View a markdown version of this page

Using a Capacity Reservation with an Aurora DB instance - Amazon Aurora

Using a Capacity Reservation with an Aurora DB instance

Reserved capacity is used only when you explicitly target the reservation. You reference the reservation by its identifier (dbcr-) or its name when you perform any of the following operations:

  • Create a DB instance in a DB cluster (CreateDBInstance) – for the writer and for each Aurora Replica

  • Modify a DB instance to move it into a reservation (ModifyDBInstance)

The DB instance's engine, DB instance class, and Availability Zone must match the reservation, and the reservation must have available capacity in that Availability Zone (the matching entry in AvailableCapacityPerAvailabilityZoneList has an AvailableCapacity greater than 0). If they don't match or no capacity is available, the operation fails with an error that says the Availability Zone doesn't have enough capacity.

When you add a DB instance to an Aurora DB cluster in the Amazon RDS console, in the Capacity reservation section, choose the reservation to launch into.

Add the --db-capacity-reservation-identifier parameter to the create or modify command. For example, to create a DB instance that uses a reservation:

aws rds create-db-instance \
    --db-instance-identifier my-db-instance \
    --db-cluster-identifier my-aurora-cluster \
    --engine aurora-postgresql \
    --db-instance-class db.r7g.large \
    --availability-zone us-east-1a \
    --db-capacity-reservation-identifier my-reservation

The --db-capacity-reservation-identifier parameter accepts either the reservation name or its dbcr- ID.

To restore into a reservation, restore the DB cluster first, and then add its DB instances with the reservation. A DB cluster restore (restore-db-cluster-from-snapshot or restore-db-cluster-to-point-in-time) creates the DB cluster without DB instances and doesn't take a Capacity Reservation.

aws rds restore-db-cluster-from-snapshot \
    --db-cluster-identifier my-restored-cluster \
    --snapshot-identifier my-cluster-snapshot \
    --engine aurora-postgresql

aws rds create-db-instance \
    --db-instance-identifier my-restored-instance \
    --db-cluster-identifier my-restored-cluster \
    --engine aurora-postgresql \
    --db-instance-class db.r7g.large \
    --db-capacity-reservation-identifier my-reservation

To move an existing DB instance into a reservation, specify the reservation in a modify request that also changes the DB instance class. The reservation must match the configuration that the DB instance has after the change. For example, to scale a DB instance to db.r7g.large and place it in a db.r7g.large reservation:

aws rds modify-db-instance \
    --db-instance-identifier my-db-instance \
    --db-instance-class db.r7g.large \
    --db-capacity-reservation-identifier my-reservation \
    --apply-immediately

A modify request that specifies a reservation without a DB instance class change fails.

When DB instances are placed into the reservation, the AvailableCapacity for the affected Availability Zone decreases. You can track this with describe-db-capacity-reservations.

Specify the DBCapacityReservationIdentifier parameter in the CreateDBInstance or ModifyDBInstance operation.