DevOps Gets a New Analytics Tool
Among the growing number of uses for machine learning-based analytics is providing insight into how the latest software release in a continuous delivery cycle affects overall IT operations. With that approach in mind, a software-as-a-service vendor has released an analytics platform for DevOps teams designed to provide greater visibility into the operational impact of continuous software delivery.
Perspica Inc., a startup based on San Jose, Calif., this week said it has integrated DevOps stacks into its operations analytics engine that includes root cause and predictive analytics for application infrastructure stacks along with monitoring, troubleshooting and diagnostic tools.
The number of DevOps platforms supported by the tool also has been expanded to include Amazon Web Services (NASDAQ: AMZN), Collectd, Docker containers, Flume, Kafka, Loggly, MongoDB, OpenTSBD, Redis and Zookeeper, the company said Thursday (June 16).
Perspica's "Incident Replay" platform is positioned as a "time machine" for tracking performance data, logs and topology. The approach allows DevOps teams to "play back" the impact of a code release on application performance and pinpoint the root causes of performance shortfalls.
"The continuous delivery model means developers have to move quickly, and they are hampered because they often can't see how the new software release is affecting operations," Perspica CEO Dan Maloney noted in a statement. "This can lead to application performance issues and even outages."
The replay capability is designed to give DevOps teams greater visibility into the impact of a new software release across the "entire development stack," the company claimed.
The operations analytics engine also is touted as allowing developers to visualize what's going on across their IT infrastructure based on performance data, logs and changes in topology. Current operations also can be compared with past performance. The result, the startup claims, is an expanded ability to detect problems when rolling out new code while limiting the impact of outages and fixing incidents as soon as they occur.
The startup's big data approach is based on the notion that hyper-scale infrastructure generates millions of "events" and performance metrics every second. Perspica and other operations analytics specialists promote their ability to ingest all this IT performance data and analyze it in real time.
The startup's approach leverages machine-learning techniques to understand network topology by determining the relationships between components and how the performance of one component affects others in the network. That automation approach is then extended to track millions of performance metrics across applications to determine what a DevOps team considers "normal behavior."
Performance anomalies are detected based on any sudden changes in the behavior of network objects, at which point a root cause analysis feature kicks in to offer "intelligent recommendations" on how to handle an incident.
Finally, a predictive analytics tool is used to spot early indications of an outage and recommend actions to prevent performance degradation.
The DevOps analytics engine is designed to operate across hybrid clouds, notes Perspica, which has attracted venture funding from several investors, including March Capital and The Fabric.
