Fleet Usage Metrics for Multi-session Fleets

The percentage of sessions in a fleet that are being used, using the following formula.

CapacityUtilization = (ActiveUserSessionCapacity / ActualUserSessionCapacity) * 100

Monitoring this metric helps with decisions about increasing or decreasing the value of a fleet's desired capacity.

The total number of session slots that are available for streaming or are currently streaming.

ActualUserSessionCapacity = AvailableUserSessionCapacity + ActiveUserSessionCapacity

The number of idle session slots currently available for user sessions.

AvailableUserSessionCapacity = ActualUserSessionCapacity - ActiveUserSessions

The total number of session slots that are either running or pending. This represents the total number of concurrent streaming sessions your fleet can support in a steady state.

DesiredUserSessionCapacity = ActualUserSessionCapacity + PendingUserSessionCapacity

The number of user sessions currently being used for streaming sessions.

The number of session slots being provisioned by WorkSpaces Applications. Represents the additional number of streaming sessions the fleet can support after provisioning is complete. When provisioning starts, it usually takes 10-20 minutes for an instance to become available for streaming.

The total number of session slots currently that are available for streaming or are currently streaming. Represents the number of concurrent streaming sessions that can be supported by the fleet in its current state.

This metric is provided for Always-On fleets only, and has the same value as the ActualUserSessionCapacity metric.

The number of instances in Drain Mode.

The number of active user sessions on instances in Drain Mode.

The number of unused session slots on instances in Drain Mode which cannot be used for user session provisioning.

The total number of instances that are available for streaming or are currently streaming. For multi-session fleets, this metric represents instance-level capacity. The relationship between instance-level and session-level capacity is as follows:

ActualUserSessionCapacity = ActualCapacity × (Maximum number of user sessions per instance)

ActualCapacity = ActualUserSessionCapacity / (Maximum number of user sessions per instance)

The total number of instances that are either running or pending. This represents the total number of instances your fleet targets in a steady state. For multi-session fleets, the relationship between instance-level and session-level capacity is as follows:

DesiredUserSessionCapacity = DesiredCapacity × (Maximum number of user sessions per instance)

DesiredCapacity = DesiredUserSessionCapacity / (Maximum number of user sessions per instance)

The number of instances currently being used for streaming sessions. For multi-session fleets, an instance is considered in use if it has at least one active user session. The relationship between instance-level and session-level capacity is as follows:

ActiveUserSessionCapacity = InUseCapacity × (Maximum number of user sessions per instance)

InUseCapacity = ActiveUserSessionCapacity / (Maximum number of user sessions per instance)

The number of instances being provisioned by WorkSpaces Applications. Represents the additional number of instances the fleet will have after provisioning is complete. When provisioning starts, it usually takes 10-20 minutes for an instance to become available for streaming. For multi-session fleets, the relationship between instance-level and session-level capacity is as follows:

PendingUserSessionCapacity = PendingCapacity × (Maximum number of user sessions per instance)

PendingCapacity = PendingUserSessionCapacity / (Maximum number of user sessions per instance)

The total number of instances currently running. Represents the number of instances that can support streaming sessions in the fleet's current state. For multi-session fleets, the relationship between instance-level and session-level capacity is as follows:

RunningUserSessionCapacity = RunningCapacity × (Maximum number of user sessions per instance)

RunningCapacity = RunningUserSessionCapacity / (Maximum number of user sessions per instance)

This metric is provided for Always-On fleets only, and has the same value as the ActualCapacity metric.

The number of session requests rejected due to lack of capacity. This metric applies the same way for both single-session and multi-session fleets.

You can set alarms to use this metric to be notified of users waiting for streaming sessions.