maintenant MCP server

<!-- TODO: replace the hero with a 10-15 s demo.gif: dashboard live, an endpoint failing, the alert landing in Discord. -->

maintenant

Drop a container. Your stack is monitored. When something breaks, fixed rules catch it, the alert is pushed to you, and your AI agent investigates through the built-in MCP server. Docker, Kubernetes, uptime, TLS, cron jobs, live logs, image updates, CVEs: auto-discovered, alerting on every one of them, from a single Go binary that fits in a few tens of MB of RAM when freshly started (about 100 MB after several days on a busy host). No PromQL, no exporters, no dashboards to build.

Try the live demo → A Pro instance monitoring a real stack, read-only. Nothing to install, nothing to sign up for.

Live demo&nbsp;&nbsp;&bull;&nbsp;&nbsp;Quick Start&nbsp;&nbsp;&bull;&nbsp;&nbsp;Why maintenant&nbsp;&nbsp;&bull;&nbsp;&nbsp;Agents&nbsp;&nbsp;&bull;&nbsp;&nbsp;Features&nbsp;&nbsp;&bull;&nbsp;&nbsp;Documentation&nbsp;&nbsp;&bull;&nbsp;&nbsp;Editions&nbsp;&nbsp;&bull;&nbsp;&nbsp;Pricing

Quick Start

# docker-compose.yml
services:
  maintenant:
    image: ghcr.io/kolapsis/maintenant:latest
    ports:
      # ⚠️  SECURITY: publishes the UI/API (no authentication of their own) on
      # every interface; put an auth reverse proxy in front, or bind "127.0.0.1:8080:8080". See https://docs.maintenant.dev/security/#reverse-proxy-setup.
      - "8080:8080"
    read_only: true
    security_opt:
      - no-new-privileges:true
    tmpfs:
      - /tmp:noexec,nosuid,size=64m
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock:ro
      - /proc:/host/proc:ro
      - /etc/os-release:/host/etc/os-release:ro
      - maintenant-data:/data
    environment:
      MAINTENANT_ADDR: "0.0.0.0:8080"
      MAINTENANT_DB: "/data/maintenant.db"
    restart: unless-stopped

volumes:
  maintenant-data:

docker compose up -d

Open http://localhost:8080. Your containers are already there, with their health, restart loops, resources and logs. Nothing to configure.

Docker socket access is automatic: the entrypoint reads the mounted socket's group and grants it to the unprivileged user, on Compose and on Swarm (where docker stack deploy silently ignores groupadd). A socket owned by group root needs DOCKERGID=0. If containers do not show up, see Troubleshooting.

Kubernetes

kubectl create namespace maintenant
kubectl apply -f deploy/kubernetes/

Shortened. The full README is on GitHub.