In order to monitor a Linux server, SOASTA CloudTest must be able to establish an SSH connection to the server. If you are not sure how to enable this, consult your Linux documentation.
If a firewall prevents SSH access, see Monitoring a Server Behind a Firewall.
About Memory Utilization on Linux
Like other Linux monitoring tools, including "free" and the GNOME Task Manager, CloudTest monitoring intentionally excludes buffered and cached memory from the total usage statistics presented.
CloudTest excludes kernel buffers and cache because Linux is
- Very aggressive about grabbing all the free RAM for kernel buffers and cache, and
- will give it back to user-level programs as soon as needed.
Without these exclusions, the memory usage would almost always display at 99-100% of the physical RAM on the machine for the reasons above.
- In the Central list, select Monitoring Server Groups.
- Click the New Monitoring Server Group icon.
The Monitoring Server / Monitor wizard appears.
- Select Monitoring Server Group and click Next.
- Type a name for the Monitoring Server Group and optionally provide a description. Click Next.
- Check System Resources and click Next.
- The Hosts/Device Agents screen appears. Add one or more servers by hostname or IP Address.
- Select the version of Linux to monitor.
- Once the Linux version is selected, select a Connection Method. Linux users can use SSH or SOASTA Conductor installed and running on each monitor server in the group.
- For SSH, enter the port number (usually the default), user name, and password to complete the server group configuration.
- For the Use Conductor on Each Machine option, ensure that the Conductor is installed and running on each machine you're monitoring.
- Click the Test Connection button to test the connection. If the connection is successful, you will see a pop-up containing some information about the Linux server. If the connection fails, you will see an error message explaining what went wrong.
- Click Next and review the information to be sure it is correct
- Check the "Create a new Monitor for this Monitor Server" box and click Finish.
- Enter a name for the Monitor and an optional description, then click Next.
- Now select the metrics to monitor. For each statistic, you can choose a "sample rate" thatndetermines how often that statistic will be collected.
Some of the statistics require an entry in the "Arguments" column. For example:
- For the "Network MBits Received" and "Network MBits Sent" statistics, you can enter the name of the network interface you want to monitor (for example,
eth1). If you do not enter a value, then the total of all network interfaces will be tracked (this excludes the loopback interface).
- For the "Per Process CPU Percentage", "Per Process Memory Usage", "Per Process Thread Count", and "Per Process Virtual Memory" statistics, you must enter the name of a process (for example,
httpd). In some cases, there may be multiple matching processes (for example, multiple Apache HTTP worker processes). When this happens, the total for all matching processes will be tracked.
- Click Next. Review the monitor configuration, then click Finish.
- To start or stop monitoring, see the topic "Starting and Stopping a Monitor"
- To automatically monitor the Linux server every time a certain test starts, see the topic "Starting a Monitor Automatically When a Test Composition Runs"