MCP N8N Chat Workflow

This lab is a continuation of the MCP lab with N8N server as an interfacing tool.

High-Level Overview

Install n8n AI workflow automation tool
Configure n8n MCP Client to connect to Nutanix MCP Server and chat (ask questions)

Install N8N

N8N is a a fair-code licensed workflow automation tool that combines AI capabilities with business process automation.

We will install and self-host N8N on the NKP Server and use it to automate a chat AI Agent node using the Nutanix MCP Server which we deployed in the previous section.

Log in to your Linux Tools VM (e.g., via VSCode as ubuntu). and open Terminal

Download N8N Manifests to deploy N8N in NKP cluster

mkdir $HOME/n8n
cd $HOME/n8n
git clone https://github.com/n8n-io/n8n-hosting.git
cd n8n-hosting/

Add the following n8n community node for use as tools with our AI Agent Node. We will do this by modifying the manifest file before deploying n8n.
```
cd n8n-hosting/kubernetes/
code n8n-deployment.yaml  # (1)  
```
1. Use vim as an alternate editor

Add the following highlighted lines to the n8n-deployment.yaml file under the env section

   env:
     < Snipped for brevity >
     - name: N8N_PROTOCOL
       value: http
     - name: N8N_PORT
       value: "5678"
     - name: NODE_TLS_REJECT_UNAUTHORIZED
       value: "false"
     - name: N8N_COMMUNITY_PACKAGES_ALLOW_TOOL_USAGE
       value: "true"

Deploy n8n

CommandCommand output

export KUBECONFIG=$HOME/_nkp_install_dir/_nkpclustername.conf
kubectl apply -f .

export KUBECONFIG=$HOME/nkp/nkpdev.conf
#
httproute.gateway.networking.k8s.io/n8n-chatapp-route created
persistentvolumeclaim/n8n-claim0 created
deployment.apps/n8n created
service/n8n created
namespace/n8n created
persistentvolumeclaim/postgresql-pv created
configmap/init-data created
deployment.apps/postgres created
secret/postgres-secret created
service/postgres-service created

Check n8n services

CommandCommand output

kubens n8n 
kubectl get all

Active namespace is "n8n".
#
NAME                            READY   STATUS    RESTARTS        AGE
pod/n8n-cf6b5b8db-ql5db         1/1     Running   0               2m
pod/postgres-57dbbcbb58-mpnq8   1/1     Running   1 (2m ago)      2m

NAME                       TYPE           CLUSTER-IP    EXTERNAL-IP   PORT(S)          AGE
service/n8n                LoadBalancer   10.99.29.36   <pending>     5678:32488/TCP   2m
service/postgres-service   ClusterIP      None          <none>        5432/TCP         2m

NAME                       READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/n8n        1/1     1            1           2m
deployment.apps/postgres   1/1     1            1           2m

NAME                                  DESIRED   CURRENT   READY   AGE
replicaset.apps/n8n-bc8878667         0         0         0       2m
replicaset.apps/n8n-cf6b5b8db         1         1         1       2m
replicaset.apps/postgres-57dbbcbb58   1         1         1       2m

If Envoy Gateway is installed as a part of NAI install use the following to create a HTTPRoute resource to access n8n

kubectl apply -f -<<EOF
apiVersion: gateway.networking.k8s.io/v1
kind: HTTPRoute
metadata:
  name: n8n-chatapp-route
  namespace: n8n                    # Same namespace as your chat app service
spec:
  parentRefs:
  - name: nai-ingress-gateway
    namespace: nai-system           # Namespace of the Gateway
  hostnames:
  - "n8n.nai.10.x.x.216.nip.io"     # Input Gateway IP address
  rules:
  - matches:
    - path:
        type: PathPrefix
        value: /
    backendRefs:
    - name: n8n
      kind: Service
      port: 5678
EOF

Now n8n will be available at the following URL n8n.nai.10.x.x.216.nip.io
Register for n8n with an email
An activation key will be sent to email address

Setup N8N Workflow with Nutanix MCP Server

This workflow is a simple example of how to use the MCP server for Nutanix in an N8N workflow. It uses the AI Agent node to interact with the MCP server via the MCP Client node. In the following workflow, we'll be using the n8n-nodes-mcp community node to interact with the Nutanix MCP server.

Add Community Node n8n-node-mcp to n8n instance under Settings

Add AI Agent node to workflow and Configure OpenAI Chat Model, Simple Memory and other settings as needed.

For the prompt, use the following:

You are an AI agent that can interact with a Nutanix cluster using the MCP server.
You can list VMs, get VM details, and perform actions like start, stop, and reboot VMs.
Use the MCP Client node to execute commands on the Nutanix cluster.

Add the NAI LLM Model deployed here and set up with credentials
- API Key: Use the one generated in NAI
- Base URL: https://nai.10.x.x.216.nip.io/api/v1
Choose Simple Memory for storing in-flight data
Search and add MCP Client Tool (via npm) tool to workflow and name it Nutanix MCP List
For Credentials - create New Credential and Configure MCP Client node to use the Nutanix MCP server SSE endpoint

SSE URL:
```
https://mcp-nutanix.${INGRESS_IP}.nip.io/sse#mcp-nutanix
```
Additional Headers: Copy and paste the following (with actual values for Nutanix Prism connectivity)
```
NUTANIX_ENDPOINT="$NUTANIX_ENDPOINT",NUTANIX_USERNAME="$NUTANIX_USERNAME",NUTANIX_PASSWORD="$NUTANIX_PASSWORD",NUTANIX_INSECURE="true"
```
Run Execute Step to test connection and view available commands
Add another MCP Client (via npm) tool node to workflow and name it something meaningful like Nutanix MCP Execute
For credential use the Nutanix MCP server SSE endpoint (previously configured)
For Tool Name, select Expression tab and provide following
```
{{ $fromAI("tool","the tool selected") }}
```
Click on Execute Step
Type vm_list in the input text box and click on Execute Step
Run the workflow to verify configuration
In the AI Chat window, ask questions about the Nutanix cluster