search

osism / kolla-operations

Repository for Grafana/Kibana dashboards and Prometheus alerting rules
https://www.osism.tech
9 stars 11 forks source link

mySQL alert is constantly alarming - proposal for finetuning #13

Closed linwalth closed 3 years ago

linwalth commented 3 years ago

We are running the dashboards and alert rules on our OSISM infrastructure, and they are working well.

It has caught my eye, though, that the mySQL alert rules are constantly firing due to "slow queries" which is based on the alert firing as soon as there has ever been more than one slow query.

https://github.com/osism/kolla-operations/blob/main/prometheus/mysql.rules#L63

Wouldn't it be better to base this on rate over time?

like so:

expr: "rate(mysql_global_status_slow_queries[5m]) > 0"

berendt commented 3 years ago

Makes sense to me. Are you preparing a PR?

linwalth commented 3 years ago

i can do that.

linwalth commented 3 years ago

PR has been created.

berendt commented 3 years ago

Closed by #14.