I tried to set up a ramp duration specified in nanoseconds, in an attempt to control the duration between user injections to a fraction of a second.
14:48:57.724 [INFO ] a.e.s.Slf4jLogger - Slf4jLogger started
GATLING_ENVIRONMENT : qa
GATLING_USERS : 100
GATLING_RAMP_UP_TIME : 6000000000000 nanoseconds
GATLING_DURATION : 60000000000 nanoseconds
GATLING_RAMP_DOWN_TIME : 1000000000 nanoseconds
GATLING_USE_PAUSES : false
14:48:58.514 [INFO ] i.g.h.e.HttpEngine - Start warm up
14:48:59.038 [INFO ] i.g.h.e.HttpEngine - Warm up done
Simulation started…
14:48:59.361 [INFO ] i.g.c.s.w.ConsoleDataWriter - Initializing
14:48:59.361 [INFO ] i.g.c.s.w.LogFileDataWriter - Initializing
14:48:59.364 [INFO ] i.g.c.s.w.ConsoleDataWriter - Initialized
14:48:59.385 [INFO ] i.g.c.s.w.LogFileDataWriter - Initialized
14:48:59.393 [INFO ] i.g.c.c.i.Injector - StoppedInjecting
14:49:02.431 [INFO ] i.g.c.c.i.Injector - All users of scenario Isolation Test - Expiration Time - EWS Interactive Session are stopped
14:49:02.431 [INFO ] i.g.c.c.i.Injector - Stopping
14:49:02.432 [INFO ] i.g.c.c.Controller - Injector has stopped, initiating graceful stop
This should have injected 100 users over 100 minutes. Instead, it injected one user and then immediately thought it was done. The ONLY change was to express the durations in nanoseconds instead of seconds.
The injection profile was rampConcurrentUsers( 1 ) to ( 100 ) during ( 6000000000000 nanoseconds )
What’s going on? Is this a bug in the injection calculation logic?
I’m running 3.3.1
