Global Threat Condition: Elevated
Agent handlers and what they do
Agent handlers are the component of ePolicy Orchestrator that handles communications between agent and server.
Each installation of ePO includes an agent handler. Beginning with ePO 4.5, agent handlers can be installed independently on systems throughout your network.
Multiple remote handlers can help you address scalability and topology issues in your network, and in some cases using multiple agent handlers can limit or reduce the number of ePO servers in your environment. They can provide fault tolerant and load-balanced communication with a large number of agents including geographically distributed agents.
How agent handlers work
Agent handlers distribute network traffic generated by agent-to-server communication by assigning managed systems or groups of systems to report to a specific agent handler. Once assigned, a managed system performs regular ASCIs to its agent handler instead of the main ePO server. The handler provides updated sitelists, policies, and policy assignment rules just as the ePO server does. The handler also caches the contents of the master repository, so that agents can pull product update packages, DATs, and other necessary information.
NOTE: When an agent checks in with its handler, if the handler does not have the updates needed, the handler retrieves them from the assigned repository and caches them, while passing the update through to the agent.
Multiple agent handlers
You can have more than one agent handler in you network. You might have a large number of managed systems spread across multiple geographic areas or political boundaries. Whatever the case, you can add organization to your managed systems by assigning distinct groups to different handlers.
Handler groups and priority
When you are using multiple agent handlers in your network, you can group and prioritize them in order to help ensure network connectivity. Configure your handler groups to meet the specific needs of your environment.
For example, You might choose to create a group of handlers in which the handlers are disperesed over a wide geographic. With handlers dispersed you can configure the handler priority so that agents first communicate to the handler nearest them.
Handler groups
When you have multiple agent handlers in your network, you can create handler groups. You can also apply priority to handlers in a group. Handler priority tells the agents which handler to communicate with first. If the handler with the highest priority is unavailable, the agent falls back to the next handler in the list. This information is contained in the sitelist.xml file in each agent. When you change handler assignments, the sitelist.xml file is updated as part of the agent-to-server commutation process. Once received, the agent waits until the next regularly scheduled communication to implement the new assignments. You can perform an immediate
agent wake-up call to update the agent immediately.
Grouping handlers and assigning priority is customizable so you can meet the needs of your specific environment. Two common scenarios for grouping handlers are:
1 Using multiple handlers for load balancing.
2 Setting up a fall back plan to ensure agent-to-server communication.
In scenario 1, using agent handlers for load balancing, you might have large numbers of managed systems in your network, for which you want to distribute the work load of agent-to-server communications, and policy enforcements.You can configure the handler list so that agents randomly pick the handler with which report.
In scenario 2, using multiple handlers to enxure communication, you might have systems distributed over a wide geographic area. By assigning a priority to each handler dispersed throughout this area, you can specify which handler agents communicate, and in what order.
This can help ensure managed systems on your network stay up-to-date by creating a fall-back agent communication, much the same as fall-back repositories ensure that new updates are available for your agents. If the handler with the highest priority is unavailable, then the agent will fall-back to the handler with the next highest priority.
In addition to assigning handler priority within a group of handlers, you can also set handler assignment priority across several groups of handlers. This can add an additional layer of redundancy to your environment to further ensure your agents can always receive the vital information they need.
