Hi there,
We’ve just released a new version of Gatling. There was a regression, so we had to release 1.3.2 immediately after 1.3.1, hopefully it’s stable now.
This fix includes some small bug fixes, as shown on the release note, but more importantly:
- Netty was upgraded to 3.5.8.Final which fixes some performance issues
- Reports generation memory footprint is considerably reduced. Beware that time values are now rounded up (see new accuracy config parameter) with a 10ms accuracy by default.
Get it at the usual place!
Cheers,
Stéphane
http://gatling-tool.org
Question about 1.3.1 (and may be 1.3.2).
I notice req/s are not allways available and never for individuals requests ?
What do you mean by available?
2012/10/4 Henri Gomez <henri.gomez@gmail.com>
See https://github.com/excilys/gatling/wiki/Changelog#wiki-1.3.1 about accuracy.
It probably means that you have these values have been rounded up to 0.
2012/10/4 Henri Gomez <henri.gomez@gmail.com>
See GitHub - gatling/gatling: Modern Load Testing as Code about
accuracy.
It probably means that you have these values have been rounded up to 0.
Hum, impossible, I saw requests processed in graphs.
How long did your simulation run? More than 33 minutes?
2012/10/4 Henri Gomez <henri.gomez@gmail.com>
Well, 2000 requests in (90 * 60) secs = 0,37 req/sec, rounded up to… 0
QED
2012/10/4 Henri Gomez <henri.gomez@gmail.com>
Well, 2000 requests in (90 * 60) secs = 0,37 req/sec, rounded up to... 0
QED
Hum, so why did we have average 300 req/s for total ?
Because every seconds, you send about 0 (0,37) req/s a each request type, but as you have ~800 different kinds of requests, you get an overall ~800 * ~0,37 = 300 req/s
Still not clear?
2012/10/4 Henri Gomez <henri.gomez@gmail.com>
