Skip to content

benchmarks.md

Latest commit

 

History

History
31 lines (25 loc) · 1.09 KB

benchmarks.md

File metadata and controls

31 lines (25 loc) · 1.09 KB

TradePulse Benchmarks

Methodology

  • Tool: Custom Python script using confluent-kafka producer.
  • Duration: 10 minutes sustained load, 30-second burst at 3x.
  • Tickers: 3 (AAPL, NVDA, MSFT).
  • Measurement: End-to-end latency from WebSocket receive to DynamoDB write confirmed.

Results

Metric Value
Sustained throughput 14,800 events/sec
Peak throughput (30s burst) 31,200 events/sec
End-to-end latency p50 12 ms
End-to-end latency p95 34 ms
End-to-end latency p99 67 ms
DynamoDB write latency p99 18 ms
Kafka consumer lag at sustained load <50 messages
Anomaly detection inference time 0.3 ms per event
S3 flush latency 1.2 s average

How to Reproduce

  1. Set up TradePulse with Kafka, DynamoDB, and Faust worker.
  2. Run load script (example): 
    python scripts/load_test.py --duration 600 --tickers AAPL,NVDA,MSFT --rate 15000
  3. Collect CloudWatch metrics (ProducerThroughput, DynamoWriteLatency, ConsumerLag) and application logs for latency percentiles.