# mcp-centreon

MCP server for Centreon IT monitoring — hosts, services, alerts, acknowledgements, and downtimes.

Integrates Claude (or any MCP-compatible AI) with [Centreon](https://www.centreon.com/) to query monitoring status, manage acknowledgements and downtimes, and retrieve performance metrics.

## Features

- **Hosts** — list, search, and inspect monitored hosts; acknowledge problems; schedule downtimes
- **Services** — list and filter services by state or host; acknowledge and schedule downtimes
- **Alerts** — browse recent events, active acknowledgements, and scheduled/active downtimes
- **Performance** — list available metrics per service and fetch time-series data
- **Platform status** — cluster-level host/service state counts; poller topology

## API version support

Targets Centreon REST API v2 (`/centreon/api/latest/`), supported in **Centreon 21.04 and later** (including 21.10, 22.x, 23.x, 24.x).

The authentication endpoint used is:

```
POST https://{host}/centreon/api/latest/authentication/providers/configurations/local
```

If you use an older Centreon installation that only exposes the legacy v1 API, the monitoring endpoints will not be available. Upgrade to Centreon 21.04+ to use this server.

## Required Centreon user permissions

A **read-only** user is sufficient for all monitoring and read operations:

| Permission | Reason |
|---|---|
| `View all objects` (or access to relevant host groups) | Needed for listing hosts and services |
| `Reach API configuration` | Required for host/service groups endpoints |
| `Reach API realtime` | Required for monitoring, events, metrics endpoints |

To use acknowledgement and downtime tools the user also needs:

| Permission | Reason |
|---|---|
| `Acknowledge resources` | `acknowledge_host`, `acknowledge_service` |
| `Schedule downtimes` | `add_downtime_host`, `add_downtime_service` |

Create the API user in **Administration > ACL > ACL Access Groups** and assign the relevant ACL menus and actions.

## Setup

### 1. Install

```bash
npm install -g mcp-centreon
```

Or run directly without installing:

```bash
npx mcp-centreon
```

Or clone and build:

```bash
git clone https://github.com/fredriksknese/mcp-centreon
cd mcp-centreon
npm install
npm run build
```

### 2. Environment variables

| Variable | Required | Default | Description |
|---|---|---|---|
| `CENTREON_HOST` | Yes | — | Centreon server hostname or IP (without `https://`) |
| `CENTREON_USERNAME` | No | `admin` | Centreon login username |
| `CENTREON_PASSWORD` | Yes | — | Centreon login password |
| `CENTREON_ALLOW_SELF_SIGNED` | No | `true` | Accept self-signed TLS certificates (`true`/`false`) |

### 3. Claude Desktop configuration

Add to your `claude_desktop_config.json`:

```json
{
  "mcpServers": {
    "centreon": {
      "command": "npx",
      "args": ["mcp-centreon"],
      "env": {
        "CENTREON_HOST": "centreon.example.com",
        "CENTREON_USERNAME": "admin",
        "CENTREON_PASSWORD": "your-password-here",
        "CENTREON_ALLOW_SELF_SIGNED": "true"
      }
    }
  }
}
```

If you installed globally via `npm install -g`:

```json
{
  "mcpServers": {
    "centreon": {
      "command": "mcp-centreon",
      "env": {
        "CENTREON_HOST": "centreon.example.com",
        "CENTREON_USERNAME": "admin",
        "CENTREON_PASSWORD": "your-password-here"
      }
    }
  }
}
```

## Available tools

### Hosts

| Tool | Description |
|---|---|
| `list_hosts` | List monitored hosts; filter by name or state (up/down/unreachable/pending) |
| `get_host` | Get full details for a host by ID |
| `list_host_groups` | List host groups with member counts |
| `acknowledge_host` | Acknowledge a host problem |
| `add_downtime_host` | Schedule a maintenance downtime for a host |

### Services

| Tool | Description |
|---|---|
| `list_services` | List services; filter by host name, service name, or state |
| `get_service` | Get full details for a service by host ID and service ID |
| `list_service_groups` | List service groups |
| `acknowledge_service` | Acknowledge a service problem |
| `add_downtime_service` | Schedule a maintenance downtime for a service |

### Alerts

| Tool | Description |
|---|---|
| `list_events` | List recent monitoring events; filter by host, state, or check type (soft/hard) |
| `list_acknowledgements` | List active acknowledgements — who acknowledged what and when |
| `list_downtimes` | List active and scheduled downtimes |

### Performance

| Tool | Description |
|---|---|
| `get_service_metrics` | List available performance metrics for a service |
| `get_metrics_data` | Get time-series data for a specific metric |

### Platform status

| Tool | Description |
|---|---|
| `get_platform_status` | Platform-level host/service counts broken down by state |
| `get_poller_list` | List pollers and remote servers with connectivity status |

## State mappings

| Code | Host state | Service state |
|---|---|---|
| 0 | Up | Ok |
| 1 | Down | Warning |
| 2 | Unreachable | Critical |
| 3 | — | Unknown |
| 4 | Pending | Pending |

## License

SEE LICENSE IN LICENSE