Load balancer
A server that spreads incoming requests across a group of backend servers, and stops sending to a backend that fails its health check. It lets you add servers behind one address, and it decides which server gets each request.
Network
In Quorum#
The Load Balancing Research unlocks the Load Balancer, which handles 2,000 requests per second per size step and adds 1 ms. By default it splits traffic by each App Server's capacity. It drops a dead host at its next health check, every 10 s by default, while DNS round robin keeps sending to a dead host. Load balancing algorithms adds least connections and power of two choices, and Health checks lets you check every 2 s or 30 s.
Read next#
Health check, Power of two choices, Vertical and horizontal scaling, Consistent hashing