Chaos in the cloud
summary
Demo uses AWS Fault Injection Simulator with JSON/YAML experiment templates, IAM safeguards and EventBridge to auto-stop; it injects a 2 minute subnet connectivity disruption against an EKS pod and measures with locust at 240 requests per second. A single-pod deployment fails completely; two replicas with cross-AZ affinity cut outage down to one load‑balancer health‑check gap (5s). New scenarios include cross-region connectivity and availability Zone power interruption, and guidance covers small templates, staging parity, least‑privilege limits and health‑check emergency brakes.
AI-written summary · about
chapters (18)
- 0:00agenda
- 0:50what is chaos engineering?
- 2:54new! the aws resilience lifecycle framework
- 3:39java chaos engineering
- 4:46chaos engineering in the cloud
- 5:51aws region design
- 7:29multi-az application
- 8:26aws fault injection simulator
- 8:43fully managed chaos engineering service
- 12:31native targets and actions
- 13:05aws fault injectioqn simulator - reference flow
- 14:47announcing two new scenarios
- 15:07cross-region: connectivity
- 16:05az availability: power interruption
- 16:46demo
- 31:16rules of thumbs
- 33:23get started
- 33:50thank you!
related talks
39mResilience & Chaos Engineering in the Cloud
42mChaos Engineering: Why Breaking Things Should be Practiced
35mImprove resilience with automated chaos engineering
57mChaos Engineering Primer
26mImprove application resilience with chaos engineering
34mChaos engineering for serverless
18mChaos Experiments under the lens of AIOps
33mBeyond the Monkeys Chaos Engineering on the Cloud