Getting your automated tests in the pipeline
summary
Claim: adding targeted end-to-end tests into CI gives measurable risk reduction without slowing delivery. Implementation uses a single automation repo, WebdriverIO and Jest, CircleCI with a Docker image with Chrome and chromedriver to avoid remote latency, parallel shards and API-first setup so tests are independent. Flakiness is caught by running new browser tests in a loop like run browser tests 10 times before adding and tagging unstable cases. Outcome: pipeline test coverage ships while releases increase from 5 to 10 per day.
AI-written summary · about
chapters (18)
- 0:00Automated testing in pipelines
- 1:05Slow and flaky tests
- 2:27End-to-end test automation at Blue Fox
- 4:13Mental models for test automation
- 5:39Adding tests to pipeline safely
- 7:02Testing your tests first
- 8:29Parallel test runs
- 11:22Test independence for parallel runs
- 12:48Atomic tests for reliability
- 13:54Domain abstractions in tests
- 15:18Shorter tests pass more
- 16:24Current state at Blue Fox
- 17:07Impacted tests with GitHub and Jest
- 18:51Flaky tests in pipeline
- 19:55Cloud test execution costs
- 20:36Automation journey best practices
- 21:42Assurance over test pyramid
- 22:25Parallel test resource issues
related talks
39mRemove your Testing Bottleneck with CI/CD Pipelines
20mThe Road to Continuous Testing
38mFreeing up Selenium
26mThe Impact of Testing on a DevOps Pipeline
40mEnd to End Automation from a Developer Standpoint
29mHow to set up a pipeline with visual unit tests for your app
38mScaling Automated Testing & Architecture from Startup to Enterprise
36mTesting In Production