https://trello.com/b/YyDNqINY/review-sdc
https://docs.google.com/document/d/1Kd_1wOL9ta_JExj_6RepGAGFlwKFvAzxEGeZm9H_24E/edit?usp=sharing
Before I scaled, the website could handle a 300 average RPS, 200ms average latency, and a 10% error rate. After I scaled I increased it to a 8500 RPS, 0% error rate, and a 137ms average latency. I achieved this through a NGINX round robin load balancer, NGINIX caching, 3 AWS EC2 proxy servers, Docker, and some server side rendering. It could handle 10,000+ RPS but the latency would increase to 1 to 1.5 seconds and my goal was to keep the average latency under 200ms. I also indexed the SQL database to keep queries under 10ms. The google speed insight before the scaling was a 33/100. After the horizontal scaling the website received a 99/100.
Total Servers:
1 Nginx
3 proxies
Low: 8100 RPS
High: 9400 RPS
Average Latency 137 ms