[snakes1100]

With 300 people online at your peak time, there is no reason to start nginx or php-fpm with that many connections, there is no way you will use them, they are also going to use resources better spent in other places.

Start them off with their default connection limits in place, consider raising them if your traffic increases.

You should adjust these values based on your sites traffic & content, NOT by what others are using, these values should reflect your needs.

Basic optimized configuration:

client_header_timeout           10m;
        client_body_timeout             10m;
        send_timeout                    10m;
        client_max_body_size            10m;
        client_body_buffer_size         10m;
        connection_pool_size            256;
        client_header_buffer_size       8k;
        large_client_header_buffers     4 32k;
        request_pool_size               8k;

        gzip                            on;
        gzip_min_length                 1100;
        gzip_buffers                    8 16k;
        gzip_comp_level                 1;
        gzip_types                      text/plain text/css application/x-javascript text/xml application/xml application/xml+rss text/javascript;
        gzip_disable                    "MSIE [1-6]\.";

        output_buffers                  1 32k;
        postpone_output                 1460;

        sendfile                        on;
        tcp_nopush                      on;
        tcp_nodelay                     on;

        server_names_hash_max_size      4096;
        server_names_hash_bucket_size   128;

        keepalive_timeout               75 20;

        ignore_invalid_headers          on;

        fastcgi_connect_timeout         60;
        fastcgi_send_timeout            180;
        fastcgi_read_timeout            180;
        fastcgi_buffer_size             8k;
        fastcgi_buffers                 4 128k;
        fastcgi_busy_buffers_size       256k;
        fastcgi_temp_file_write_size    256k;
        fastcgi_intercept_errors        on;
        index                           index.php;

        proxy_buffering                 on;
        proxy_buffer_size               4k;
        proxy_buffers                   8 32k;

The above is for a site pumping a lot of content, with a lot of ads, with multiple sites hosted, in the header etc, again base these settings off your site.

With 2500 online:

nginx.conf
worker_processes                        8; <-- 1 per core
worker_rlimit_nofile                    4096;

events {
                                        worker_connections 1024;
                                        use epoll;
}


php-fpm.conf
listen.backlog = -1
pm = dynamic
pm.max_children = 50
pm.start_servers = 15
pm.min_spare_servers = 10
pm.max_spare_servers = 25
pm.max_requests = 500

With 4000 online, different server/forum

nginx.conf
worker_processes                        16; <-- 1 per core
worker_rlimit_nofile                    4096;

events {
                                        worker_connections 2048;
                                        use epoll;
}


php-fpm.conf
listen.backlog = -1
pm = dynamic
pm.max_children = 10
pm.start_servers = 35
pm.min_spare_servers = 15
pm.max_spare_servers = 50
pm.max_requests = 1024