BGTech blog.bgyani.co.in
Visual Explainer

Health Checks in Load Balancer

This simulation shows how a load balancer actively probes servers, counts failures, marks unhealthy nodes down, and forwards user traffic only to healthy backends in real time.

Controls
Health Check

Probe Behavior

Traffic

Client Requests

Behavior: the client continuously sends requests to the load balancer. The load balancer forwards requests only to healthy servers and skips any server marked down.
Server State

Failure / Recovery

Current Mode

Live Logic

Check Type: HTTP
Interval: Normal
Threshold: 2
Routing: Round-robin across healthy servers only
Live Flow
Traffic Probe Success Probe Failure
Traffic Source Load Balancer Backend Servers Client Request packets move to LB LB Decision Logic Healthy-only routing Round robin among UP servers Health Check Engine Type: HTTP Interval: Normal Threshold: 2 Server A Healthy Failures: 0 Server B Healthy Failures: 0 Server C Healthy Failures: 0
Healthy Servers
3
Traffic State
ON
Probe Interval
Normal
Last Routed To
Server A
What you are seeing: the LB keeps sending active probes. A failed server is not removed instantly unless it reaches the selected failure threshold. Once a failed node starts responding again, the next successful check resets its counter and the server becomes eligible for traffic again.