/ /

Troubleshoot MongoDB MCP Server

Use the information on this page to help troubleshoot issues with the MongoDB MCP Server.

Log Files

If you have issues with the MCP Server, examine the MCP Server logs for information that can help troubleshoot the problem.

By default, the MCP Server writes the logs to disk and also sends the logs to the MCP client. You can change the loggers option to specify where the logs are sent. For details about the MCP Server options, see MongoDB MCP Server Configuration Options.

The following table shows the default directories where the MCP Server stores the log files:

Operating System	Default Log Directory
Windows	`%LOCALAPPDATA%\mongodb\mongodb-mcp\.app-logs`
macOS and Linux	`~/.mongodb/mongodb-mcp/.app-logs`

To change the log path, set one of these:

logPath option in a JSON configuration file or on a command line when you start the MCP Server
MDB_MCP_LOG_PATH environment variable

Connection String Issues

Ensure that your configuration file is correct and that you're using a valid connection string for your MongoDB deployment.

For connection string formatting and examples, see Connection Strings.

HTTP Transport Issues

If you're using HTTP transport and experiencing connection issues:

Verify that your client connection URL includes the /mcp endpoint path:
http://<server-address>:3000/mcp
If connecting to a remote server (different computer, container, or cloud instance), ensure the server is started with --httpHost=0.0.0.0 to accept remote connections:
npx -y mongodb-mcp-server@latest --transport http --httpHost=0.0.0.0 --readOnly
Ensure that the specified port is open and accessible from your client computer.

Atlas Tools vs Database Tools Configuration Guide

Overview

The MongoDB MCP Server provides three main categories of tools: Atlas tools for managing Atlas infrastructure, Local Atlas tools for managing local deployments similarly to what the Atlas CLI offers, and Database tools for data operations. The following table compares the configuration of each tool.

Configuration Comparison Table

Configuration Aspect	Atlas Tools	Local Atlas Tools	Database Tools
Authentication Requirements	Required: Atlas API credentials • `apiClientId` (Service Account ID) • `apiClientSecret` (Service Account Secret)	Required: Docker runtime • No additional credentials needed • Automatically manages local deployments	Optional: Multiple connection methods • Pre-configured connection string • Atlas credentials for `atlas-connect-cluster` • Runtime connection via `connect` tool • Local deployment via `atlas-local-connect-deployment`
Connection Method	API-based connection to Atlas Management API • No direct database connection needed • Uses Atlas REST API endpoints	Local deployment management • Uses Docker containers • No external network access required	Direct connection to MongoDB database. Multiple options available: • Configured connection string • Runtime connection prompt • Atlas cluster discovery • Local deployment connection
Permission Requirements	Service Account Roles (Atlas project-level): • `Project Read Only` (minimum for read operations) • `Project Owner` (for cluster creation/management) • `Organization Owner` (for project creation)	System Requirements: • Docker runtime permissions • Local file system access • Container management capabilities	Database User Roles (when applicable): • `readAnyDatabase` (read-only mode) • `readWriteAnyDatabase` (full operations) • Custom roles for specific databases/collections
Configuration Options	Environment Variables: `MDB_MCP_API_CLIENT_ID` or `MDB_MCP_API_CLIENT_SECRET` Command Line: • `--apiClientId` • `--apiClientSecret`	Environment Variables: • `DOCKER_HOST` (optional) Command Line: • No specific MCP configuration required • Automatically detects Docker runtime	Environment Variables (optional): `MDB_MCP_CONNECTION_STRING` Command Line (optional): • Connection string as positional argument: `--connectionString` Runtime Methods: • Interactive connection via LLM prompts • Atlas credential-based discovery
Read-Only Mode Impact	• Read operations: ✅ Available • Cluster inspection: ✅ Available • Cluster creation: ❌ Disabled • User management: ❌ Disabled	• Local deployment creation: ✅ Available • Deployment inspection: ✅ Available • Container management: ✅ Available • No read-only restrictions	• Read operations: ✅ Available • Query operations: ✅ Available • Aggregations: ✅ Available as long as they don't have a `$out` or `$merge` stage. • Insert/Update/Delete: ❌ Disabled • Index creation: ❌ Disabled
Tool Categories	Can be disabled with: `--disabledTools atlas` or `MDB_MCP_DISABLED_TOOLS="atlas"`	Can be disabled with: `--disabledTools atlas` or `MDB_MCP_DISABLED_TOOLS="atlas"`	Can be disabled with: `--disabledTools mongodb` or `MDB_MCP_DISABLED_TOOLS="mongodb"`
Network Requirements	Outbound HTTPS to Atlas API: • Standard internet connection	Local Docker Network: • Docker daemon access • Local port availability • No external network requirements	Connection to MongoDB cluster (when applicable): • Atlas: `mongodb+srv://` (port 27017) • Sharded clusters (port 27016) • Self-managed: Custom host/port • Local deployments: Docker network • VPC/network access as required
Available Operations	• Organization management • Project management • Cluster lifecycle (create, inspect, delete) • Database user management • Access list configuration • Performance Advisor recommendations • Alert management	• Local deployment creation and management • Container lifecycle operations • Local cluster configuration • Development environment setup • Testing deployment management	• CRUD operations (find, insert, update, delete) • Aggregation pipelines • Index management • Collection and database administration • Schema inspection • Vector search operations • Data export
Error Scenarios	Missing API Credentials: Atlas tools return: "Atlas API credentials not configured" Insufficient Permissions: HTTP 403 errors from Atlas API	Docker Unavailable: "Docker daemon not running" or "Docker not installed" Container Issues: Port conflicts, resource constraints	Connection Failures (when configured): Network timeouts, authentication failures Missing Database Permissions: MongoDB authorization errors
Temporary Credentials	Not applicable: Uses persistent service account credentials	Not applicable: Local deployments use container-managed authentication	Auto-generated database users (when applicable): When using Atlas API credentials without connection string: • Random username/password • 4-hour expiration by default • Automatic role assignment

Database Connection Methods

The Database tools support multiple connection methods to accommodate different use cases and security requirements:

Pre-configured Connection String: Configure connectionString via CLI, environment variable, or configuration file for persistent database access.
Atlas Cluster Connect: Use atlas-connect-cluster tool with Atlas API credentials to automatically discover and connect to Atlas clusters.
Runtime Connection: Use the connect tool to provide connection strings interactively during LLM conversations - no pre-configuration required.
Local Deployment Connection: Use atlas-local-connect-deployment with Docker to create and connect to local Atlas deployments - requires Docker but no configuration.

Security Recommendations:

Production databases: Use pre-configured connection strings via environment variables or Atlas API credentials for secure, automated connections.
Development/testing: Runtime connection or local deployments provide flexibility without persistent configuration.
Chat security: Avoid runtime connection methods unless you're confident chat history won't be stored or shared with third parties.

Learn More

Back

Memory Overflow

Security