- 최신
- 최다 투표
- 가장 많은 댓글
You are billed when the instance is in running state. You can find additional details here
--Syd
My understanding is that billing should start and stop based on the API request. So for EC2, from the time an instance is launched (created) until it is terminated or stopped.
Billed per second; Minimum of 1 Minute (On-demand, Reserved, Spot) - Amazon linux - Ubuntu - EBS
Billed per hour - Windows - Commercial Linux (Red Hat EL, SUSE)
This would be similiar for RDS but would depend on the OS. From the time a DB instance is launched until it is stopped or deleted.
Does this help? More info can be found here: https://aws.amazon.com/pricing/
but how do you define created? Is it the moment the instance is initializing or the moment it is done initializing and is ready to accept requests?
Not exactly. The instances are billed only when they are in the "running" state. If they are in any other state - they are not charged. For example, when you want to shut it down, it will be in "terminating" state for a while, as soon as the state changes - you are no longer billed for it (since it's not "running" anymore).
With that said, note that sometimes EC2 instances may be in the "running" state, however they may be unreachable (for example, you cannot SSH or access it due to your own misconfiguration) or they may be failing health checks... You are still billed for the EC2 running instance in this case. It's in your interest to fix any possible issues you might have with the resource itself that is not related to underlying infrastructure (things like networking misconfiguration or lost access keys, etc)
+1 to this. Worth noting that sometimes EC2 instances may be in the "running" state, however they may be unreachable (for example, you cannot SSH or access it due to your own misconfiguration) or they may be failing health checks... You are still billed for the EC2 instance as long as it's running. It's in your interest to fix any possible issues you might have with the resource itself that is not related to underlying infrastructure (things like networking misconfiguration or lost access keys, etc)