Is there a Gatling graph similar to JMeter's "Response Time Percentiles"?

Hi,

great past 2 weeks spent on learning curve to replace many of my JMeter performance tests with Gatling equivalent.

I faced some issues, but I always managed to find solutions through the code or the issues or this mailing list.

Yet there’s is one more thing I’m missing though from JMeter: a graph similar to JMeter’s “Response Time Percentiles”, where I would get percentiles in X axis and response time in Y axis.

I understand that somehow this is embedded in Gatling’s “Response Time Percentiles over Time”, but the information is color based and the graph is focusing on the time spent doing the load test (the X axis is time, not percentiles). It’s actually quite hard to present this to anyone used to JMeter (or even LoadRunner) graphs.

Is there a way to get such relevant graph? Or was this kind of graph discarded by choice from the inception of Gatling?

Rgds
Christian

Hi,

Percentiles usage (see response time distribution over time) landed in Gatling way before they landed in JMeter.
Percentiles distribution charts came later from Git Tene’s amazing work, particularly with HdrHistogram.
Gatling OSS HTML reports were designed before that, and we have no plans for additional features in there atm.

Percentiles distributions are available in FrontLine, our Enterprise version, though.

percentiles-distribution.png

Regards,

Hi Stephane,

Frontline is unfortunately not in my scope of evaluation… I’m focusing on finding a pure OSS replacement to some of our JMeter tests.

Yes I’m also aware of Gil Tene’s work (and coordinated omission). HdrHistogram-2.1.9.jar is included in gatling-charts-highcharts-bundle-2.3.1

Is there an easy way to use simulation.log with HdrHistogram? Especially when using grouping?

Rgds
Christian

Not AFAIK.
Feel free to build your own parser but beware that the log file’s format is internal (not a public API) and can change without further notice. For example, it’s currently text based but we might consider switching to a binary format some day.