34 Configuring Monitoring for Enterprise Manager High Availability

Configuration With Cloud Control

Cloud Control comes preconfigured with a series of default rules to monitor many common targets. These rules can be extended to monitor the Cloud Control infrastructure as well as the other targets on your network to meet specific monitoring needs.

Console Warnings, Alerts, and Notifications

The following list is a set of recommendations that extend the default monitoring performed by Enterprise Manager. Use the Incident Rules link to adjust the default rules provided on the Configuration/Rules page:

Ensure the Agent Unreachable rule is set to alert on all Management Agents unreachable and Management Agents clear errors.
Ensure the Repository Operations Availability rule is set to notify on any unreachable problems with the Management Service or Management Repository nodes. Also modify this rule to alert on the Targets Not Providing Data condition and any database alerts that are detected against the database serving as the Management Repository.

Modify the Agent Upload Problems Rule to alert when the Management Service status has hit a warning or clear threshold.

Configure Additional Error Reporting Mechanisms

Enterprise Manager provides error reporting mechanisms through e-mail notifications, PL/SQL packages, and SNMP alerts. Configure these mechanisms based on the infrastructure of the production site. If using e-mail for notifications, configure the incident rule through the Cloud Control console to notify administrators using multiple SMTP servers if they are available. This can be done by modifying the default e-mail server setting on the Notification Methods option under Setup.

Component Backup

Backup procedures for the database are well established standards. Configure backup for the Management Repository using the RMAN interface provided in the Cloud Control console. Refer to the RMAN documentation or the Maximum Availability architecture document for detailed implementation instructions.

In addition to the Management Repository, the Management Service and Management Agent should also have regular backups. Backups should be performed after any configuration change.

Troubleshooting

In the event of a problem with Cloud Control, the starting point for any diagnostic effort is the console itself. The Management System tab provides access to an overview of all Management Service operations and current alerts. Other pages summarize the health of Management Service processes and logged errors These pages are useful for determining the causes of any performance problems as the summary page shows at a historical view of the amount of files waiting to be loaded to the Management Repository and the amount of work waiting to be completed by Management Agents.

Upload Delay for Monitoring Data

When assessing the health and availability of targets through the Cloud Control console, information is slow to appear in the UI, especially after a Management Service outage. The state of a target in the Cloud Control console may be delayed after a state change on the monitored host. Use the Management System page to gauge backlog for pending files to be processed.

Notification Delay of Target State Change

The model used by the Management Agent to assess the state of health for any particular monitored target is poll based. Management Agents immediately post a notification to the Management Service as soon as a change in state is detected. This infers that there is some potential delay for the Management Agent to actually detect a change in state.