To start using Capacity Blocks, you first find and purchase an available Capacity Block that matches your reservation size, duration, and timing needs. Then, when the reservation begins, you can use the Capacity Block by launching instances that target the reservation ID. Thirty minutes before the reservation expires, we begin to terminate any instances that are still running in the Capacity Block.
Capacity Blocks are delivered as targeted
Capacity Reservations in a single Availability Zone.
To run instances in a Capacity Block, you must specify the reservation ID when launching your
instances. If you stop instances on your own and the Capacity Block expires, you can't restart
them until you target another Capacity Block in the active
state.
By default, Capacity Blocks deliver low-latency, high-throughput network connectivity between the instances inside the Capacity Block, so there is no need to use a cluster placement group with a Capacity Block.