Load Testing · Middle · Question 28 of 76
During a spike test, autoscaling adds servers, but errors happen in the first minutes. Why?
- 1Load balancers block all traffic while new servers are added
- 2New instances take time to start and warm up, so the spike hits first
- 3Autoscaling always removes servers before it adds new ones
Show the answer
Correct answer: New instances take time to start and warm up, so the spike hits first
Explanation
Scaling reacts to metrics with a delay, and new instances need to boot and warm caches. Spike tests reveal this gap; pre-scaling or buffers close it.