Set up Load Balancing

1. By default, Porter uses a round-robin load balancing algorithm for distributing requests across multiple instances.

2. To modify the load balancing algorithm, add ingress annotations to your web service in the porter.yml file:

   services:
     - name: ${SERVICE_NAME}
       type: web
       ingressAnnotations:
         nginx.ingress.kubernetes.io/load-balance: "ewma"  # or "round_robin"

3. Choose between two available algorithms:

   • round_robin - Distributes requests sequentially across available instances

   • ewma - (Exponentially Weighted Moving Average) Routes based on server response times

Usage

The load balancing configuration works automatically with your autoscaling settings. In the example configuration:

• Minimum instances: 1

• Maximum instances: 5

• Scaling triggers: 70% CPU or memory utilization

When autoscaling creates new instances, the load balancer automatically includes them in the request distribution according to the configured algorithm.