-
Bug
-
Resolution: Unresolved
-
P1: Critical
-
None
-
master
-
None
How does our CI handle timestamps for things we record? I'm looking at e.g. https://testresults.qt.io/coin/integration/qt/qtbase/tasks/1760867358, where:
- The integration started on 10/19/25, 11:49 AM
- As part of that integration we have a android-arm64-on-macos build that starts before the integration, at 10/19/25, 11:14 AM
- As part of that integration we have a android-armv7-on-windows build that starts long before the integration, at 10/19/25, 5:09 AM
- etc
This data seems to propagate to what's recorded in the Postgres database used by Grafana as well, giving quite strange results:

- Are the timestamps of jobs (tasks, workitems, etc) not recorded by the coin central scheduler, using a fixed reference?
- Is it up to each coin worker agent to pull the timestamp out from the local time of that particular VM?
Looking at the agent log, I'm seeing:
agent:2025/10/19 03:09:49 agent.go:325: Welcome. The agent (id: 1760341733-146052) is listening on 10.215.207.60:49684
agent:2025/10/19 03:09:49 agent.go:326: The local time is: 2025-10-19 03:09:49.6143302 +0000 UTC m=+0.053260601
- relates to
-
COIN-1270 Workitems still "running" after task completes
-
- Open
-