Closed elisabettai closed 1 year ago
It looks like that the rule code_cpu_seconds_total_v2:nonidle_increase_over_nodes_12weeks_v2
disappeared from prometheus rules: https://monitoring.osparc.io/prometheus/rules
Did we maybe deleted it by mistake? Luckily we still have the rule node_cpu_seconds_total:nonidle_increase_over_nodes_12weeks
@mrnicegyu11, do you know what is the difference between the v2 and v1 rule?
I vaguely remember the v2 rules but I sadly dont remember clearly anymore if me or sylvain introduced them and for what reason :( I need to check if and why the vanished, please open an issue in the osparc-ops issue tracker if you want us to follow this up
@mrnicegyu11 this is the issue I transferred from osparc-simcore repo.
Could you re-create the rule node_cpu_seconds_total_v2:nonidle_increase_over_nodes_12weeks_v2/3600
that disappeared?
@sanderegg doesn't remember creating it.
Also, do you remember what is the difference between this "_v2" rule and the one that still exists (node_cpu_seconds_total:nonidle_increase_over_nodes_12weeks
)?
Hi @mrnicegyu11 or @Surfict, I'd very soon need the missing data to be restored as you explained in the PR.
In order to make past data available in these new rules, please follow https://jessicagreben.medium.com/prometheus-fill-in-data-for-new-recording-rules-30a14ccb8467`
Ideally, it would be nice to have the data by the end of this week (for the quarterly reports). I hope that works for you.
done
Running this query in prometheus (osparc.io):
node_cpu_seconds_total_v2:nonidle_increase_over_nodes_12weeks_v2/3600
Shows only values up to Dec.1:
Also I don't see that rule in "Status > Rules" in Prometheus. Might this be the problem?
Do you have an idea on how we can report a reasonable number for this quarter? As for reference, for last quarter (Sep-Nov.2022) we reported 10’800 CPU-hours