Web API server to support the demands of production traffic for a 'Questions and Answers' widget of a mock e-commerce website. API endpoints include reading, updating, and inserting new data.
- ETL process to import and manage 12 million records of data
- Optimized database queries and tuning of Postgres database for fast response times
- Deployed across multiple AWS EC2 t2.micro instances to support horizontal scaling
- Added NGINX load balancer and reverse proxy with optimized configurations for maximum throughput and latency
- Configured NGINX caching to increase data retreival performance for heavy read requests
- Load testing done on Loader and with k6
- Created my own tool to programatically run load tests and capture results for accurate measurements
Maintaining 1,000 simultaneous clients per second making continuous requests over a 30 second duration, application can support on average across all endpoints
- 9.5k successful responses per second
- 232ms latency
- less than 1% error rate
2,000 simultaenous clients per second
- 6.7k successful responses per second
- 478ms latency
- less than 1% error rate