Overview
Minimise costs – only pay for what we need
Reactive scaling – Scale to meet demand based on metrics
Proactive scaling – Scale to meet expected demand
Horizontal Scaling
Vertical Scaling
Verne Scaling
Horizontal Scaling
Horizontal Scaling is achieved by adding or removing nodes of the same size to or from the cluster. Scaling up and down can be triggered by different methods including;
Manual Trigger – If the cluster is manually configured setting the desired node count by manually adding or removing nodes
AWS Resource Thresholds – trigger adding or removing nodes based on AWS Metrics such as CPU
Cluster Resource Thresholds – trigger adding or removing nodes based on cluster Metrics such as CPU
Scheduled – trigger adding or removing nodes based on business hours
Vertical Scaling
Vertical Scaling is achieved by adding larger (or smaller) nodes and removing the retired nodes. Scaling up and down can be triggered by different methods including;
Manual – If the cluster is manually configured setting the desired node size
AWS Resource Thresholds – trigger adding or removing nodes of the desired size based on AWS Metrics such as CPU
Cluster Resource Thresholds – trigger adding or removing nodes of the desired size and cluster Metrics such as CPU
Scheduled – trigger adding or removing nodes based on business hours
