I must have read over 100 pages of stuff and searched google for many hours.
I cant find a definitive answer. I run a site with a lot of image traffic.

What things cause nginx to throw a 503??? Ive tried a ton of configs.

At the linux level I did this:
http://fasterdata.es.net/fasterdata/host-tuning/linux/

changes sysctl.conf to have:
# increase TCP max buffer size setable using setsockopt()
# 16 MB with a few parallel streams is recommended for most 10G paths
# 32 MB might be needed for some very long end-to-end 10G or 40G paths
net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
# increase Linux autotuning TCP buffer limits
# min, default, and max number of bytes to use
# (only change the 3rd value, and make it 16 MB or more)
net.ipv4.tcp_rmem = 4096 87380 16777216
net.ipv4.tcp_wmem = 4096 65536 16777216
# recommended to increase this for 10G NICS
net.core.netdev_max_backlog = 30000

-------------------------------------------------------------------

I run nginx with fastcgi to php on the back end. Config snippets I have added include:

events {
worker_connections 8192;
use epoll;
# multi_accept on;
}

client_header_buffer_size 1k;
large_client_header_buffers 8 16k;
proxy_buffering off;

client_body_timeout 65;
client_header_timeout 65;
keepalive_timeout 0 0;
send_timeout 65;

## TCP options
tcp_nodelay on;
tcp_nopush on;

server {
listen 80 default rcvbuf=64K backlog=5000;
# was hoping backlog would help... it did not.
# This is at the top server level... I have many virtual hosts underneath which do no specify backlog (causes a conflict).
# Is a single backlog directive correct at the top server level???

# Below here all the virtual host files are loaded.

ANY IDEAS on how to make it foolproof against 503's? The server unix load is only about a 2 on a quad cpu box.