Ask HN: What is your high performance server setup?
https://news.ycombinator.com/item?id=32439601
After reading this, I am wondering, how do you configure your linux server for high throughput API? I am struggling with 10k concurrent request with node.js on c6i.4xlarge and one ElastiCache redis cluster. I can't imagine handling 50k/100k concurrent request with single server. Profile your server and figure out what the bottleneck in your app is:
https://nodejs.org/en/docs/guides/simple-profiling/ Also, consider building the ElasticSearch index on another server than the one serving live traffic. Are you maxing out all cores? Perhaps try running multiple processes. Yes (using pm2). I have a very good performance (req/sec) for 100/200 concurrent request. But it is a disaster for high concurrent requests.