video1.0<iframe src="https://www.loom.com/embed/78ed42225a484ee7a13aa14e70058da6" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/78ed42225a484ee7a13aa14e70058da6-e33b5fa9ed7a813a.gif392.033This Loom introduces the new DB Marlin MCP service, which lets AI clients query DB Marlin through a middle layer between DB Marlin APIs and large language models like Claude. The demo shows configuration in an MCP-enabled extension pointing to a DBM endpoint, with 22 read-only tools for tasks such as instance activity, top SQL, blocking sessions, deadlocks, host metrics, and table and index details. Using Claude, the presenter asks for database performance for the last hour across four instances, then drills into specific time ranges and statements, including reproducing charts from two load tests over the last six hours. The Loom also demonstrates tuning a delete statement by suggesting an index on orders.customer_id based on performance findings, and mentions the orders table size as about 6.22 rows and 1.4 megabytes.