Aggregation

The flow query engine supports rendering Top-N metrics from pre-aggregated documents stored in Elasticsearch. Use these to help alleviate compute load on the Elasticsearch cluster, particularly for environments with large volumes of flows (>10,000 flows/sec).

To use this functionality you must enable the Kafka forwarder as described in Configure Kafka forwarder and set up Nephron to process the flows.

Nephron currently requires an Apache Flink cluster to deploy the job.

Set the following properties in $OPENNMS_HOME/etc/org.opennms.features.flows.persistence.elastic.cfg to control the query engine to use aggregated flows:

Table 1. Optional parameters for flow aggregation
Property	Default	Description
alwaysUseRawForQueries	true	Use raw flow documents to respond to all queries instead of aggregated flows.
alwaysUseAggForQueries	false	Use aggregated flow documents to respond to all queries instead of raw flows.
timeRangeDurationAggregateThresholdMs	120000 (2 minutes)	Queries with time range filters that have a duration greater than this value will use aggregated flows when possible.
timeRangeEndpointAggregateThresholdMs	604800000 (7 days)	Queries with time range filters that have an endpoint that is older than this value will use aggregated flows when possible.