A technical walk-through of the most important PerformanceGuard concepts.

When you start to use PerformanceGuard, make sure that you are familiar with PerformanceGuard concepts and terminology. For example, it's important to know that the PerformanceGuard web interface is your main configuration interface. This—along with many other concepts and terms—is covered in the following.

Key Components

• PerformanceGuard agent: An application that's installed on computers in order to gather resource- and network performance data. It runs as a service (that is it runs in the background without any user interface), and is often simply referred to as an agent.
• PerformanceGuard frontend server: A server that receives raw data from PerformanceGuard agents. It runs as a service, it can be installed on more than one computer if required. The frontend server continuously listens for data from agents, but it never actively contacts agents. You can have multiple frontend servers if required, for example if you expect to receive data from many agents, and you want to distribute the load.
• PerformanceGuard frontend server database: A database that contains configuration, entities and data from a frontend server. If you use multiple frontend servers, each frontend server must have its own database.
• PerformanceGuard backend server: A server that receives data from the frontend server(s). The backend server aggregates data and deletes old data as required. The backend server also offers a notification module that can automatically send e-mail notifications to specified recipients. The backend server runs as a service.
• PerformanceGuard backend server database: A database that contains configurations and aggregated data received from the frontend server(s).
• PerformanceGuard web server: Presents the PerformanceGuard web interface (see the following). It can also use a REST (REpresentational State Transfer) interface to present data in other applications. The web server runs as a service and by default listens on port 80 and 81.
• PerformanceGuard web interface: A web-based user interface with which you can view computer and network performance. If you are logged in as administrator, the PerformanceGuard web interface lets you configure your PerformanceGuard solution and control your PerformanceGuard agents.

Servers—and how PerformanceGuard Handles them

A server is a physical computer on which a server software program runs. A server software program is a program designed to process incoming requests from client computers and return responses.

Only servers reported by PerformanceGuard agents are known by PerformanceGuard. When PerformanceGuard learns about a new server from agents, the new server is placed on the PerformanceGuard discovered servers list.

Agents on PerformanceGuard systems typically register contact with thousands of servers. Some servers are relevant to know about, others are of little interest, but you can quickly gather a lot of data about all of them. That's why in order to use data about interesting new servers in PerformanceGuard graphs and reports, you must move those servers from the discovered servers list to the PerformanceGuard monitored servers list. In other words, monitored servers are servers that you have actively selected that you want to know more about.

Ports—and how PerformanceGuard Handles them

In TCP/IP and UDP/IP networks, a port is an endpoint to a logical connection, and the way that a client program specifies a specific server program on a computer in a network. Some ports have numbers that are preassigned to them by the Internet Assigned Numbers Authority (IANA), and these are called well-known ports. Port numbers range from 1 to 65536, but only ports numbers 1 to 1024 are reserved for privileged services and designated as well-known ports.

Only predefined ports and ports reported by PerformanceGuard agents are known by PerformanceGuard. Just like servers (see the previous) new ports are placed on the PerformanceGuard discovered ports list. In order to use data about such new ports in PerformanceGuard graphs and reports, you must move the ports from the discovered ports list to the PerformanceGuard monitored ports list.

PerformanceGuard automatically places a few ports, such as port 80 (HTTP), on the monitored ports list. If you want to monitor more ports, you must manually move them from the discovered ports list.

Which ports does PerformanceGuard itself use?  

For a list of ports used for communication between the PerformanceGuard components, see Ports and Port Numbers.

Services—and how PerformanceGuard Handles them

In PerformanceGuard terminology, a service is defined as one port on one server.

If a server runs multiple server software programs, or a server software program handles requests on more than one port, PerformanceGuard sees this as multiple services.

If the same type of server software program runs on different servers, PerformanceGuard also sees this as multiple services.

Applications

In PerformanceGuard, applications are used to group performance measurements in order to create reports that closely match business needs.