AWS News Blog Reports Faster Scaling for ECS – What Changed?

The AWS News Blog recently announced that Amazon Elastic Container Service (ECS) now supports high‑resolution metrics that update every 20 seconds instead of the previous 60‑second interval. This change makes auto scaling much faster — according to AWS testing, the time to trigger a scale‑up dropped by 76%, and the total time to provision new tasks improved by 72%.

For any business running containerised applications on AWS, faster scaling means your system can react to sudden traffic surges without waiting a full minute for fresh data. Instead of relying on slower, one‑minute metrics that often force you to keep extra capacity “just in case,” you can now rely on near‑real‑time signals to add or remove resources quickly.

Why High‑Resolution Metrics Matter for Cloud Workloads

In traditional cloud setups, auto scaling policies often feel like driving a car with foggy windows — you can see the road ahead, but the delay makes you guess. With 20‑second resolution, the fog clears. Your application can respond to load changes almost as they happen, reducing the risk of downtime or slow user experiences during unexpected demand.

This isn’t just a technical tweak; it changes how you plan capacity. Many companies over‑provision compute resources because they worry that slow scaling won’t keep up. Faster metrics let you run closer to your actual needs, cutting waste. The trade‑off is that high‑resolution CloudWatch metrics incur additional costs, while the standard 60‑second metrics remain free. So you need to weigh the performance gain against the extra monitoring expense.

What This Means for Australian SMBs

Australian small and mid‑sized businesses often run seasonal or event‑driven applications — think e‑commerce during the Christmas rush, ticketing for a local festival, or software that gets hit with Monday‑morning traffic. Until now, slower scaling meant you either paid for idle capacity or risked slow pages after a spike. The new ECS high‑resolution metrics give you a third option: scale fast enough to handle spikes without padding.

Because the feature works across AWS Fargate, ECS Managed Instances, and EC2, you can adopt it without changing your underlying infrastructure. For SMBs with limited DevOps teams, the simpler setup — just flipping a few settings in the console — means you don’t need custom engineering to get aggressive scaling behaviour. The cost of CloudWatch high‑resolution metrics is modest for most workloads, especially when compared to the savings from not over‑provisioning.

What You Can Do Now

  • Review your current ECS scaling policies. If you use target tracking with standard 60‑second metrics, test the new high‑resolution option in a non‑production environment first.
  • Estimate the CloudWatch cost increase. High‑resolution metrics (20 seconds) cost more than standard metrics. Use the AWS Pricing Calculator to see the difference for your expected metric volume.
  • Enable high‑resolution metrics on existing services. In the ECS console, update your service to generate 20‑second metrics, then change the scaling policy to use the new metric names (e.g., ECSServiceAverageCPUUtilizationHighResolution).
  • Right‑size your baseline task count. After enabling faster scaling, try reducing the minimum number of tasks you keep running. The quick scale‑out should handle spikes, lowering your baseline compute costs.
  • Monitor performance for a week. Check whether your application responds to traffic spikes within the new 86‑second (or faster) window. Adjust scaling thresholds if needed.

At MS&VG, we help Australian SMBs get the most out of AWS without the steep learning curve. Our cloud consultants can guide you through enabling faster auto scaling, optimising costs, and ensuring your applications stay responsive during peak demand.