Setting Alerts

Set alerts to monitor the condition of an HPC cluster.

Setting an Alert for Nodes that are Offline

Set an alert so that an email is sent to a specified email address when nodes go offline.

You can set an alert for a specific node, a subset of nodes, or all nodes in an HPC cluster.

  1. Click the Monitor tab.
  2. Click the Alerts tab from the PBS Professional menu located on the left-hand side of the web page.
  3. Click Edit.
  4. For Email Address, enter the email address where the alert will be sent.
  5. Choose one of the following options to set an alert when nodes go offline:
    • Click on the name of the HPC cluster in the Available Nodes list to set an alert for all nodes in the cluster.
    • Expand the cluster to view the individual nodes in the Available Nodes list and click on the name of node to set an alert for just that node.
    The nodes that have been selected now appear in the Selected Nodes list.
    Tip: To remove a node offline alert, click on either the name of the HPC cluster or the node name in the Selected Nodes list.
  6. Choose one of the following options:
    • Click Subscribe to set an alert for the first time.
    • Click Update to update an existing alert.

Setting an Alert for CPU Utilization

Set an alert so that an email is sent to a specified email address when CPU utilization either drops below a specified percentage and/or rises above a specified percentage at a PBS cluster.

The percentage of CPU utilization for an HPC cluster is calculated as:

(Total NCPUs utilized in cluster *100)/Total available NCPUs in cluster

  1. Click the Monitor tab.
  2. Click the Alerts tab from the PBS Professional menu located on the left-hand side of the web page.
  3. Click Edit.
  4. For Email Address, enter the email address where the alert will be sent.
  5. To set an alert for when CPU utilization drops below a specified percentage:
    1. Enable Below for the cluster.
    2. Enter the percentage.

    Alert for Drop in CPU Utilization
    Tip: Enable Below for All to set an alert for all clusters.
    Figure 1. Alert for Drop in CPU Utilization
  6. To set an alert for when CPU utilization rises above a specified percentage:
    1. Enable Above for the cluster.
    2. Enter the percentage.

    Alert for Rise in CPU Utilization
    Figure 2. Alert for Rise in CPU Utilization
    Tip: Enable Above for All to set an alert for all clusters.
  7. Choose one of the following options:
    • Click Subscribe to set an alert for the first time.
    • Click Update to update an existing alert.
    Tip: To remove a CPU utilization alert disable the Below or Above check box for a cluster.

Setting an Alert for when a Cluster is Unavailable

Set an alert so that an email is sent to a specified email address when a cluster becomes unavailable.

A cluster can become unavailable when the PBS Professional Server is down or if the system on which the PBS Professional Server is hosted is unreachable.

  1. Click the Monitor tab.
  2. Click the Alerts tab from the PBS Professional menu located on the left-hand side of the web page.
  3. Click Edit.
  4. For Email Address, enter the email address where the alert will be sent.
  5. Enable the check box next to the cluster name to set an alert for when a cluster becomes unavailable:

    Alert for when the Cluster becomes Unavailable
    Figure 3. Alert for when the Cluster becomes Unavailable
    Tip: To remove the alert disable the check box next to the name of the cluster.
  6. Choose one of the following options:
    • Click Subscribe to set an alert for the first time.
    • Click Update to update an existing alert.

Unsubscribe from All Alerts

Remove all alerts by unsubscribing.

Warning: Unsubscribing from alerts will remove all alerts.
  1. Click the Monitor tab.
  2. Click the Alerts tab from the PBS Professional menu located on the left-hand side of the web page.
  3. Click Edit.
  4. Click Unsubscribe.
  5. Confirm the action by clicking Confirm.
    All alerts are removed.