Truenas Aiops MCP server
{
"mcpServers": {
"aiops-tools-truenas-aiops": {
"command": "uvx",
"args": [
"truenas-aiops"
],
"env": {
"TRUENAS_AIOPS_MASTER_PASSWORD": "<truenas_aiops_master_password>"
}
}
}
}{
"mcpServers": {
"aiops-tools-truenas-aiops": {
"command": "uvx",
"args": [
"truenas-aiops"
],
"env": {
"TRUENAS_AIOPS_MASTER_PASSWORD": "<truenas_aiops_master_password>"
}
}
}
}{
"mcpServers": {
"aiops-tools-truenas-aiops": {
"command": "uvx",
"args": [
"truenas-aiops"
],
"env": {
"TRUENAS_AIOPS_MASTER_PASSWORD": "<truenas_aiops_master_password>"
}
}
}
}{
"servers": {
"aiops-tools-truenas-aiops": {
"command": "uvx",
"args": [
"truenas-aiops"
],
"env": {
"TRUENAS_AIOPS_MASTER_PASSWORD": "${input:truenas_aiops_master_password}"
}
}
},
"inputs": [
{
"id": "truenas_aiops_master_password",
"type": "promptString",
"description": "Truenas aiops master password",
"password": true
}
]
}{
"mcpServers": {
"aiops-tools-truenas-aiops": {
"command": "uvx",
"args": [
"truenas-aiops"
],
"env": {
"TRUENAS_AIOPS_MASTER_PASSWORD": "<truenas_aiops_master_password>"
}
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsoninstall uvx truenas-aiopsversion 0.10.0
What it does
TrueNAS AIops exposes 25 MCP tools for managing TrueNAS SCALE storage operations with built-in governance. It supports both REST v2.0 and JSON-RPC/WebSocket APIs, records every operation in an audit log, encrypts credentials, and tracks reversible operations with undo capability.
- Diagnose and triage degraded storage pools by checking pool health, error counters, and capacity, then initiate scrubs or corrective actions.
- Automate routine storage tasks like snapshot creation, dataset management, and service restarts while maintaining an encrypted audit trail of all operations.
- Monitor TrueNAS system health across multiple appliances by querying alerts, disk SMART data, replication status, and system information via a unified API layer.
Tools this server exposes
| Tool | Does |
|---|---|
| overview | [READ] Fetch high-level system status and summary information. |
| system_info | [READ] Get detailed system configuration and specifications. |
| pool_health_rca | [READ] Diagnose pool health with worst-first analysis of degradation, errors, and capacity. |
| alert_and_capacity_rca | [READ] Identify active alerts and datasets approaching full capacity. |
| pool_list | [READ] List all storage pools and their basic properties. |
| pool_get | [READ] Retrieve detailed information for a specific pool. |
| pool_status | [READ] Check pool operational status, topology, and scan details. |
| scrub_status | [READ] View the status of a running or completed pool scrub. |
| pool_capacity | [READ] Query used and available capacity metrics for a pool. |
| scrub_start | [WRITE] Initiate an integrity scrub on a pool (medium risk). |
| dataset_list | [READ] List datasets within a pool. |
| dataset_get | [READ] Get detailed properties of a specific dataset. |
| dataset_create | [WRITE] Create a new dataset with specified properties. |
| snapshot_list | [READ] List snapshots for a dataset or pool. |
| snapshot_create | [WRITE] Create a snapshot with automatic undo tracking (medium risk). |
| snapshot_delete | [WRITE] Delete a snapshot with before-state capture (high risk). |
| disk_list | [READ] List all disks and their basic properties. |
| disk_smart | [READ] Retrieve SMART data and health status for a disk. |
| alert_list | [READ] Fetch all active and historical alerts. |
| service_list | [READ] List configured services and their states. |
| service_restart | [WRITE] Restart a service (high risk, requires double confirmation). |
| replication_list | [READ] List replication tasks and cloud sync configurations. |
| secret_set | [WRITE] Store an encrypted secret in the credential vault. |
| secret_list | [READ] List stored secrets (names only, not values). |
| secret_rm | [WRITE] Remove a stored secret. |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.