Adding alerts for DC and node message latency

Add alerts for notification when datacenter and node messaging latencies reach unacceptable levels for your environment.

Add alerts for notification when datacenter and node messaging latencies reach unacceptable levels for your environment. Messaging latency metrics are available for DSE versions 5.1.0 and later.

Procedure

  1. Click Alerts from the main OpsCenter monitoring menu.
    The Active Alerts dialog appears.
  2. In the Active Alerts dialog, click Manage Alerts.
    The Manage Alerts Rules dialog appears.

    Manage Alert Rules dialog to select cluster and add alerts

  3. If applicable (multiple clusters are monitored in the OpsCenter instance), select the cluster from the Cluster list.
  4. Click Add Alert.
    The Add Alert dialog appears.
  5. In the Notify me when list, select Advanced > Cassandra > Datacenter Messaging Latency.

    DC message latency alert dialog

  6. Specify the notification criteria:
    1. Choose above or below. Default: above.
    2. Enter a value for ms/op (milliseconds per operation).
    3. Select the elapsed time for how long the condition should be in the defined state before alerting. Default: 5 minutes.
    4. Select the percentile. Choose: min (default), 50th, 90th, 99th, or max.
    5. Select the Source Datacenter.
    6. Select the notification frequency from Notify every. Default: 5 minutes.
    7. Click Save Alert.
      The alert is added to the Manage Alert Rules dialog.

      Populated alerts Disable, Edit, or Delete in Manage Alert Rules dialog

  7. Repeat this procedure for the Node Messaging Latency alert. Select Source Datacenter is not applicable to node-level alerts.

    Node message latency add alert dialog

    When the alert criteria is met, the fired alert is recorded in the Event Log and sent via any other alert mechanism configured for your environment, such as email or posted URLs.

    Fired alerts shown in the Event Log for cross-node latency