-
-
Notifications
You must be signed in to change notification settings - Fork 598
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
More time reported than actual usage #626
Comments
Hi, I also sometimes get active times reported that are unrealistic. I looked into the sql lite db today below are the entries for today. Now I don't know what the normal storage for this data is supposed to look like, but this looks strange. Also some completely duplicate entries like at 09:47:35.425000+00:00 Server Logs has similar entries as in your description and also entries like
This is on windows v0.11.0, using aw-server. I have now switched to aw-server-rust and will watch if this makes a difference.
|
Seems wrong to merge ActivityWatch/aw-watcher-afk#51 in the watcher if the error seems to be in the server IMO. Let's hope that ActivityWatch/aw-core#105 fixed it so we can revert the watcher change. |
@johan-bjareholt I think ActivityWatch/aw-watcher-afk#51 addresses a different problem than the one observed in this issue, which ActivityWatch/aw-core#105 does not fix.
We certainly need better tests for catching bugs that lead to overlapping events like this, especially during collection. Basically all occurrences of "Gap was of negative duration" should be considered bugs, and it's still unclear to me if the bug is in the server or the watcher (wild guess: perhaps something weird with the queue?). There might be more things at play here, but ActivityWatch/aw-watcher-afk#51 significantly reduces the things that can go wrong (making this easier to debug in the future), at practically no cost. You could argue it's just sweeping the problem under the rug (which it kinda is) but it was a very pragmatic PR that quickly solved that particular aspect of the problem, making it easier to look into the others (like #602, which affects aw-server-rust as well). |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
I am getting a related issue on my windows pc, the time tracker reports way less time than I spent active in a day. I have no idea what could be causing it. I'm on the latest version. |
I am on the latest ActivityWatch version.
I have searched the issues of this repo and believe that this is not a duplicate.
OS name and version: Arch Linux
ActivityWatch version: v0.11.0 (rust) and master (aw-server-python)
Describe the bug
Incorrect time reported by aw-server (long stretch of AFK time was incorrectly included).
I exported the buckets and imported into aw-server-rust, which reported the correct time.
I'm assuming this is originally an issue with negative gaps arising from the heartbeat, the resulting events of which are not handled correctly in aw-server-python during flooding.
To Reproduce
Not yet sure how to reproduce.
Expected behavior
Time would be correctly reported.
Documentation
Note how aw-server-python spams a lot more than aw-server-rust.
aw-server-python logs from a standard Activity query:
aw-server-rust logs from a standard Activity query:
Screenshots
aw-server-python, with bad data:
aw-server-python, with bad data and no AFK filtering:
aw-server-rust, with bad data (imported from aw-server-python):
I also tried importing data from testing buckets into the prod instance (after renaming the buckets their hostnames), but did not find any issue there.
Additional context
Something similar has probably been reported before, I'm noting this instance in particular as I have data for it.
The text was updated successfully, but these errors were encountered: