Configuring the analytics message queue

The analytics message queue is embedded in the data ingestion pipeline. It is most useful when offloading data from the analytics subsystem to a third-party system. It ensures that data collection continues in the event of a failure.

About this task

The analytics message queue is an optional service that provides a reliable ingestion pipeline, especially when offloading analytics to a third party system. The message queue ensures that the analytics pipeline will not be blocked when either analytics storage or the offload endpoint becomes unavailable. Data collection will continue internally in the analytics subsystem, and will be buffered if analytics offload has been configured. With the message queue enabled, data storage will continue even if the offload endpoint is unresponsive. Additional benefits include better handling of traffic spikes and a larger buffer to handle ElasticSearch issues such as red indices, mapping issues, etc.

Procedure

Complete the following steps to configure an analytics message queue:

  1. Enable the message queue using the APICUP installer and install the analytics subsystem. The command is apicup subsys set analytics enable-message-queue=true.
  2. Set an optional mq-storage-class, or use the storageClass object.

    This parameter is used in Kubernetes installations only. It is not used in VMware environments (OVA files).

  3. Install the Analytics subsystem to activate the message queue.
  4. Optionally, configure analytics third-party offload if required. See Configuring analytics offload