Power Platform Community Forum Thread Details

## Environment
- Power Automate (standard license, no premium)
- Scheduled flow running every 5 minutes

## What the flow does
1. Get list of joined teams (standard Teams action)
2. For each team → **Graph HTTP**: `GET /v1.0/teams/{teamId}/channels?$select=id`
3. For each channel → **Graph HTTP**: `GET /v1.0/teams/{teamId}/channels/{channelId}/messages/delta?$filter=lastModifiedDateTime gt {startTime}`
4. **Graph HTTP**: `GET /v1.0/me/chats?$orderby=lastMessagePreview/createdDateTime desc&$top={n}`
5. For each chat → **Standard action** ("Get messages in a chat") with time filter and Top=10

## Problem
About once per day, one of the Graph HTTP actions (or the standard "Get messages in a chat" action) gets stuck in **Running** state for several hours. The timeout setting on each action is set to **PT1M**, but it is never triggered — the action simply hangs.

When the run is stuck, attempting to cancel it results in the run staying in **Canceling** state indefinitely as well.

Because the scheduled flow has a concurrency limit, subsequent runs queue up behind the stuck run and the entire flow stops processing until someone manually intervenes.

## What I have already tried
- Set **timeout to PT1M** on all Graph HTTP actions → does not fire
- Set **retry policy to None** on all Graph HTTP actions → problem still occurs
- Reduced **parallel branch concurrency** to minimum → no improvement
- **Copying the flow** and re-enabling it resolves the issue temporarily, but the hang recurs within roughly 24 hours

## Key observations
- The problem is **not limited to Graph HTTP connector actions** — the standard Teams action "Get messages in a chat" also gets stuck in the same way
- Since both connector types are affected, this does not appear to be a connector-specific issue
- The hang occurs roughly once per day, not on every run
- Copying the flow gives temporary relief, suggesting some kind of per-flow state corruption, but this is just a guess
- **The frequency of the schedule appears to affect how long before the hang occurs**: when running every 5 minutes, the flow stops after approximately 8–10 hours; when reduced to every 10 minutes, it runs for approximately 30 hours before hanging. This suggests the issue may be related to cumulative run count or resource exhaustion rather than elapsed time alone.

## Questions
1. Why does the PT1M timeout setting not fire when an action is in this hung state?
2. Is there a known issue with Graph API-related actions (both Graph HTTP connector and standard Teams connector) hanging indefinitely in scheduled flows?
3. Is there any reliable workaround short of using a premium HTTP connector or an external service (Azure Functions, etc.)?

Any insight would be greatly appreciated.

Categories:

General topics

Using flows

@CU12041057-0

 
Why the PT1M timeout not fired 
Timeout settings in Power Automate typically govern the response time from the service. If the action is "hung," it’s often because the request hasn't even successfully reached the Graph gateway, or the underlying connection is stuck in a TCP "wait" state. The engine thinks it’s still "trying" to initiate, so the timeout timer never actually kicks in.
 
Vishnu WR
LinkedIn - My Tech Space 
 
Please ✅ Does this answer your question if my post helped you solve your issue. This will help others find it more readily. It also closes the item. If the content was useful in other ways, please consider answering Yes to Was this reply helpful? or give it a Like ♥