It's often unclear how a complex autoscaling policy affects a SolrCloud cluster and its resources. Autoscaling suggestions and violations indicate what to do in order to balance the cluster (according to the rules) but when resources become tight it's sometimes not clear whether they really result in desired improvements. Deploying modified rules carries a risk of causing unwanted resource re-allocations, which may be costly and disruptive.
This session provides some details on the autoscaling simulation framework. It also presents how to use the command-line autoscaling tools for testing and fine-tuning arbitrary autoscaling policies, and how to use autoscaling snapshots to run reproducible simulations of long-term effects of autoscaling policies on cluster resources.
ACTIVATE Conference: www.activate-conf.com
Lucidworks: www.lucidworks.com