Microagents

The agent interfaces the managed objects on its local machine using microagents. Microagents communicates with managed objects such as operating system subsystems, agent components, log files, event logs or applications. Each microagent exposes a set of methods to the agent that the agent uses to collect information and take action.

A set of default microagents is included when you install TIBCO Hawk software. The complete list varies by operating system, but the following standard microagents are installed on all platforms:

•

The Self microagent gathers version information for the local agent, the security policy in effect, and available microagents. It also activates or deactivates the collection of diagnostics for technical support.

•

The SysInfo microagent gathers basic information on the operating system, hardware architecture, computer name and IP address of the agent machine.

•

The Process microagent counts instances of a process on the agent machine and gathers process usage statistics.

•

The FileStat microagent gathers information on a disk file or files on the agent machine.

•

The Rendezvous microagent gathers information on TIBCO Rendezvous daemon activity on an agent machine and on TIBCO Rendezvous licensing. This micro agent is available only when Hawk is configured to use Rendezvous as a message transport mechanism between Agent and HMA. This micro agent will not be visible when TIBCO DataGrid is configured as the message transport.

•

The RuleBaseEngine microagent gathers information about the rulebases and takes actions on rulebases that affect the TIBCO Hawk agent.

•

The Logfile microagent responds asynchronously each time a new line is added to a log file you specify. You can create tests that search for specific strings in a log file line.

•

The Custom microagent executes any command-line executable that can run on the current agent machine. The microagent can also return text or numeric results from the executable.

This microagent is typically used to give a TIBCO Hawk agent access to information it cannot obtain from other microagents. It is a useful tool for extending agent capabilities.

For detailed information about these microagents with their methods, arguments and results, as well as descriptions of platform-specific microagents, see the TIBCO Hawk Microagent Reference.

Microagent methods can be accessed interactively from TIBCO Hawk WebConsole or through rulebases downloaded to the agent. Chapter 3, Monitoring with Rulebases, talks about how to use a method as the data source of a rule.

You can also access methods programmatically using the Console API. For more information, see the TIBCO Hawk Programmer’s Guide.

When you start up an instance of TIBCO Hawk WebConsole, it automatically discovers machines running TIBCO Hawk agents on your network depending on the Hawk message transport mechanism you have configured.

Each agent has a set of default microagents, that are loaded when the agent is started. If you install and start an adapter or gateway, or instrument an application with AMI, microagents for these objects are dynamically added to the agent. In TIBCO Hawk WebConsole, you can view microagents and their methods for any discovered TIBCO Hawk agents. For detailed information about how to get the list of microagent for an agent, see the TIBCO Hawk WebConsole User’s Guide.

You can invoke a microagent method in TIBCO Hawk WebConsole and immediately view the results. Invoking is useful when you want to test a method before using it in a rule, or to check a return value for troubleshooting purposes.

Microagent methods perform a wide variety of tasks, and are grouped into categories according to the impact that invoking the method has on a managed system. These categories are used by the TIBCO Hawk agent to display appropriate methods for the current context:

Microagent methods can be synchronous or asynchronous.

To view microagent method results over time, you can subscribe to the method in TIBCO Hawk WebConsole and view results as they are returned. Subscribing provides continuous updates of method results and displays a moving window of the result history. Creating a subscription is useful when you want to test a range of return values before specifying boundaries in a rule, or to identify general patterns of activity.