Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

If I have a site with say 5 million unique visitors a month (probably 10+ million bot crawls) what size of a server would I need for nginx solely as a load balancer (using reverse proxy to apache on other servers)?

I've read somewhere that some use atom servers as frontend load balancers but I'm not sure about that.

share|improve this question
up vote 3 down vote accepted

I'm not sure how relevant unique IPs is for a load balancer; hits per second and bandwidth seem more useful for me. As to that, I'm running my whole stack (nginx, php, postgres) on one server, and nginx is taking ~3% of the CPU to handle 1000 hits/sec @ 200Mbps. I'm fairly sure that any server you buy new today will be able to fill a gigabit pipe with ease if all it's doing is taking network data from one port and pushing it out to a different one.

That said, if you specifically want load balancing and can live without nginx's advanced web serving features, haproxy is probably a better fit.

share|improve this answer
Interesting! What are you using to pass nginx requests to php? That's one of the main reasons why I'm still using bloated apache, reliability. I've read a few posts about issues with passing nginx to php so I'm curious to know what you are using. – Joe Nov 7 '11 at 21:27
I'm using a few copies of "php5-cgi -b /tmp/php-1.socket" (incrementing the number each time), with each process monitored and restarted on crash by runit; then nginx connects to the sockets and talks fastcgi through them – Shish Nov 10 '11 at 9:41

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.