PASS Summit 2027: North America and Europe|Register now

Connecting Redgate Monitor to AI assistants with MCP

MCP (Model Context Protocol) is the open standard that lets AI assistants like Claude, GitHub Copilot and Cursor connect to external tools and data. For database monitoring, that means a DBA or engineer can ask their AI assistant about the state of the estate, such as recent alerts, server health, top queries and blocking, and get answers drawn from live monitoring data.

Guest post

This is a guest post from Carly Meichen.

Redgate Monitor, from Redgate Software, includes a read-only MCP server that connects your self-hosted Redgate Monitor to the AI assistant your team already uses to your Redgate Monitor data, including alerts, server health, top queries, wait statistics and blocking. It's generally available from Redgate Monitor v14.36, and it works alongside Redgate Assistant, the AI assistant built into Redgate Monitor, which is also generally available.

What MCP is and why it matters for database teams

AI assistants are good at reasoning and writing, but out of the box they know nothing about your databases. Ask one why a server slowed down overnight and, without access to your monitoring data, the best it can do is offer generic advice.

MCP fixes that. It defines how an AI assistant, acting as an MCP client, connects to an MCP server that exposes tools and data. The assistant decides when a tool would help answer your question, calls it, and uses the result in its reply. Because the protocol is open and standard, a vendor can build one MCP server and any MCP-compatible assistant can use it, without a custom integration for each one.

For database teams, that turns monitoring data into something you can ask about in plain language, from wherever you already work, instead of knowing which screen to open or piecing the answer together yourself.

What to look for in a database monitoring tool with MCP support

Not every MCP integration is equal. When you evaluate one, ask:

  • Is the MCP server generally available or in preview, and on which deployment options?
  • What data does it expose, such as alerts, server health, query performance or blocking, and on which database platforms?
  • Is it read-only, or can an assistant change things?
  • Where does your data go, and how is access authenticated, scoped, rate limited and logged?
  • Which assistants has the vendor actually tested?

Why connect Redgate Monitor to your own AI assistant

Get monitoring data to the people who need it

Not everyone who needs answers from Redgate Monitor works in it every day. A developer checking whether their release slowed a query, an engineer on call, or a team lead who wants a quick status summary can ask from the assistant they already use. Each access token is scoped to the servers that person should see. MCP sits alongside Redgate Monitor rather than replacing it: deep investigations still happen in the UI.

Reason across your whole toolchain

An AI assistant can connect to more than one MCP server at once. Pair Redgate Monitor with the MCP servers for tools like Jira or Octopus Deploy, and one conversation can link a performance problem to the deployment that preceded it, or pull monitoring evidence into the ticket you're already working on. Your assistant reasons across the whole picture rather than one tool at a time.

Keep AI on your own security terms

Many teams can only use an AI service their company has approved, or need query data to stay inside their own security boundary. The MCP server has no AI model of its own. With the MCP server, your Redgate Monitor data goes to the assistant and AI model your team chooses, not to a Redgate-hosted AI service. The server itself is built for control:

  • Read-only. The tools can only read data. They can't change anything in Redgate Monitor or on your databases.
  • HTTPS only. Connections are encrypted, and plain HTTP is rejected.
  • A dedicated, scoped token. Access uses a least-privilege token with an MCP role, scoped to the servers you choose and revocable at any time.
  • Rate limited. Each token is limited to 60 tool calls per minute by default, and you can change that.
  • Logged. Every request to the MCP server is logged.

What you can do with the Redgate Monitor MCP server

The Redgate Monitor MCP server is generally available for self-hosted Redgate Monitor 14.36 and later, on a Redgate Monitor Standard or Redgate Monitor Enterprise license. It isn't available on legacy perpetual licenses or on Redgate Monitor SaaS.

Once it's connected, you can ask your assistant questions like:

  • "Summarize the state of my monitored estate and flag anything that isn't being monitored normally."
  • "Show me the most resource-intensive queries on this server over the last two hours, and what makes each one expensive."
  • "What were the top waits on this SQL Server instance overnight, and which queries contributed to them?"
  • "Pull the query plan for this query and explain why it's slow."
  • "Is anything blocking on this instance right now, and what's at the head of the chain?"
  • "Which alerts were raised in the last 24 hours, and what are the details of the critical ones?"

Server lists, metrics, top queries, query details and alerts work across SQL Server, PostgreSQL, Oracle, MySQL and MongoDB. Query plans cover SQL Server and PostgreSQL. Wait statistics and blocking analysis cover SQL Server.

Which assistants work with it

Redgate has tested the MCP server with Claude Code, GitHub Copilot in VS Code, Cursor, Gemini CLI and AWS Kiro. Any MCP client that can connect to a remote MCP server over HTTPS with a bearer token should work. It's been tested with Claude models and is expected to work with equivalent models from other providers.

