[Bug] [API Error] Workflow instance duration keeps counting after stopped by serial discard (end time is empty)

[honkerjha]

Search before asking

• I had searched in the issues and found no similar issues.

What happened

When a workflow instance is stopped due to serial discard mechanism, its status changes to "STOPPED", the end time remains empty (displayed as "-"), but the running duration continues to accumulate indefinitely. This is clearly a bug.
Additionally, I want to confirm: if this issue is not fixed, will it cause database field overflow and further lead to the collapse of the entire scheduling system?

What you expected to happen

After a workflow instance is stopped by serial discard, its running duration should be set to the empty.

How to reproduce

1)Create a workflow (including Shell script and stored procedure execution) with a scheduled trigger (trigger every 3 minutes), and set the workflow to serial discard mode.
2)Shut down all DolphinScheduler modules: worker, master, api, alert, etc.
Wait for a period of time (e.g., 20 minutes).
3)Restart all DolphinScheduler modules and check the workflow instances.
4)Observe that the historical instances generated during the downtime are marked as "STOPPED", their end time is displayed as "-", and the running duration keeps increasing continuously.

Anything else

No response

Version

3.4.1

Are you willing to submit PR?

• Yes I am willing to submit a PR!

Code of Conduct

• I agree to follow this project's Code of Conduct

[SbloodyS]

This bug has not been fixed. The remaining time is calculated in real time when the page is opened, and will not cause any oom.