Overview#

A controlled environment for creating infrastructure failures and observing how applications and monitoring systems respond.

Reading about failure modes is not the same as watching one happen. This lab exists so that the interesting failure scenarios are reproducible instead of theoretical.

Scenarios#

• Pod termination
• Resource exhaustion (CPU and memory)
• Broken dependencies
• Node failure
• Increased network latency
• Bad deployments
• Application crashes

The loop I repeat for every scenario.

Inject failure
   ↓
Observe signals
   ↓
Compare to expectation
   ↓
Adjust alerts / probes
   ↓
Repeat