So the short answer is that on-demand servers are only charged when it is running and not in a stopped state. You can verify this programmatically or through the console what the current state of your EC2 instance is. That being said, you are charged for any attached storage like EBS volumes.

Assuming that your instance only starts up when the failover is started/detected you would then get charged for the instance usage.

Example:

t2.large on-demand in US-East-1

• EC2 - $0.0928/hr
• EBS - $0.10/GB/month if gp2
• Route53 Health Checks: $0.50/mo per health check plus another $0.75 per million health check queries assuming that you have that sort of setup.

Also remember that EC2 on-demand are charged based on the "per instance-hour" consumed. You are charged for at least 1 hour each time the instance is started.

Here is a thread that discusses "an instance hour": https://stackoverflow.com/questions/273489/can-someone-explain-the-concept-of-an-instance-hour-as-used-by-cloud-computing#:~:text=An%20instance%20hour%20is%20simply,you%20used%20it%20or%20not.

I would advise to utilize the official AWS Pricing Calculator https://calculator.aws/#/

Put in your entire setup and review what the actual costs would be.

Let me know if that answers your query.