Configuring a Win 2003-based server

To notify you when the computer is low on resources. Programs in Windows Server 2003 define the performance data that it collects from three aspects of objects, counters, and instances. A performance object is any resource, program, or service that can be measured. You can use System Monitor and Performance Logs and Alerts to select performance objects, counters, and instances to collect and display performance data for system components or installed software.

You can set an alert on the counter so that when the value of the selected counter exceeds or falls below the specified setting, you can send a message, start the program, or start the log.

How to add counters to System Monitor

Click Start, point to Administrative Tools, and then click Performance. If you select an object on a remote computer, there may be a brief delay when System Monitor refreshes the list to reflect the existing objects in the computer.

Right-click the details pane of System Monitor, and then click Add Counters.

To monitor any computers that are running the monitoring console, click Use local computer counters. If you want to monitor a specific computer regardless of where the monitoring console is running, click Select counters from computer, and then specify the computer name.

Under Performance objects, click the object that you want to monitor. By default, the Processor object is selected.

Click All counters, or click Select a counter from the list, and then click a list item.

Click Add.

How to define counters and alert thresholds

Click Start, point to Administrative Tools, and then click Performance.

Double-click Performance Logs and Alerts, and then tap Alerts.

Right-click the alert, click New Alert Settings, type a name for the alert, and then click OK.

Click the General tab, type a descriptive comment for the alert, and then click Add.

For each counter or group of counters that you want to add to the log, perform the following steps:

To monitor counters from the computer running the Performance Logs and Alerts service, click Use local computer counters.

Or, to monitor counters from a specific computer, regardless of where the service is running, click Select counters from computer, and then specify the name of the computer that you want to monitor.

Under Performance objects, select the objects you want to monitor.

Then select one or more counters to monitor.

To monitor all instances of the selected counter, click All instances. Note that the binary log can contain instances that are not available at the start of the log but subsequently become available. Or, to monitor a specific instance of a selected counter, click Select an instance from a list, and then click one or more instances to monitor.

Click Add, and then click Close.

In the alarm that will be set, if the value is box, click less than or greater than. In the Limit to box, specify the value that triggers the alert.

In the Data sampling interval box, specify the value and unit of measure for the update interval.

Click the Actions tab to determine the action that is raised when an alert is triggered. To record an alert in the Event Viewer log, click to select the entry in the Application event log check box.

To send an alert message to your computer, click to select the Send network information to check box, and then type the name of the computer that you want to receive the alert information for.

Click to select the Start Performance data log check box to start the log file.

If you want to start the program when the alert condition is reached, click "Execute this program." You can either type the path of the program directly or click Browse to manually select the program you want to use.

Click the Schedule tab, and then configure the appropriate settings to start or stop logging manually or at the scheduled time.

How to select the data to monitor

Start monitoring the activities of the following components in the following order:

Memory

Processor

Disk

Internet

The following list shows the minimum counters recommended for server monitoring. Note that when you examine a specific resource, you should include additional counters for the related performance object.

Component: Disk

Performance aspects being monitored: usage

Counters to monitor:

Physical Disk/Disk Reads/sec (Read/sec)

Physical Disk/Disk Writes/sec (Write/sec)

Logical Disk/% free space (available)

You must carefully interpret the disk Time counter. Because the _total instance of the counter may

Does not accurately reflect usage on multiple-disk computers, it is important to use the% Idle time counter at the same time. Note that these counters cannot display values that exceed 100%.

Component: Disk

Performance aspects being monitored: bottlenecks

Counters to monitor: Physical Disk/Avg.Disk Queue Length (average Disk Queue Length) (all instances)

Component: Memory

Performance aspects being monitored: usage

Counters to monitor:

Memory/Available Bytes (free bytes)

Memory/cache Bytes (cache bytes)

Component: Memory

Performance aspects being monitored: bottlenecks or vulnerabilities

Counters to monitor:

Memory/pages/sec (Pages/sec)

Memory/page reads/sec (pages Read/sec)

Memory/Transition FAULTS/SEC (Transmission error/sec)

Memory/Pool Paged Bytes (paged memory pool bytes)

Memory/Pool Nonpaged Bytes (nonpaged memory Pool bytes)

The following components, although not specialized memory object counters, are useful for memory analysis:

Paging File/% Usage (usage) object (all instances)

Cache/Data Map Hits (Datastore hit)%

Server/Pool Paged Bytes (paged memory pool bytes)

Server/Pool Nonpaged Bytes (nonpaged memory Pool bytes)

Component: Network

Performance aspects that are being monitored: throughput

Counters to monitor:

Protocol transfer counters (varies by network protocol), for TCP/IP, including:

Network interface/Bytes total/sec (total bytes/sec)

Network interface/PACKETS/SEC (Packets/sec)

Server/Bytes total/sec (total bytes/sec) or Server/Bytes transmitted/sec (Bytes Sent/sec)

Server/Bytes received/sec (Bytes Received/sec)

You may need to monitor network and server throughput for other objects as described in the "Monitoring Network Activity" section

Component: Processor

Performance aspects being monitored: usage

Counters to monitor: Processor/% Processor time (processor times) (all instances)

Component: Processor

Performance aspects being monitored: bottlenecks

Counters to monitor:

System/Processor Queue Length (Processor queue lengths) (all instances)

Processor/Interrupts/sec (Interrupts/sec)

System/Context Switches/sec (Context Switch/sec)