How to calculate uptime % or Downtime % in Prometheus Grafana

Question

How to calculate uptime in % already in prometheus grafana using below query process_uptime_seconds{application="$application", instance="$instance", job!="jobid"}

Answer 1

For up could you do something like this?

(1 - avg_over_time(up[1d])) * 86400

That would give you the number of seconds down over the last day.

Answer 2

You wouldn't want to use that metric to calculate uptime. The fact that the number reset, just means that the process restarted.

Instead I would recommend using the up metric that Prometheus creates automatically. Then you would want to group it by application and ignore the instance and jobid.

Something like:

sum(up{application="$application"} == 1) by (application) /
 sum(up{application="$application"}) by (application)

That way you are dividing how many are succeeding in being up by the total of that application.