SAS wrote:
Hi Team,
In my load test, I have a series of web service requests and they are
responding in less than 20 milliseconds except the first one that is
taking around
15+ milliseconds additional time. This happens even with multiple
iterations and multiple threads.
I've checked the response times in Dynatrace and even for the first
request, it shows 20 milliseconds whereas JMeter shows 35 milliseconds.
Appreciate any solution to this.
Thanks
SASTRY
Most probably it's the time to establish the initial TCP connection
<https://www.sciencedirect.com/topics/computer-science/three-way-handshake>
and perform the SSL handshake
<https://www.ssl.com/article/ssl-tls-handshake-overview/>. Each thread
during 1st iteration starts the session and depending on your HTTP
Request sampler and JMeter configuration it's either re-used or
re-created on subsequent iterations.
You don't need to compare JMeter to Dynatrace, you need to compare
JMeter with the real browser (or other API client application/upstream
system) and replicate the behaviour of that browser/application when it
comes to establishing and keeping the session.
JMeter's behaviour is controllable by:
1. "Use Keep-Alive" box in the HTTP Request
<https://jmeter.apache.org/usermanual/component_reference.html#HTTP_Request>
sampler
2. "https.sessioncontext.shared" JMeter property
<https://www.blazemeter.com/blog/jmeter-properties-customization>
3. Use Loop Controller
<https://jmeter.apache.org/usermanual/component_reference.html#Loop_Controller>
to orchestrate the iterations instead of Thread Group
---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscr...@jmeter.apache.org
For additional commands, e-mail: user-h...@jmeter.apache.org
