Using DM 10.1 CU4/10.1.11
When creating a new Aggregation Rule, it seems that you cannot define which Agent's Aggregation Engine should be used when defining a rule.
Upon examining the rule's XML, I can see that it has a tag <AggregationElement id="<DMA ID>/<El ID>"/>. However we have seen numerous times that this ID is not tied to the same agent hosting the elements for which it is aggregating data.
Is this due to the agent logged into at the time of creation? Or tied to the view in some way? Most importantly, should we be concerned if this occurs for reliability (i.e. if the aggregating agent's is disconnected temporarily, does a different agent pick up this rule) or network load reasons (transferring unnecessary data throughout the cluster)?
I believe since DELT was introduced you no longer have a correspondence between the DMAid in the element ID and the agent hosting that element, the info about the hosting DMA is in the "DataMiner" column:
I undertsand aggregation rules can be view-specific, but normally the Views are in Sync (once a view is created on one DMA you can see it also when logging in from a different DMA of the same cluster).
I'd expect the aggregation to be in sync with the Views file: element data would come in the aggregation as long as the element is actively polled // the hosting agent is active. Worth checking with a DMA in STOP as there are different things that need SYNC (for Views, Forced Sync for files, for correlation rules in Brain.xml, or syncRules.xml - the latter pevents sync across DMAs with different versions - cannot recall if the behaviour is exactly the same across all the supported releases).
HTH,
A.