1C MCP Fleet Monitor
A single-container web dashboard for monitoring and controlling a fleet of 1C MCP servers: status, indexing progress, VRAM and license checks
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Controls container lifecycle: start, stop, restart and recreation with a different profile
- Can delete a graph project in Graph-beta and reset the index database
Install
Manual install
docker build -t mcp-1c-monitor:latest .
docker run -d --name mcp_1c_monitor --restart unless-stopped --gpus all --env-file "$env:USERPROFILE\.mcp-secrets\monitor.env" -p 127.0.0.1:8090:8090 -v /var/run/docker.sock:/var/run/docker.sock mcp-1c-monitor:latestBuild and run on Windows with PowerShell; the env file is kept outside the repo. On Linux add --add-host=host.docker.internal:host-gateway.
This is third-party code. Review the repository files before installing.
What it does
The tool polls the Docker API and MCP ports of seven servers from the onerpa.ru lineup (HelpSearch, GraphMetadata, CodeMetadata, SSLSearch, TemplatesSearch, SyntaxCheck, 1CCodeChecker) plus a Neo4j database, refreshing each card every 5 seconds: container state, a real MCP probe via initialize and tools/list over Streamable HTTP, license key status, indexing progress parsed from logs, and start, stop, restart and browser log-view buttons. A separate GPU panel shows used and free VRAM and lets you switch four model-backed servers between GPU and CPU without reindexing: index volumes are preserved, only the container profile changes. A data panel exposes what is inside the indexes: collection stats for CodeMetadata, and graph stats, project lists and background task status for Graph-beta, with buttons to reindex and to delete a graph project.
Who it is for. For 1C developers and devops engineers who already run a fleet of onerpa.ru MCP servers and want one dashboard for their state instead of separate docker ps calls and logs.
Good fit when
- You run several onerpa.ru MCP servers and want their readiness and indexing progress in one place
- You need to quickly switch model-backed servers between GPU and CPU to free VRAM for a local LLM
- You need to check the servers' license keys without going into each container
Not a fit when
- You have not deployed the onerpa.ru 1C MCP servers themselves: the monitor is useless without them
- You need an MCP server for an agent, not a monitoring web panel: the monitor itself does not expose tools to an agent
Example request
Show the status of all fleet servers and the CodeMetadata indexing progressLimitations
The tool is built for a specific onerpa.ru MCP server lineup, documented outside this repo, and is useless without them. It needs access to the Docker socket and, for the GPU panel, GPU access via --gpus all. The project is small, essentially a single app.py file, with no tests visible in the repo tree.
How to disable. Stop and remove the container: docker stop mcp_1c_monitor && docker rm mcp_1c_monitor.
Security check
- Controls container lifecycle: start, stop, restart and recreation with a different profile
- Can delete a graph project in Graph-beta and reset the index database
README in short
The README describes a single container that polls Docker and the MCP ports of the server fleet, showing readiness, indexing progress parsed from logs, license status and control buttons. It gives a quick start via docker build and docker run with example flags, an HTTP API table, an explanation of how readiness is determined given the images lack standard health endpoints, a GPU/CPU switching section with specific CodeMetadata indexing tuning parameters, and a data panel section covering stats, reindex and graph project deletion tools. MIT license.
FAQ
Is this an MCP server?
No, it is a web monitoring and control panel for already-deployed onerpa.ru 1C MCP servers; the monitor itself exposes no tools to an agent.
Do I need to reindex after switching GPU/CPU?
No, indexes live in volumes and survive container recreation; only embedding processing speed changes.
Related
An MCP server built into the Netdata agent: metrics, logs, alerts and live process, service and container data for an AI assistant
GitHub's official MCP server: code, issues, pull requests, Actions and security alerts straight from the agent
Agent Skills for Google products
Agent Skills for Google products and technologies
Official Google skill collection for working with Google Cloud, BigQuery, GKE, ads and analytics from an agent
AWS's official MCP server suite: docs, IaC, containers, serverless, databases, cost and monitoring