Monitoring & Observability
AWS Fault Injection Service
Chaos engineering as a managed service.
Official docsOverview
FIS runs controlled fault-injection experiments (terminate EC2, throttle API, inject latency in RDS) with stop conditions tied to CloudWatch alarms.
When to use it
- Validate resilience and runbooks
- Game days
- Pre-prod chaos tests
Setup
- Create experiment template (actions + targets + stop condition).
- Run from console or CLI.
How to use
Run
aws fis start-experiment --experiment-template-id EXT123QA use cases
- Inject AZ failure during a load test; verify Route 53 + Multi-AZ RDS keep SLOs.