AWS › Auto Scaling Groups
How can predictive scaling in Auto Scaling Groups improve upon reactive scaling policies?
AdvancedAnswer
Predictive scaling uses machine learning to analyze your application's historical load patterns and automatically forecasts future capacity needs, proactively scaling out ahead of expected demand increases rather than waiting for a metric like CPU utilization to actually cross a threshold, which is especially valuable for applications with predictable daily or weekly traffic patterns where reacting after the fact could mean brief periods of degraded performance.
aws autoscaling put-scaling-policy --auto-scaling-group-name my-asg --policy-name predictive-scaling --policy-type PredictiveScaling --predictive-scaling-configuration file://predictive-config.jsonReal-world example A food delivery application with predictable spikes every evening during dinner hours uses predictive scaling to launch additional capacity just before the expected surge begins, rather than reacting only after CPU utilization has already climbed and users have started experiencing slower response times.