The wider AI capability set in Redgate Monitor

The MCP server is one part of Monitor AI, the set of AI features built into Redgate Monitor. Each one brings AI to the monitoring data you already have, in a different way.

  • Redgate Assistant is the AI assistant built into Redgate Monitor, generally available from v14.40. You ask a question about your estate inside Redgate Monitor and get a plain-language answer grounded in your own metrics, queries and waits. It also powers query analysis and alert analysis in Redgate Monitor. It runs on a Redgate-hosted AI service.
  • The MCP server is for teams who want to use their own assistant instead, because their security policy requires an approved model, or because they want to reason across Redgate Monitor and other tools in one place.
    You can use both. They draw on the same monitoring data.
  • Dynamic Alert Thresholds use machine learning to learn what normal looks like for each server and adjust alert thresholds to match, so routine spikes don’t bury the alerts that matter. It runs locally on your Base Monitor, so the data it learns from never leaves your environment.

How database monitoring tools compare on MCP support

MCP support has spread quickly across monitoring tools during 2026. Here's where the main options stand as of November 2026, based on each vendor's public documentation and announcements. In every case, MCP access comes with the vendor's own product, so its cost depends on how that product is licensed.

Tool MCP status Where it runs What you need Worth knowing
Redgate Monitor Generally available (14.36 and later) Self-hosted only A Redgate Monitor Standard or Redgate Monitor Enterprise license Database-specific tools for waits, blocking and query plans. Read-only, with scoped, revocable tokens
Datadog Generally available since March 2026 Cloud-hosted A Datadog account Covers the wider Datadog platform, with an optional Database Monitoring toolset. Supports read and write permissions
Dynatrace Remote MCP server generally available Cloud-hosted, with a separate server for Dynatrace Managed A Dynatrace environment Covers the wider Dynatrace observability platform
SolarWinds Available in SolarWinds Observability Self-Hosted 2026.4 and later Self-hosted, online environments only SolarWinds Observability Self-Hosted, with the SolarWinds AI assistant and Platform Connect enabled Authenticates with a SolarWinds username and password
Quest Foglight Available in Foglight Cloud, with on-premises support announced Cloud-hosted (on-premises announced) A Foglight Cloud account Read-only access to monitoring data
Quest Spotlight No MCP server documented at the time of writing Not applicable Not applicable Not applicable
IDERA SQL Diagnostic Manager No MCP server documented at the time of writing Not applicable Not applicable Not applicable

last updated - November 2026

The differences come down to three things. Platforms like Datadog and Dynatrace connect your assistant to a broad observability estate, which suits teams that have standardized on them. Redgate Monitor's MCP server is built specifically for databases, so your assistant can ask about waits, blocking and query plans directly. And access models vary, from dedicated read-only tokens to a user's own username and password.

Frequently asked questions

Does Redgate Monitor have an MCP server?

Yes. Redgate Monitor has a generally available, read-only MCP server (v14.36) that connects self-hosted Redgate Monitor to MCP-compatible AI assistants. It gives your assistant access to monitored servers, metrics, alerts, top queries, query plans, wait statistics and blocking.

Which AI assistants can connect to Redgate Monitor?

Redgate has tested the MCP server with Claude Code, GitHub Copilot in VS Code, Cursor, Gemini CLI and AWS Kiro. Any MCP client that can connect to a remote MCP server over HTTPS with a bearer token should work.

Is the Redgate Monitor MCP server available now?

Yes. It's generally available for self-hosted Redgate Monitor v14.36, on a Redgate Monitor Standard or Redgate Monitor Enterprise license. It isn't available on Redgate Monitor SaaS.

Which databases does the Redgate Monitor MCP server support?

Server lists, metrics, top queries, query details and alerts work across SQL Server, PostgreSQL, Oracle, MySQL and MongoDB. Query plans cover SQL Server and PostgreSQL. Wait statistics and blocking analysis cover SQL Server.

Is it safe to connect an AI assistant to Redgate Monitor?

The MCP server is read-only, so an assistant can't change anything in Redgate Monitor or on your databases. Your data goes to the assistant and model your team chooses, not to a Redgate-hosted AI service. Connections are HTTPS only, access uses a dedicated token scoped to the servers you choose, each token is rate limited, and every request is logged.

Do I need Redgate Assistant to use the MCP server?

No. The MCP server works with your own AI assistant. Redgate Assistant is the AI assistant built into Redgate Monitor. You can use either, or both.

 

Next steps

Explore AI in Redgate Monitor · Read the MCP setup documentation · See the Redgate Monitor roadmap · Start a free Redgate Monitor trial

This document contains proprietary information and is protected by copyright law.
Copyright © 2026 Red Gate Software Limited. All rights reserved

Tools in this post

Redgate Monitor

Real-time multi-platform performance monitoring, with alerts and diagnostics

Find out more

Loading comments...