Skip to content

Asynchronous Performance & Stress Test Benchmark

This document outlines the performance benchmark and stress test results for the asynchronous processing pipeline, demonstrating data integrity, fault tolerance, and resource efficiency under constrained environments.

Performance Test Results

Tested Configuration

The benchmark was executed under restricted container/server resources to evaluate system stability under high concurrent pressure:

  • Virtual Users (VUs): 100 concurrent users

  • Application Server: 1 Gunicorn Worker

  • Concurrency Model: 2 Threads per worker

  • Resource Constraint: 0.7 CPU Cap

Benchmark Results & Key Metrics

1. Data Ingestion Integrity

  • Total Requests: 3,516 continuous requests ingested.

  • Error Rate: 0.00%

  • Result: Achieved 100% data ingestion integrity without dropping records or corrupting payloads.

2. Guaranteed Fault Tolerance

  • Behavior Under Load: Even when resource constraints caused processing latency to shift (with the P99 latency reaching 7.0s), the architecture held firm.

  • Gateway Stability: The Nginx gateway and Flask middleware sustained the heavy load completely without triggering a single 502 Bad Gateway or 504 Gateway Timeout error.

3. Resource Efficiency & Decoupling

  • CPU Usage: Peak CPU utilization remained low at 38.4%.

  • Result: Core business logic execution remained decoupled and stable, proving that the system is safe from cascading infrastructure-wide crashes during traffic surges.