Kubernetes Cluster: Vertical Scaling
Vertical scaling for the Kubernetes Cluster is represented via two implementations:
- platform-managed automatic vertical scaling - allows dynamically allocating resources for the Kubernetes nodes, which are used (and charged) only when they are needed
- Kubernetes-managed Vertical Pod Autoscaler (beta) - adjusts pod memory/CPU requests and limits on the go
The combination of PaaS vertical scaling and VPA can help minimize resource utilization and thus reduce cluster maintenance costs. In addition, it makes deployments highly available and fault-tolerant.