Skip to content
Nov 3, 2026 - Version 2026.10 has been released
For the complete documentation index, see llms.txt. This page is also available as Markdown.

Diagnose a live performance spike

Run a real-time analysis to find what is slowing an instance down right now.

When an instance slows down in the moment, ask for a real-time (Live360) analysis to find what is happening and why.

Real-time analyses rely on the Live360 module, the only part of the platform that connects directly to the PostgreSQL instance instead of reading from the repository. Access to it is a flag set on each user profile, under Users & Roles. Without that flag, the assistant does not offer real-time analyses and redirects you to the closest historical one.

Example prompts:

  • Real-time snapshot:

    • "What's happening right now on prod-eu-2?"

    • "Show me the live sessions on prod-eu-1."

    • "Is a vacuum or index build running on prod-eu-1 right now, and how far along?"

  • Diagnose a slowdown:

    • "prod-eu-1 feels slow, find out what's wrong."

    • "Users get 'too many connections' errors on prod-eu-1: diagnose it."

    • "Is the server hosting prod-eu-1 CPU-bound?"

Security

Opening a live analysis to the assistant never opens your database to it. Every real-time analysis transits through the platform and runs only a few read-only queries via the Datasentinel agent and the Datasentinel collection user. No modification is ever made on the PostgreSQL cluster, and no PostgreSQL credential is ever known to the AI assistant.

This walkthrough was recorded with the ChatGPT Desktop app.

Continue from the menu. A live snapshot ends with a short menu of next steps tied to what it found: untangle the blocking chain, fetch the execution plan of the query hogging the sessions, check whether the host itself is the limit, or switch to the historical analysis of the same window once the spike is over. Pick an option to keep the investigation going without retyping the context.

Last updated