Exchange Server 2010 : Monitoring Events, Services, Servers, and Resource Usage (part 2) - Monitoring Exchange Messaging Components & Using Performance Alerting

4/18/2012 3:54:39 PM

3. Monitoring Exchange Messaging Components

When you are troubleshooting or optimizing a server for performance, you can use performance monitoring to track the activities of Exchange messaging components. Performance Monitor graphically displays statistics for the set of performance parameters you've selected for display. These performance parameters are referred to as counters. Performance Monitor displays information for only the counters you're tracking. Thousands of counters are available, and these counters are organized into groupings called performance objects.

When you install Exchange Server 2010 on a computer, Performance Monitor is updated with a set of objects and counters for tracking Exchange performance. These objects and counters are registered during setup in the Win32 performance subsystem and the Windows registry. You'll find dozens of related performance objects for everything from the Microsoft Exchange Availability Service to the Microsoft Exchange Journaling Agent to Microsoft Exchange Outlook Web App.

You can select which counters you want to monitor by completing the following steps:

In the Exchange Management Console, select the Toolbox node, and then double-click Performance Monitor. This opens the Exchange Server Performance Monitor, which is a custom console with several preselected objects and counters. You also can start Performance Monitor by selecting the related option on the Administrative Tools menu.
Select the System Monitor entry in the left pane, as shown in Figure 3.
The Performance Monitor tool has several views and view types. Ensure that you are viewing current activity by clicking View Current Activity on the toolbar or pressing Ctrl+T. You can switch between the view types (Line, Histogram Bar, and Report) by clicking the Change Graph Type button or pressing Ctrl+G.
To add counters, click Add on the toolbar. This displays the Add Counters dialog box shown in Figure 4.

Figure 3. Track performance objects and counters to monitor server performance.

Figure 4. Select the counters you want to monitor.
In the Select Counters From Computer combo box, enter the Universal Naming Convention (UNC) name of the Exchange server you want to work with, such as \\MailServer18, or leave it at the default setting of <Local computer> to work with the local computer.
Note:

You need to be at least a member of the Performance Monitor Users group in the domain or the local computer to perform remote monitoring. When you use performance logging, you need to be at least a member of the Performance Log Users group in the domain or the local computer to work with performance logs on remote computers.
In the Available Counters panel, Performance Objects are listed alphabetically. If you select an object entry by clicking it, all related counters are selected. If you expand an object entry, you can see all the related counters and you can then select individual counters by clicking them. For example, you can expand the entry for the MSExchangeIS object and then select the Active User Count, Client RPCs Attempted/Sec, Connection Count, and RPC Operations/Sec counters.
When you select an object or any of its counters, you see the related instances, if any. Choose All Instances to select all counter instances for monitoring. Or select one or more counter instances to monitor. For example, you can select the Exchange ActiveSync or Exchange Content Indexing instances of MSExchangeIS Client.
When you've selected an object or a group of counters for an object as well as the object instances, click Add to add the counters to the graph. Repeat steps 5 through 8 to add other performance parameters.
Click OK when you're finished adding counters. You can delete counters later by clicking their entry in the lower portion of the Performance window and then clicking Delete.

4. Using Performance Alerting

In Windows Server 2008, Data Collector Sets are used to collect performance data. When you configure Data Collector Sets to alert you when specific criteria are met, you are using performance alerting. Windows performance alerting provides a fully automated method for monitoring server performance and reporting when certain performance thresholds are reached. You can use performance alerting to track the following:

Memory usage
CPU utilization
Disk usage
Messaging components

Using notifications, you can then provide automatic notification when a server exceeds a threshold value.

4.1. Tracking Memory Usage

Physical and virtual memory is critical to normal system operation. When a server runs low on memory, system performance can suffer and message processing can grind to a halt. To counter this problem, you should configure performance alerting to watch memory usage. You can then increase the amount of virtual memory available on the server or add more random access memory (RAM) as needed.

