What happens to my existing provisioned instance when AWS runs out of instance capacity?

Question

I am using the g4dn.xlarge instance types (on-demand), distributed across the AZs: us-east-1a,us-east-1b,us-east-1c,us-east-1d,us-east-1f. They work fine until one fine evening I cannot power on any of these instance types and get a message via the API:
"Temporary insufficient AWS capacity for instance type g4dn.xlarge in the following zones: us-east-1a,us-east-1b,us-east-1c,us-east-1d,us-east-1f"

I am not using any Capacity reservations here. I would like to know if these instance types were given to someone else when I tried to power them on? How does this works as I see this issue recovers on its own?

Accepted Answer

When you stop an EC2 instance you free up the compute capacity that was assigned to your instance for anyone else wanting to launch that type of instance in the particular AZ.  This means that you will only be able to start a stopped instance if there is available on-demand capacity in the particular AZs.

As you've identified, unless you pay extra for a Capacity Reservation, I'm not aware that a stopped instance has any priority over available capacity when restarting.

You may find these help pages useful:
* https://aws.amazon.com/premiumsupport/knowledge-center/ec2-insufficient-capacity-errors
* https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/troubleshooting-launch.html#troubleshooting-launch-capacity

What happens to my existing provisioned instance when AWS runs out of instance capacity?

Relevant content