Add circuit breaker mechanism to prevent cascading backend failures

[Prithvish7]

Currently, Routrix mainly depends on periodic health checks to decide whether a backend is alive or not. However, in real scenarios a backend might still be marked as alive but keep failing at request level (timeouts, connection errors, slow responses).In such cases, the load balancer can continue sending traffic to that backend until the next health check marks it down, which may lead to repeated failures and cascading issues. To improve reliability, I propose adding a circuit breaker mechanism for backends. I’d like to work on this issue if it aligns with the project goals.