PostgreSQL AIops MCP server

This is a helper for people who look after a PostgreSQL database. It lets your AI look at what the database is doing, explain why it is slow, and, if you allow it, make careful changes. It is handy for database administrators and anyone who has to keep a Postgres server healthy.

What this MCP server does

You ask your AI a question about your PostgreSQL database. The AI sends that request to this helper program. The helper connects to your database, reads the system information, and works out the answer. Then it sends back a clear explanation or a list, and if you asked for a change, it makes the change and records how to undo it. Everything it does is written to an audit log.

Click to zoom

What you can do with it

• Find the slowest queries and explain why they are slow
• Check which tables are bloated and need vacuuming
• See which queries are blocking each other and who started it
• List unused or missing indexes
• Check replication lag and WAL status
• Create or drop an index with a preview and an undo
• Terminate a runaway query or backend

Try asking your AI

• “Why is my database slow right now?”
• “Show me the top 5 queries by total time and explain the worst one”
• “Which tables have the most dead rows and should be vacuumed?”
• “Is there a lock chain blocking anything? Who is the root blocker?”

What it gives back to you

You get back plain answers in the chat: a ranked list of slow queries with the reason each one is slow, a table of bloated tables with a recommendation, or a lock chain showing who is blocking whom. For changes, you get a preview first, then a confirmation that the change was made, plus a note that it can be undone. The audit log keeps a record of every call.

Before you start

Who it's for

Database administrators, backend engineers, and anyone responsible for keeping a PostgreSQL server fast and healthy.