The SearchAPI uses Gatling to executes load tests.

The application.conf file that provides all configuration to load tests.

You can override each config above using Java property. For example, if you can override gatling.users property you must use -Dgatling.users=<value>.

For more details about how to do override, see How To Run section.

To build this project, you need to execute these commands:

git submodule update --init --recursive

make build

The vivareal/load-test:load-test docker image will be created.

Gatling works with a Simulation concept and for SearchAPI we creates the SearchAPIv2Simulation.

There some steps when you run the load tests:

1. If not already built, builds and runs load-test docker image.
2. Executes your simulations.
3. Uploads you simulation report on Amazon S3.
4. Notifies report status on Slack with link to access them.

There are two parameters to use:

• LT_ENDPOINT*: load tests target endpoint.

• LT_EXTRA_ARGS: gatling extra args

The environment variables with * are required. You can override each config using LT_EXTRA_ARGS.

To run local, you simple use make run-local with the target ip to load test, for example:

make run-local LT_ENDPOINT="<TARGET_IP>"

To run using Kubernetes, you simple use make run:

make run LT_ENDPOINT="<TARGET_IP>"

Yon can use K8S_RUN_ARGS to configure kubectl run params.

To run using Gradle too and you simple use gatlingRun task.

./gradlew gatlinRun

The upload and notification process is separated and to do this you must use uploadReport.

./gradlew gatlinRun uploadReport

No tests currently implemented

load-test is a tool/lib project and the deploy is subjective according to the project that uses.

You may use docker to deploy: make push

To push successful docker image, make sure you setup docker credentials.

Just implement the simulation in package com.vivareal.search.simulation and sent it in gatling.simulations.include parameter:

./gradlew gatlingRun -Dscenario.users=1 -Dgatling.rampUp=30 -Dgatling.maxDuration=60 -Dgatling.simulations.include=**/SimpleRequestSimulation.scala

Or you can send it in LT_EXTRA_ARGS:

make run LT_EXTRA_ARGS="-Dgatling.simulations.include=**/SimpleRequestSimulation.scala"

It's a Simulation that downloads a csv file from Graylog API based on a query and executes the resulting URIs requests in circle and during a configured time.

All above params can be overriden using System Properties.