Reserve BI Engine capacity

You purchase BigQuery BI Engine capacity by creating a reservation. The reservation is attached to a project and region you identify when the reservation is created. BI Engine uses this capacity to cache data. For information about the maximum reservation size for BI Engine, see Quotas and limits.

When you use BigQuery BI Engine, your charges are based on the BI Engine capacity you purchased for your project. Reservations are charged per GiB/hour, priced per region according to on-demand pricing. If you have a capacity-based commitment, you will be eligible for discount of up to 100% based on the active commitment. For details see capacity-based pricing.

Reservation project: BI Engine reservations are managed by the billing project's ID. When purchasing a reservation, ensure that you specify the billing project ID and the region that will be used to query the data. This is not necessarily the same project that contains the dataset.

Required roles

To get the permissions that you need to create and delete reservations, ask your administrator to grant you the BigQuery Resource Admin (roles/bigquery.resourceAdmin) IAM role on the project. For more information about granting roles, see Manage access to projects, folders, and organizations.

You might also be able to get the required permissions through custom roles or other predefined roles.

Create a reservation

To reserve on-demand BI Engine capacity, follow these steps:

Console

1. On the BigQuery page, in Administration, go to the BI Engine page.

   Go to BI Engine

   Note: If prompted to enable BigQuery Reservation API, click Enable.

2. Click Create reservation.

3. On the Create reservation page, for Step 1:

   • Verify your project name.
   • Choose your location. The location should match the location of the datasets you are querying.

   • Adjust the slider to the amount of memory capacity you're reserving. The following example sets the capacity to 2 GB. The current maximum is 250 GB. You can request an increase of the maximum reservation capacity for your projects. Reservation increases are available in most regions, and can take from 3 days to one week to process.

     

4. Click Next.

5. Preferred tables (optional). Preferred tables let you limit BI Engine acceleration to a specified set of tables. All other tables use regular BigQuery slots.

   In the Table Id field, specify the table that you want to accelerate using the pattern: PROJECT.DATASET.TABLE.

   Replace the following:

   • PROJECT: your Google Cloud project ID
   • DATASET: the dataset
   • TABLE: the table that you want to accelerate

6. Click Next.

7. For Step 3, review your reservation details, and then click Create.

After you confirm your reservation, the details are displayed on the Reservations page.

SQL

Use the ALTER BI_CAPACITY SET OPTIONS DDL statement to create or modify a BI Engine reservation.

1. In the Google Cloud console, go to the BigQuery page.

   Go to BigQuery

2. In the query editor, enter the following statement:

   ALTER BI_CAPACITY `PROJECT_ID.LOCATION_ID.default`
   SET OPTIONS (
     size_gb = VALUE,
     preferred_tables =
       ['TABLE_PROJECT_ID.DATASET.TABLE1',
       'TABLE_PROJECT_ID.DATASET.TABLE2']);

   Replace the following:

   • PROJECT_ID: the optional ID of the project that will benefit from BI Engine acceleration. If omitted, the default project is used.
   • LOCATION_ID: the location where data needs to be cached, prefixed with region-. Examples: region-us, region-us-central1.
   • VALUE: the INT64 size of the reservation for BI Engine capacity in gigabytes, 1 to 250 GB. You can request an increase of the maximum reservation capacity for your projects. Reservation increases are available in most regions, and can take from 3 days to one week to process. Setting VALUE replaces the existing value if there is one. Setting to NULL clears the value for that option.
   • TABLE_PROJECT_ID.DATASET.TABLE: the optional list of referred tables to which acceleration should be applied. Format: TABLE_PROJECT_ID.DATASET.TABLE or DATASET.TABLE. If the project is omitted, then the default project is used.

3. Click Run.

For more information about how to run queries, see Run an interactive query.

bq

Use the bq update command and supply the data definition language (DDL) statement as the query parameter:

bq --project_id=PROJECT_ID update \
    --bi_reservation_size=SIZE \
    --location=LOCATION \
    --reservation

Replace the following:

• PROJECT_ID: the ID of your project
• SIZE: the reservation memory capacity in gigabytes, 1 to 250 GB. You can request an increase of the maximum reservation capacity for your projects. Reservation increases are available in most regions, and can take from 3 days to one week to process.
• LOCATION: the location of the dataset you are querying

Estimate and measure capacity

To estimate capacity requirements for a BI Engine reservation, follow these steps:

1. View the TOTAL_LOGICAL_BYTES view to determine the logical size of the table, and use that for your initial BI Engine reservation. For example:

   SELECT
     SUM(TOTAL_LOGICAL_BYTES) / 1024.0 / 1024.0 / 1024.0 AS logical_size_gb
   FROM
     `region-us.INFORMATION_SCHEMA.TABLE_STORAGE`
   WHERE
     TABLE_NAME IN UNNEST(["Table1", "Table2"]);

   For example, for queries against a set of tables that contain a total of 200GiB of data, as a best practice you can start with a 200GiB BI Engine reservation. More selective queries that only use a subset of available fields or partitions could start with a smaller reservation size.

2. Run all of the queries that need optimization and that were created in the same project and region as the BI Engine reservation. The goal is to approximate the workload that you need to optimize. The increased load requires more memory to handle queries. Data is loaded into BI Engine after the query is received.

3. Compare your BI Engine RAM reservation to the number of bytes used, reservation/used_bytes in the Cloud Monitoring bigquerybiengine metrics.

