Mcp Victoriametrics
Let your coding agent query PromQL/MetricsQL and VictoriaMetrics docs against your live vmsingle or vmselect cluster without leaving the chat.
Overview
mcp-victoriametrics is an Operate-phase MCP server that connects AI coding agents to your VictoriaMetrics metrics API and documentation via authenticated stdio tools.
What is this MCP server?
- Stdio MCP package (OCI ghcr.io/victoriametrics-community/mcp-victoriametrics:1.18.0) wired to VM_INSTANCE_ENTRYPOINT
- Supports single-node and cluster instance types via VM_INSTANCE_TYPE
- Bearer-token auth (VM_INSTANCE_BEARER_TOKEN) plus optional VictoriaMetrics Cloud VMC_API_KEY
- Exposes VictoriaMetrics API integration and bundled documentation to the agent
- Pairs with VictoriaMetrics-Community MCP siblings for traces, logs, and anomaly workflows
- Server schema version 1.18.0 in registry metadata
- Stdio transport via OCI image ghcr.io/victoriametrics-community/mcp-victoriametrics:1.18.0
- Three required env vars: VM_INSTANCE_ENTRYPOINT, VM_INSTANCE_TYPE, VM_INSTANCE_BEARER_TOKEN
Community signal: 178 GitHub stars.
What problem does it solve?
Debugging production metrics from chat means copy-pasting PromQL into dashboards or terminals instead of letting the agent query your VictoriaMetrics instance directly.
Who is it for?
Indie builders who already run VictoriaMetrics (single or cluster) and want agent-assisted incident triage, query drafting, and doc-aware metrics exploration.
Skip if: Teams without VictoriaMetrics, or builders who only need log search or distributed traces—use the VictoriaLogs or VictoriaTraces MCP servers instead.
What do I get? / Deliverables
After you register the MCP with entrypoint, instance type, and bearer token, your agent can reason over live VictoriaMetrics metrics and docs in the same session where you fix code.
- Registered MCP server talking to your VictoriaMetrics instance
- Agent-callable tools backed by VictoriaMetrics API and documentation
- Faster metrics-aware debugging without leaving the coding session
Recommended MCP Servers
Journey fit
Observability for shipped products belongs in Operate—metrics are how solo builders answer “is prod healthy?” after launch. Monitoring is the canonical shelf: this MCP is for time-series metrics APIs, not traces, logs, or deploy tooling.
How it compares
Metrics-focused VictoriaMetrics MCP integration, not a standalone monitoring SaaS or a generic agent skill.
Common Questions / FAQ
Who is mcp-victoriametrics for?
Solo and small-team builders operating VictoriaMetrics who want Claude Code, Cursor, or Codex to call the metrics API and read VictoriaMetrics documentation through MCP.
When should I use mcp-victoriametrics?
Use it during Operate when you are investigating alerts, writing or validating PromQL/MetricsQL, or explaining metrics behavior while shipping fixes.
How do I add mcp-victoriametrics to my agent?
Add the OCI package with stdio transport to your agent MCP config, set VM_INSTANCE_ENTRYPOINT, VM_INSTANCE_TYPE, and VM_INSTANCE_BEARER_TOKEN (and VMC_API_KEY if using VictoriaMetrics Cloud), then restart the agent.