𝐃𝐢𝐫𝐞𝐜𝐭 𝐋𝐚𝐤𝐞 𝐆𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬: 𝐈𝐝𝐞𝐧𝐭𝐢𝐟𝐲 𝐏𝐨𝐭𝐞𝐧𝐭𝐢𝐚𝐥 𝐅𝐚𝐥𝐥𝐛𝐚𝐜𝐤𝐬 𝐰𝐢𝐭𝐡 𝐒𝐞𝐦𝐚𝐧𝐭𝐢𝐜 𝐋𝐢𝐧𝐤 𝐋𝐚𝐛𝐬

𝐅𝐚𝐥𝐥𝐛𝐚𝐜𝐤:

Power BI semantic models in Direct Lake mode read Delta tables directly from OneLake. However, if a DAX query on a Direct Lake model exceeds the SKU limits, the query can fall back to Direct Query mode, affecting query performance.

𝐔𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝𝐢𝐧𝐠 𝐃𝐢𝐫𝐞𝐜𝐭 𝐋𝐚𝐤𝐞 𝐆𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬:

Direct Lake guardrails define the resource boundaries within which Direct Lake mode operates optimally. When these limits are reached, Power BI switches to Direct Query to process DAX queries.
Key guardrails include: 𝐭𝐡𝐞 𝐧𝐮𝐦𝐛𝐞𝐫 𝐨𝐟 𝐩𝐚𝐫𝐪𝐮𝐞𝐭 𝐟𝐢𝐥𝐞𝐬, 𝐫𝐨𝐰 𝐠𝐫𝐨𝐮𝐩𝐬, 𝐨𝐫 𝐫𝐨𝐰𝐬 𝐩𝐞𝐫 𝐃𝐞𝐥𝐭𝐚 𝐭𝐚𝐛𝐥𝐞.

𝐃𝐢𝐫𝐞𝐜𝐭 𝐋𝐚𝐤𝐞 𝐆𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬 𝐂𝐡𝐞𝐜𝐤:

👉Consult the Microsoft documentation for detailed guardrails based on your Fabric SKU: https://learn.microsoft.com/en-us/fabric/get-started/direct-lake-overview#fallback

👉Also, you can use the SemanticLinkLabs and run the 𝐠𝐞𝐭_𝐝𝐢𝐫𝐞𝐜𝐭_𝐥𝐚𝐤𝐞_𝐠𝐮𝐚𝐫𝐝𝐫𝐚𝐢𝐥𝐬() function to quickly view these guardrails, helping you understand when Direct Lake semantic models will fall back to Direct Query.

𝐈𝐝𝐞𝐧𝐭𝐢𝐟𝐲 𝐏𝐨𝐭𝐞𝐧𝐭𝐢𝐚𝐥 𝐅𝐚𝐥𝐥𝐛𝐚𝐜𝐤 𝐓𝐚𝐛𝐥𝐞𝐬 𝐰𝐢𝐭𝐡 𝐒𝐞𝐦𝐚𝐧𝐭𝐢𝐜 𝐋𝐢𝐧𝐤 𝐋𝐚𝐛𝐬:

In the screenshot below 👇👇, I used the 𝐒𝐞𝐦𝐚𝐧𝐭𝐢𝐜 𝐋𝐢𝐧𝐤 𝐋𝐚𝐛𝐬, via the 𝐠𝐞𝐭_𝐥𝐚𝐤𝐞𝐡𝐨𝐮𝐬𝐞_𝐭𝐚𝐛𝐥𝐞𝐬() function to check if any lakehouse tables will hit the Direct Lake guardrails based on the SKU used by my Fabric workspace capacity.
The function retrieves the tables of a lakehouse and their respective properties relevant to Direct Lake guardrails👇

link: https://semantic-link-labs.readthedocs.io/en/stable/sempy_labs.lakehouse.html#sempy_labs.lakehouse.get_lakehouse_tables


❗Note: You can disable fallback to DirectQuery mode if you want to process DAX queries in pure Direct Lake mode only.