Load test REST APIs using gatling
0 likes1,441 views
Gatling can be used to load test REST APIs by writing Scala simulation scripts. The scripts define test scenarios by chaining together the steps of invoking APIs concurrently. Feeders can inject runtime values. Tests are run by executing the Scala scripts from sbt or triggered from Jenkins. Gatling generates test reports in the target directory with results like response times and errors.
Load test REST APIs using gatling
- 1. Load Test APIs using Gatling Jayaram Sankaranarayanan
- 2. Gatling Simulation Scripts written in Scala Gatling runtime interprets Scala scripts and runs the tests Virtual Users invoking series of APIs concurrently Define test scenarios Feeders to inject runtime values for tests Record & Replay
- 3. First, a bit of Scala
- 4. • Immutability • vals rule • no one messes with my data • Functions • def a function • parameters, yes with default • return more than one, really ? • no side effects • function composition • functions as first class citizens • Object Oriented Scala • Classes, Objects, Traits • Collections • No more looping • How to do => What to do Lets see in Action - ScalaREPL
- 5. Now , get back to Load Testing REST APIs using Gatling
- 6. Pre-Requisites • Install sbt • configure gatling plugin • build and run tests • Install ScalaIDE • so that we can write code
- 7. Gatling Simulation • Create Scenarios • define execution steps • supports http/s or Websockets • chain the steps together • Inject pause times for simulating real life events • Trigger Scenario • Rampup Virtual Users • Execute scenarios • Specify either the duration / iteration count • Feeders • Use feeders to inject runtime varying values • feed values for every iteration of the scenario
- 8. Run Test • Using sbt • sbt test • sbt ‘testOnly com.myorg.api.load.MyApiLoadTest’ • Trigger From Jenkins • Use Jenkins SBT plugin and Gatling Plugin • Archive Load test reports • Report failure using TextFinderPlugin • Test Report • reports in <Project>/target/gatling/*
- 9. gatling.io