Managed Prometheus Alerts¶

You can manage your Prometheus alerts with the Robusta Alerts UI, instead of managing them as PrometheusRule YAMLs in-cluster.

This lets your team create and customize Prometheus alerts with templates, without needing to know PromQL.

This guide covers how it works, and also the steps involved to use this feature.

How it works¶

Robusta manages the list of alerts in the Robusta UI and syncs them to your clusters as managed PrometheusRule files (CRDs).

You can disable/enable individual alerts and change thresholds via the UI.

Robusta uses the enabledManagedConfiguration Helm value to determine whether to sync alerts from the Alerts UI to your cluster, by default in generated_values.yaml it is set to true.

enabledManagedConfiguration: true

Important to mention that Robusta stores its managed rules in PrometheusRules custom resources that start with robusta-prometheus.rules--. If left in the cluster, you might have double alerts.

Enable Alerts UI for your cluster¶

Choose the appropriate instructions below, based on whether you use the Prometheus bundled with Robusta or your own Prometheus.

Robusta’s Prometheus

Make sure that the enabledManagedConfiguration value is set to true in generated_values.yaml:

enabledManagedConfiguration: true

Then perform a Helm Upgrade.

External Prometheus

First, ensure you have the Prometheus operator installed by running the following command:

kubectl get crd | grep prometheus

To make sure Prometheus picks up Robusta's rule files, add the following to the Kube Prometheus Stack configuration:

prometheus: # collect rules from all namespaces and ignore label filters
    ruleNamespaceSelector: {}
    ruleSelector: {}
    ruleSelectorNilUsesHelmValues: false

Finally, make sure that the following snippet is in Robusta’s Helm values file named generated_values.yaml:

enabledManagedConfiguration: true # Enables managed alerts

Then perform a Helm Upgrade.

Disabling the Feature¶

If you choose to stop using the Robusta Alerts UI, set enabledManagedConfiguration: false in generated_values.yaml and do a Helm Upgrade.

Modify and run the following command for all the Robusta rule custom resources present in your cluster.

kubectl delete prometheusrules.monitoring.coreos.com robusta-prometheus.rules--Value -n NameSpace