Distributed performance testing at the push of a button
summary
Automated pipelines create and tear down distributed test clusters at the push of a button: Terraform spins up a test cluster, GitLab pipelines upload a jar, Cucumber drives JMeter to assemble and run distributed plans on remote slaves, reports are published and the cluster is destroyed to save costs. The project takes about two months and eight people to build, test updates take under a minute, early runs hit ~20 req/s but the framework scales by adding slaves and tests run green with aggregated JMeter and Cucumber reports.
AI-written summary · about
chapters (7)
- 0:00Start
- 4:10Concept An executable test suite
- 5:25Concept Laying the pipes Test pipeline
- 7:45Costs and benefits How to create a test framework that all teams can and will use?
- 8:54Costs and benefits Benefits using Cucumber
- 10:37Costs and benefits Benefits of do-it-yourself
- 11:18Costs and benefits Benefits of reusing product tooling
related talks
45mBuilding loose-coupled continuous testing platform
26mOn-demand Test Infrastructure on ECS as code
46mPerformance Testing Framework
45mHow to Build a Performance Testing Pipeline from Scratch
27mAdvanced Multipurpose Infrastructure: A Self-Managed Cloud Environment with HashiCorp Terraform
29mEphemeral Cloud Infrastructure for Acceptance Tests with Terraform Enterprise & GitLab
47mTest infrastructure on steroids
43mInfinite Space for Your Tests
