Lightning Talk: A Culture of Resiliency through Chaos
summary
Claims a repeatable chaos program that finds real failures by running experiments under productionlike load, not toy faults. Uses Elasticsearch statefulset (3 replicas), a load‑testing client, and an API call to kill pod via API during load test while watching Elastic monitoring and golden signals; failures spike when one replica loses its shard, then recovery happens as Kubernetes recreates the pod and increasing replica count eliminates the outage. Describes GitHub/Slack game‑day automation and a simple resiliency scoring cycle.
AI-written summary · about
chapters (6)
- 0:00Introduction to resiliency
- 2:47Chaos engineering basics
- 3:51Methodology for chaos testing
- 6:45Chaos demo setup
- 7:27Elasticsearch failure demo
- 9:58Summary and takeaways
related talks
39mPracticing Cloud-Native Resliency Is a Challenge: Decoding It with Chaos...
29mEngineering Chaos for Resilient Applications: A Gameday Guide
39mResilience & Chaos Engineering in the Cloud
05mLightning Talk: Chaos Engineering In Practice
30mBringing Order to Chaos: Make Your Systems More Resilient with Chaos Engineering
20mKeynote: Finding the Joy in Chaos Engineering
26mBreaking Systems for Resilience
46mEmbracing Collaborative Chaos: Chaos Days on Large Platforms