Cloud SQL for PostgreSQL pricing

This page contains information about pricing for Cloud SQL.

Cloud SQL for PostgreSQL pricing is composed of the following charges:

CPU and memory pricing

For dedicated-core instances, you choose the number of CPUs and the amount of memory you want, up to 96 CPUs and 624 GB of memory. Pricing for CPUs and memory depends on the region where your instance is located. Select your region in the dropdown on the pricing table.

Read replicas and failover replicas are charged at the same rate as stand-alone instances.

HA prices are applied for instances configured for high availability, also called regional instances. Learn more about high availability.

Storage and networking pricing

Storage and networking prices depend on the region where the instance is located. Select your region in the dropdown on the pricing table.

HA prices are applied for instances configured for high availability, also called regional instances. Learn more about high availability.

Network Egress Pricing

When network traffic leaves a Cloud SQL instance, the charge applied depends on the destination of the traffic, and in some cases, whether a partner is involved.

Internet egress is network traffic leaving a Cloud SQL instance to a client that is not a Google product, such as using a local server to read data from Cloud SQL.

Destination Price
Compute Engine instances and Cloud SQL cross-region replicas Within the same region: free
Between regions within North America: $0.12/GB
Between regions outside of North America: $0.12/GB
Google Products (except Compute Engine and traffic to Cloud SQL cross-region replicas) Intra-continental: free
Inter-continental: $0.12/GB
Internet egress using Cloud Interconnect $0.05/GB
Internet egress (not using Cloud Interconnect) $0.19/GB
If you pay in a currency other than USD, the prices listed in your currency on Cloud Platform SKUs apply.

Instance pricing

Instance pricing applies only to shared-core instances. Dedicated-core instances, which can have up to 96 vCPUs and 624 GB of memory, are charged by the number of cores and amount of memory they have.

Instance pricing is charged for every second that the instance is running (the activation policy is set to ALWAYS). Cloud SQL uses seconds as the time unit multiplier for usage. This means that each second of usage counts toward a full billable minute. For more details, see Billing on partial seconds.

HA prices are applied for instances configured for high availability, also called regional instances. Learn more about high availability.

In the following table:

  • Select your region from the dropdown menu to see the price for that region
  • Use the slider to choose Monthly or Hourly pricing
  • Find the machine type you want to use to view pricing details

*Shared CPU machine types (db-f1-micro and db-g1-small) are not covered by the Cloud SQL SLA.

Billing on partial seconds

Milliseconds of usage are rounded to the nearest second. If usage is under half a second, (499ms or less), it rounds down to zero and does not count toward billable usage.

For example:

  • If you use an instance for 499ms, you are not billed for that second.
  • If you use an instance for 500ms or 1.49 seconds, in both cases, you are billed for 1 second.
  • If you use an instance for 1.5 seconds or 2.49 seconds, in both cases you are billed for 2 seconds.

What's next?