Performance Counters

This section discusses the performance chart (in the details pane for a cluster or member) and default performance counters that are charted in Application Center.

The Performance View

You can view the performance counter data for a cluster or member for several time frames: 15 minutes (the last 15 minutes), 1 hour (the last hour), 1 day (the last 24 hours), 1 week (the last 7 days), and 3 months (the last 3 months, including the current month).

Application Center uses two types of data manipulation: data summarization and data aggregation. Data summarization refers to how the data is summed up for logging. Aggregation refers to how the data is summed up for the performance chart.

 Note   The different time frames do not affect the logging of event and performance counter data. This data is collected and stored, if Application Center Event and Performance Logging is enabled for your cluster. Application Center updates the performance chart every 10 seconds.

Time Periods

The following table lists the time frames that you can use to view performance counters. These time frames also represent different data resolutions; that is, the shorter time periods collect data more often.

Time period	Data interval	Data summarized	Typical usage
Fifteen Minutes	10 seconds	Every minute	Monitoring and troubleshooting performance issues and testing performance adjustments.
Two Hours	1 minute	Every 15 minutes	Monitoring and troubleshooting performance issues and testing performance adjustments.
One Day	15 minutes	Every 2 hours (even hours)	Gathering usage and availability data as the day proceeds. Most monitoring is done by using this time frame.
One Week	2 hours	Every day (at midnight, local time)	Gathering usage and availability data for the whole week, to date. This time frame is useful for analyzing performance and load trends.
Three Months	1 day	No summarization	Gathering usage and availability data for the last three months, to date. This time frame is useful for analyzing usage and load trends and high-level capacity planning, such as hardware upgrades and purchases.

Default Performance Counters

The following table presents the default performance counters (listed alphabetically) that are used by Application Center and includes their name, a short description, the units used, and how Application Center aggregates the the data in the performance chart. Aggregation methods include:

• Average—an average of the values gathered for the last time period.

• Sum—the total for the values over the time period.

• Min Value—the smallest value over the time period.

• Max Value—the largest value over the time period.

• Last—the last value over the time period.

• None—no aggregation was done. This applies to the Fifteen Minutes view.

 Notes

For example, ASP Requests Queued shows an aggregation for server:avg - cluster:sum. This means that when viewing the counter for a member at the Two Hours resolution, the value given for a member is the average number of ASP requests queued over the last minute (the collection interval). When viewed for a cluster, the value represents the summed values of the individual members' averages over the last minute (the collection interval).

For example, for a cluster of 8 members and the member values for this counter are 5, 6, 2, 8, 5, 5, 5, and 7, the cluster value is 43, which is the sum of the individual average values. This means that over the last minute, there were a total of 43 ASP requests in the queue every second over the entire cluster.

Counter	Description	Units	Aggregation
ASP Errors	Number of errors generated by ASP applications per second.	Errors/sec	server:avg - cluster:sum
ASP Requests Queued	The number of requests waiting for service from the queue. This number should be small, except during heavy traffic periods. Large numbers of queued requests indicate a performance bottleneck.	Requests	server:avg - cluster:sum
ASP Requests Queued (max)	The number of requests waiting for service from the queue. This number should be small, except during heavy traffic periods. Large numbers of queued requests indicate a performance bottleneck.	Requests	server:max value - cluster:max value
ASP Requests	Number of requests executed per second.	Requests/second	server:avg - cluster:sum
ASP Request execution time	Number of milliseconds that it took the most recent ASP request to complete.	milliseconds	server:avg - cluster:avg
ASP Request Wait Time	Number of milliseconds that the most recent ASP request was waiting in the queue.	Milliseconds	server:avg - cluster:avg
Memory Available	The amount of physical memory available to processes running on the member. It is calculated by summing space on the Zeroed, Free, and Stand by memory lists. This figure should be at least five percent of total memory at all times.	Bytes	server:avg - cluster:avg
Memory Page Faults	The number of times, per second, that the computer reads the page file on the disk or from memory that is not assigned to the working set. Most CPUs can handle large numbers of page faults without consequence; however, if disk reads are high, there might be performance degradation.	Faults/sec	server:avg - cluster:avg
Physical Disk Queue Length	The number of requests outstanding on the disk at the time the performance data is collected, which includes requests in service at the time of the reading. Multi-spindle disk devices can have multiple requests active at one time, but other concurrent requests are waiting for service. This counter might reflect a transitory high or low queue length, but if there is a sustained load on the disk drive, it is likely that this will be consistently high. Requests are experiencing delays proportional to the length of this queue minus the number of spindles on the disks. For good performance, this difference should average less than two.	Requests (interger)	server:avg - cluster:sum
Inetinfo Private Bytes	The number of bytes of memory taken up by a particular process (Inetinfo in this case, which is part of IIS).	Bytes	server:avg - cluster:avg
Processor Utilization	Peak utilization can be 100 percent, but utilization should not be sustained at high level (for example, greater than 90 percent). Processor bottlenecks are characterized by a high value for Processor:% Processor Time while the network adapter card remains well below capacity. On a multi-processor computer, the System:% Total Processor Time counter can mask a processor bottleneck because it represents the the total processor time for all processors divided by the number of processors. If one processor is overloaded and creating a bottleneck while the other processors are idle, System:% Total Processor Time might be quite low, but the overworked single processor is blocking the computer. On a multi-processor computer, examine the Processor:% Processor Time counter determine up any imbalance.	Percentage	server:avg - cluster:avg
Processor Privileged Time	The percentage of non-idle processor time spent in privileged mode. Privileged mode is a processing mode designed for operating system components and hardware-manipulating drivers. It allows direct access to hardware and all memory. The alternative, user mode, is a restricted processing mode designed for applications, environment subsystems, and integral subsystems. The operating system switches application threads to privileged mode to gain access to operating system services. This includes time spent servicing interrupts and deferred procedure calls (DPCs). A high rate of privileged time might be caused by a large number of interrupts generated by a failing device. This counter displays the average busy time as a percentage of the sample time.	Percentage	server:avg - cluster:avg
Processor User Time	The percentage of non-idle processor time spent in user mode. User mode is a restricted processing mode designed for applications, environment subsystems, and integral subsystems. The alternative, privileged mode, is designed for operating system components and allows direct access to hardware and all memory. The operating system switches application threads to privileged mode to gain access to operating system services. This counter displays the average busy time as a percentage of the sample time.	Percentage	server:avg - cluster:avg
Log Database Total Memory	The total amount of dynamic memory that Application Center Events and Performance Logging is currently consuming.	Kilobytes	server:avg - cluster:avg
Context Switches	Indicates excessive locking in code, perhaps contention for resources. If this value is too high, add another computer or check with Microsoft Support Online for the latest patches.	Switches/sec	server:avg - cluster:avg
TCP Connections Established	The number of times TCP connections have made a direct transition to the Syn-sent state from the Closed state.	Connections (interger)	server:avg - cluster:sum
Web Service Current Connections	Number of current client connections to the W3Svc Service.	Connections (interger)	server:avg - cluster:sum
Web Service Get Requests	The number of HTTP requests that are using the GET method per second. GET is the most common method used on the Web.	Requests/sec	server:avg - cluster:sum
Web Service Bytes Total	The sum of Bytes Sent/sec and Bytes Received/sec. This is the total rate of bytes transferred by the W3Svc Service.	Bytes/sec	server:avg - cluster:sum
Web Service ISAPI Requests	The number of ISAPI extension requests that are being processed simultaneously by the W3Svc Service per second.	Requests/sec	server:avg - cluster:sum

Request Forwarding Performance Counters

The following table lists the request forwarding performance counters (listed alphabetically) that are used by Application Center. These counters are not available in the performance chart by default, but are included in the Windows 2000 Performance Viewer under Application Center HTTP Forwarding:

Counter name	Description
Total Requests	Total requests processed by the request forwarder filter.
Total Forwarded Requests	Total requests forwarded to another member.
Total Dynamic Requests	Total dynamic requests processed by the request forwarder filter. These requests are for files that have been configured as request forwarding exceptions in the Advanced Load Balancing settings in the cluster_name Properties dialog box. The files are typically static files.
Total Publishing Requests	Total publishing requests forwarded to the controller.
Total Coherent Session Requests	Total requests forwarded to an originating member.
Total Web Administration Requests	Total IIS Administration site requests forwarded to the controller.
Total Application Center Administration Requests	Total Application Center Administration site requests forwarded to the controller.
Total Failed Requests	Total requests that were forwarded but did not complete successfully.
Total Requests/sec	Total requests processed by the request forwarder filter, per second.
Total Forwarded Requests/sec	Total requests forwarded to another member, per second.
Total Dynamic Requests/sec	Total dynamic requests processed by the request forwarder filter, per second.
Total Publishing Requests/sec	Total publishing requests forwarded to the controller, per second.
Total Coherent Session Requests/sec	Total requests forwarded to an originating member, per second.
Total Web Administration Requests/sec	Total IIS Administration site requests forwarded to the controller, per second.
Total Application Center Administration Requests/sec	Total Application Center Administration site requests forwarded to the controller per second.
Total Failed Requests/sec	Total requests that were forwarded but did not complete successfully, per second.

Related Topics

• For more information about additional sample counters, see Samples Available.

Did you find this information useful? Please send your suggestions and comments about the documentation to acdocs@microsoft.com.