DataDog: Triggering an alert when EC2 instance changing its state

I want to create DataDog alerts when an EC2 instance changes its state:

1. Started - send an email with INFO notification "Instance i-80989890d UP"
2. Stopped - send an email with INFO notification "Instance i-80989890d DOWN"
3. Not accessible- send an email when threshold reached with CRITICAL status
4. Idle - create a forecasting alert when EC2 does not utilize its resources and might be replaced with a less expensive instance

This is forbidden to use CloudWatch by customers, so only DataDog is eligible.