query-performance-analyzer 0.1.0

A library with the Lakehouse Framework

Query Load testing

This project is a benchmarking tool that compares query performance across different SQL Warehousing solutions.

Current support out of the box for:

• Synapse Serverless
• Databricks SQL Warehousing

It supports concurrent and sequential execution modes to evaluate performance, caching effects, and scalability.

---

📂 Project Structure

.
├── run.py              # Main script to run benchmarks
├── config.py           # Configuration file (credentials, queries, concurrency settings)
├── requirements.txt    # Python dependencies (not provided, create manually)
└── README.md           # Documentation

---

⚙️ Setup

1. Clone the repository

git clone <repo-url>
cd <repo-folder>

2. Install dependencies

pip install -r requirements.txt

Dependencies include:

• pyodbc
• azure-identity
• databricks-sql-connector

3. Configure config.py

The project now uses a structured configuration format with backward compatibility:

New Structured Format (Recommended):

CONFIG = {
    "CONNECTIONS": {
        "Synapse": {
            "SYNAPSE_SERVER": "your-synapse-server",
            "SYNAPSE_DATABASE": "your-db",
            "SOURCE": "Synapse"
        },
        "Databricks": {
            "DATABRICKS_TOKEN": "your-token",
            "DATABRICKS_SERVER": "your-databricks-server", 
            "DATABRICKS_HTTP_PATH": "your-http-path",
            "SOURCE": "Databricks"
        }
    },
    "QUERIES": {
        "External": ["SELECT ...", ...],
        "Managed": ["SELECT ...", ...],
        "Synapse": ["SELECT ...", ...]
    },
    "EXPERIMENTS": {
        "ConcurrentComparison": {
            "CONNECTIONS": ["Databricks", "Synapse"],
            "QUERY_PLANS": ["External", "Managed", "Synapse"],
            "CONCURRENCY_LEVEL": 8,
            "RUN_SEQUENTIAL": True
        }
    }
}

▶️ Running the Benchmark

Use the new structured runner for custom experiments:

import runner
from config import CONFIG

# Run a custom experiment
results = runner.run_experiment("Experiment1", CONFIG)

# Or run queries directly
queries = CONFIG["QUERIES"]["External"]
connection_config = CONFIG["CONNECTIONS"]["Databricks"]

results = runner.run_threading_queries(
    queries=queries,
    parallel_runs=4,
    sequential_runs_per_query=2,
    source="Databricks",
    config=connection_config,
    experiment="MyTest"
)

---

✅ Use Cases

• Performance benchmarking between Databricks and Synapse
• Identifying query latency differences

---

📝 Notes

• Requires Azure credentials (DefaultAzureCredential) for Synapse authentication
• Requires Databricks ODBC driver
• Make sure queries are aligned: len(DATABRICKS_QUERIES_EXTERNAL) == len(DATABRICKS_QUERIES_MANAGED) == len(SYNAPSE_QUERIES)

Improvements

• Create Warehouse
• Create APP
• Command line support
• non PY config