Working with Resource Alerts :: SAS(R) Environment Manager 2.3: User's Guide

Overview of Alerts

If you want to identify a type of event for notification or further action, you can create an alert. Alerts are a user-defined type of event that indicates a critical condition in a selected resource. When an alert occurs, it must be acknowledged, and alerts are listed until they are marked as being fixed. You can define escalation schemes to identify the actions that happen if an alert is not fixed within a specified time.

Alerts are logged by the agents and all events throughout the entire deployment are displayed in the Alert Center. To access the Alert Center, select Analyze

Event Center.

You can filter the alerts, for example, so that only the most recent ones or the ones of a specified type are displayed. Click the icon to acknowledge an alert. Select the check box next to an alert, and then click Fixed to fix the alert. You can also click on the entry in the Alert Definition column to display the Alert Details page, where you can view details about the alert and mark the alert as fixed (with comments).

Defining an Alert

To define an alert, follow these steps:

Select ResourcesBrowse or use a dashboard portlet to locate the resource for which you want to create an alert.
There are three icons on the left of the entry for the resource in the table . Click on the alert icon , which is on the right of the group. The Alerts page for the resource appears.
Click New to display the New Alert page.
In the Alert Properties area, specify the name and priority for the alert and whether it is active.
In the Condition Set area, specify the conditions that must be met in order for the alert to be triggered. You can specify up to three conditions for each alert. Use these fields to specify the condition that triggers the alert

Metric

specifies that the alert is triggered based on the value of a metric that is monitored for the resource. You can specify that the condition is based on comparison to a fixed value, a percentage of a value, or a change in value. If you want to create an alert for a metric that is not listed, you must first enable collection of that metric.

Update the metric collection settings for the resource type (choose Monitoring Defaults from the Manage page) or for the specific resource (click Metrics on the resource’s Monitor page).

Inventory Property

specifies that the alert is triggered based on a change in the value of a resource property (such as version number). This condition is available only for certain types of resources (such as platforms and SAS Metadata Servers).

Control Action

specifies that the alert is triggered when an action meets a specified condition (such as the action of stopping a failed server ). This condition is available only for servers that can be controlled through control actions. See Controlling Resources Using Control Actions for more information.

Events/Logs Level

specifies that the alert is triggered when a selected type of log entry (such as Error) and an optional accompanying text string appears in the log.

If you are defining an alert based on events from SAS server logs, the available values in this field do not match the logging levels available in SAS server logs. The four SAS Environment Manager levels must match the six levels in SAS server logs. Selecting Error in this field matches both the Fatal and Error levels in SAS server logs. Selecting Debug in this field matches both the Trace and Debug levels in SAS server logs.

Config changed

specifies that the alert is triggered when a configuration file changes (you can choose to specify the name of the configuration file).
In the Enable Action field, specify whether the alert is triggered only once, or periodically as long as the alert condition persists.
Click OK to define the alert and display the Alert Configuration page.
On the Alert Configuration page, you can specify an escalation scheme and identify the users and roles that should be notified when the alert occurs. To create an escalation scheme, see Defining an Escalation Scheme.
Click Return to Alert Definitions when you are finished.

Defining an Escalation Scheme

An escalation scheme is a series of actions that take place when an alert is not acknowledged or fixed within a certain period of time. An escalation scheme can be applied to multiple alerts. You can define an escalation scheme to perform any of these actions:

send an e-mail or SMS message
make an entry in a system log
issue an SNMP notification

To define an escalation scheme, select Manage

Escalation Schemes Configuration (in the Server Settings area).

For information about the information required when defining an escalation scheme, refer to the Help for the page.

Example: Defining an Alert for SAS Work Directory Space

This example provides information for setting up an alert to be triggered whenever the SAS Work directory reaches 90% of its capacity. The alert should be issued once every two hours until the condition is cleared. When the alert is triggered, users with the Operations role should be notified.

Locate the service SAS Home Directory 9.4 SAS work directory. The service is under the SAS Home Directory 9.4 server.
Navigate to the Resource Detail page for the service. On the Detail page, select AlertConfigure to display the Alert Configuration page. Click New to display the New Alert Configuration page.
Name the alert, select the priority, and specify that the alert should be active.
In the If Condition area, select the Metric radio button, then select Use Percent in the Metric field.
To specify 90% capacity, enter .9 in the absolute value field. To specify that the alert is triggered whenever the used capacity exceeds 90%, specify and select > (Greater than) from the comparison menu.
In the Enable Action(s) field, specify 1 for the number of times the alert is issued, 2 for the timer period, and select hours for the time period units. These values specify that the alert is issued one time every two hours while the alert conditions are met.
Click OK to define the alert and display the Configuration page for the new alert.
Select Notify Roles, and then select Add to List.
Select the check box beside Operations in the Roles list and use the arrow control to move the role to the Add Role Notification list.
Click OK to close the Role Selection page and then Return to Alert Definitions to complete the process of defining the alert.

Example: Defining an Alert for a SAS Server Log File

This example provides information for setting up an alert to be triggered whenever a warning message for the I/O Subsystem appears in the log of the SAS Metadata Server. The alert should be issued every time an error appears in the log.

Follow the procedure in Creating Events Based on SAS Server Logs to create an event from the SAS Metadata Server log file. Add the entry
```
level.warn.2=.*I/O Subsystem.*
```
to the sev_logtracker_plugin.properties file for the SAS Metadata Server.
Locate the server SASMeta – SAS Metadata Server in the Resource page.
Navigate to the Resource Detail page for the server. On the Detail page, select AlertConfigure to display the Alert Configuration page. Click New to display the New Alert Configuration page
Name the alert, select the priority, and specify that the alert should be active.
In the If Condition area, select the Event/Logs Level radio button, then select Warn in the Event/Logs Level field.

In the match substring field, enter I/O Subsystem. These values specify that an alert is issued whenever an event is found for a Warn message from the log containing the string “I/O Subsystem.”
In the Enable Action(s) area, select the Each time conditions are met radio button. This specifies that the alert is triggered each time an I/O Subsystem warning appears in the log.
Click OK to define the alert.

Example: Defining an Alert for Available Memory

This example provides information for setting up an alert to be triggered whenever the free memory on a SAS Web Application Server falls below 90% of 1.3 GB. The alert should be issued once every fifteen minutes until the condition is cleared.

Locate the server in the Resource page. The Resources page lists SAS Web Application Server using the server type SpringSource tc Runtime.
Navigate to the Resource Detail page for the server. On the Detail page, select AlertConfigure to display the Alert Configuration page. Click New to display the New Alert Configuration page
Name the alert, select the priority, and specify that the alert should be active.
In the If Condition area, select the Metric radio button, then select Heap Memory Free in the Metric field.
To specify the memory condition for the alert, enter 90 in the % field and then select 1.3 GB (Min Value). Select < (Less than) from the comparison menu.
In the Enable Action(s) field, specify 1 for the number of times the alert is issued, 15 for the timer period, and select minutes for the time period units. These values specify that the alert is issued one time every 15 minutes while the alert condition is met.
Click OK to define the alert.