Explanation
Why the waste happens and who it affects.
Fleets often remain oversized due to conservative default configurations or lack of schedule-based scaling. Tuning the scaling policies to better reflect usage patterns ensures that streaming infrastructure aligns with actual demand.
Billing model
The pricing dimensions that drive this cost.
Amazon WorkSpaces Applications (formerly AppStream 2.0) streaming instances are billed according to fleet type. Always-On fleets are charged for every running instance whether or not it is in use (Windows hourly; Linux in one-second increments with a 15-minute minimum). On-Demand fleets charge the running rate only while an instance is in a streaming session and a smaller stopped-instance fee otherwise. Elastic fleets are charged in one-second increments, with a 15-minute minimum, for the duration of each streaming session. For Always-On and On-Demand fleets, minimum instance counts and scheduled provisioning settings directly influence total cost.
How to detect
4 checks to find it in your estate.
- Identify WorkSpaces Applications fleets where the minimum instance count consistently exceeds active session demand over a representative period
- Review fleet usage patterns and idle capacity during off-peak hours or low-concurrency windows
- Check whether auto scaling policies are based on real usage or static assumptions
- Assess whether scheduled scaling is configured to adjust fleet size based on predictable usage cycles
How to fix
4 ways to remove the waste.
- Adjust minimum instance counts in auto scaling policies to reflect observed demand
- Implement schedule-based scaling to reduce instance counts during predictable low-usage periods and increase during peak hours
- Evaluate whether suitable Always-On fleets can move to On-Demand or Elastic fleets, which do not charge the full running rate for idle capacity
- Regularly review and update scaling policies based on current usage data to ensure ongoing efficiency
Documentation
Vendor references for pricing and configuration.