Hello,

I have a apache cluster with more than 10 nodes, based on ldirector and heartbeat. The problem is that I cannot predict if my nodes will handle the traffic in the next day (hosting a website based on daily campaigns). So I decided to limit the number of active connection on the nodes (from apache), but this is only a temporary solution. I want to create a page that will appear to users that are getting over the limit. Has anyone made this before? Can you tell me how is it possible (I don't want a how to, just a starting point to study)? I think squid can do it, but I don't know how to search for it.

To give you an example of what I want, you can see the same thing on deviantart.

Thank you.

What error code user receives when there's overload? I presume it's 500 or 502(you'll have to check that in your logs if you don't know). Look into ErrorDocument directive.

Hey,

Thanks for the reply. No, there is no message, just keeps on loading. It puts the connection in a queue (http://httpd.apache.org/docs/2.2/mod...#listenbacklog). Also, I'm planning to migrate to lighty so I need a universal solution, but I don't know how to search what I need.

Thanks.

there are a number of modules within nginx which can threshold maximum numbers of connections to a server it's front-ending. I looked throught the modules now, and TBH I can't immediately see anything for global connections, only per-source connections, but i'm sure I'm missing something.