Page tree

Collection Agents are primarily used for collecting data locally. However, if you need to collect large volumes of data remotely (more than 4 GB data per day or more than 500 data collectors), then it is recommended that you use a Collection Agent for your data collection. This is important to avoid the Collection Station resources getting overburdened. The Collection Station needs to be primarily reserved for handling incoming data as opposed to being used for data collection. To be able to use a Collection Agent for remote data collection, you need to perform certain configurations.

The following information describes the architecture and configurations necessary for using a Collection Agent for remote data collection:

Recommended deployment architecture

You can use a Collection Agent to collect data remotely over both LAN and WAN networks.

In the following diagram, you can see two scenarios of using the remote Collection Agent – locally and remotely.

In the remote data collection architecture (over WAN), you can see that the Collection Agent is deployed in the UK data center while the Collection Station is deployed in the US data center. The Collection Agent performs remote data collection (by way of remote data collectors) from various target hosts present in the same data center. And the Collection Agent passes the collected data to the Collection Station over WAN.

In the local data collection architecture (over LAN), the Collection Agent, target hosts, and Collection Station are in the same data center. Notice that even though the Collection Station and Collection Agent are in the same data center, the Collection Agent performs data collection while the Collection Station only receives the data collected. This is important to prevent the Collection Station from getting overburdened with data collection work. The Collection Station is primarily required to play the role of a receiver as opposed to collector.

Deployment architecture for remote Collection Agent

Configurations required to enable a Collection Agent for remote data collection

 

In the following table, the configuration location paths must be preceded by one of the following environment variables that refer to the Collection Agent installed location:

  • Windows: %BMC_ITDA_REMOTE_AGENT_HOME%
  • Linux: $BMC_ITDA_REMOTE_AGENT_HOME

In the following table, the "Recommended value" column follows the hardware configuration requirements described in the  Sizing and scalability considerations page.

Note

The following table provides a good estimate of the required values. However note that these values are subject to change based on variable factors such as the data patterns, data collector types, and the total number of data collectors used for data collection.

Configuration property and locationRecommended value

Initial memory required for the remote Collection Agent

wrapper.java.initmemory 3072 MB
custom\conf\services\collectionagent.conf

Maximum memory required for the remote Collection Agent

wrapper.java.maxmemory
 3072 MB
custom\conf\services\collectionagent.conf

Thread pool size for remote Collection Agent

collection.thread.pool.size
50
\collection\custom\conf\agent.properties

Note: If you plan to create one of the following kinds of data collectors, then in addition to the recommended value, you need one additional thread per data collector: