Skip to main content

MCP Server Deployment Workflow

This guide provides a step-by-step walkthrough for adding and deploying a new MCP server in NexaStack. Following this workflow ensures proper configuration of resources, deployment type, and monitoring of the deployment pipeline.

Goal​

Learn how to:

  • Add a new MCP server to NexaStack.
  • Configure deployment resources and parameters.
  • Upload MCP packages or connect to a Git repository.
  • Monitor the deployment pipeline and verify successful server onboarding.

Step 1: Login to the Platform​

  1. Open the NexaStack login page.
  2. Login with correct credientials
Security Reminder

Always log in via a secure HTTPS connection and never share your credentials publicly.

Step 2: Navigate to MCP Servers Section​

  1. From the main sidebar, go to MarketPlace.
  2. Click on the MCP Servers tab to view all available MCP servers.
  3. Ensure the Add New MCP Server button is visible on the page.

MCP Servers Menu

Step 3: Start Adding a New MCP Server​

  1. Click Add New MCP Server.
  2. The Deploy New MCP Server form will open. Verify that all sections are displayed correctly.

Deploy New MCP Server Form

Step 4: Upload MCP Server​

You can deploy the MCP server using either method: Upload Zip or Git Repository.

Git Repository Method​

  1. Click Git Repository as the deployment method.
  2. Fill in the following details:
FieldDescription
MCP Server NameName of the MCP server to create
Repository URLGit repository URL containing MCP server code
UsernameGit account username (if private repo)
TokenAccess token or password for repo authentication
BranchSelect a branch (auto-populated)
  1. Ensure the repository follows a mono-architecture structure (all MCP server components in one repo).
  2. The platform automatically detects the MCP server structure to validate deployment.
  3. Click Proceed Next to continue.

Git Repository Form

Repository Validation

Make sure your repository contains all MCP server components. Branch selection is important; the platform will auto-detect the MCP server structure.

Step 5: Upload Zip Method​

  1. Click Upload Zip.
  2. Select the MCP zip file (e.g., mcp-odoo.zip) from your local machine.
  3. Click Upload File to send it to the server.
  4. Wait for the upload to complete and verify the file appears in the list.
  5. Click Proceed Next to continue.

Upload MCP Zip

File Validation

Ensure the zip file contains all necessary configuration and application files. Incomplete packages may cause deployment failures.

Step 6: Enter Resource Requirements​

  1. The Resource Requirements page will appear.
  2. Enter CPU and Memory Requests (minimum required resources):
ResourceExample ValueUnit
CPU10millicores (m) or cores
Memory128Mi or Gi
  1. Enter CPU and Memory Limits (maximum allowed resources):
ResourceExample ValueUnit
CPU150millicores (m) or cores
Memory512Mi or Gi
  1. Verify all input fields are filled correctly.

Resource Requirements

Resource Planning

Set CPU and memory limits based on expected server load to avoid throttling or resource exhaustion.

Step 7: Select Cluster​

  1. Choose a Cluster (or another available option based on your infrastructure).
  2. This determines how and where the MCP server will be provisioned and managed.

Deployment Type Selection

No Cluster Available?

If you don’t have a cluster yet, review your preferences and onboard a cluster accordingly.
Follow this step-by-step guide to set up your cluster:
👉 Onboard a Cluster

Step 8: Start Deployment Pipeline​

  1. Click Start Pipeline to initiate the MCP server deployment.
  2. The platform will launch the deployment workflow.
  3. Monitor the status of the deployment:
  • Pending: Server is waiting to start.
  • Running: Deployment in progress. s
  • Succeeded: Deployment completed successfully.
  • Failed: Deployment encountered errors (check logs for details).

MCP Deployment Pipeline

Monitoring Tip

Use the console logs and pipeline dashboard to verify each step of the deployment. This helps identify issues early if the deployment fails.

Best Practices​

  • Validate Zip Files: Always check your MCP zip file before upload.
  • Resource Planning: Allocate CPU and memory carefully to prevent under/over-provisioning.
  • Select Correct Deployment Type: Ensure you choose the type that matches your infrastructure.
  • Monitor Logs: Watch the deployment logs to catch errors immediately.
  • Testing: After deployment, test the MCP server functionality before production use.

Next Steps​

  • Access the deployed MCP server via the server URL provided in the deployment dashboard.
  • Deploy AI agents or services on the MCP server.
  • Monitor server performance and resource usage in NexaStack dashboards.

MCP Deployment Completed

MCP Server Deployed

You have successfully added and deployed a new MCP server in NexaStack.
The server is now available for model deployments and agent integration.