4. Adjust your reservation capacity based upon the results. In many use cases, a smaller reservation can accelerate the majority of your queries, conserving money and resources. For more information about Monitoring for BI Engine, see BI Engine monitoring.

The following factors affect BI Engine reservation size:

• BI Engine only caches the frequently accessed columns and rows that are required to process the query.
• When a reservation is fully used, BI Engine tries to offload the least recently used data to free up capacity for new queries.
• If multiple computationally intensive queries are using the same dataset, then BI Engine loads additional copies of the data to redistribute and optimize response times.

Modify a reservation

To modify an existing reservation, complete the following steps:

Console

To specify a set of tables for acceleration in an existing reservation, follow these steps:

1. In the Google Cloud console, go to the BigQuery page.

   Go to BigQuery

2. In the BigQuery navigation menu, click BI Engine.

   If your project is configured for preferred tables, a set of tables is displayed in the Preferred Tables column.

   

3. On the row for the reservation that you want to edit, click the icon in the Actions column, and then select Edit.

4. Adjust the GiB of Capacity slider to the amount of memory capacity you're reserving. Click Next.

5. Preferred tables: To specify a set of tables for acceleration in an existing reservation, in the Table Id field, specify the table that you want to accelerate using the pattern: PROJECT.DATASET.TABLE.

   Replace the following:

   • PROJECT: your Google Cloud project ID
   • DATASET: the dataset
   • TABLE: the table that you want to accelerate

   Changes can take up to ten seconds to take effect. Only tables in the preferred tables list can use the BI Engine acceleration.

   Click Next.

6. Confirm your modified reservation. If you agree, click Update.

SQL

You can use the ALTER BI_CAPACITY SET OPTIONS DDL statement to create or modify a BI Engine reservation.

1. In the Google Cloud console, go to the BigQuery page.

   Go to BigQuery

2. In the query editor, enter the following statement:

   ALTER BI_CAPACITY `PROJECT_ID.LOCATION_ID.default`
   SET OPTIONS (
     size_gb = VALUE,
     preferred_tables =
       [`TABLE_PROJECT_ID.DATASET.TABLE1`,
       `TABLE_PROJECT_ID.DATASET.TABLE2`]);

   Replace the following:

   • PROJECT_ID: optional ID of the project that will benefit from BI Engine acceleration. If omitted, the default project is used.
   • LOCATION_ID: the location where data needs to be cached, prefixed with region-. Examples: region-us, region-us-central1.
   • VALUE: the INT64 size of the reservation for BI Engine capacity in gigabytes, 1 to 250 GB. You can request an increase of the maximum reservation capacity for your projects. Reservation increases are available in most regions, and can take from 3 days to one week to process. Setting VALUE replaces the existing value if there is one. Setting to NULL clears the value for that option.
   • TABLE_PROJECT_ID.DATASET.TABLE: optional list of preferred tables to which acceleration should be applied. Format: TABLE_PROJECT_ID.DATASET.TABLE or DATASET.TABLE. If the project is omitted, then the default project is used.

3. Click Run.

For more information about how to run queries, see Run an interactive query.

Delete a reservation

To delete a capacity reservation, follow these steps:

Console

1. On the BigQuery page, in Administration go to the BI Engine page.

   Go to BI Engine

2. In the Reservations section, locate your reservation.

3. In the Actions column, click the icon to the right of your reservation and choose Delete.

4. In the Delete reservation? dialog, enter Delete and then click DELETE.

SQL

Sets the options on BI Engine capacity.

1. In the Google Cloud console, go to the BigQuery page.

   Go to BigQuery

2. In the query editor, enter the following statement:

   ALTER BI_CAPACITY `PROJECT_ID.LOCATION_ID.default`
   SET OPTIONS (
     size_gb = 0);

   Replace the following:

   • PROJECT_ID: optional ID of the project that will benefit from BI Engine acceleration. If omitted, the default project is used.
   • LOCATION_ID: the location where data needs to be cached, prefixed with region-. Examples: region-us, region-us-central1.

   When you delete all capacity reservations in a project, BI Engine is disabled for that project.

3. Click Run.

For more information about how to run queries, see Run an interactive query.

bq

Use the bq update command and supply the DDL statement as the query parameter.

bq --project_id="PROJECT_ID" \
update --reservation
    --bi_reservation_size=0 \
    --location=LOCATION

Replace the following:

• PROJECT_ID: the ID of your project
• LOCATION: the location of the dataset you are querying

Verify BI Engine information

You can get information about your BI Engine capacity by querying the INFORMATION_SCHEMA tables.

Verify reservation status

To verify the status of your reservation, including a set of preferred tables, view the INFORMATION_SCHEMA.BI_CAPACITIES view using a SQL query. For example:

SELECT
  *
FROM
  `<PROJECT_ID>.region-<REGION>.INFORMATION_SCHEMA.BI_CAPACITIES`;

In the Google Cloud console, the result of this SQL query looks similar to the following:

View reservation changes

To view the history of changes for a particular reservation, use the INFORMATION_SCHEMA.BI_CAPACITY_CHANGES view using a SQL query. For example:

SELECT
  *
FROM
  `<PROJECT_ID>.region-<REGION>.INFORMATION_SCHEMA.BI_CAPACITY_CHANGES`
ORDER BY
  change_timestamp DESC
LIMIT 3;

In the Google Cloud console, the result of this SQL query looks similar to the following:

What's next

• Learn more about BI Engine.
• Learn about BI Engine pricing.
• Analyze data with Looker Studio.
• Monitor BI Engine