This example shows how to create alert rules for Prometheus, using PrometheusRule CRD. Example defines two rules for two different type of event that can occur.
First rule fires an alert if application MyApp has disappered from Prometheus target discovery. Second rule fires an alert if application MyApp's failure rate measured on a time window of 2 minutes was higher than 10% in the last 10 minutes.
Alert rule conditions are defined based on PromQL expressions, with expr field
in your rule definitions. To learn more about PromQL please refer to Prometheus
documentation
-
Run furyctl to get packages:
furyctl install -
Add new rules based on conditions expressed with PromQL, add annotations to inform user about alert and specify an interval with
forfield. -
Run
make buildto see output of kustomize with your modifications. -
Once you're satisfied with generated output run
make deployto deploy it on your cluster.