VictoriaMetrics MCP Server: Query Time-Series Data Through Natural Language

Stop switching between your code editor, terminal, and VictoriaMetrics UI just to check system metrics. This MCP server puts your entire VictoriaMetrics cluster at your fingertips through Claude, letting you write PromQL queries, analyze trends, and debug performance issues without leaving your development workflow.

Why This Changes Your Metrics Workflow

If you're running VictoriaMetrics in production, you know the drill: performance issue hits, you need to check CPU utilization across pods, memory usage spikes, or request latency patterns. Usually that means:

• Opening the VictoriaMetrics UI in another tab
• Writing PromQL queries from scratch
• Copying results back to your analysis
• Repeating this cycle for every metric you need

Now you can ask Claude: "Show me CPU usage for the auth service in the last hour" or "What's causing those memory spikes in production?" and get real data back instantly.

Complete VictoriaMetrics Integration

This MCP server gives you six essential tools that cover everything from data ingestion to advanced querying:

Query Operations:

• vm_query - Get current metric values with PromQL
• vm_query_range - Analyze time-series data over specific periods
• vm_labels - Discover available metrics and dimensions
• vm_label_values - Find specific label values for filtering

Data Ingestion:

• vm_data_write - Write custom metrics programmatically
• vm_prometheus_write - Import existing Prometheus exposition format data

Each tool handles the VictoriaMetrics API complexity for you, so you can focus on understanding your data rather than debugging HTTP requests.

Real-World Use Cases

Incident Response: When alerts fire, ask Claude to pull related metrics, compare current values to historical baselines, and identify correlations between services. No more fumbling with Grafana dashboards while systems are down.

Performance Analysis: Building a new feature? Have Claude query request latency, error rates, and resource usage patterns to validate your optimization efforts or spot regressions before they hit users.

Capacity Planning: Need to justify infrastructure costs? Ask for month-over-month growth trends, peak usage patterns, and resource utilization across your entire stack.

Custom Metrics: Publishing business metrics from your application? Use the write APIs to send custom data points and immediately query them for analysis.

Installation Takes 30 Seconds

Add this to your Claude Desktop configuration:

{
    "mcpServers": {
        "victoriametrics": {
            "command": "npx",
            "args": ["-y", "@yincongcyincong/victoriametrics-mcp-server"],
            "env": {
                "VM_URL": "http://your-vm-instance:8428"
            }
        }
    }
}

Or install via Smithery:

npx -y @smithery/cli install @yincongcyincong/victoriametrics-mcp-server --client claude

That's it. No complex authentication setup, no API keys to manage, no additional services to deploy. Point it at your VictoriaMetrics instance and start querying.

Your Metrics, Your Language

The real power isn't in the individual tools—it's in how Claude can chain them together. Ask complex questions like "Find all services with high error rates in the last 24 hours, show me their CPU and memory usage, and compare to last week's baseline." Claude will automatically use the right combination of label discovery, range queries, and data analysis to give you actionable insights.

For teams running microservices, container orchestration, or any system generating time-series data, this MCP server transforms how you interact with your observability stack. Your metrics become as accessible as having a conversation.