Home icon

Optimize compute resources on Amazon ECS with Predictive Scaling

Containers Blog



AWS has introduced Predictive Scaling for Amazon ECS, a new feature that uses machine learning to anticipate and proactively scale container services based on historical demand patterns.

  • Uses advanced ML algorithms to forecast demand and scale tasks before traffic spikes
  • Works alongside existing reactive scaling policies like target tracking
  • Helps improve application availability and performance
  • Operates in two modes: Forecast Only and Forecast And Scale
  • Learns from up to 14 days of historical data to generate hourly forecasts

Ideal for applications with consistent daily or weekly demand patterns, Predictive Scaling helps optimize compute resources by scaling tasks proactively and reducing over-provisioning costs.



Go to article

The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.

Related articles

Nov 12
2024
Optimizing compute-intensive tasks on AWS
Jun 11
2026
Optimize EC2 costs with AWS Compute Optimizer right sizing
Jun 24
2024
Proactive scaling of Amazon ECS services using Amazon ECS Service Connect Metrics
Jan 27
2024
Optimize costs by automating AWS Compute Optimizer recommendations

The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.