Windows Performance Counters

< a >windows Performance Counter Analysis

Object	Counter	Analysis
Processor	%processor time	Recommended threshold Value 85%
Memory	Available bytes	The recommended threshold value is less than 4MB to add memory; In addition, it is recommended that you have at least 10% of the physical memory value
Pages reads/sec	Page reads/sec refers to the number of times a disk is read to parse a hard page error, and if the value continues to be large, it indicates a possible low memory recommended threshold of 30 (5). ), a large value indicates that disk reads rather than cache reads
Pages writes/sec	Page writes/sec refers to the number of times a page is written to disk in order to free up physical memory space
Pages input/sec	Page input/sec refers to the number of pages that are read from the disk to resolve a paging error
Pages output/sec	Pages Output/sec is the number of pages written to disk to free physical memory space if the value is much greater than the pages input/sec, there may be a memory leak
Pages/sec	Pages/sec refers to the number of page recommended thresholds that are read from or written to disk to resolve hard page errors 20
Network interface (for TCP/IP)	Bytes received/sec	The data combined with bytes total/sec look
Bytes sent/sec	The data combined with bytes total/sec look
Bytes total/sec	Recommended do not exceed 50% of bandwidth
Packets/sec	Based on the actual amount of data, no recommended thresholds, the data combined with bytes total/sec look
Physical Disk	Disk reads/sec	Depending on the size of the hard disk manufacturer, check the specified transfer speed of the disk to verify that the speed does not exceed the specification
Disk writes/sec	Depending on the size of the hard disk manufacturer, check the specified transfer speed of the disk to verify that the speed does not exceed the specification and that the above two values are added, should be less than the maximum capacity of the disk device
%disk time	Recommended threshold Value 90%
Current Disk Queue Length AVG. Disk Queue Length (if you are using a RAID device, the%disk time counter can display a value greater than 100%.) If greater than 100%, AVG is used. Disk Queue Length counter determines the average number of system requests that are waiting for disk access	No more than the number of disks 1.5~2 if the top two values are always high, consider upgrading the disk drive or moving some files to another disk or server

some things to note1. If monitoring does not exceed 4 hours, it is reasonable to update once every 15 seconds; If you will monitor the system for 8 hours or longer, set the interval less than 300 seconds 2. Personally think that the test report results also have to attach a map reference, because the single minimum, maximum and average does not explain problem 3. Unlike the physical disk counter data, the data for the logical disk counter is not collected by the operating system by default. To obtain performance counter data for a logical drive or storage volume, you must type DISKPERF–YV at the command prompt. By default, the operating system uses the Diskperf–yd command to contain physical drive data. For more information about using command diskperf, type diskperf-。。 at a command prompt 4. Generally, determining whether performance is acceptable is a subjective judgment that changes significantly as the user's environment changes. 5. Low memory is the most common cause of severity problems in computer systems. Workstation responsiveness is most likely due to memory and processor problems, and servers are more susceptible to disk and network problems. 6. At the start of the program, the Process/%processor time value of each program quickly climbs, decreases, and then stabilizes. Note the peak value of the processor during program startup is important; You may want to temporarily ignore the high boot values in the monitoring data to get a more precise picture of the typical program using the processor. 7. When memory is reduced, the operating system begins to supplement by obtaining memory from the working set of the less active program (working set), so you will see an increase in the working set of the program and a decrease in the value of the other programs. If there is not enough memory in the system to meet the requirements of all active programs, memory paging will occur and program performance will be affected. 8. If a memory leak occurs, the value of the Process/private bytes counter and the Process/working set counter is often elevated, while the available bytes is lowered. 9. If the process is missing, modify the disable under the registry hkey_local_machine/system/currentcontrolset/services/perfproc/performance The performance counters value is 0. < two > performance analysis

1. Memory Analysis Method

The memory analysis is used to judge whether the system has memory bottleneck and to improve the performance of the system by means of increasing memory.

Counters to use for memory Analysis: Memory category and Physical Disk category. The main methods and steps of memory analysis:

(1) First view Memory/available MBytes index

If the data for this metric is small, the system may have memory problems and you need to continue with the following steps for further analysis.

Note: In Unix/linux, the corresponding metric is free (KB)

(2) Note the values of pages/sec, Pages read/sec, and page faults/sec

The operating system can use disk better to improve the amount of memory available on the system or to improve the efficiency of RAM. These three indicators directly reflect the frequency of disk swapping by the operating system.

If the PAGES/SEC technology continues above hundreds of, there may be a memory problem. The PAGES/SEC value is not necessarily large nine indicates a memory problem, possibly due to the program running the memory-mapped file. Page Faults/sec describes the number of pages that occur per second, and the more page failures, the more times the operating system reads to memory. This matter needs to see the pages read/sec count value, the threshold of the counter is 5, if the count is more than 5 (there is a problem, the passing people know can be the next ...), you can determine the existence of memory problems.

