click on the "Decimal" radio button. If you are using direct machine connectivity (i.e. you are using the machine runtime app to connect your machine to the cloud), try increasing the value to 30. If you are using the on-premises data gateway (not recommended), change the value to 15.
-
Restart the power automate service by going to task manager (CTRL-SHIFT-ESC), selecting the "Services" tab, selecting "UIFlowService" in the list, right clicking and selecting restart.
