Limit Chaos Experiment Blast Radius
Define a steady-state hypothesis and abort conditions so a chaos experiment auto-stops before it takes down the whole fleet.
Task
Add chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5% to the test suite so regressions are caught automatically on every pull request.
When to use this
A refactor passes existing tests but breaks chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5%, shipping a silent regression to users.
Prerequisites
- A test runner configured (Vitest or Jest)
- Source code with testable units
Steps
1. Configure chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5% in the project configuration file with the recommended defaults.
chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5%
Expected output: The configuration is accepted and chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5% runs without errors on the first invocation.
2. Run the check locally to verify chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5% produces the expected result before pushing.
pnpm test
Expected output: Local run matches CI behavior; no surprises after push.
3. Add the gate to the CI pipeline so chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5% blocks merge on failure.
- name: Run add chaos experiment blast radius
run: pnpm test
Expected output: Pull requests with failing chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5% are blocked from merging.
Troubleshooting
Problem
The check passes locally but fails in CI
Solution
Pin dependency versions and the runtime to match CI exactly; reproduce with the same Docker image or act.
Problem
Configuration is rejected with a schema error
Solution
Check the tool version matches the config schema; older versions may not support newer fields.
Limitations
- The gate only runs on events that trigger the workflow; locally pushed commits bypass it.
- Configuration drift is possible if someone edits settings outside version control.
Cover the regression path
Start by measuring the problem. Define a steady-state hypothesis and abort conditions so a chaos experiment auto-stops before it takes down the whole fleet. The most common mistake is adding the gate without first confirming it catches the issue you care about. Run chaos run experiment.json –hypothesis steady-state.json –abort-on error-rate>5% against the current codebase and inspect the output before wiring it into CI.
chaos run experiment.json --hypothesis steady-state.json --abort-on error-rate>5%Keep tests maintainable
Once the gate is in place, make failures actionable. A failing chaos run experiment.json –hypothesis steady-state.json –abort-on error-rate>5% should point to the exact file and line, not a generic error. Pin the tool version in the lockfile so a upstream release does not silently change behavior. Review the gate output in every pull request just like a test result.