Billing dimensions

Aura has a usage-based billing model (also known as consumption-based billing or metered billing). Users pay for the actual amount of a product or service consumed, rather than a fixed fee.

AuraDB

For AuraDB, compute and storage are tracked.

  1. Compute: measured in GB-hours, based on RAM capacity. For example, if your AuraDB instance has 4GB of RAM, and runs for 1 hour, you are charged for 4 GB-hours of compute usage.

    • Compute is billed both in a running state and a paused state.

    • Compute is billed for the primary database and for any secondary databases. For secondary databases, the total running capacity is the number of secondaries multiplied by the primary DB capacity.

  2. Storage: each AuraDB instance has a free storage quota of twice the RAM capacity. You can purchase additional storage beyond the quota, which is billed the same as compute in GB-hours.

    • Each secondary DB benefits from the same free storage quota as the primary DB.

AuraDB examples

Primary database only

You have set up an 8GB AuraDB Business Critical instance. Every hour the primary database is charged for 8 GB-hours of usage, at the Business Critical hourly price.

Primary and secondary databases

You have set up a 4GB AuraDB Business Critical instance, with 2 secondaries. Every hour the primary DB and its secondaries are running, you consume:

  • 4 GB-hours of primary db compute

  • 2 secondaries x 4 GB = 8 GB-hours of secondary db compute

  • You are charged for a total of 4 GB-hours of primary database compute and 8 GB-hours of secondary database compute.

Primary database and additional storage

You have set up a 4GB AuraDB Business Critical instance, and requested a total of 32 GB of storage. This includes a free storage quota of twice the RAM size, so 2 x 4 GB = 8 GB. This means, every hour you consume:

  • 4 GB-hours of primary database compute

  • 32 GB - 8 GB = 24 GB-hours of storage

  • You are charged for a total of 4 GB-hours of primary database compute and 24 GB-hours of storage.

AuraDS

For AuraDS, only compute is tracked:

  • Measured in GB-hours, based on RAM capacity.

  • Billed both in a running state and in a paused state

  • AuraDS includes a free quota of twice the RAM capacity for storage, without the possibility to purchase more storage.

  • Secondaries are not available for AuraDS

Aura Graph Analytics

Aura Graph Analytics does not require a persistent instance run at all times.

You can start Graph Analytics sessions that execute graph algorithms in-memory. You specify the allocated memory size when the session is launched.

The billing dimension for Graph Analytics sessions is compute and it is measured in GB-minutes. The duration of a session is calculated in minutes and multiplied by the RAM capacity used by the session.

Graph Analytics example

You launch a session with a size of 4GB of RAM, and it runs for 25 minutes. You are then billed 25 minutes x 4GB = 100 GB-minutes of usage.

Note: The minimum billed duration of a session is 10 minutes. In the example, if the session ran for 8 minutes and had a 4GB size, you are billed for 10 minutes x 4GB = 40 GB-minutes.

GraphQL API

Data APIs like GraphQL run on top of an existing AuraDB instance but consume additional memory. You are billed for compute, measured in MB-hours, based on the RAM capacity of the GraphQL additional memory.

GraphQL example:

You set up a GraphQL API with 256MB of RAM, every hour the database is running triggers a usage record of 256 MB-hours.

Query log forwarding

With log forwarding, you can stream logs directly to a cloud project owned by your organization. For query log forwarding, you are billed for data egress. Egress is measured in GB, and is added to the customer usage periodically.