How are my project systems being utilized?

Regardless of the number of hosts in a project, in practice it is common to find shortages of free hosts. At the same time, there are almost always hosts which are under-utilized or even completely idle, which could be re-allocated or consolidated. Finding and reallocating these hosts allows more efficient use of your project infrastructure.

Project-Level Analytics

The Analytics tab in your project shows the following metrics. These metrics are the base available across all operating systems that TeamForge Lab Management supports.

Load Average
Processes
CPU
Logged-in Users

Note: More metrics, including host-specific and operating-system-specific metrics, are available for individual hosts when you click on the host name in the results table.

To see a metric, click the name of the metric. You can see these time ranges:

Daily: Approximately the past 24 hours
Weekly: Approximately the past 7 days
Monthly: Approximately the last 30 days
Yearly: Approximately the last 365 days

If a given host says "no value" in any of its columns, this means that TeamForge Lab Management has been unable to collect this data over the requested time interval. If the machine is up, and not collecting data, contact a TeamForge Lab Management administrator to investigate why data collection is not working properly. If the machine is currently down, or was down during the requested time range, you will not be able to get any performance data for those times. There is no way to retroactively "catch up" data if collection is not working properly.

The data used to build each graph and chart presented to you can be exported in CSV (Comma-Separated Value) format, suitable for opening in any spreadsheet application. This allows you to build your own visualizations of the data to complement the ones TeamForge Lab Management creates.

Beating The System: Dealing With the Possibility of Users Generating "Fake" Load To Make Machines Seem Busier

Since we publish the metrics for determining how busy TeamForge Lab Management thinks machines are, it is possible for irresponsible users to generate automated jobs which simulate a busy machine, even if the machine is really not being used for anything.

We encourage all users and administrators of the project to make sure people in your projects understand that this type of behavior is not acceptable and may lead to loss of privileges or other actions against them. Presumably, if someone has a good reason for wanting to keep a machine, it is good for their project, and your organization if they do so.

As administrators, we strongly recommend you take the time to understand and listen to your users' concerns about how machines are allocated. You may just need more machines in your project, or you may be overzealous about de-allocating machines once they drop below a certain usage threshold. De-allocating machines that seem to be not busy may seem efficient, but if those machines took their users a long time to set up, that might be counterproductive. Perhaps you can reach a middle ground and use virtual machines, or a smaller virtual or physical machine, for that user.

Finally, we always recommend talking to your users and trying to understand how they are really using their machines. The statistics that TeamForge Lab Management gathers are a starting point for managing your software development and testing infrastructure, but not the final word.