Postgres MCP Pro

Instructions

Implementation Reference

• src/postgres_mcp/server.py:407-420 (handler)

  The implementation of the `analyze_workload_indexes` tool, which delegates to either `DatabaseTuningAdvisor` or `LLMOptimizerTool` via `TextPresentation`.

  async def analyze_workload_indexes(
      max_index_size_mb: int = Field(description="Max index size in MB", default=10000),
      method: Literal["dta", "llm"] = Field(description="Method to use for analysis", default="dta"),
  ) -> ResponseType:
      """Analyze frequently executed queries in the database and recommend optimal indexes."""
      try:
          sql_driver = await get_sql_driver()
          if method == "dta":
              index_tuning = DatabaseTuningAdvisor(sql_driver)
          else:
              index_tuning = LLMOptimizerTool(sql_driver)
          dta_tool = TextPresentation(sql_driver, index_tuning)
          result = await dta_tool.analyze_workload(max_index_size_mb=max_index_size_mb)
          return format_text_response(result)

• src/postgres_mcp/server.py:405-405 (registration)

  Registration of the `analyze_workload_indexes` tool using `@mcp.tool`.

  @mcp.tool(description="Analyze frequently executed queries in the database and recommend optimal indexes")