You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I recently switched to jmeter 5.6.2 and listener version 2.7, and I have noticed a decline in performance.
When the backend listener is disabled, the graph in the jmeter HTML report appears smooth (screenshot 1) and is at 7k hits per second. Whereas with using listener it's inconsistent and reaches only 5k at most (screenshot 2).
I was able to obtain the desired performance and a better graph by changing some parameters: 20k async queue size, 4k batchSize, 1k flushInterval, and recordSubSamples false (screenshot 3). But before that when using jmeter 5.5 and listener 2.6 I could use default settings and would achieve similar results to screenshot 1.
Some additional info about the environment:
distributed test (5 VMs)
1000 threads
VMs are Standard_D16as_v5, no issue with memory or CPU during the test
InfluxDB version 2.7.1 ran in a container instance
java 21
Could you please let me know if you have any insights into what might be causing this issue? I've not seen this issue with other tests so far, I'm assuming it's because they're sending less requests.
The text was updated successfully, but these errors were encountered:
Hello,
I recently switched to jmeter 5.6.2 and listener version 2.7, and I have noticed a decline in performance.
When the backend listener is disabled, the graph in the jmeter HTML report appears smooth (screenshot 1) and is at 7k hits per second. Whereas with using listener it's inconsistent and reaches only 5k at most (screenshot 2).
I was able to obtain the desired performance and a better graph by changing some parameters: 20k async queue size, 4k batchSize, 1k flushInterval, and recordSubSamples false (screenshot 3). But before that when using jmeter 5.5 and listener 2.6 I could use default settings and would achieve similar results to screenshot 1.
Some additional info about the environment:
distributed test (5 VMs)
1000 threads
VMs are Standard_D16as_v5, no issue with memory or CPU during the test
InfluxDB version 2.7.1 ran in a container instance
java 21
Could you please let me know if you have any insights into what might be causing this issue? I've not seen this issue with other tests so far, I'm assuming it's because they're sending less requests.
The text was updated successfully, but these errors were encountered: