Flow run completed and tasks in "Running" state

[lucasbelo777]

Bug summary

Flows runs are keeping the tasks in running state even with the flow run already completed

The logs says the task finished in state completed but seems as running in the UI and in the database

Version info

Version:             3.4.13
API version:         0.8.4
Python version:      3.12.11
Git commit:          2b0ea9be
Built:               Thu, Aug 14, 2025 09:13 PM
OS/Arch:             linux/aarch64
Profile:             ephemeral
Server type:         ephemeral
Pydantic version:    2.11.7
Server:
  Database:          postgresql
Integrations:
  prefect-kubernetes: 0.6.4
  prefect-redis:     0.2.4

Additional context

No response

[desertaxle]

Thanks for the issue @lucasbelo777! Do you have an MRE that you can share so that we can reproduce this issue?

[lucasbelo777]

yes.

Here is an example with a dummy flow:

Code:

@task()
async def do_work(wait_time: int):
    logger.info("doing some work")
    await asyncio.sleep(wait_time)
    return {"message": "dummy message"}


@prefect_flow(name="dummy_flow")
async def flow(report_id: str, wait_time: int = 5):
    await do_work(wait_time)

[lucasbelo777]

Also flows finishing in state running.

[desertaxle]

@lucasbelo777 I haven't been able to reproduce this issue yet, but I suspect it might be related to the fact that you're running an ephemeral server. The server might shut down before it receives the completed task run event.

Could you please let me know how often this issue occurs, and do you have any logs from a run where you observed the issue?

[fortunerains]

mark I have met the same problem

[fortunerains]

@lucasbelo777 I haven't been able to reproduce this issue yet, but I suspect it might be related to the fact that you're running an ephemeral server. The server might shut down before it receives the completed task run event.@lucasbelo777 我目前还未能复现这个问题，但我怀疑这可能与你正在运行一个临时服务器有关。服务器可能在接收到完成的任务运行事件之前就关闭了。

Could you please let me know how often this issue occurs, and do you have any logs from a run where you observed the issue?能否告知这个问题发生的频率，以及你是否有一个观察到问题的运行日志？

It will appear when I call the submit 1000 subtask now and I use local prefect server deployed docker（postgresql + service+ server）

[fortunerains]