Skip to content

Changelog

All notable changes to pgEdge Load Generator will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[1.0.0-beta1] - 2026-01-05

Added

  • Automatic size maintenance for the wholesale application: old orders are periodically deleted to maintain the database at the target size specified during initialisation. This prevents unbounded growth during long-running tests and simulates real-world data archival practices.
  • New --no-maintain-size flag for the run command to disable automatic cleanup if unbounded growth is desired.
  • SizeMaintainer interface allowing other applications to implement size maintenance in future.

[1.0.0-alpha2] - 2025-12-12

Changed

  • Set application_name connection parameter for each worker connection, making it easier to identify load generator connections in pg_stat_activity (e.g., "pgedge-loadgen - client 1")
  • Removed unnecessary long-lived connection pool from the run command; workers now create dedicated connections only

Fixed

  • Documentation links in README

[1.0.0-alpha1] - 2025-12-11

Added

  • Initial release of pgedge-loadgen

  • Seven fictional applications for load testing:

    • wholesale - TPC-C based OLTP workload
    • analytics - TPC-H based OLAP workload
    • brokerage - TPC-E based mixed workload
    • retail - TPC-DS based decision support
    • ecommerce - pgvector product catalog with semantic search
    • knowledgebase - pgvector knowledge base with article similarity
    • docmgmt - pgvector document management with content similarity

  • Four usage profiles for temporal patterns:

    • local-office - Business hours with breaks
    • global - 24/7 global enterprise
    • store-regional - Regional e-commerce with evening peak
    • store-global - Global retail with multi-region peaks

  • Two connection modes:

    • pool - Rapid connection reuse for web applications
    • session - User session simulation with think time

  • Multiple embedding generation options for pgvector apps:

    • Random vectors (default)
    • OpenAI embeddings
    • Sentence Transformers
    • pgedge-vectorizer service

  • YAML configuration file support

  • Real-time statistics reporting
  • Graceful shutdown with summary statistics
  • Cross-platform binaries (Linux, macOS, Windows)