Creating Chaos in the University with Linkerd and Chaos Mesh
summary
A university project builds a Kubernetes demo that deliberately injects failure: Linkerd traffic-split 50/50 sends half the requests to a working Apache podset and half to an error-injector, while Chaos Mesh pod-kill every 6s and 12s repeatedly kills Apache pods. The test runs on a two-node cluster (2 vCPU, 4 GB RAM each), shows the Linkerd dashboard dropping success rate to ~50% and Grafana latency spikes, and verifies automatic pod restarts so the system stays available while revealing vulnerabilities.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:09Chaos Engineering
- 6:26Technology
- 7:41Demonstration
- 12:47Traffic Split Results
- 15:10Chaos Engineering Test
- 18:40Lessons
related talks
32mUnraveling the Chaos: Raising Workload Resilience on K8s with Chaos Mesh
01h 22mKeynotes Día 2
30mBringing Order to Chaos: Make Your Systems More Resilient with Chaos Engineering
45mThe Glorious Art of Destroying Things on Purpose: an Introduction to Chaos Engineering
37mBringing order to the chaos!
32mImplementing Chaos Engineering in Kubernetes
19mChaos Engineering at Scale
01h 02mPractical Chaos Engineering breaking things on purpose to make them more resilient against failure