What if you could save most of the hosting cost of your Solr servers when they are underutilized?
High-end Solr configurations consist of dozens, sometimes hundreds of nodes. Salesforce, for instance, is rumored to run over 800 Solr nodes. The problem is that search is a highly-volatile consumer of resources, with enormous swings between the low and high utilization periods. Now that so many organizations run their applications in the cloud, that volatility has become an opportunity to generate substantial savings. Unfortunately, until today, it has been a wasted opportunity.
Not anymore with SearchStax AutoScaling. This is a new feature added to the SearchStax Solr Management Platform. AutoScaling can modify the number of nodes in a configuration based on rules (average QPS over the last 5 minutes, average CPU load over the last 10 minutes, etc.) and a schedule. Boundaries can be specified to prevent runaway situations.
To start with, this feature is available to SearchStax customers with Platinum and higher support plans. The typical AutoScaling user runs clusters consisting of five nodes or more. Large Sitecore, Drupal, or Adobe Experience Manager properties are often good candidates. AutoScale is an extremely powerful feature with the potential to not only deliver substantial savings but also ensure high-throughput without loss in latency in extraordinary situations. SearchStax Support and Solutions teams can usually configure autoscale in a few hours after a couple of conversations.