Apache Monitor

To monitor the Apache server, enable the mod_status module as described in Preparing Servers for Monitoring.

The monitor has the following settings:

• Port – The port used to communicate with the server. Defaults to 80.

• Connect via https – Specifies whether ReadyAPI will connect to the server using https.

• Use client SSL Auth – Specifies whether the client authorization during the SSL handshake will be used.

• Keystore location – The location of the Apache keystore.

• Keystore password – The password to the Apache keystore.

• Client auth password – The password used to get authorized on the client.

The Apache monitor provides information about how users interact with the Apache server. ReadyAPI displays the following statistics in charts by default:

• Basic – Busy Workers – The percentage of busy workers (individual threads handling requests) on the server. Monitor these metrics to get relative usage of workers. This value should not exceed 80% in your soak and peak tests.

• Extended – Req Per Sec – The number of requests handled per second by the server. This value is useful when you simulate the VUs load profile to measure how many requests the server receives per second.

• Extended – CPU Load – The amount of the CPU used by the Apache server. Monitor this metric and compare it to the metrics of Unix or windows monitors if you have a CPU bottleneck. If the returned value takes a large percent of total – the Apache server is causing the issues. If not, search for another reason for the heavy CPU load.

Here are some other frequently used metrics:

• Basic – Idle Workers – The number of idle workers available on the server. As long as there are idle workers, the client does not need to wait for the server to finish the current work before responding to the request. If you hit 0 idle workers, increase the number of allowed workers in the settings. Each worker needs hardware resources such as memory and CPU, so you should monitor these as well. If you have too many workers and not enough hardware resources, the slowdown will be more significant as compared to the slowdown you will have if you have not enough workers.

• Extended – Uptime – The period of time during which the server has been running. Most servers need to be regularly rebooted for installation of updates and for maintenance. A server with long uptime may cause performance issues.

When you connect to the Apache server, the following metrics are available: