By integrating the machine learning capabilities of Prelert, recently acquired, into the Elastic Stack, customers can leverage machine learning without requiring complex expertise or separate development. Elastic's new machine learning capabilities feature time-series data solutions that automatically detect anomalies, efficiently perform root cause analysis, and reduce false positives in real-time applications. This technology is suitable for businesses seeking to monitor infrastructure issues, cyberattacks, and business challenges in real time.

The increasing demand to secure real-time insights for operational use is diminishing the practicality of traditional data analysis approaches. While it is possible to generate statistical models using third-party machine learning, developing real-time operational systems for existing tasks is challenging. Identifying the correct statistical models for diverse and distinct datasets is not only costly but also requires sophisticated data science skills. In addition, manually created rules are unstable and result in many false positives.

Elastic's first unsupervised machine learning capabilities, available as a feature of X-Pack in version 5.4, automatically detect anomalies in time-series data such as log files, application and performance metrics, network flows, and financial transaction data. Because Elastic's new machine learning capabilities leverage existing data stored in Elasticsearch, users can immediately apply workflows such as logging, security analysis, and metric analysis to operations. Furthermore, they enable the creation of sophisticated machine learning tasks through the familiar and user-friendly Kibana UI, while minimizing complexity and cumbersome integrations.

"Elastic's vision is to eliminate complexity so that users can easily deploy machine learning on the Elastic Stack for logging, security, metrics, and more," said Shay Banon, founder and CEO of Elastic. "Unsupervised machine learning capabilities provide an excellent user experience. Because they can detect anomalies in their own time-series data, search and analysis will naturally scale."