Is there a way to not let errors be taken into account when generating the response times table? It makes no sense that errors are taken into account in the current response times table. Sometimes I get timeouts which screws with the averages etc… Is there an option I’m missing?
Thanks,
Joerek
Not in the OSS version, though you can filter by status in FrontLine (*, OK only, KO only).
In the end, caller/user did experience those timeouts, so IMO it makes sense to account for them in the stats.
Well, to take errors within the response times is a bit tricky. Either you have very fast responses on errors, which could make the average response time very low. Or timeouts, which is just a setting in gatling how long to wait for the response, which is usually higher than a browser default or ‘user’ default behaviour. Right now I have a case where a group isn’t finished because of an error, but the total time of the group is still taken into account for the group response time, which makes no sense as the subsequent requests are not done.
It would be great to see an option in gatling.conf like gatling.charting.filterErrorsFromResponseTimes = true/false (default = false, as it is right now).