I have been experimenting with MCP servers in a few ways, including the Redgate Monitor MCP Server. It took me a few tries, and some help from engineers to get connected. I’ll cover how I did this in Claude and Visual Studio Code in this post, and try to warn you about a few places I made mistakes.
This is part of a series of posts on Redgate Monitor. Click to see the other posts.
Creating A Token
I’m not going to bore you with the setup and configuration of the Redgate Monitor instance. That procedure might change by release, so I’ll assume you’ve got the MCP server enabled.
To access the MCP server, you need to give your AI agent a token. This authorizes the MCP server to connect to Redgate Monitor and read data. These are read only tokens, and designed for the MCP and the tools that are available.
In the configuration, there is an Access Tokens area. Click here.
Click Create Token on the next screen. This is in the upper right.
On the next screen, give your token a name, description, and set an expiration date. The default is one year.
Make sure you select the MCP type. I messed this up the first time.
Below this, pick the servers the MCP token can access. I’ve selected just the Production serer group, but you can set this as needed. I’d make sure the name and/or description lets you know what can be accessed.
You will get a review screen, to verify what you’ve done. Accept that and you’ll get a dialog with the token value. Make sure you save this securely somewhere.
Once you’ve created this, it’s in a list, and you can click the “view details, which shows you what the token can access, as you can see below. You cannot get the token itself anymore.
Configuring VS Code
I’m going to show this working in VS Code. I opened the Command Palette and selected the MCP: Open User Configuration, as shown.
Inside the config, I pasted this inside the “servers” key. This should be a sub element, and this goes alongside any other MCP servers.:
"redgate-monitor": {
"type": "http",
"url": "https://<your-monitor-host>/mcp",
"headers": {
"Authorization": "Bearer ${input:redgate-mcp-token}"
}
}Then inside the “inputs” key, add this:
{
"id": "redgate-mcp-token",
"type": "promptString",
"description": "Redgate Monitor MCP token",
"password": true
}The save this. When you start the MCP server, it will ask you for your token. Paste the value in and then you should connect. You do this from the MCP: List Servers item in the command palette. This should give you output like this, with a connection message and tools discovered:
If I open a chat, and add the MCP as a tool, I can ask questions. Here’s a question I asked and the initial responses from the LLM.
Once this is done, I’ll get a list of the alerts. I did approve some PoSh commands where the LLM stored results from Redgate monitor and then tried to process those after writing them to a temp file.
This is the raw LLM trying to use tools and fumbling around. With a few skills and guidance, this will run smoother, quicker, and more efficiently in your environment.
Summary
The Redgate Monitor MCP server is a great way to start using AI to analyze all the data in your database estate that Redgate Monitor collects. There were 7 tools when I started, but within a few weeks there were 11. By the time you read this, there are likely going to be even more.
Give Redgate Monitor a try and learn how to use MCP servers to link your LLM to Redgate Monitor and start analyzing systems in a more natural manner.
Redgate Monitor is a world-class monitoring solution for your database estate. Download a trial today and see how it can help you manage your estate more efficiently.