Cloud-Native Microservices With Kubernetes - 2nd Edition

The Frontiers of Horizontal Scaling

Horizontal scaling, even coupled with cluster autoscaling, is not a magic solution to all problems. Here is a scenario:

Your application needs 1.5Gi of memory to handle the current load. Your initial requests and limits were configured to 512Mi:

[...]
        resources:
          requests:
            memory: "512Mi"
            cpu: "250m"
          limits: