nixCraft

Hello,
I need to create a Control Unit for the Apache server (v1.3.41) which will minimize the response times (server side) by altering online the MaxClients and KeepAliveTimeout Directives (will alter and do a graceful restart every time they need to be altered). The problem is that i cant find a way to extract the proper metrics such as Hits/sec Requests/sec Simultaneous Connections/sec Server Response time Per Connection etc. Anyone has any idea where to look?

See following docs it has more info on metrics:

1. mod_perl: Performance Tuning
2. Performance-tuning Apache

__________________
Vivek | My personal blog
Linux Evangelist
Play hard stay cool

Well i think you misunderstood me. I dont want to optimize my server the
old fashioned way. The point is i'm creating a software that will optimize these Directives online (aka while the server is doing it's job) to the optimal for that particular time window values. To do that i need to have a reference
value of how the server is coping, one metric that tells me that, is the (server side) response time, aka the delta time from the moment the request arrives at the TCP queue until the first response of the server to that client. If i sample that, lets say 30 times a minute, i get a very nice picture about the server's status. That's something mod_status cant tell though . Anyway.. gonna figure it somehow.

Are you undertaking a high-resolution SNMP replacement? Or are you trying to combat server load problems?

Why not write a shell script to poll uptime and netstat every 5 seconds? Log it to a file. Write something else to display the file nicely, like php and gd.

I find that if I'm dealing with a request flood or a DoS, higher resolution monitoring doesn't necessarily help me, accurately identifying the IP that's bothering me is what's important.