Consistent TPS for multiple requests in same scenario

[Zuhair]

Hi there,

I have one requirement that every request should run in consistent TPS for a whole test.
Below is my code snippet:

  val updateLineItemScenarioFromV3: ScenarioBuilder = scenario("Update Line Item Scenario")
      .feed(Feeder.tlv4_1EntDetailsV3UptLineChkDate)
        .exec(session => {
          session.set("accessToken", MessagePasser.getMessage(1))
        })
        .exec(session => {
          val endDate = s"2025-10-10T23:59:59Z"
          session.setAll(Map(
            "endDate" -> endDate,
          ))
        })
        .exec(UpdateLineEndDate).exec(flushHttpCache)
        }

And my injection profile:

 UpdateEntitlement.updateLineItemScenarioFromV3.inject(constantUsersPerSec(10).during(1 minutes)).throttle(reachRps(10) in (0 seconds),holdFor(1 minutes))

When I execute for 1 minute, I got my requirement of consistently 10 requests per second.

Now, I have to get date from the “getEntitlementById” API and pass it to the “UpdateLineEndDate” API. So, adding one more request in my scenario.

  val updateLineItemScenarioFromV3: ScenarioBuilder = scenario("Update Line Item Scenario")
      .feed(Feeder.tlv4_1EntDetailsV3UptLineChkDate)
        .exec(session => {
          session.set("accessToken", MessagePasser.getMessage(1))
        })
        .exec(getEntitlementById)
        .exec(session => {
          val endDate = s"${updateLineEndDate}T23:59:59Z"
          session.setAll(Map(
            "updatedEndDate"-> updateLineEndDate,
            "endDate" -> endDate,
          ))
        })
        .exec(UpdateLineEndDate).exec(flushHttpCache)
        }

And my injection profile:

 UpdateEntitlement.updateLineItemScenarioFromV3.inject(constantUsersPerSec(20).during(1 minutes)).throttle(reachRps(20) in (0 seconds),holdFor(1 minutes))

When I re-executed for 1 minute, I observed consistently 20 requests per second.

I also want consistent tps for each API’s:
getEntitlementById - 10 TPS
UpdateLineEndDate - 10 TPS
But for the individual API’s , it got dispread.

I am not sure about the execution behind the scenes. So, please provide me some resolution to reach consistent TPS in each API’s.

Thank You in advance!

[Samir]

There is no problem with the load test results.

The graphs have a time resolution of 1s for your duration. If a request occurs at t-1ms, it is counted in one bucket, and if it occurs at t+1ms, it is counted in the next bucket.

With a low amount of RPS (10 RPS), it doesn’t take much for a request to switch to the next bucket, which is reflected in the results.

[Zuhair]

Hi, @Samir

Thanks for the information. But is there any way to make “getEntitlementById” and “UpdateLineEndDate” requests(both come under the same scenario) with 10 RPS each in a single second(bucket)?

[Samir]

Hi,

Here some link to help you: Gatling Simulation scripting reference