Describe the bug
Running avg_over_time query fails when run against certain intervals of time
Other functions _over_time seem to behave properly without failing.
The job does not need to exist, no data is needed to reproduce the error. The problem is independent of the producer, collector or log format.
Alternative query from @slim-bean that exhibits the same behaviour
avg by (http_route) (avg_over_time({job="nojob/nojob"} |json | keep http_route, response_time | unwrap response_time [$__auto]))
Use Grafana UI and set a range of time last 5 min, get empty (or values)
Describe the bug Running
avg_over_time
query fails when run against certain intervals of time Other functions_over_time
seem to behave properly without failing. The job does not need to exist, no data is needed to reproduce the error. The problem is independent of the producer, collector or log format.To Reproduce Steps to reproduce the behavior:
avg_over_time({job="nojob/nojob"} | json | keep http_route, response_time | unwrap response_time [$__auto])
avg by (http_route) (avg_over_time({job="nojob/nojob"} |json | keep http_route, response_time | unwrap response_time [$__auto]))
aggregation operator '"sum"' without grouping
Payload:Expected behavior I expect values being returned, or empty if no data is found
Environment:
Screenshots, Promtail config, or terminal output n/a