Fabric start guide

DX Fabric combines a Software catalog, Scorecards, and Self-service workflows into a single platform for managing and improving your software ecosystem. This guide walks you through setting up Fabric for the first time, whether you are using it standalone or as part of the DX Platform. By the end, you will have a populated catalog and be able to create scorecards and self-service workflows. For a conceptual overview, see Fabric Overview.

Before you begin

Follow the instructions in the DX Start Guide before setting up Fabric. This will set up your DX instance with:

  • A team structure in DX
  • Connections to your source control management and other tools
  • User linking

You do not have to perform a snapshot in order to begin using Fabric.

Step 1: Install the DX CLI

The DX CLI is an AI-native command-line tool that lets you manage your software catalog, configure Scorecards, and access DX API endpoints directly from your IDE or terminal. It also includes a built-in agent skill so your AI agents can interact with your DX data.

See CLI for instructions on installation.

Step 2: Set up the Software Catalog

Populate the catalog

There are a few ways to populate the Software catalog:

  1. Auto-discover entities
    Once you’ve set up connections to third party tools like your source code management tool, DX scans those connections and automatically discovers entities within those tools. Upon opening the catalog, DX displays these auto-discovered entities as suggestions, which you can accept or ignore.

Auto-discovered entities can be filtered by source, and also may come pre-populated with owner teams, based on deployment data, and aliases, such as the repository link.

  1. Create entities via API
    There are a number of catalog API endpoints that allow you to create entities and set up custom properties programatically.

If you need to bring more data into DX directly via API, the DX Postgres instance includes a custom schema for creating tables and indexes. Provision a DB User from the DX Admin settings, and use the provided credentials to programmatically connect to, and populate, your custom table.

  1. Config-as-code

You can manage your Software catalog (as well as Scorecards) through the DX Terraform provider.

  1. Import components from another catalog

You can import components from Compass or import components from Backstage into DX.

For assistance with these connections, please reach out to your account representative.

  1. Manual creation of entities

In the Fabric UI, you can click +Create to manually add an entity to your Software catalog.

Configure the catalog

Custom entity types

The catalog is made up of entities, each entity representing a piece of software within your organization. There are three default entity types you’ll see upon loading the catalog for the first time: Services, Domains, and Teams.

You can configure your catalog by adding custom entity types that reflect the software in your organization, such as an API entity type.

Custom properties
You can add custom properties to each entity type to store all relevant metadata. Custom properties support types such as JSON, OpenAPI spec, computed SQL values, and more.

Aliases
For each entity type, various aliases are available. Toggling on these aliases allows users to connect a third party tool of the alias type onto the entity. Example aliases include Github repository links, Datadog service, and PagerDuty service.

Relations
Relations define how entity types are connected to one another. They are defined at the entity type level - for example, you would define a relation that says entities of type API can depend on entities of type Service. Once an API entity is created, a relation can then be created between that entity and a Service entity.

Entity reports
Entity reports allow you to create custom dashboars that appear directly on individual entity pages within the Software catalog. Built on the same foundation as  Data Studio Custom Reports, entity reports give you access to SQL queries, flexible visualizations, and dynamic variables—but scoped to the specific entity you’re viewing rather than displayed as a standalone page.

Step 3: Set up Scorecards

Scorecards let you evaluate custom rules against your Software catalog entities using SQL-based checks. Scorecards can be created directly in the DX UI, via the CLI, or via the DX API.

Create a scorecard

After creating a new scorecard, refine it by:

  • Selecting which entities to evaluate (entities can be filtered by type or SQL query)
  • Selecting a scorecard type - either Levels or Points-based
  • Defining custom levels to use for grading
  • Adding checks via SQL that run against all data in your DX instance

Explore example scorecards to get started creating scorecards.

Evaluate a scorecard

Once created, scorecards automatically evaluate the entities they apply to. Scorecards display results in two views:

  • Entity view — See pass/fail results for each individual entity scored by the scorecard.
  • Team view — See results rolled up by group or team across your organizational hierarchy. Drill down into any level to view individual entity results for that team. Note: entities must have an owner assigned to appear in team-level rollups.

Both views support filtering by various attributes, and the Trends tab lets you track check pass rates over time.

Create initiatives

Initiatives are actionable tasks created based on scorecard results to help track progress towards scorecard completion and drive action across your organization. Each initiative targets one or more specific scorecard checks.

Initiative creation form

Track initiatives

Once an initiative is created, nudges can be configured to be sent out via Slack or MS Teams to responsible team members.

Progress can be tracked through high-level reporting that breaks down completion rate by team, entity, or check. Additionally, initatives can be synced with your project management software such as Jira for tracking.

Step 4: Set up self-service

With Self-service workflows, you can automate essential workflows and surface them directly in DX to be triggered by both engineers and AI agents.

Define the workflow

  • Give your workflow a name, icon, and description so users understand what it does.
  • Set the scope of your workflow to specify if it should apply on an existing entity or globally.
  • Choose an execution type
    • Event-driven execution: The workflow makes callbacks to DX during execution, populating a real-time log so users can track progress.
    • Simple execution: The workflow completes immediately once the HTTP request is sent, with no progress tracking.
  • Set permissions on who can run the workflow
  • Define input parameters that users or agents will be prompted to fill in before running the worfkflow. These values are passed along in the HTTP request body.
  • Configure the HTTP request that runs when the workflow is triggered

Run the workflow

When a workflow is triggered (either from the DX UI or via the DX CLI by an AI agent), the user will be prompted to select an entity (if scoped) and fill in any required parameters.

For event-driven workflows, a run history view shows real-time logs and links produced by the HTTP request as the automation progresses.

Next steps

To continue building out Fabric for your team, see the Fabric documentation, specifically: