Issue with Azure Virtual Desktop Scaling Plan

We are currently using following combinations to optimise costs on Azure Virtual Desktop

• Start VM on connect
• Scaling Plan

We would like to achieve the following requirement to further optimise the costs.

• 1st virtual machine starts when 1st user logins.
• 2nd virtual machine starts when 2nd user logins
• 3rd virtual machine starts when 3rd user logins

When the user logout, the VM should be powered off.

We have used the following Scaling Plan configurations to achieve this

Total Session Hosts: 5

Pool Type: Pooled

Maximum Session Limit: 1

Load balancing algorithm: Depth-first

Ramp-up

Start time (12 hour system) : 01:00pm

Load balancing algorithm: Depth-first

Minimum percentage of hosts (%): 0

Capacity threshold (%): 100

Peak Hours

Start time (12 hour system) : 01:05pm

Load balancing algorithm: Depth-first

Capacity threshold (%): 100

Ramp-down

Start time (12 hour system) : 01:10pm

Load balancing algorithm: Depth-first

Minimum percentage of active hosts (%): 0

Capacity threshold (%): 100

Off-Peak Hours

Start time (12 hour system) : 01:15pm

Load balancing algorithm: Depth-first

Capacity threshold (%): 100

With this scaling plan configurations we are able to achieve start first VM when a first user logins, which takes approximately 2 mins, but when second user logins, the start of 2nd VM takes more than 4 mins and sometimes user receives Disconnect (no resources available) message. This is same for any subsequent user logins (3rd 4th user etc).  Also when any of the users logout, the VM should get turn off, but that does not happen for subsequent user logins (2nd, 3rd, 4th etc). It only works for first user login/logout.

Could you please help why this scaling plan works for first user and not for any subsequent user logins?