Databricks Status

/databricks-status — Databricks Workspace Status

You are checking the operational status of a Databricks workspace: clusters, recent job runs, SQL warehouses, current user.

This skill is a thin wrapper around the databricks MCP tools.

Prerequisites

• Databricks MCP server databricks configured in .mcp.json

Usage

/databricks-status                          # default: clusters + warehouses + user
/databricks-status clusters                 # only clusters
/databricks-status runs [--limit N]         # recent job runs (default 10)
/databricks-status warehouse                # SQL warehouses only
/databricks-status user                     # whoami

Arguments via $ARGUMENTS:

• positional what — one of clusters | runs | warehouse | user (default: all)
• --limit N — for runs, how many to show (default 10)

Behavior

Step 1: Resolve User (always)

mcp__databricks__get_current_user()

Show the authenticated user and home workspace path — verifies that the MCP profile is bound to the expected workspace.

Step 2: Clusters

mcp__databricks__list_clusters()

Render as a table:

Highlight in bold the cluster_id matching overlays/{env}.yaml databricks.cluster_id if defined (when an --env is supplied or guessed from the active project).

States to display: RUNNING, TERMINATED, PENDING, RESIZING, RESTARTING, ERROR, UNKNOWN.

Step 3: SQL Warehouses

mcp__databricks__list_warehouses()

Render:

If databricks.sql_warehouse_id is defined in any overlay, mark the matching warehouse with [configured].

Step 4: Recent Runs

mcp__databricks__manage_job_runs(action="list", limit={N}, completed_only=False)

Render:

Status values: SUCCESS, FAILED, RUNNING, PENDING, CANCELED, TIMEDOUT, INTERNAL_ERROR.

The run page URL helps the user jump to the Databricks UI for log inspection.

Logging

Status checks are read-only. Do NOT append to ops.log for status calls (would create noise).

Notes

• list_clusters() returns all clusters visible to the user (may include shared/other-user clusters). Use overlay highlighting to disambiguate the "right" one for this project.
• list_warehouses() lists all warehouses the user can access.
• For lineage queries, use /databricks-lineage (separate skill).
• For SQL exec, use /databricks-query.