Skip to main content

Postgres copilot

Postgres copilot is the current milestone (2025) on our journey to Self-Driving Postgres. We're not fully autonomous yet — the system watches, diagnoses, and prepares pull requests for performance improvements, security patches, and zero-downtime upgrades. You review, approve, and merge. Human in the loop on every change.

Each recommendation is tested and verified on a clone of your real data, backed by expert validation.

How it works​

╔════════════╗        ╔═ Postgres copilot ════════════════════════╗
║    Your    ║░       ║  ┏━━━━━━━━━━━━━━┓     ┏━━━━━━━━━━━━━━┓   ║░
║  Postgres  ║░──────▶║  ┃  Monitoring  ┃────▶┃ Health check ┃   ║░
║  database  ║░       ║  ┗━━━━━━━━━━━━━━┛     ┃   & Issues   ┃   ║░
╚════════════╝░       ║                       ┗━━━━━━━━━━━━━━┛   ║░
 ░░░░░░░░░░░░░░       ╚══════════════════════════════════════════╝░
      ▲                ░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░
      │                                               │
      │                                               ▼
      │     ╔═════════════════════════╗     ╔═════════════════════╗
      └─────║ GitHub PRs / GitLab MRs ║░◀───║ AI tool (Cursor, …) ║░
            ╚═════════════════════════╝░    ╚═════════════════════╝░
             ░░░░░░░░░░░░░░░░░░░░░░░░░░░     ░░░░░░░░░░░░░░░░░░░░░░░
  1. postgres_ai monitoring runs in your infrastructure, collecting metrics 24/7
  2. Health checks automatically detect issues and generate actionable recommendations
  3. Issues are created with specific fixes (e.g., CREATE INDEX ...)
  4. AI tools (Cursor, GitHub Copilot, etc.) can turn Issues into pull requests
  5. You review and merge — human approval on every change

Components​

postgres_ai monitoring​

Open-source (Apache 2.0), enterprise-grade Postgres observability. Designed for senior DBAs, SREs, and AI systems who need rapid root cause analysis.

  • Emergency dashboard — "Shallow but wide" analysis to identify problems in 1 minute
  • Query analysis — Comprehensive metrics for performance patterns and bottlenecks
  • Wait event analysis — Similar to AWS RDS Performance Insights
  • Health check reports — Actionable recommendations based on real-world experience

Install postgres_ai monitoring →

Issues​

When Postgres copilot detects a problem, it creates an Issue with:

  • Clear description of the problem
  • Severity and impact assessment
  • Specific fix recommendation (tested on your data)
  • Supporting evidence and metrics

Issues can be worked on by AI coding assistants to generate pull requests automatically.

Learn about working with Issues →

Expert validation​

Every recommendation is backed by human Postgres expertise. We don't just throw metrics at an LLM — our AI is built on methodologies from 20 years of production incidents.

What you get​

  • 24/7 monitoring that catches issues before they page you
  • Actual solutions, not just alerts:

    CREATE INDEX idx_orders_customer_id ON orders(customer_id); — tested on a clone, validated by experts, ready to merge

  • Monthly health checks — deep-dive analysis from senior DBAs
  • Direct Slack access to people who've seen everything

Pricing​

$500/month per cluster.

Monitoring is free, forever, open source. Postgres copilot adds the AI layer, expert validation, monthly health checks, and Slack access.

Get started​

  1. Install postgres_ai monitoring
  2. Set up Postgres copilot
  3. Start receiving Issues and fixes

Live demo → (login: demo / password: demo)