How to create a DevOps Data Kit?

Introduction

A Data Kit is a Salesforce container that bundles related Data 360 configuration so you can move it between orgs as one unit. That configuration includes data streams, data lake objects, calculated insights, and data graphs. Data Kits are separate from your other Salesforce metadata.

Salesforce and Gearset both recommend deploying Data 360 configuration through a DevOps Data Kit, for these reasons:

Some Data 360 metadata types can only be deployed through a Data Kit.
A Data Kit keeps related components together, so they activate in the target org in the order set by the kit's publishing sequence.
Gearset reads the Data Kit's manifest, builds the deployment package from it with dependencies included, and activates the Data Kit in the target org after deployment.

Prerequisites

Before you create a DevOps Data Kit, check that:

Data 360 is enabled in both your source and target orgs.
The user Gearset connects with has the Data Cloud Architect permission set in both orgs. Without it, Gearset can't see your Data Kits. For the full list, see Permissions required to deploy Data 360 metadata.
A data space with the same name exists in both orgs. A DevOps Data Kit deploys to the same data space in the target org that you create it in.

Why does it have to be a DevOps Data Kit?

Gearset deploys Data Kits through the Metadata API, so the Data Kit you deploy must be a DevOps Data Kit. When you choose the Data Kit deployment method, Gearset only lists the DevOps Data Kits in your source org.

Salesforce offers two types of Data Kit, and Gearset deploys only one of them.

DevOps Data Kit: use this type to move Data 360 configuration between your own orgs, for example from a sandbox to production. This is the type Gearset deploys. You can create it in any data space, and it deploys to the data space with the same name in the target org.
Standard Data Kit: Salesforce designed this type for packaging Data 360 solutions to share with customers, usually through a managed package. Gearset doesn't list standard Data Kits, so you can't deploy one with Gearset.

For more detail, see Salesforce's Data Kits article.

How do I create a DevOps Data Kit?

In your source org, go to Data Cloud Setup.
Under Developer Tools, click Data Kits.
Click New.
For Data Kit Type, select DevOps.
For Data Space, select the data space that holds the components you want to deploy.
Enter a Data Kit Name and a Data Kit API Name.
Optionally, enter a Description of what the Data Kit contains.
Click Save.

On the Data Kit Contents tab, click Add next to a component type, select the components you want to deploy, and save.
Repeat step 9 for each component type your configuration uses.
Click the Publishing Sequence tab and check the order of the components.
To change the order, click Edit Sequence, and move each component below any component it depends on.

Salesforce orders components by creation date until you define a sequence. Until then, the tab shows You haven't defined the publishing sequence for the data kit. Use the Edit Sequence button to define the sequence. Gearset follows this sequence when it activates the Data Kit in the target org, so a wrong order can make activation fail.

What should I include in my DevOps Data Kit?

Add every Data 360 component your configuration needs. When you pick a Data Kit, Gearset retrieves its components and traces their dependencies, and its problem analyzers flag anything still missing before you deploy. However, Gearset finds dependencies of what's in the Data Kit. If a piece of configuration was never added to the kit, nothing links it to the kit, so Gearset has no dependency to find.

Salesforce doesn't add some dependencies automatically, so add these yourself:

Calculated insights: add the insight's dependencies too, such as child calculated insights, data model objects, data lake objects, and data graphs.
DLO-to-DMO mappings: add the output data lake objects explicitly, so their mappings carry over.

For details, see Salesforce's Considerations for Data Kits in Data 360.

What's next?

Once your DevOps Data Kit is ready, deploy it through a pipeline or from org to org. For both workflows, see How to deploy Data Kits with Gearset.