action #161138
closedFix graphs/data on job age dashbaord
0%
Description
Observation¶
Graphs on https://stats.openqa-monitor.qa.suse.de/d/7W06NBWGk/job-age?orgId=1&from=now-30d&to=now show negative values. That leads to Grafana appending "ago" to durations displayed at the very top of the dashboard. You can also see the negative values if you select the table view in edit mode of the affected graphs/panels. I think all panels that are using one of the job_age_p*
fields are affected.
Acceptance criteria¶
- AC1: Negative values are no longer written to the
job_age_p*
fields so the graphs/panels using them don't show "ago" or negative values in the table view anymore (because it makes no sense for an age to be negative).
Suggestions¶
- Checkout the Telegraf config for how those values are computed and fix the computation
Out of scope¶
- Fixing/removing negative values that are already in InfluxDB (although it may be worth a try if it is easy)
Updated by tinita about 2 months ago
- Category set to Regressions/Crashes
- Status changed from New to In Progress
- Assignee set to tinita
- Target version set to Ready
Updated by tinita about 2 months ago · Edited
- Status changed from In Progress to Feedback
Updated by tinita about 2 months ago · Edited
I think it broke with this commit:
1c134577e710ea470f50bc01b80ab01e881e2b8b
Wed Sep 22 17:15:22 2021 +0200
"telegraf: Ensure all now() calls in SQL to be in UTC"
Updated by tinita about 2 months ago
- Status changed from Feedback to Resolved
Looks good now for the last 2 hours: https://stats.openqa-monitor.qa.suse.de/d/7W06NBWGk/job-age?orgId=1&from=now-2h&to=now&viewPanel=11 (currently it says 51 minutes)
Updated by okurz about 1 month ago
- Status changed from Resolved to Feedback
the top-left panel on https://stats.openqa-monitor.qa.suse.de/d/7W06NBWGk/job-age?orgId=1&from=now-30d&to=now shows "1 hour ago" so … not fixed?
Updated by tinita about 1 month ago
okurz wrote in #note-5:
the top-left panel on https://stats.openqa-monitor.qa.suse.de/d/7W06NBWGk/job-age?orgId=1&from=now-30d&to=now shows "1 hour ago" so … not fixed?
That's because my fix is only 6 days old.
https://stats.openqa-monitor.qa.suse.de/d/7W06NBWGk/job-age?orgId=1&from=now-6d&to=now looks good
Updated by tinita about 1 month ago · Edited
For the record:
Out of scope¶
Fixing/removing negative values that are already in InfluxDB (although it may be worth a try if it is easy)
I thought about that but that would have involved timezone aware calculations and I wasn't up for that
Updated by tinita about 1 month ago
It might be as easy as adding 7200 to every value during daylight saving months, and 3600 for the rest of the time.