A compute group is one or more compute profiles with the same policy for assigning access and administration privileges to their users. It also sets the number of concurrent queries that can run on a compute cluster, by specifying the query strategy, STANDARD, ANALYTIC, or ANALYTIC_L1.
If a compute profile specifies automatic scaling for a group, the system brings compute clusters in the group up and down, based on query backlog and CPU usage levels of active compute clusters. A compute group that has no compute profiles has no compute clusters and therefore does not accept work from the primary cluster.
A compute group is a database object that no database or user owns. An organization administrator creates a compute group for each department, application, or project and gives a few users in the department, application, or project the privileges to manage its compute profiles. These users are called compute cluster administrators.
How to Use an Existing Compute Group
- You must have the COMPUTE GROUP privilege on the compute group.
- The compute group must be your default compute group or your session compute group.
The compute cluster administrator grants you the COMPUTE GROUP privilege and sets your default compute group.
If the compute cluster administrator has assigned you to the compute group, you can make the compute group your session compute group (see SET SESSION COMPUTE GROUP).
You can also select a compute group to use when running SQL in the Editor. See Configuring the compute group option.
How to Create a New Compute Group
You must have the CREATE COMPUTE GROUP privilege (granted by the compute cluster administrator).
A compute group is standard (by default) or analytic. Standard compute groups belong to Standard Compute Clusters. Analytic compute groups belong to Analytic Compute Clusters.
- With the CREATE COMPUTE GROUP: SQL command
- Using the VantageCloud Lake Console (see Creating a Compute Group: Console or Step: Creating Compute Groups).