Spikes Test
Spike testing validates how an application behaves and recovers when subjected to sudden, extreme changes in load.
Spike testing validates system resilience when traffic changes instantaneously, not gradually. The goal is not to see how much load the system can take long-term, but
- How it absorbs shock
- How it fails (if it does)
- How fast it recovers
Why Spike Testing
| Spike Testing | Chaos Engineering |
|---|---|
| Predictable load spikes | Random failures |
| Focus on traffic | Focus on fault tolerance |
| Pre-release | Often in production |
When Spike Testing Is Critical
Spike testing is essential when:
- Traffic is event-driven
- User behavior is unpredictable
- Business impact of downtime is high
What Spike Testing Actually Validates
Application Layer:
- Thread management
- Request queuing
- Garbage collection behavior
- Dimeter, SIP, HTTP-2 messages or API timeouts
- Retry logic
Infrastructure Layer
- Auto-scaling trigger thresholds
- Load balancer behavior
- Container startup times
- VM provisioning delay
Data Layer
- Connection pool limits
- Lock contention
- Cache eviction storms
- Read/write latency spikes