View a markdown version of this page

Create a dataset on the Data Hub page - AWS Billing

Create a dataset on the Data Hub page

You can create a dataset on the Data Hub page at any time, whether or not an agent exists. To create a dataset, you need the Data Hub actions for the person who creates it, described in Identity-based policy examples.

  1. Open the AWS Billing and Cost Management console in the account that will hold the dataset, and choose Data Hub in the navigation pane.

  2. Choose Create dataset.

  3. For Name, enter a name of 1 to 64 characters, using lowercase letters, digits, and underscores and starting with a letter or a digit. AWS FinOps Agent reads the name when it lists datasets.

  4. For Schema name, choose cost_usage.

  5. For Tags, add a key and an optional value if you want them. A dataset carries up to 50 tags.

  6. Choose Create dataset, then watch the status of the new row reach Active.

Important

Provisioning takes up to 24 hours, and Status reads Active when agents can query the dataset.

When provisioning does not finish, Status reads Failed and the status reason on the dataset detail page states why. A dataset that has not finished provisioning still counts against the quota of one dataset per account, so read the status reason, delete the dataset, and create it again. If the same reason repeats, contact AWS Support.

You can change the name and the tags after creation. Choose Edit on the dataset detail page. A change to the name or the tags leaves the provisioned data in place. The schema, the coverage, and the account are fixed at creation, so changing any of those means deleting the dataset and creating a new one, which loads the history again.

On the Data Hub page, choose the dataset name to open its detail page, which reports the following.

Field What it reports
Status The dataset status, which reads Provisioning, Active, or Failed, with the status reason beside it when provisioning did not complete. An Active dataset returns to Provisioning when AWS finds a problem with the data, and queries are rejected until it is Active again.
Schema name cost_usage.
Data The period the dataset covers, which is 13 months of resource-level cost and usage in the CUR 2.0 schema.
Last refreshed When the data behind an answer was last updated.
Tags The tags on the dataset, with Manage tags to change them.
Note

The dataset detail page does not list the agents that read the dataset. In the delegated administrator case those agents live in another account, which the management account cannot enumerate, so the page reports the dataset and leaves consumers out.