nginx HTTP load balancer

Load balancing across multiple application instances optimizes resource utilization

Methods

• round-robin: requests are distributed in a round-robin fashion, assigning sequentially (default)
• least-connected: next request is assigned to the server with the least number of active users
• ip-hash: a hash function is used to determin what server should be selected (based on IP)
  • good for "sticky" configurations, of sending clients to the same app server

Default Config

http {
    upstream myapp1 {
        server srv1.example.com;
        server srv2.example.com;
        server srv3.example.com;
    }

    server {
        listen 80;

        location / {
            proxy_pass http://myapp1;
        }
    }
}

• round-robin by default, or define least_conn directive at the top of the upstream block
• ip_hash directive in the upstream group configuration applies the ip hash algorithm
• weight= is a directive (optional parameter of the server directive), increases likelihood of matching in load balancing (default weight=1)

Health checks

Reverse proxy implementation in nginx includes in-band server health checks

• if response from a particular server fails with an error, load balancer will avoid selecting this server for subsequent requests
  • max_fail directive (for server) sets number of consecutive unsuccessful attempts during fail_timeout (default 1)