Mcp Victoriametrics

VictoriaMetrics MCP Server is an MCP server that exposes read-only VictoriaMetrics APIs and embedded documentation to LLM clients and autonomous agents. It connects directly to single-node VictoriaMetrics instances, multi-tenant VictoriaMetrics clusters, and VictoriaMetrics Cloud deployments using standard HTTP or token authentication. Platform engineers, site reliability engineers, and DevOps teams use this server to examine time-series metrics, inspect cluster status, and troubleshoot performance issues directly within AI interfaces. The server provides capabilities equivalent to the VMUI exploration interface, enabling clients to query Prometheus-compatible metrics, retrieve series metadata, and visualize time-series trends. It also allows engineers to analyze query execution traces, prettify complex MetricsQL expressions, evaluate alerting and recording rules, and audit metric cardinality. Additionally, the server includes local search over offline VictoriaMetrics documentation and tools to debug relabeling configurations, retention filters, and downsampling policies. By packaging these observability endpoints into standard Model Context Protocol tools and resources, the server facilitates interactive telemetry analysis across diverse operating environments through standard input/output, Server-Sent Events, or Streamable HTTP transport modes.

Category: Databases & Data Stores

Tags: metrics, monitoring, time-series, victoriametrics

Visit Mcp Victoriametrics

How to install and configure Mcp Victoriametrics

1. Download the latest binary release from GitHub and place it in your system PATH, or deploy using the official Docker container: bash docker run -d --name mcp-victoriametrics \ -e VM_INSTANCE_ENTRYPOINT=https://play.victoriametrics.com \ -e VM_INSTANCE_TYPE=cluster \ -e MCP_SERVER_MODE=sse \ -e MCP_LISTEN_ADDR=:8080 \ -p 8080:8080 \ ghcr.io/victoriametrics/mcp-victoriametrics 2. For local integration with Claude Desktop using stdio mode, add the binary to your configuration file: json { "mcpServers": { "victoriametrics": { "command": "mcp-victoriametrics", "env": { "VM_INSTANCE_ENTRYPOINT": "http://localhost:8428", "VM_INSTANCE_TYPE": "single" } } } } 3. Alternatively, connect via remote Streamable HTTP mode by targeting the public testing endpoint https://play-mcp.victoriametrics.com/mcp or your custom server address.

What you can do with Mcp Victoriametrics

• Querying and visualizing time-series metrics from single-node or cluster VictoriaMetrics deployments using MetricsQL expressions. - Investigating high-cardinality label values and analyzing metric usage statistics to optimize storage and indexing overhead. - Tracing, prettifying, and explaining complex MetricsQL queries to diagnose slow evaluation times and resolve syntax issues. - Validating and testing alerting configurations, recording rules, relabeling logic, and retention policies before applying them to production. - Searching offline VictoriaMetrics documentation to verify parameter names, architectural concepts, and API specifications within AI prompts.

Key facts

• https://github.com/VictoriaMetrics-Community/mcp-victoriametrics
• Databases & Data Stores, Monitoring & Observability
• metrics, monitoring, time-series, victoriametrics

Part of MCP Servers

How do I install VictoriaMetrics MCP Server?

You can install the server by downloading the precompiled release binary from GitHub and placing it in your system PATH, by running the official Docker container from ghcr.io/victoriametrics/mcp-victoriametrics, or by compiling it from source using Go and make build. It is also deployable on Kubernetes using an official Helm chart.

What can VictoriaMetrics MCP Server do?

It exposes read-only endpoints from VictoriaMetrics to MCP clients. Supported operations include executing MetricsQL queries, listing metrics and label values, analyzing label cardinality, tracing slow queries, testing alerting rules, and debugging relabeling filters. It also includes an embedded copy of VictoriaMetrics documentation for offline semantic queries.

Which MCP clients work with VictoriaMetrics MCP Server?

VictoriaMetrics MCP Server works with any client that implements the Model Context Protocol, including Claude Desktop, Cursor, and custom agentic frameworks. It supports multiple transport layers such as standard input and output (stdio), Server-Sent Events (SSE), and Streamable HTTP.

How do I connect to VictoriaMetrics Cloud?

To connect to VictoriaMetrics Cloud, configure the server with the VMC_API_KEY environment variable using the key obtained from your VictoriaMetrics Cloud Console. When this variable is set, you do not need to configure the VM_INSTANCE_ENTRYPOINT or VM_INSTANCE_TYPE settings, allowing automated integration with managed deployments.

Can I try VictoriaMetrics MCP Server without running an instance?

Yes, you can test the server using the public endpoint provided at https://play-mcp.victoriametrics.com/mcp. This instance runs in Streamable HTTP mode and connects directly to the public VictoriaMetrics Playground. It allows you to test queries and MCP integration inside clients like Claude Desktop without deploying a local database instance.