Skip to main content
Knowledge graphs capture the structure, relationships, and meaning of your data. Maia Team automatically uses your knowledge graphs when working on tasks in Mission Control, or building pipelines in Designer. For more information, read Context Engine in chats. With a knowledge graph selected, Maia Team understands your business terminology and builds pipelines relevant to your data landscape. By using knowledge graphs, you can enrich the answers Maia Team provides with more knowledge about your business data, semantics, and processes. You can add information to knowledge graphs from a number of sources, including your data warehouses, pipeline run history in Maia, and external sources such as SaaS applications. We recommend you add metadata, such as column descriptions and tags, to your data—that is, your databases, schemas, or datasets. This metadata will help Maia Team understand your data landscape and provide better responses to your prompts.

Prerequisites

To add a new knowledge graph, you need an account role with the Create knowledge graph permission. Adding a knowledge graph includes crawling data to add information to the knowledge graph, so that it contains relevant business logic and information when it is created. Each crawler type has its own prerequisites:
  • Warehouse data crawlers harvest your warehouse and structured sources supported via connectors to populate the graphs. This helps keep an accurate view of the data landscape. To run a Warehouse data crawler, your project must contain a schema that is populated with data.
  • Pipeline execution crawlers harvest your pipeline executions for your chosen project and environment to build an operational understanding of how data flows through your organization and workflow. To run a Pipeline execution crawler, your project must contain at least one pipeline.
  • External source crawlers harvest data from external applications to add business objects to your knowledge graph. To run an External source crawler, you need a connection to the source you want to crawl.

Create a knowledge graph

To create a knowledge graph, add a crawler to the graph, and run the crawler for the first time:

Add a knowledge graph

  1. On the Context Engine dashboard, click Add knowledge graph.
  2. Click Build a graph to create a new knowledge graph.
    If the information you’re adding relates to an existing knowledge graph, we recommend clicking Add to a graph to add more context to this knowledge graph. This will help Maia Team to understand your business logic by keeping related content in one place.
  3. Select the type of crawl you want to perform for the new knowledge graph, and then follow the steps in the corresponding tab below:
    • Data warehouse: Populate the knowledge graph with data from tables, columns, and schemas in a data warehouse.
    • Pipeline executions: Populate the knowledge graph with data from the pipeline run history in a specific project and environment, including lineage data.
    • External source: Populate the knowledge graph with data about business objects from an external application.
    • Harvest documents: Populate the knowledge graph with unstructured content from documents, such as runbooks or support materials.
  1. Select the type of cloud data warehouse you want to harvest data from.
  2. In Project, select a project that is connected to the warehouse you want to harvest data from.
  3. In Environment, select the environment you want to use to harvest data.
  4. (Optional) Click Manage next to the list of users in the selected project to view and edit the users who will be collaborators on this knowledge graph. To remove a collaborator, click the trash can icon next to their name.
  5. (Optional) In the Invite individual to collaborate drop-down, search for and click a Maia user to add them as a collaborator on this knowledge graph.
  6. Click Next.
  7. In the Select data step, select the data to crawl:
    • If you’re crawling a Snowflake warehouse, select the database and schemas to crawl.
    • If you’re crawling a Databricks warehouse, select the catalog and schemas to crawl.
    • If you’re crawling an Amazon Redshift warehouse, select the schemas to crawl.
    • If you’re crawling a Google BigQuery warehouse, select the datasets to crawl.
  8. (Optional) Select or deselect the Include data from pipeline executions checkbox to include or exclude pipeline executions in the selected project and environment in the crawl.
  9. Click Next.
  10. In the Schedule crawl step, choose how often this crawl should run. The crawl will run immediately after you create this knowledge graph, then follow this schedule.
    1. Select Standard or Advanced schedule settings.
    2. In Timezone, select the timezone for the crawl schedule.
    3. If you selected Standard settings, use the repeat drop-downs to set how often the crawler runs.
    4. If you selected Advanced settings, enter a cron expression to set how often the crawler runs.
  11. Click Run.
You’ll be taken to the Crawlers tab of the new knowledge graph. The knowledge graph is created with a default name, which you can change by clicking the edit icon next to its name. You can now use the tabs on the left to perform the following actions. For more information, read Managing knowledge graphs.
  • Crawlers: Add and edit crawlers to ingest data, and monitor the status and run history of all crawlers in this knowledge graph.
  • Documents: Upload documents to ingest unstructured data.
  • Projects: Manage the projects that have access to this knowledge graph.
  • Access: Manage the users who can contribute to this knowledge graph.