Note: In the Unix/linux system, the index is (page) Si and (page) so.

(3) Analyzing performance bottleneck according to the value of physical disk counter

The analysis of the physical disk counter includes the analysis of page reads/sec and%disk time and aerage disk Queue length. If the pages read/sec is low and the value of%disk time and average disk Queue length is high, there may be a disk bottleneck. However, if the queue length increases while the pages read/sec does not degrade, there is not enough memory.

Note: In the Unix/linux system, the corresponding metric is reads (writes) per sec, Percent of the "disk is busy and average number of transactions waiting for SE Rvice.

2. Processor Analysis method

(1) First look at the System/%total Processor time performance counter count value

The value of this counter reflects the overall processor utilization of the server, which, for multiprocessor systems, alerts the average utilization of all CPUs. If the value continues to exceed 90%, the entire system is facing a processor bottleneck and needs to be increased by increasing the processor.

Note: In multiprocessor systems, the data itself is small, but the put direct load situation is extremely uneven and should be seen as a bottleneck in the system-generated processor.

(2) Next look at the processor/%processor time and Processor/%user time and processor/%privileged time for each CPU

Processor/%user time is the CPU that is consumed by the non-core operation of the system, and if the value is large, consider whether the value can be lowered by means of a friendly algorithm. If the server is a database server, the reason for the Processor/%user time value is likely to be that the database is sorted or the function operation consumes excessive CPU times, at which point you can consider optimizing the database system.

(3) Research system processor bottleneck

View the value of the System/processor Queue Length counter, when the value of the counter is greater than the total number of CPUs + 1 o'clock, which indicates that a processor block has been generated. When the processor's%process time is high, it typically blocks with the processor, but when the processor is blocked, the value of the processor/%process time counter is not necessarily large, and the reason for the processor blockage must be found.

%doc time is another concern, and the lower the counter the better. In a multiprocessor system, if this value is greater than 50% and processor/%precessor time is very high, adding a NIC may improve performance.

3. Disk I/O analysis methods

(1) Calculate the I/O number of the MUI disk

The I/O number of the MUI disk can be compared to the I/O capability of the disk, and if the calculated number of per disk I/O exceeds the disk's nominal I/O capability, the disk's performance bottleneck does exist.

Per-disk I/O calculation method

RAID0 calculation method: (reads +writes)/number of disks

RAID0 calculation method: (reads +2*writes)/2

RAID0 calculation method: [reads + (4*writes)]/number of disks

RAID0 calculation method: [reads + (2*writes)]/number of disks

(2) Analysis with processor/privileged time merger

If the physical disk counter, only%disk time is large, the other values are moderate, the hard drive may be a bottleneck. If several values are large and the value lasts more than 80%, a memory leak may occur.

(3) Analysis according to disk Sec/transfer

In general, the definition of this value is less than 15ms excellent, between 15~30ms is good, 30~60ms between acceptable, more than 60ms you need to consider the replacement of the hard disk or the hard drive RAID mode.

4. Process Analysis Methods

(1) View the%processor time value of the process

The%processor time for each process reflects the processor times consumed by the process. By comparing the processor time consumed by different processes, you can see which processes are consuming the most processor time during performance testing, which can be optimized for applications.

(2) View the page invalidation generated by each process

You can use the ratio of page invalidation (obtained by Prcess/page failures/sec counter) and System page invalidation (which can be obtained by Memory/page failures/sec counter) per process to determine which process has generated the most page invalidation. This process is either a process that requires a lot of memory, or a very active process that can be focused on.

(3) Understanding of the process of process/private Bytes

Process/private Bytes is the number of current bytes allocated by the process that cannot be shared with other processes. This counter is used primarily to determine whether a process has a memory leak during performance testing. For example, for a Web application on top of IIS, we can focus on monitoring the private Bytes of the Inetinfo process, if the private Bytes counter value of the process is increasing during the performance test, or after the performance test has stopped, The private bytes of the process continues at a high level, indicating that there is a memory leak in the application.

Note: In the Unix/linux system, the corresponding indicator is resident Size

5. Network Analysis method

Network Interface/bytes Total/sec is the rate at which bytes are sent and received, which can be used to determine whether the network link speed is a bottleneck, by comparing the value of the counter with the bandwidth of the current network.

RAID0 calculation method: [reads + (2*writes)]/number of disks

(2) Analysis with processor/privileged time merger

If the physical disk counter, only%disk time is large, the other values are moderate, the hard drive may be a bottleneck. If several values are large and the value lasts more than 80%, a memory leak may occur.

(3) Analysis according to disk Sec/transfer

In general, the definition of this value is less than 15ms excellent, between 15~30ms is good, 30~60ms between acceptable, more than 60ms you need to consider the replacement of the hard disk or the hard drive RAID mode.