Will data in Insight be populated across cluster or only stored locally when we compose a cluster with several servers?
If Insight is enabled on a node (recommended to be a Publisher node) within a cluster, all nodes in the cluster will send data to this node for storage. As a result, it’s guaranteed to get a cluster-wide view of request processing, which would help in creating filters and use insight data for our custom requirements.
If you enable Insight on a given node within a cluster, you should be able to reliably compute bandwidth, session timers etc. as long as that node does not go down. If you need redundancy here, enable Insight on two nodes.