Gremlin

Identify the target Kubernetes cluster.

Select a 'Node Shutdown' attack.

Set the blast radius to 1 node.

Run experiment and monitor pod migration time.

Verify that traffic remains uninterrupted via the Load Balancer.

Configure a 'Blackhole' attack on all traffic to us-east-1.

Execute the attack.

Check Route 53 health checks.

Verify secondary region database promotion.

Monitor global latency for users.

Select the application-to-database network path.

Apply 'Latency' fault (e.g., +500ms).

Monitor application response times.

Verify that the circuit breaker opens and returns a cached response.

Stop attack and check recovery time.

Select target service group.

Apply a 'CPU Stress' attack at 90% load.

Observe AWS CloudWatch scaling metrics.

Confirm new instances are spun up and healthy.

Terminate attack and verify downscaling.

Target the egress traffic to the 3rd party IP range.

Apply a 100% packet loss 'Blackhole' attack.

Verify the 'Checkout' page still loads with a 'Service Unavailable' message for that specific feature.

Confirm the rest of the application remains functional.

Select a target host.

Run 'Disk' attack to fill /var/log to 95%.

Check if log rotation kicks in or if the app crashes.

Verify alerting triggers in the monitoring system.

Cleanup and restore disk space.

Apply a 'DNS' attack to redirect internal calls to a non-existent IP.

Monitor if the security team receives a 'DNS resolution failure' alert.

Evaluate the time to detection vs. time to remediation.

Document the response gap.