You configure a memory alert by completing the following steps:

In the Exchange Management Console, click the Toolbox node, and then double-click Performance Monitor. This opens the Exchange Server Performance Monitor.
Expand the Performance Logs And Alerts and the Data Collector Sets nodes, and then select User Defined. You should see a list of current alerts (if any) in the right pane.
Right-click the User-Defined node in the left pane, point to New, and then choose Data Collector Set.
In the Create New Data Collector Set Wizard, type a name for the Data Collector Set, such as Memory Usage Alert. Select the Create Manually option and then click Next.
On the What Type Of Data Do You Want To Include page, the Create Data Logs option is selected by default. Select the Performance Counter Alert option and then click Next.
On the Which Performance Counters Would You Like To Monitor page, click Add. This displays the Add Counter dialog box. Because you are configuring memory alerts, expand the Memory object in the Performance Object list. Select Available Mbytes by clicking it, and then click Add.
Expand the Paging File object in the Performance Object list. Click %Usage. In the Instances Of Selected Object panel, select _Total, and then click Add. Click OK.
On the Which Performance Counters Would You Like To Monitor page, you'll see the counters you've added. In the Performance Counters panel, select Available Mbytes (as shown in Figure 5 ), set the Alert When list to Below, and then enter a Limit value that is approximately 5 percent of the total physical memory (RAM) on the server for which you are configuring alerting. For example, if the server has 2 GB of RAM, you set the value to 100 MB to alert you when the server is running low on available memory.
In the Performance Counters panel, select %Usage. Set the Alert When list to Above, and then type 98 as the Limit value. This ensures that you are alerted when more than 98 percent of the paging file is being used.
Click Next and then click Finish. This saves the Data Collector Set and closes the wizard.
In the left pane, select the related Data Collector Set and then double-click the data collector for the alert in the main pane. This displays the data collector Properties dialog box.

Figure 5. Configure the alert threshold.
On the Alerts tab, use the Sample Interval options to set a sample interval. (See Figure 6 .) The sample interval specifies when new data is collected. Don't sample too frequently, however, because you'll use system resources and might cause the server to seem unresponsive. By default, Performance Monitor checks the values of the configured counters every 15 seconds. A better value might be once every 10 to 30 minutes. Generally, you'll want to track performance periodically over several hours at a minimum and during a variety of usage conditions.

Figure 6. Set the sample interval.
If you want to log an event rather than be alerted every time an alert limit is reached, on the Alert Action tab, select the Log An Entry In The Application Event Log check box. Selecting this option ensures that an event is logged when the alert occurs but does not alert you via the console.
Click OK to close the Properties dialog box. By default, alerting is configured to start manually. To start alerting, select the User Defined node in the left pane, click the alert in the main pane to select it, and then click the Start button on the toolbar.

To manage an alert, select the User Defined node in the left pane, right-click the alert in the main pane, and then select one of the following options:

Delete Deletes the alert
Properties Displays the alert's Properties dialog box
Start Activates alerting
Stop Halts alerting

4.2. Tracking CPU Utilization

You can use a CPU utilization alert to track the usage of a server's CPUs. When CPU utilization is too high, Exchange Server can't effectively process messages or manage other critical functions. As a result, performance can suffer greatly. For example, CPU utilization at 100 percent for an extended period of time can be an indicator of serious problems on a server. To recover, you might need to use Task Manager to end the process or processes with high CPU utilization, or you might need to take other corrective actions to resolve the problem, such as closing applications you are running while logged on to the server.

You'll also want to closely track process threads that are waiting to execute. A relatively high number of waiting threads can be an indicator that a server's processors need to be upgraded.

