Unix Monitor

Net-SNMP must be installed and configured on the server. See Preparing Servers for Monitoring.

The monitor has the following settings:

• Port – The port used to communicate with the server. Defaults to 161.

• SNMP – The SNMP version to use. ReadyAPI supports versions 1, 2c and 3.

The other settings depend on the SNMP protocol version you use to communicate with the server.

• SNMP version 1 and 2c

  • Community string – SNMP Community String. Must match the SMNP string specified in the SNMP config file on the server.

• SNMP version 3

  • Login and Auth password – The user account and password for the SNMP 3 connection.

  • Encryption type – The data encryption algorithm to be used.

  • Encryption password – The password that will encrypt the traffic between the monitor and ReadyAPI. This password must match the SNMP 3 password you specify on the server side.

Unix server monitors return information about hardware usage of your server. This helps you find out how close to the hardware limit your test has come. When you create these monitors, ReadyAPI displays the following statistics in charts by default:

• Misc – Number of processes – Measures the number of processes on the computer. Some services create multiple processes to handle user requests. This metric helps you see how well these processes are cleaned out of the computer memory after they are no longer needed.

• CPU – Usage – Measures the usage of your server CPU. 80% or more of CPU usage sustained during soak or peak tests may be the reason to upgrade the server.

Here are some other frequently used metrics:

• Disk – Total Size and Disk – Used – Measure how much disk space is used during the test run. If you are using over 80-90% of the disk space in your tests, you may need to add additional storage.

• RAM – Total Size and RAM – Used – Measure how much RAM is used during a test run. If there is not enough RAM on your server, Unix machines swap the least used memory blocks to the virtual memory on the hard drive. This may cause performance bottlenecks, since hard drive access is slower than RAM access. This behavior is acceptable for spike tests, but your soak and peak tests should rely on RAM.

The available metrics are a subset of ucdavis (including the diskio module) and mib-2 SNMP MIBs statistics. Depending on the server configuration, you may be able to access different metrics. The following list shows the metrics you can access on a standard Ubuntu server: