SO this is what I wanted to query: rate(count(github_workflow_run_status)[30m])
but it doesn't work, maybe because count() is an aggregate function(?). Probably better would be if there's an inbuilt cumulative counter for workflows of different status:
github_workflow_run_status_total{state="success"} - incremented on success
github_workflow_run_status_total{state="started"} - incremented when job is started (or if an otherwise-unknown job is seen)
github_workflow_run_status_total{state="error"}- incremented on err
github_workflow_run_status_total{state="queued"} - incremented when job is queued, not decremented when it leaves the queue
github_workflow_run_status_total{state="in_progress"} - incremented when job is in progress, NOT decremented when it completes
If jobs go from error/success back to started, then let them start incrementing again.
SO this is what I wanted to query:
rate(count(github_workflow_run_status)[30m])but it doesn't work, maybe because
count()is an aggregate function(?). Probably better would be if there's an inbuilt cumulative counter for workflows of different status:github_workflow_run_status_total{state="success"}- incremented on successgithub_workflow_run_status_total{state="started"}- incremented when job is started (or if an otherwise-unknown job is seen)github_workflow_run_status_total{state="error"}- incremented on errgithub_workflow_run_status_total{state="queued"}- incremented when job is queued, not decremented when it leaves the queuegithub_workflow_run_status_total{state="in_progress"}- incremented when job is in progress, NOT decremented when it completesIf jobs go from error/success back to started, then let them start incrementing again.