refactor(main): count MCP tools at INFO, per-tool details at DEBUG

Production log noise: every container start was emitting 14 INFO lines
listing every tool name and full description, which is dev-time
debugging info that doesn't belong in INFO.

Now:
  - INFO: one line per service with the count
    'MCP service \042spark_executor\042: 14 tool(s) registered'
  - DEBUG: per-tool name + description, prefixed with the service name
    so multi-service deploys are grep-friendly
    '  spark_executor._prepare_submit_job_prepare_submit_job_post  -  ...'

Operators can flip SPARK_EXECUTOR_LOG_LEVEL=INFO for production (no
per-tool spam) or DEBUG for dev (full inventory).

146/146 still pass. Live verified: at INFO level only the count line
shows; at DEBUG level both the count and per-tool lines show.

main.py

@@ -50,13 +50,16 @@ def _log_mcp_tools(mcp_server, service: McpService) -> None:
     scannable. Each tool is logged on one line: `  N. tool_name  -  desc`.
     """
     tools = mcp_server.tools
-    logger.info(f"MCP tools registered ({service.name}, {len(tools)}):")
+    # Single INFO line per service summarizing how many tools it exposes.
+    # Per-tool name + description is DEBUG-level so production logs (INFO)
+    # stay scannable but operators can flip to DEBUG to see the full list.
+    logger.info(f"MCP service {service.name!r}: {len(tools)} tool(s) registered")
     for i, tool in enumerate(tools, 1):
         desc = (tool.description or "").strip()
         if "### Responses" in desc:
             desc = desc.split("### Responses", 1)[0].rstrip()
         desc = " ".join(desc.split())
-        logger.info(f"  {i:>2}. {tool.name}  -  {desc}")
+        logger.debug(f"  {service.name}.{tool.name}  -  {desc}")
 
 
 @asynccontextmanager