Object |
Counter |
Describe |
Measure value |
System |
%total Processor Time The average percent of time that all processors on the |
System are busy executing non-idle threads, reflecting overall processor utilization. For example, on multiprocessor systems, when all processors are always busy, this value is 100% ; when 1/4 100% busy state, this value is 25% |
If the value lasts higher than 80%~90% The system faces a processor bottleneck. You can upgrade cpu cpu improved |
Process queue length |
Thread waiting to be assigned cpu The length of the column that the resource is queued for, which does not include being occupied cpu The thread of the resource. This value is an instantaneous count, not the average for a period of time |
If the value is always greater than the number of processors +1 Span style= "font-family: Arial" > You need to upgrade cpu cpu |
Processor (CPU Analysis ) |
%processor time |
This counter is particularly common and represents CPU utilization. The observable processor is saturated, the reference value is <80% |
If the value continues to be higher than 95%, then the bottleneck of the system is CPU. Can add a processor or replace a higher-performance processor |
%user time |
This counter represents the CPU cost of some database operations |
When using aggregate functions, sorting or using a large number of algorithms or complex operations, this value is larger, can be solved by increasing indexes, simplifying table joins, horizontally splitting large tables, and optimizing algorithms. |
%priviliaged time |
The percentage of time that the CPU spends processing threads in privileged mode. such as memory management,SQL Server I/O requests, etc. |
If the value and the physical Disk counter value are too high, consider improving the performance of the I/O subsystem |
%DPC time |
The lower the counter value, the better the time the processor consumes on the network processing |
%dpc time is greater than 50%if the process\%processor time value is high, Indicates that the current network resource is saturated. Adding a network card may improve performance |
Memory ( memory Analysis ) |
Available Mbytes |
The remaining available physical memory (in megabytes) for the system, and the reference value is greater than or equal to 10% |
If the value is too low (4MB or less), there may be memory problems or some programs do not release memory, you can improve by adding memory or modifying programs |
Pages/sec |
Represents the number of pages that are read from or written to the hard disk in order to resolve hardware errors or page faults, and the reference value is 00~20 |
If the value remains at the order of hundreds of, the page exchange activity needs to be studied further. It may be necessary to increase memory, but it may also be caused by a program running a memory-mapped file |
Pages reads/sec |
Represents the number of pages read from the hard disk in order to resolve a hard error. The lower the value, the better the threshold value is 5 . |
If the value is low, but the value of %disk time and Average disk Queue length is high, a disk bottleneck may occur, and if the queue length is increased, the Page Reads/sec does not degrade, there may be insufficient memory |
Page faults/sec |
This counter represents the number of error pages processed by the CPU per second, including hard and soft errors |
|
Process ( Process Analysis ) |
Private Bytes |
The number of bytes that the process could not share with other processes |
If the value continues to be large, a memory leak is most likely to occur |
Work set |
Processing the most recently used memory page of a thread |
If the value of the Memory\Available bytes counter continues to decrease, but process\private bytes and Process \working The value of the Set counter continues to rise, and memory leaks are most likely to occur |
Physical disk ( disk i/o analysis ) |
%disk time |
Percentage of time that the disk drive is servicing a read or write request |
If only %disk time 80% |
Average disk queue length |
Average number of Read and write requests queued to the selected disk during the sampling interval. The value should not exceed the number of disks 2 times |
If %disk time and avg disk queue length page reads/sec avg disk queue length page reads/sec |
Average Disk Sec/read |
Average time to read data on disk in seconds (s) |
If the value is greater than, the disk has a bottleneck and can be improved by changing the hard drive or by changing the RAID mode of the hard drive |
Network Interface ( Web analyst ) |
Byte total/sec |
Indicates the speed at which bytes are sent and received in the network, and the comparison between this counter and the network bandwidth can be used to determine whether the network has bottlenecks |
Generally this counter / bandwidth <50% |