Q: You have a fleet of EC2 instances behind an ALB. One instance keeps getting traffic even though it's unhealthy. What's wrong?

The ALB health check is likely misconfigured:

• Health check path wrong — ALB checks / but the app serves health at /health. Returns 404 → ALB marks it unhealthy → but wait, 404 might be in the success codes range.
• Success codes misconfigured — if success codes include 404 or 5xx, unhealthy instances appear healthy.
• Health check threshold — UnhealthyThreshold might be set very high, so the instance needs to fail many times before being marked unhealthy.

Fix: Review ALB target group health check settings. Set correct path, success codes (200-299 only), and reasonable thresholds.

• Health check port/protocol wrong — checking wrong port.
• Security Group — ALB can't reach the instance's health check port → health checks time out → AWS treats timeout differently than explicit failure in some configs.