Basic PostgreSQL Tests with Diesel ORM

This is simple set of tests for me to better understand how PostgreSQL performs under light load, with different operations and configurations. I decided to use Diesel ORM for the logic, because I'm familiar with it, and it presumably adds little overhead.

Quick Start

# Baseline PostgreSQL
bash configs/baseline/run-benchmark.sh

# Optimized PostgreSQL  
bash configs/optimized/run-benchmark.sh

# Aggressive PostgreSQL (⚠️ NO DATA SAFETY)
bash configs/aggressive/run-benchmark.sh

# Load Balanced PostgreSQL (HAProxy + 2 Read Replicas)
bash configs/load-balanced/run-benchmark.sh

Results

Specs:

• AMD Ryzen 5 7640U
• 64GB RAM
• 2TB NVMe

Batch Insert Performance

Individual Operations Performance

What's Tested

• Bulk Inserts: 1M records × 10 batches = 10M total records
• Single Writes: Individual insert operations
• Count Queries: COUNT(*) performance as table grows
• Text Exact: Exact string matching on varchar fields
• Text LIKE: Pattern matching on varchar fields
• JSON Filtering: Queries on JSONB fields

Configurations

1. Baseline: Standard PostgreSQL with default settings
2. Optimized: Tuned single instance (512MB shared_buffers, optimized WAL)
3. Aggressive: Extreme performance, no safety (fsync=off, autovacuum=off)
4. Load-Balanced: HAProxy + Primary + 2 Read Replicas for parallel reads

Consistent colors across charts: 🔴 Write, 🟢 Count, 🔵 Text Exact, 🟠 Text LIKE, 🟣 JSONB Filter

---

Note: AI was involved generating parts of this repository, including bash scripts and config README's.