Evaluated against the current usage in panel 2. Percent actions use the documented Application Auto Scaling rounding: results above 1 round down, fractions between 0 and 1 become 1 (and mirrored for scale-in), so a percent policy always moves capacity by at least one session. New desired capacity is clamped to your min / max user sessions.
| Instances needed | roundup(user sessions / max sessions per instance) - doc examples: min 12 sessions at 5 per instance = roundup(12/5) = 3 instances; max 52 sessions = roundup(52/5) = 11 instances. |
| CapacityUtilization | (ActiveUserSessionCapacity / ActualUserSessionCapacity) × 100 |
| ActualUserSessionCapacity | AvailableUserSessionCapacity + ActiveUserSessionCapacity, and ActualCapacity × max sessions per instance at the instance level. |
| DesiredUserSessionCapacity | DesiredCapacity × max sessions per instance - capacity and scaling policies are configured in user sessions; the service calculates and launches the instances. |
| Percent policy action | Doc example: current capacity 4, action "Add 25% capacity" gives desired capacity 5. Rounding for non-integer results: >1 rounds down (12.7 becomes 12), 0 to 1 becomes 1, 0 to -1 becomes -1, <-1 rounds up toward zero (-6.67 becomes -6). |
| Scaling metrics | Capacity Utilization (scale on % used), Available Capacity (keep a session buffer), Insufficient Capacity Error (rejected session requests > 0). |
Sources:
Scaling concepts ·
Multi-session fleet metrics ·
Application Auto Scaling step scaling.
Personal helper tool, not an official AWS calculator. Billing is per instance, so the instance counts are the cost line.
For Foxy and Hooky ♡