Influxdb3 MCP Server

What InfluxData/influxdb3_mcp_server MCP server does

InfluxData/influxdb3_mcp_server MCP server connects MCP clients to InfluxDB 3 deployments and exposes database operations as agent-callable tools. Supported deployment types include Core, Enterprise, Cloud Dedicated, Cloud Clustered, and Cloud Serverless, with some capabilities restricted to particular editions.

The server covers the main stages of working with a time-series database: checking connectivity, finding databases and tables, inspecting schemas, querying data, writing line protocol, and changing database configuration. It also provides token-management operations. Destructive actions include deleting databases, deleting tokens, and regenerating the operator token, so clients should apply appropriate authorization and confirmation controls.

The listing belongs to the 🗄️ Databases category and is suited to agents that need direct access to InfluxDB metadata and data rather than a general-purpose analytics abstraction.

How it works

An MCP client calls the server's tools and can also read exposed resources. Available resources include InfluxDB configuration, current connection and health status, the database list, and custom database context supplied by the user. The server also includes prompts, although the provided material does not describe their individual names or behavior.

For investigation workflows, an agent can enumerate databases, list tables or measurements, inspect columns and types, then issue a bounded SQL or InfluxQL read. Structured query responses can include row counts, truncation status, warnings, and correlation metadata such as request and query identifiers. Read-only workflows can use INFLUX_MCP_TOOL_PROFILE=readonly; this removes access to mutation, administration, token-management, and host-level tools.

Setup and configuration

Running the InfluxData/influxdb3_mcp_server MCP server requires an InfluxDB 3 instance and credentials. Core, Enterprise, and Cloud Serverless connections use an instance URL and token. Cloud Dedicated and Cloud Clustered connections use a cluster ID and tokens.

For npm or npx-based use, the documented prerequisites are Node.js 20.11 or newer and npm 9 or newer. Docker is also supported as a setup route. The read-only tool profile is enabled with the INFLUX_MCP_TOOL_PROFILE environment variable set to readonly.

Tools and capabilities

InfluxData/influxdb3_mcp_server MCP server provides tools for:

• Listing databases, tables, and measurements, and describing measurement schemas.
• Running SQL, bounded read-only SQL, and bounded read-only InfluxQL queries.
• Writing line protocol and creating, updating, or deleting databases.
• Investigating a database through discovery and sampling operations.
• Checking connection health and retrieving help or troubleshooting guidance.
• Creating and managing Core or Enterprise tokens, including admin and resource tokens.
• Managing database tokens for Cloud Dedicated and Cloud Clustered deployments.
• Loading optional database-specific context and documentation.

Schema descriptions use conservative column categories, and read-only query tools are designed to bound data access. Cloud-specific database configuration and token tools are not interchangeable with Core or Enterprise token operations.

Limitations and notes

Tool availability varies by deployment. Admin and resource token operations, token deletion, and operator-token regeneration are listed for Core and Enterprise only. Cloud database-token operations are listed for Cloud Dedicated and Cloud Clustered only. Database deletion and operator-token regeneration are irreversible or dangerous according to the tool descriptions.

The provided material does not specify a license, detailed transport configuration, or exact credential variable names beyond the read-only profile variable. Enterprise preview user authentication can use a JWT bearer credential in the read-only workflow instead of an apiv3_ token.