What's the difference between <code>probe_success</code> and <code>up</code>? I see various examples where alerting is done based on either of them (eg. site down, instance down). Am I missing something?

<code>up</code> indicates whether Prometheus could talk to and successfully scrape a target, such as the blackbox exporter. <code>probe_success</code> is a metric exposed by the blackbox exporter indicating if a probe succeeded. For alerting you need both, as if the blackbox exporter is down or timing out then that's indicated by <code>up</code> and if the probe itself is failing that'll be indicated by <code>probe_success</code>.

Difference between probe_success and up?

1 Answers

up indicates whether Prometheus could talk to and successfully scrape a target, such as the blackbox exporter.

probe_success is a metric exposed by the blackbox exporter indicating if a probe succeeded.

For alerting you need both, as if the blackbox exporter is down or timing out then that's indicated by up and if the probe itself is failing that'll be indicated by probe_success.