I created REPLICA service and added 2 policies for scaling it. One for scale-in, second is for scale-out. Configured alarms for them and started testing service.
When CPUUtilization value pass the alarm limits and alarm's state becomes ALARM. It's history says 'Successfully executed action...'. In service i see that desired count
value becomes 2, but only one task is running.
What the problem? How should i configure scaling to run additional tasks?
Scaling policies
According to screenshots, It seems your scaling policies are working fine. However, Your RUNNING tasks do not equal the number of the desired count.
I suggest you check the 'Events Tab' to see the exact reason why your tasks were not able to run.