One of the great things about Splunk as both an Operations Management tool and as an Application Performance Management tool is the ease with which an astonishing variety of data sources can be fed into the Splunk data store. Splunk automatically indexes this data based upon time stamps, and stores it in a back end data store that scales out horizontally on commodity servers with commodity storage. This means that Splunk one of the very few management solutions that can scale out to accept the tsunami of management that is generated across the infrastructure and application stack in a modern dynamic or cloud based environment.
The Splunk Architecture
The wealth of data sources that can be collected an indexed by Splunk are shown in the left portion of the image below. The scaled out architecture that is how Splunk can keep up with the management data tsunami is shown in the rest of the diagram.
Now we come to the part about the good news and the bad news. The good news is that Splunk is able to be your management data store across your physical hardware, virtualization layer, operating system layer, application infrastructure layer (middleware) and the layer comprised of the applications themselves.
The bad news is that until today, if you wanted to pull all of the data together than pertained to a particular application, you had to be an expert in the topology of that application (where does it run), the virtual and physicalÂ infrastructureÂ that supports that application (what is it dependent on) and on how to tie disparate data sources together in Splunk to create a cohesive view or dashboard. Organizations with a few (or one) mission critical application that was of such high value that it warranted a dedicated support team could easily justify the investment in learning required to pull this off. Organizations with thousands of business critical and performance critical applications saw this as an infinitely high cliff.
The Prelert Anomaly Detective for Splunk
The Prelert Anomaly Detective automatically learns the normal patterns of the Splunk data. It then automatically identifies anomalous behavior in the Splunk data and uses the ability of the Splunk Query Language to find cross-correlated data and events.
The Prelert Anomaly Detective allows for a significant advance in how customers use Splunk and its data. Today most customers use Splunk as a forensics tool to find the problem, after some other tool or user has reported the problem. The combination of the Prelert Anomaly Detective with Splunk allows Prelert to notify customers of anomalies that the customer did not even know to go look for and that can easily be leading indicators of problems that have not yet been reported.
The complete Prelert announcement is here – “Prelert Introduces Anomaly Detective, an Advanced Predictive Analytics Solution for Splunk Enterprise Environments“