Elastic Pool Billing when a Pool is Created or Terminated

When an elastic pool is created or terminated, the leader is billed for the full hour for the elastic pool. In addition, individual instances that are either added or removed from the pool are billed for any compute usage that occurs while the instance is not in the elastic pool (in this case the billing applies to the individual Autonomous Database instance).

• Pool Creation Example: Assume there is an Autonomous Database instance with 4 ECPUs that is not part of any elastic pool. At 2:15pm, if you create an elastic pool with this instance with a pool size of 128 ECPUs, the instance becomes a pool leader. Assuming the Autonomous Database idles between 2-3pm, and there are no other Autonomous Database instances in the pool, billing for the hour between 2-3pm is as follows:

  The bill for the period 2-3pm is: (4 * 0.25) + 128 = 129 ECPUs

  Where the (4 * 0.25) is the billing for compute for the fifteen minutes before the Autonomous Database instance created the elastic pool, and 128 ECPUs is the billing for the elastic pool for the hour when the elastic pool is created.

• Pool Termination Example: Assume an Autonomous Database instance with 4 ECPUs is the leader of an elastic pool and the pool size is 128 ECPUs. At 4:30pm, if you terminate the elastic pool, the database becomes a standalone Autonomous Database instance that is not part of any elastic pool. Assuming the Autonomous Database idles between 4-5pm, and there are no other Autonomous Database instances in the pool, billing for the hour between 4-5pm is as follows:

  The bill for 4-5pm is: (4 * 0.5) + 128 = 130 ECPUs

  Where the (4 * 0.5) is the billing for compute for the thirty minutes after the Autonomous Database instance terminates the elastic pool, and 128 ECPUs is the billing for the elastic pool for the hour when the elastic pool was terminated.

Elastic Pool Billing with Built-in Tools

For either the pool leader or the members, compute resources that are allocated to the built-in tools, OML, Graph, or Data Transforms, are separate and do not count towards the elastic pool total allocation. For billing purposes, the elastic pool leader is billed for any built-in tool ECPU usage by either the leader or elastic pool members, in addition to the elastic pool ECPU usage.

For example, assume there is an elastic pool with a pool size of 128 ECPUs; if in a given billing hour the aggregated peak ECPU utilization of the pool leader and the members is 80 ECPUs for the billing hour, and during this hour the combined total ECPU utilization for instances using built-in tools is 30 ECPUs, the leader is charged for the pool size (128 ECPUs), plus the built-in tool ECPU usage (30 ECPUs), for a total of 158 ECPUs for that hour.

Elastic Pool Usage Details in OCI Usage Reports and OCI_USAGE_DATA View

You can obtain a detailed breakdown of elastic pool usage in the Oracle Cloud Infrastructure (OCI) usage reports and this information is also shown in the OCI_USAGE_DATA view. See Cost and Usage Reports and OCI_USAGE_DATA View for more information.

The following table shows the product/resource column values in an OCI usage report. The OCI usage report provides details on elastic pool usage for the pool leader and for pool members a given billing hour (similar information is available in OCI_USAGE_DATA view):

SELECT billed_quantity FROM OCI_USAGE_DATA
   WHERE resource_name='PIC_ADBS_DB_ECPU_PEAK' and
                    resource_id=OCID_of_the_pool_member and
                    interval_usage_start=start_time and
                    interval_usage_end=end_time

SELECT billed_quantity FROM OCI_USAGE_DATA
   WHERE resource_name='PIC_ADBS_DB_ECPU_PEAK' and
                    resource_id=OCID_of_the_pool_leader and
                    interval_usage_start=start_time and
                    interval_usage_end=end_time

SELECT billed_quantity FROM OCI_USAGE_DATA
   WHERE resource_name='PIC_ADBS_ELASTIC_POOL_DB_ECPU' and
                    resource_id=OCID_of_the_pool_leader and
                    interval_usage_start=start_time and
                    interval_usage_end=end_time

Notes for elastic pool billing information in the OCI usage report and the OCI_USAGE_DATA view:

• Elastic pool aggregate peak ECPU usage is shown for terminated databases.

• Elastic pool aggregate peak ECPU usage is shown for non-pool databases that were part of an elastic pool during a billing hour when the database was a member of an elastic pool.

See Billing Information: Autonomous Transaction Processing ECPU Compute Model and Track Oracle Cloud Infrastructure Resources, Cost and Usage Reports with Autonomous Database Views for more information.

Local Autonomous Data Guard Standby Database Billing

When you add a local standby, a total of two times (2 x) the primary's ECPU allocation is counted towards the pool capacity (1 x for the primary and 1 x for the standby). Meaning a local standby multiplies the primary's peak usage by 2.

For example, if you create an elastic pool with a pool size of 128 ECPUs, with a pool capacity of 512 ECPUs, adding the following Autonomous Database instance uses the elastic pool capacity:

• 1 instance with 256 ECPUs with local Autonomous Data Guard enabled, for a total of 512 ECPUs allocation from the pool.

  When using this instance the CPU utilization is 256 ECPUs, however the overall peak ECPU utilization will be reported as 512 because of the local standby database 2 x multiplication factor. And the billing is based on 4 x the pool size (512 ECPUs).

Cross-Region Autonomous Data Guard Standby Database Billing

Enabling Cross-region Autonomous Data Guard for a leader or for member has no effect on the elastic pool capacity. A cross-region Autonomous Data Guard peer database has its own OCID and the cross-region peer is billed independently from the elastic pool.

Note the following:

• Cross-region Autonomous Data Guard peer ECPUs do not use pool capacity and billing for Autonomous Data Guard cross-region peer databases happens on the peer instance.

• When the leader of an elastic pool enables cross-region Autonomous Data Guard, the cross-region peer database ECPU allocation does not count towards the elastic pool capacity. Billing for cross-region Autonomous Data Guard is on the cross-region instance, which is not part of the elastic pool (elastic pools do not operate across regions).

• When a member of an elastic pool enables cross-region Autonomous Data Guard, the cross-region peer ECPU allocation does not count towards the pool capacity. Billing for cross-region Autonomous Data Guard is on the cross-region instance, which is not part of the elastic pool (elastic pools do not operate across regions).

For example, if you create an elastic pool with a pool size of 128 ECPUs (with a pool capacity of 512 ECPUs), adding the following Autonomous Database instances of different sizes uses the entire elastic pool capacity:

• A pool that contains the following instances:

  • 1 instance with 128 ECPUs with cross-region Autonomous Data Guard enabled (using a total of 128 ECPUs from the pool).

  • 64 instances with 2 ECPUs each with both local and cross-region Autonomous Data Guard enabled (using a total of 256 ECPUs from the pool).

  • 128 instances with 1 ECPU, each with cross-region Autonomous Data Guard enabled (using 128 ECPUs from the pool).

Pool Leader Operations

The following operations are valid only for the pool leader:

Pool Member Operations

The following operations are valid for a pool member or for the pool leader: