Sarvam Transliteration API Load Testing 🚀

API Docs | Dashboard Features | Google Sheets Report

---

A complete solution for load testing the Sarvam Transliteration API:

• Automated load test scripts
• Streamlit dashboard for configuration & live monitoring
• Tools for aggregating and analyzing results
• Designed for both quick and advanced load sweeps
• Clear reporting and troubleshooting guidance

---

📑 Table of Contents

• Project Structure
• Requirements
• Setup
• .env File Setup
• Running Tests
• Dashboard Features
• Google Sheets Report
• Troubleshooting
• Notes

---

🗂️ Project Structure

File/Folder	Description
API_Requests.py	Locust script for load test tasks (Hindi, Tamil, Bengali)
dashboard.py	Streamlit dashboard for config, live metrics, and results
run_tests.sh	Bash script for load sweep (varying users/spawn rates)
process_results.py	Aggregates CSV outputs into summary.csv
<prefix>_stats.csv	Per-test CSV stats (e.g., simple_test_stats.csv)
<prefix>_stats_history.csv	Per-test CSV history
summary.csv	Aggregated results from all tests
requirements.txt	Python dependencies
.env	(Optional) API keys and environment variables
Report/	Final PDF report

---

🛠️ Requirements

All dependencies are listed in requirements.txt:

streamlit
requests
plotly
pandas
locust
python-dotenv

Install all dependencies:

pip install -r requirements.txt

---

⚙️ Setup

1. Clone the Repository

   git clone https://github.com/justAbhinav/Sarvam_AI_API_Load_testing
   cd Sarvam_AI_API_Load_testing

2. Install Dependencies

   pip install -r requirements.txt

3. Set Up Environment Variables

   You can set your Sarvam API key in your shell or in a .env file (recommended for convenience and security).

---

🔐 .env File Setup

Create a .env file in the project root:

SARVAM_API_KEY=your_api_key_here

• Replace your_api_key_here with your actual Sarvam API key (Get it here).
• The project uses python-dotenv to load environment variables automatically.

Alternatively, export the variable in your shell:

export SARVAM_API_KEY="your_api_key_here"

---

🚦 Running Tests

1. Free Required Ports

   The dashboard and Locust use ports 8089 and 8090. Ensure these ports are free:

   lsof -i :8089
   lsof -i :8090
   kill -9 <pid>

2. Load Sweep Tests

   Run the load sweep to test different user counts and spawn rates:

   chmod +x run_tests.sh
   ./run_tests.sh

3. Process Results

   Aggregate all CSV outputs into a single summary.csv:

   python3 process_results.py

4. Run the Dashboard

   Launch the Streamlit dashboard for interactive test configuration and live monitoring:

   streamlit run dashboard.py

   • Open http://localhost:8501 in your browser.
   • Choose Simple (1 user, 1 spawn rate, 1m) or Custom test type.
   • Configure concurrency, spawn rate, run time (e.g., 30s), and CSV prefix as needed.
   • Click Run Test to view live metrics and final results.

---

📊 Dashboard Features

• Test Configuration

  • Simple Mode: Fixed parameters (1 user, 1 spawn rate, 1m duration, simple_test prefix)
  • Custom Mode: User-defined concurrency, spawn rate, run time, and CSV prefix

• Live Metrics

  • Median, average, and p95 latency (ms)
  • Requests per second (RPS)
  • Error rate (%)

• Final Results

  • Total requests, RPS, average/p50/p75/p95 latency, error rate
  • Language-specific p95 latency (Hindi, Tamil, Bengali) with a bar chart

  Note: There is a bug where you need to refresh the page for final results. All CSV files are saved correctly in the dashboard directory.

• CSV Output

  • Generates <prefix>_stats.csv and <prefix>_stats_history.csv per test

---

📈 Google Sheets Report

A detailed Google Sheets report is available for further analysis and visualization:

Sarvam Transliteration API Load Test Results

• Tabs:
  • Summary Dashboard: Visualizations of key metrics
  • Raw Data: Imported CSV data
  • Sweep Configurations/Notes: Test parameters and observations

---

🛠️ Troubleshooting

• Port Conflicts
  • Free ports 8089/8090 as shown in Setup.
  • Check for running Locust processes:

    ps aux | grep locust
    kill -9 <locust_pid>

---

📝 Notes

• The dashboard uses a single Locust instance on port 8089 for live metrics and a headless instance on port 8090 for CSV generation.
• If live metrics fail, verify API connectivity.
• Ensure your API key is valid and your network connection is stable.

---

For questions or contributions, please open an issue or submit a pull request.