The following real-time performance metrics are displayed in each graph
1. Key Performance Indicators (KPI) monitor a fixed set of performance parameters from the user perspective. They include:
- Users - The number of instantiated and active VUs, iterating through their respective test cases. Some VUs can complete the test earlier than others and become inactive. VUs which completed all their iterations before the test ends are excluded from the active user count. If the test is configured to complete only after all VUs complete their iterations, then the User graph will show a gradual declining the number of VUs at the end of the test.
- Req/Sec - The number of requests being sent per second.
- Avg. Response(s) - The average response time of the sent requests. Every datapoint on the response time graph reflects the average value during the checkpoint intervals. If no responses were received between two datapoints, the response time could not be determined, so such datapoints are skipped to reflect an accurate response time curve.
- KB Received/Sec - The number of kilobytes currently received per second.
- Errors/Sec - The number of errors currently received per second.
- Pages/Sec - The number of pages currently sent per second.
- Pending Requests - The number of currently pending requests. Pending requests are those that are issued, but StresStimulus do not yet receive the response. Generally, a higher number of pending requests indicates a slower server response. This parameter can be used to gauge performance change in several test runs.
- Total Transactions/Sec - The total number of transactions completed per second. A completed transaction is one that executes all child objects. If transactions are nested, then each transaction is counted separately toward the total completed transactions.
2. Pages monitor the average response time of every page in the test
3. Transactions monitor the average response time of every transaction in the test case
4. Test Cases monitor the average time of the iteration
5. Windows Servers counters monitor a set of performance parameters configured in the Other Options -> Monitoring -> Windows Servers Monitoring section. The counters are collected from remote Windows servers involved in the test, including web, application, and database servers. Hundreds of different parameters, such as CPU load, memory, and disk usage, can be monitored. Monitoring server metrics can help to identify potential bottlenecks limiting the application’s performance.
6. Linux/Unix Servers counters monitor a set of performance parameters configured in the Other Options -> Monitoring -> Linux Unix Servers Monitoring section. The counters are collected from the remote Linux or Unix servers involved in the test using the SNMP protocol.
5. Load Agents counters monitor a set of performance parameters configured in the Other Options -> Monitoring -> Load Agents Monitoring section. The counters are collected from client testing machines. This use used to make sure that the testing machines are not overloaded.
Graphs display instant performance characteristics, and the performance counter's data plotted with the frequency defined by the Graph Sample Rate property located in the Other Options -> Advanced Options section.
- No labels