Scalability Testing

Prerequisites

  • a working version of k6: https://k6.io/docs/get-started/installation/

Test setup

  • The tests use a number of environment variables, to simplify the execution of the tests two scripts are provided.

    • run_test.bat which is preconfigured for the local docker environment. It assumes that the gateway runs on port 8080 and a user with the name and password “test” is registered in keycloak.

    • deployed_test.bat which is preconfigured for the deployed environment. It runs against https://orange.informatik.uni-stuttgart.de/graphql and also is preconfigured for a user with the name and password “test”. To properly function this script also requires a VPN connection.

  • To change any of the environment variables just change the respective variable in one of the scripts.

  • The following variables are defined:

Name Function
K6_HOST URL to the Keycloak. Used to fetch a token.
K6_REALM Name of the Realm in Keycloak. Used to fetch a token.
K6_CLIENT_ID Client ID of the Application. Used to fetch a token.
K6_USERNAME Username of a valid user. Used to fetch a token.
K6_PASSWORD Password of the user. Used to fetch a token.
K6_GATEWAY_URL URL to the Graphql Api/Gateway. Used to make queries.

  • The token.js script is used to acquire an authentication token, and needs to be imported into any script.

  • To execute a test just run the run_test.bat or deployed_test.bat via a shell and add the name of the script as an argument e.g. .\run_test.bat getmedia.js

  • To write the test results to a file use --out csv=<filename>.csv more info: https://k6.io/docs/results-output/real-time/

Writing and running tests

K6 uses test script written in javascript. The most important part are the options. They define how a test run should look like. More info: https://k6.io/docs/using-k6/k6-options/ The getmedia.js has a variety of test scenarios already. All other tests currently only have one scenario. https://k6.io/docs/using-k6/scenarios/

  • The five_hundred_users scenario tries to make 500 requests as fast as possible.

  • The ramping_users scenario increases the number of virtual users over a period of 3 minutes. The test will make as many requests as possible.

  • the constant_arrival_rate scenario makes a constant number of requests per second over a period of 1 minute.

  • The ramping_arrival_rate scenario increases the number of requests per second every 30 seconds up to 500 requests per second.

The const query defines the query which should be used. Mutations are also possible. See createCourse.js. The function setup sets up the tests by running the getToken function from the token.js script. The default function runs the actual request. const headers is used define the authorization header. const res is the actual request. Those two shouldn’t require any modifications.

Once the tests have run they will print a number of metrics to the console: https://k6.io/docs/using-k6/metrics/