--- title: "MDLH: Setting up the MDLH" url: "https://atlan.com/demos/mdlh-setting-up-the-mdlh/" excerpt: "Connect your query engine to live Atlan metadata in minutes — skip the weeks of API pipeline engineering your team was about to build." description: "Go from zero to running SQL against your full metadata estate in a single session. Walk through every setup step — enabling the Lakehouse, grabbing connection credentials, and running the Snowflake commands that make it all queryable. No pagination logic, no rate limit handling, no schema maintenance. By the end, your metadata sits right alongside your other data sources, ready to query." format: "Video" duration: "PT4M13S" video: "https://videos.ctfassets.net/nwa1c00rtgxb/4kJ56QXBbDZPsmi2qPwop1/b4b97c659ab2556a29e86a1b251fddfe/MDLH_-_Setting_up_the_MDLH.mp4" thumbnail: "https://images.ctfassets.net/nwa1c00rtgxb/4CjvVQRW0CANC075KJb3XO/00bc7c3c5c037339a8b409f31d013f13/mdlh-setting-up-the-mdlh.webp" content_purpose: ["Product Overview"] target_persona: ["General", "Data Analyst", "Data Engineer", "IT Administrator"] journey_stage: ["S2 - Discovery", "S3 - Solution Design", "S4 - Business Case", "C1 - Onboarding", "C2 - First Value", "C3 - Adoption"] use_case_context: ["Training"] product: ["Enterprise Data Graph - Integration"] published: "2026-03-13" updated: "2026-06-02" content_type: "video transcript" transcript_source: "sheet" --- # MDLH: Setting up the MDLH Transcript of the video at https://atlan.com/demos/mdlh-setting-up-the-mdlh/ Your catalog has owners, lineage, descriptions, and tags — but you still can't run SQL against any of it. The Metadata Lakehouse changes that. Connecting it to Snowflake takes a few minutes. We’ve seen teams build full API extract pipelines just to get Atlan metadata into a queryable form — pagination logic, rate limit handling, schema maintenance every time something changes. Weeks of engineering to answer questions the Lakehouse handles in a single query. The setup here takes less time than it took to scope those pipelines. Without the Lakehouse, bulk metadata access means going through the API — paginating through thousands of assets, managing rate limits, and rebuilding extracts when schemas change. The Lakehouse changes the model entirely: instead of extracting metadata to query it elsewhere, we connect our query engine directly to where it lives. Every asset, tag, lineage path, and governance attribute becomes tables and attributes. By the end of this, that connection will be live. By the end of this walkthrough, we'll have the Metadata Lakehouse enabled, Snowflake connected, and our first query running against live Atlan metadata. Atlan handles all the infrastructure — on our end, it's two steps in the Atlan UI and two SQL commands in Snowflake. Quick note before we start: we’re using Snowflake as the compute layer in this walkthrough, but the Metadata Lakehouse is built on Apache Iceberg — which means it’s compatible with any engine that supports the Iceberg spec. If our team runs on Databricks, Spark, or Trino, the same Lakehouse sits underneath. The connection steps will differ on our side, so be sure to check the documentation. Let’s walk through the setup. To begin the set up, we must be an administrator. We go into the Workflows tab and then marketplace, which is where we find the Atlan Lakehouse tile. We Click it and hit Enable. We’re prompted to enter our tenant name, we Confirm it, and Atlan kicks off the setup. In the background, Atlan is provisioning the Lakehouse — building out Iceberg-formatted tables for every asset type in our workspace: tables, columns, dashboards, lineage paths, glossary terms, tags, and governance attributes. This part is async, we don’t even have to worry about it. — setup can take a few hours depending on the size of our data estate. We'll get an email when it's done. The connection details won't be available until that setup completes, so we have to wait for the confirmation before moving on. Once we get the confirmation email, we come back to the Marketplace and reopen the Atlan Lakehouse tile. By clicking on the view connection details, We see our catalog URI, catalog name, and the OAuth credentials we'll need to connect our query engine. Atlan pre-fills all of these details for us. We're not configuring anything manually — we're just copying what Atlan has already set up. Now let’s take those details over to Snowflake. Now it’s time to set up Snowflake. At the bottom, we have tabs for the different computes that we could look to connect to our lakehouse. For Snowflake, we are provided with the exact Snowflake commands we need to create the integration and linked databases. Now we open our Snowflake console and navigate to a new SQL query. We'll need a role with ACCOUNTADMIN privileges to run this — if we don't have that, get our Snowflake admin involved before starting. We paste the SQL and run. The script creates two things: a catalog integration that points at the Atlan Lakehouse, and a linked database called context_store. As noted, we can update the script to match our naming conventions before running, but be sure to keep the updates consistent throughout the script. Within seconds, we'll see context_store appear under our Databases in Snowflake — right alongside our other data sources, like it was always there. Here's what that actually means: every asset in Atlan — every table, column, dashboard, glossary term, every tag and every lineage relationship — is now queryable from Snowflake. not a snapshot pushed to you. It's live metadata, available through standard SQL. You're treating your metadata like data. We've got the Metadata Lakehouse running and connected to Snowflake. From here, everything in our Atlan workspace is queryable — asset metadata, lineage relationships, glossary terms, tag coverage, usage data. No pipelines, no extracts, no waiting for someone to build a report for us. If you want to see what kinds of queries you can actually run against the Metadata Lakehouse — metadata completeness, lineage exports, glossary analysis — check out our use case videos. Drop a comment with what you’d want to query from your catalog and why — we read them all.