# PostgreSQL community introduces radar and pg_statviz for agentless observability

> **Key takeaway:** New open-source tools radar and pg_statviz provide agentless diagnostic snapshots and in-database time-series analysis for PostgreSQL.

**Published:** 2026-11-05T19:57:00+00:00 
**Source:** Planet PostgreSQL 
**Category:** databases-storage 
**Canonical URL:** https://fosswire.org/news/postgresql-community-introduces-radar-and-pgstatviz-for-agen.html 

## Summary
During the PGConf.EU 2026 conference in Valencia, new open-source diagnostic and observability tools for PostgreSQL were presented to the community. The featured tools, radar and radar-analyst, aim to solve the problem of post-incident support by providing agentless snapshots of database metadata and host configurations.

Additionally, the presentation of pg_statviz introduces a method for performing in-database time-series analysis using existing cumulative statistics. These developments, presented alongside discussions on community onboarding and observability, represent a push toward more accessible and deterministic database monitoring and troubleshooting workflows.

## Technical breakdown of new PostgreSQL observability tools
- **radar (Diagnostic Tool):** A single static Go binary designed to collect a snapshot of PostgreSQL metadata, including catalogs, statistics, and configurations, into a compressed archive without accessing row data or passwords.
- **radar-analyst (Analysis Engine):** An automated tool that processes radar snapshots using deterministic rules to provide verdicts and fixes, with an optional LLM layer used only for summarizing findings rather than generating them.
- **pg_statviz (Time-Series Analysis):** A tool that captures PostgreSQL's internal cumulative statistics and stores them in local database tables, enabling time-series querying and charting without requiring external exporters or agents.
- **AI Integration:** Both radar-analyst and pg_statviz utilize vision-capable or text-based LLMs to interpret charts or write briefs, though the core logic remains grounded in deterministic rule-based findings.

## Why it matters
These tools reduce the dependency on complex monitoring agents and external exporters for database troubleshooting. By providing agentless, metadata-only snapshots and in-database time-series capabilities, they offer sysadmins and DBAs a more lightweight and secure way to perform diagnostics and observability.