Improve resilience with automated chaos engineering
summary
Automated chaos engineering uses AWS Fault Injection Simulator plus Lambda, EventBridge and Step Functions to run reproducible experiments (example: stop an EC2, restart after one minute) and to schedule daily or trigger on auto scaling and to gate deployments with an experiment stage in CodePipeline. The claim is that recurring, event-driven and CI/CD experiments catch regressions faster than one-off tests; demos include stop conditions (three alarms) that abort experiments and a pipeline rollback when an experiment fails.
AI-written summary · about
chapters (21)
- 0:00Introduction
- 0:50What is chaos engineering
- 1:58How chaos engineering works
- 2:37Oneoff experiments
- 5:03Use case example
- 6:22AWS fault injection simulator
- 9:45Good practices for deploying
- 10:52Service dependencies
- 12:47Manual experiments
- 13:58Automated experiments
- 14:39How our system evolves
- 15:26Scheduling an experiment
- 19:43Event triggered experiments
- 20:28Event triggered serverless application
- 23:53Continuous delivery experiments
- 24:52Continuous delivery experiment demo
- 29:24Recap
- 29:57Automated chaos experiments
- 30:42Summary
- 31:58Additional resources
- 32:19Questions
related talks
29mImprove Resilience with automated chaos engineering
42mChaos Engineering: Why Breaking Things Should be Practiced
26mImprove application resilience with chaos engineering
26mA proactive approach to resilience in modern apps
34mChaos engineering for serverless
57mChaos Engineering Primer
18mChaos Engineering: Break It On Purpose
39mResilience & Chaos Engineering in the Cloud