You configure a CPU utilization alert by completing the following steps:

In the Exchange Management Console, click the Toolbox node, and then double-click Performance Monitor. This opens the Exchange Server Performance Monitor.
Expand the Performance Logs And Alerts and Data Collector Sets nodes, and then select User Defined. You should see a list of current alerts (if any) in the right pane.
Right-click the User-Defined node in the left pane, point to New, and then choose Data Collector Set.
In the Create New Data Collector Set Wizard, type a name for the Data Collector Set, such as CPU Utilization Alert. Select the Create Manually option and then click Next.
On the What Type Of Data Do You Want To Include page, the Create Data Logs option is selected by default. Select the Performance Counter Alert option and then click Next.
On the Which Performance Counters Would You Like To Monitor page, click Add. This displays the Add Counter dialog box. Because you are configuring CPU alerts, expand the Processor object in the Performance Object list. Click % Processor Time. In the Instances Of Selected Object panel, select _Total and then click Add.
Expand the System object in the Performance Object list. Click Processor Queue Length, and then click Add. Click OK.
On the Which Performance Counters Would You Like To Monitor page, you'll see the counters you've added. Select % Processor Time. Then set the Alert When list to Above, and type 98 as the Limit value. This ensures that you are alerted when processor utilization is more than 98 percent.
In the Performance Counters panel, select Processor Queue Length. Then set the Alert When list to Above, and type 3 as the Limit value. This ensures that you are alerted when more than three processes are waiting to execute, which can be an indicator that a server's processors need to be upgraded.
Click Next and then click Finish. This saves the Data Collector Set and closes the wizard.
Finish configuring the alert by following steps 11 through 14 under Section 4.1.

4.3. Tracking Disk Usage

Exchange Server uses disk space for data storage, logging, tracking, and virtual memory. To ensure there is always ample disk space available, Exchange Server monitors free disk space. If free disk space drops below specific thresholds, Exchange will gracefully shut itself down. When Exchange is in this state, it is likely that data could get lost. To prevent serious problems, you should monitor free disk space closely on all drives used by Exchange Server.

You'll also want to track closely the number of system requests that are waiting for disk access. A relatively high value for a particular disk can affect server performance and is also a good indicator that a disk is being overutilized or that there may be some problem with the disk. To resolve this problem, you'll want to try to shift part of the disk's workload to other disks, such as by moving databases, logs or both.

You configure disk usage alerting by completing the following steps:

In the Exchange Management Console, click the Toolbox node, and then double-click Performance Monitor. This opens the Exchange Server Performance Monitor.
Expand the Performance Logs And Alerts and Data Collector Sets nodes, and then select User Defined. You should see a list of current alerts (if any) in the right pane.
Right-click the User-Defined node in the left pane, point to New, and then choose Data Collector Set.
In the Create New Data Collector Set Wizard, type a name for the Data Collector Set, such as Disk Usage Alert . Select the Create Manually option and then click Next.
On the What Type Of Data Do You Want To Include page, the Create Data Logs option is selected by default. Select the Performance Counter Alert option and then click Next.
On the Which Performance Counters Would You Like To Monitor page, click Add. This displays the Add Counter dialog box. Because you are configuring disk alerts, expand the LogicalDisk object in the Performance Object list. Click % Free Space. In the Instances Of Selected Object panel, select all individual logical disk instances except _Total and then click Add.
Expand the PhysicalDisk object in the Performance Object list. Click Current Disk Queue Length. In the Instances Of Selected Object panel, select all individual physical disk instances except _Total and then click Add. Click OK.
On the Which Performance Counters Would You Like To Monitor page, you'll see the counters you've added. Select the first logical disk instance, set the Alert When list to Below, and then type 5 as the Limit value. This ensures that you are alerted when available free space is less than 5 percent. Repeat this procedure for each logical disk.
In the Performance Counters panel, select the first physical disk instance, set the Alert When list to Above, and then type 2 as the Limit value. This ensures that you are alerted when more than two system requests are waiting for disk access. Repeat this procedure for each physical disk.
Click Next and then click Finish. This saves the Data Collector Set and closes the wizard.
Finish configuring the alert by following steps 11 through 14 under Section 14.1.