The APM power pack for Webservers uses the Apache mod_status module to poll for metrics (http://httpd.apache.org/docs/2.0/mod/mod_status.html). Consequently, the following performance metrics are available for Apache web servers (as well as IBM HTTP Server, and Oracle HTTP Server web servers):
Bytes Transferred Per Interval -Total number of bytes the workers (server) serve per interval. Available only if ExtendedStatus is enabled.
Current Number of Busy Workers -The number of workers serving requests.
Current Number of Idle Workers -The number of idle workers.
Current Percentage CPU Load -The current percentage CPU used in total by all workers in the web server. Available only if ExtendedStatus is enabled. Available only for Apache servers on UNIX machines.
Requests Per Interval -Number of requests the workers (server) serve per interval. Available only if ExtendedStatus is enabled.
Current Status of Worker Processes/Threads (%)
The following metrics are available for Current Status of Worker Processes/Threads (%):
Closing connection - Percentage of workers closing connection already created.
DNS Lookup - Percentage of workers performing a DNS lookup for IP address to host name resolution.
Gracefully finishing - Percentage of workers doing a graceful finish operation before actually closing the connection.
Idle cleanup of worker - Percentage of workers doing an idle cleanup.
Keepalive (read) - Percentage of workers keeping some of the existing connections alive before timing out and closing it.
Logging - Percentage of workers involved in logging the activities to log files or custom log listeners.
Open slot with no current process -This percentage indicates the free slots for workers which can be filled up with a process due to incoming requests.
Reading Request - Percentage of workers reading a request in their process.
Sending Reply - Percentage of workers sending a reply in their process.
Starting up - Percentage of workers starting up in a new process to serve a request.
Waiting for Connection - Percentage of workers that started and are waiting for a connection so that request processing can begin.