Stress Testing

Stress testing pushes the system beyond its normal operating capacity to determine where it breaks and how it recovers. Unlike load testing (expected load), stress testing intentionally overloads the system.

Goals of stress testing:

• Find the system's breaking point
• Identify how the system fails (graceful degradation vs crash)
• Verify recovery behavior after overload
• Discover resource leaks under extreme conditions

Types:

• Sustained stress: Maintain high load for extended periods
• Spike stress: Sudden, extreme load increase
• Gradual stress: Slowly increase load until failure

What to look for:

• Memory leaks (memory usage grows without releasing)
• Connection pool exhaustion
• Database deadlocks
• Disk space depletion
• Error cascade (one failure triggers others)
• Data corruption under extreme load

Good systems degrade gracefully: they slow down, queue requests, or return meaningful errors — they don't crash, corrupt data, or become unresponsive without explanation.