Gatling Scenario Issue

[Amature]

Hi Gatling Team,

I am using Gatling to Performance Test for an application but i am facing few issues.

1. The Setup have been designed for having 100 constant users but when i actually run the test the number of users is not what i gave, please find the sample code for setup mentioned below,

setUp(scn.inject((rampUsers(100) over (100 seconds)),constantUsersPerSec(200) during(600 seconds))).protocols(httpProtocol)

2. You can see the active users is going beyond 20K mark though i intended to keep it only 100 , kindly find it in the snapshot attached.

3. Few changes i have made is to increase the heap size allocation

set JAVA_OPTS=-server -XX:+UseThreadPriorities -XX:ThreadPriorityPolicy=42 -Xms512M -Xmx512M -Xmn100M

connectTimeout = 60000

pooledConnectionIdleTimeout = 60000

requestTimeout = 60000

for ref i have been reading the following
http://gatling.io/docs/2.1.7/quickstart.html#gatling-scenario-explained

4. I am getting the following errors
   java.net.ConnectException: No buffer space available (maximum connections reached?): connect

java.net.ConnectException: Connection timed out: no further information: /

Kindly let me know you suggestion,
Thanks in Advance

[Amature]

Please note i am using the latest version → gatling-charts-highcharts-bundle-2.1.7

[slandelle]

You don’t understand how Gatling injection works and are confusing injection, and concurrent number of users inside your system.
see http://gatling.io/docs/2.1.7/general/simulation_setup.html

It seems you actually want a closed workload model. But it’s very likely that this model is ill suited for your application.

[Amature]

Hi Stephane,

you are correct i was trying to achieve a closed workload model, and thank you for the link for the documentation link.
But is there any way we can have 500 users to ramp up in 500 seconds and then after 10 mins we can injects 1 users for every 20 seconds. (this is the ideal scenario which i was trying to achieve that)

Few things which i fail to understand are mentioned below,

1. I have used the following scenario statement, but i can see a constant 55 users during my entire test and 55 transactions in a second, which is no where near the load i am intending to simulate.
   setUp(scn.inject((rampUsers(500) over (500 seconds)),constantUsersPerSec(5) during(600 seconds))).protocols(httpProtocol)

Please find the snapshot attached.