An application is running on Amazon EC2 instances behind a Classic Load Balancer. The instances run in an Auto Scaling group across multiple Availability Zones. Occasionally multiple incoming…

Question

Options

• AHealthyHostCount
• BBackendConnectionErrors
• CSurgeQueueLength
• DUnHealthyHostCount

• A
  3% (1)
• B
  12% (4)
• C
  79% (26)
• D
  6% (2)

Explanation

SurgeQueueLength measures the number of pending requests queued by the Classic Load Balancer waiting for a healthy backend instance - when this queue fills up, the load balancer starts dropping requests and returning 5xx errors, which is exactly what the SpillOverCount metric captures. By triggering Auto Scaling actions when SurgeQueueLength grows, you add capacity before the queue overflows, preventing those errors entirely.

Why the distractors are wrong:

• A (HealthyHostCount) tells you how many healthy instances exist, but scaling on this alone doesn't account for actual request pressure - you might have plenty of healthy hosts that are still overwhelmed.
• B (BackendConnectionErrors) reflects connection failures between the load balancer and instances, which is a symptom of a different problem (network/app issues), not queue overflow.
• D (UnHealthyHostCount) is useful for detecting failed instances but doesn't signal queue saturation - you could have zero unhealthy hosts and still overflow the surge queue under high traffic.

Memory tip: Think of SurgeQueue → SpillOver as a bucket overflowing - SurgeQueueLength is the water level, SpillOverCount is the spill. Scale on the water level before it overflows, not after.

No community discussion yet for this question.