Random load balancing is one of the simplest strategies: each incoming request is directed to a randomly chosen target. While it doesn’t account for load or performance differences among backends, it can be surprisingly effective in distributing traffic in environments where the backends have similar capabilities or where simplicity is paramount.

​
How It Works

  1. Random Selection Each time a request arrives, the gateway picks one of the available targets at random.

  2. Even Distribution Over Time With enough requests, traffic typically evens out across targets, though short-term fluctuations can occur.

  3. Low Overhead Random load balancing requires minimal overhead since no real-time tracking of connection counts or performance metrics is necessary.

​
Example: Creating an Upstream with Random Load Balancing

Below is an example of how to create an Upstream using the random algorithm. The configuration includes two targets, each connecting to different services.

curl -X POST http://localhost:8080/api/v1/gateways/{gateway-id}/upstreams \
  -H "Content-Type: application/json" \
  -d '{
    "name": "random-upstream",
    "algorithm": "random",
    "targets": [
      {
        "host": "api.serviceA.com",
        "port": 443,
        "protocol": "https",
        "priority": 1,
        "credentials": {
          "header_name": "Authorization",
          "header_value": "Bearer serviceA-key"
        }
      },
      {
        "host": "api.serviceB.com",
        "port": 443,
        "protocol": "https",
        "priority": 1,
        "credentials": {
          "header_name": "Authorization",
          "header_value": "Bearer serviceB-key"
        }
      }
    ],
    "health_checks": {
      "passive": true,
      "threshold": 3,
      "interval": 60
    }
  }'