At the moment we got 2 ways to do performance testing: branch testing and cloud testing
Use it to test your feature branch and compare results with master.
Start a new run:
- go to load-testing job
- click 'Build with Parameters'
- enter your github username & branch name, simulation classes, uncheck 'INGEST RESULTS' if you don't need it in Kibana stats
- click 'Build'
Job will do this for you:
- build Kibana from branch source code
- start Kibana and ES using FTR
- start Metricbeat monitoring
- run performance simulations
- ingest results and monitoring stats to Kibana stats
Use it to test Kibana on Cloud Staging as a part of new deployment.
Start a new run:
- go to load-testing-cloud job
- click 'Build with Parameters'
- enter your stack version, simulation classes, uncheck 'INGEST RESULTS' if you don't need it in Kibana stats
- click 'Build'
You can use existing deployment config files to define your desired stack setup or create add a new one to
kibana-load-testing/config/deployment/
path and specify it in 'DEPLOY_CONFIG' input field
Job will do this for you:
- create a new deployment
- run performance simulations
- delete deployment
- ingest results and to Kibana stats
- branch testing
- when: daily
- what: 7.x, master
- cloud testing
- when: daily
- what: latest 7.x and 8.0 snapshots
Open finished job run and click 'Google Cloud Storage Upload Report' You will find archive with html reports
Go to Kibana-Stats cluster and open 'Load Testing' space to access test results.