AWS Fault Injection Service
Chaos engineering experiments
Runs controlled fault-injection experiments (stopping instances, spiking CPU, blocking network, slowing APIs…) against AWS resources to verify that the system tolerates failures well, with stop conditions based on CloudWatch alarms.
- Cost$ Low
- Difficulty●●●●●
- Costs money when idle?No
- Serverless
Think of it asLike a fire drill: deliberately ringing the alarm to see whether everyone evacuates following the correct procedure.
When to use it
✓ Use it when
You want to verify high availability, auto scaling, and failover architectures before a real incident happens.
✗ Avoid it when
You only need a theoretical assessment of application resilience → Resilience Hub; systems that don't yet have monitoring and alarms.
Common uses
- Randomly terminate instances in an Auto Scaling group
- Simulate an AZ outage
- Congest the network between services to test retries/timeouts
What does it cost?
- Pricing model: Charged per minute each action in the experiment runs.
