Alerts Page

The Alerts page allows you to toggle CockroachCloud alerts, send test alerts, and view the email recipients and alert history for your CockroachCloud Organization. To view the Alerts page, log in and click Alerts.

If alerts are enabled, CockroachCloud will send alerts to specified email receipients when the following usage metrics are detected:

Storage Utilization:

  • Cluster-wide available disk capacity is 20% or less.
  • Node-level available disk capacity is 10% or less.

CPU Utilization:

  • Cluster-wide CPU usage is 80% or greater on average for at least 60 minutes.
  • Node-level CPU usage is 90% or greater on average for at least 90 minutes.

Memory Utilization:

  • Cluster-wide available memory is 20% or less on average for at least 60 minutes.
  • Node-level available memory is 10% or less on average for at least 90 minutes.

If you receive an alert repeatedly, you may need to optimize your workload or scale your CockroachCloud cluster.

Tip:

When scaling your cluster, we recommend first scaling VMs to include more than 2 vCPUs each. If this doesn't sufficiently improve performance, then add more nodes.

Configure alerts

To enable alerts:

  • Under Configure CockroachCloud alerts, toggle the Alerts are on switch.
  • Under Add Email Recipients, add at least one email address and click Add.

Note that alerts are enabled for all clusters in your Organization. To filter alerts on specific clusters, you can use an email alias to send alerts to a monitoring tool such as Alertmanager or PagerDuty.

To send a test alert:

  1. Under Send test email, select a cluster and an alert type.
  2. Add at least one email address and click Add.
  3. Click Send test email.

Repeated alerts

If the condition triggering an alert does not change, the alert will repeat at a specific cadence:

Storage Utilization:

  • Cluster-wide alerts: every 60 minutes
  • Node-level alerts: every 60 minutes

CPU Utilization:

  • Cluster-wide alerts: every 60 minutes
  • Node-level alerts: every 90 minutes

Memory Utilization:

  • Cluster-wide alerts: every 60 minutes
  • Node-level alerts: every 90 minutes
YesYes NoNo