Add pg-status: PostgreSQL instance monitoring utility

[gpas45]

Summary

Add pg-status, a read-only Go utility for monitoring PostgreSQL and Postgres Pro instances on a host. The tool discovers instances via systemd, reports their status, and collects detailed statistics from running servers.

Key Changes

• New utility: pgsql/pg-status/main.go — complete implementation with ~574 lines

  • Instance discovery via systemd units (supports both Postgres Pro 1C and standard PostgreSQL naming schemes)
  • Status reporting: systemd state, port, data directory, disk usage
  • Server statistics collection via psql: version, uptime, connections, database sizes, cache hit ratio, transaction counts
  • Color-coded output with automatic terminal detection and NO_COLOR support
  • Direct port querying mode (-p flag) for non-systemd scenarios
  • Privilege handling: works from root (via runuser/sudo) or postgres user for peer authentication

• Documentation: pgsql/pg-status/README.md — usage guide with examples and dependencies

• Tests: pgsql/pg-status/main_test.go — unit tests for utility functions (human, duration, engineLabel, unitBin)

• Module file: pgsql/pg-status/go.mod — Go module definition (no external dependencies, uses only stdlib)

• Index update: pgsql/README.md — added reference to new utility

Notable Implementation Details

• Instance discovery: Combines multiple sources (systemd list-units, .wants symlinks, drop-in directories) with deduplication
• Port detection: Multi-fallback strategy (active socket via ss, postgresql.conf, PGPORT environment variable)
• Data directory discovery: Reads from process cmdline, systemd Environment, and EnvironmentFiles
• Formatting utilities: Human-readable byte sizes (B/KiB/MiB/GiB/TiB/PiB) and duration formatting (days/hours/minutes)
• SQL queries: Efficient single-query approach for scalar statistics and separate query for database listing
• Privilege model: Automatic detection of execution context with appropriate command prefixing for privilege escalation

https://claude.ai/code/session_01KVXrszT51udmuntdfAHMVo