For many organizations, Exchange is the communications lifeblood without which business comes to a standstill. Without it, customers become detached from support and sales personnel. As such, keeping these Exchange servers in good health is of prime importance.
Exchange includes some built-in resource monitors that will alert you when some component of your server has gone awry and needs attention. To get started, open the Exchange System Manager. Once open, expand the Administrative Groups > First Administrative Group > Servers branch and right-click the server for which you would like to configure monitoring. From the resulting shortcut menu, choose the Properties option and, from the Properties window, select the Monitoring tab.
With this tab, you can configure a number of different monitors, including:
- Available virtual memory: Determine at what point warning and critical states are initiated for virtual memory. When virtual memory on your Exchange system gets too low, Exchange may stop working. Use this option to alert an administrator via e-mail when available virtual memory drops below a level that specify. With this option, you need to tell the monitoring service for how many minutes virtual memory needs to be below the target percentage before an administrator is notified.
- CPU utilization: Consistently high CPU utilization on your Exchange server may indicate a need to add additional resources to your Exchange environment. Again, you can tell the Exchange monitor at what point warning and critical states are initiated and an administrator can be notified when these targets are hit. Like virtual memory, CPU utilization needs to stay peaked beyond your target percentages for a certain number of minutes (that you specify) before any action is taken.
- Free disk space: Allows you to define warning and critical states that are based on the amount of disk space available on your Exchange system. Exchange is very disk intensive, and having enough space to grow is critically important.
- SMTP queue growth: If a queue continually grows, it can signal that one of your connectors is on the fritz and not sending messages. For this monitoring object, you can specify the number of minutes of continual growth after which the queue is placed into a warning or critical state.
- Windows 2000 service: Allows you to watch specific Windows services and throw an error and notify an administrator when a service stops.
- X.400 queue growth: Same as SMTP queue growth, but watches the X.400 queue.
To add one of these items, click the Add button. From the Add Resource window, choose the resource from the list above that you want to monitor. For any selection, you will have to make some choices about warning levels and timing. (For example, if you choose CPU utilization, you're asked to provide a duration and CPU thresholds.) Click OK when you’re finished.
Now, in order to notify an administrator via e-mail when one of your warning thresholds is violated, you need to tell the Exchange System Manager who should get these reports. Here's how:
- In the manager, go to Tools > Monitoring and Status > Notifications.
- Right-click Notifications and choose New > E-mail Notification
- In the resulting window, provide the e-mail address of an administrator and decide under which condition--warning or critical--you want to email the administrator. You can set up separate notifications for each state.
- Also, specify the name of the mail server that should be used to send the notification. If possible, use a different server that the monitored system to send the message.
- Click OK when you’re finished.
