docs: document MCP server connection status

guides/ai-agents/mcp-servers.mdx

@@ -83,6 +83,17 @@ The agent pulls the data from your warehouse, generates the summary, and appends
 
 The agent searches Confluence for the canonical definition, then answers using both the doc and your data.
 
+## Connection status
+
+Each MCP server shows its current runtime status in the servers list so you can tell at a glance whether agents can reach it. If a server fails to connect or returns an error during use, Lightdash records the latest error message alongside the status. Use this to debug issues like expired credentials, network errors, or invalid endpoints without having to reproduce the failure in a chat.
+
+If a server stays in an error state:
+
+- Check the **URL** is reachable and correct.
+- For **Bearer token** auth, verify the token has not expired or been revoked.
+- For **OAuth**, click **Reconnect** to refresh the authorization.
+- Re-test by attaching the server to an agent and sending a message; the status updates after the next connection attempt.
+
 ## Security
 
 - All credentials (bearer tokens and OAuth tokens) are encrypted at rest.