These are very obvious and important metrics to report on any kind of Performance related tests and many of you already aware of it but those who are new to Performance Testing domain will get some good insight.
1. Response Time (RT) : No Performance Test Report can be prepared without reporting the RT, this is soul of any of the Performance Report. One can directly judge the applications behavior and Performance based on this number.
Common Observations one can deduce by analyzing Response Time :
2. Transaction Per Second(TPS) or Throughput : Transaction finished per unit of time shows how fast our application is on given processor/hardware. As the number of users increases processor along with other application components also has to perform and process applications request faster.
Common Observations one can deduce by analyzing TPS :
3. CPU Utilization : There is always a direct relation of number of users an application can scale up, based on CPU Utilization. More number of users accessing the application causes increase in CPU Utilization (if application has not hit any bottleneck; will cover on this later). Ideally we should not stretch CPU usage by 70-80%.
Common Observations one can deduce by analyzing CPU Utilization :
4. Memory Utilization : Memory usage mostly depends on type of application we are using it may not have direct relation with number of users accessing the application as modern systems are coming with number of techniques like cache, in memory and shared library usage.But its always important to monitor and report Memory usage especially at Start of the test , During the test and at End of the test.It is more helpful if memory has been recorded for each processes. Memory analysis is very complex metrics to analyze as there are many components to look after, like Virtual Memory,Physical Memory at process level or used/free memory at server level .
Common Observations one can deduce by analyzing Memory Utilization :
5. Disk IO Utilization : Disk Utilization shows whether our application do more Write or Read operation, how much percentage of disk is used during the test.
Common Observations one can deduce by analyzing Disk IO Utilization :
Related Article :
http://www.speforums.com/resources/...tools-commands-no-external-tools-required.24/
1. Response Time (RT) : No Performance Test Report can be prepared without reporting the RT, this is soul of any of the Performance Report. One can directly judge the applications behavior and Performance based on this number.
Common Observations one can deduce by analyzing Response Time :
- If RT number is higher than SLA mentioned then application is under Performing.
- If RT increases exponentially with increase in number of users then there are chances of improvement in code.
- Comparison of RT with old release or tests shows the performance impacts because of development changes in application.
2. Transaction Per Second(TPS) or Throughput : Transaction finished per unit of time shows how fast our application is on given processor/hardware. As the number of users increases processor along with other application components also has to perform and process applications request faster.
Common Observations one can deduce by analyzing TPS :
- As the number of users increases TPS should also increase linearly, if not then there is as a scalability issue with the application. we can easily say application is scalable till the level TPS has increased.
- Reduced TPS can also be cause of Network or Connection being limitations.
3. CPU Utilization : There is always a direct relation of number of users an application can scale up, based on CPU Utilization. More number of users accessing the application causes increase in CPU Utilization (if application has not hit any bottleneck; will cover on this later). Ideally we should not stretch CPU usage by 70-80%.
Common Observations one can deduce by analyzing CPU Utilization :
- CPU Utilization should increase as the number of users increases, if not we might have hit the bottleneck which needs to observe at various level.
- CPU Utilization indicates that whether application is stressed or under utilized.
- This is very important metrics to do the Capacity Planning.
4. Memory Utilization : Memory usage mostly depends on type of application we are using it may not have direct relation with number of users accessing the application as modern systems are coming with number of techniques like cache, in memory and shared library usage.But its always important to monitor and report Memory usage especially at Start of the test , During the test and at End of the test.It is more helpful if memory has been recorded for each processes. Memory analysis is very complex metrics to analyze as there are many components to look after, like Virtual Memory,Physical Memory at process level or used/free memory at server level .
Common Observations one can deduce by analyzing Memory Utilization :
- Look for Difference in memory consumption Before and After the test to check if any memory leak is present. Mostly if memory is not coming to its original position after the test indicates that memory has not been freed by application.
- Its because of memory you might observer page in / page out or SWAP during the test, these are clear signs of physical memory(RAM) being the constraint factor.
5. Disk IO Utilization : Disk Utilization shows whether our application do more Write or Read operation, how much percentage of disk is used during the test.
Common Observations one can deduce by analyzing Disk IO Utilization :
- Even if CPU and Memory is under utilized Disk can be a limiting factor if either read or write is more, it causes application to wait and hence Response of the application may reduce.
- This metrics is especially important for DB servers as DB always handles read and write of data.
Related Article :
http://www.speforums.com/resources/...tools-commands-no-external-tools-required.24/
Last edited: