Best Practices Product 6 Min Read

Introducing New Capabilities for Operationalizing your Big Data Applications at Scale

Best Practices Product 6 Min Read
By: Bala Venkatrao, VP, Products

The Unravel mission from the outset has been to provide Application and Operations teams with the best full-stack and intelligent Application Performance Management (APM) platform for their Big Data Applications. Our Innovation agenda continues to be focused on making the life of operators easier while making big data applications highly reliable and predictable at scale. Our upcoming release, Unravel 4.4 is packed with new innovations and capabilities to further address these needs for operating teams and application owners. Some of the highlights of this release include the following:

Drive Operations efficiency with Automation

Unravel’s roots goes back to research done by Dr Shivnath Babu (co-founder and CTO) at Duke University on self-tuning systems. With the new “Auto-Tuning” capabilities and “Sessions” frameworks, Unravel can now tune jobs automatically to a desired goal (e.g speedup, SLA, resource efficiency etc.). We use various AIOps optimization techniques under the hood to converge operation metrics to recommendation and actions and deliver on a business outcome. See preview demo View Preview Demo here!

Be Proactive with Capacity Planning

Big Data clusters tend to accumulate lots of data and Operators are constantly under pressure to add more capacity to meet the demands. Traditional approaches to capacity planning are very labor intensive and further exacerbated via ad-hoc Excel models and tribal knowledge of cluster administrators. With the new Capacity Planning reporting, Unravel provides an intuitive and data-driven approach to big data visualization and allows you to predict (with a degree of confidence backed by our AI/ML algorithms encompassing millions of data runs) the growth of the cluster and make an informed decision around capacity and resource requirements..

Big Data Applications at Scale

Unify performance monitoring with new NoSQL Support

Big Data applications leverage a myriad of systems such as Hadoop, MPP SQL, Spark, Kafka, and NoSQL as part of their application architecture. Unravel’s goal is to provide a unified, full-stack perspective for the complete application. In this release, we are extending our capabilities to NoSQL and specifically “HBase”. In the first release, we provide an APM centric view for HBase usage and help identify anomalies and outlier issues (e.g table/region hotspotting) that could be adversely affecting the application, while providing remediation techniques.

new NoSQL Support

Improve Application Performance with Cluster Optimization

Unravel is unique in providing actionable insights/recommendation on a per application basis. We were first in the industry to pioneer the approach where we provide tuning recommendations of precise configuration parameters. Working with many of our enterprise customers, we have now extended this capability to provide configuration tuning recommendation at the entire cluster level. With these global recommendations, cluster administrators can now improve the performance of all jobs/applications running on the cluster.

Cluster Optimization

New actionable insights with Small Files Analysis

Hadoop and small files do not go hand in hand!. Very many small files results in significant resource utilization and also tend to put a lot of pressure on the NameNode causing performance issues that affect the overall cluster. With this new report in Unravel, we provide a comprehensive view of all directory with small files and various statistics around them.

The above just highlights some of the key features in Unravel 4.4. This release also includes several other updates including enhanced SQL recommendations (for Impala), Auto Actions support for Impala and Kafka, revised dashboard for Kafka and many many UI fixes and enhancements!

We look forward to you trying the latest version of Unravel and providing us your feedback on our big data management platform.

Register for our webinar on webinar on Wednesday, Sep 5th. Where I will be coving all these new capabilities Register now.

If you also plan to attend #StrataData New York this year? Visit us at booth #940

Bala