Preview: Machine Learning-Based Amazon DevOps Guru

Amazon Web Services (AWS), has launched its DevOps Guru managed operation service. This uses machine learning to detect operational problems and recommends specific actions to correct them.
Amazon DevOps Guru is now available in preview. It collects and analyzes application metrics, logs and events to identify abnormal operating patterns. This includes things like database I/O overload, memory leaks and under-provisioned compute capacities. AWS announced DevOps Guru in its ongoing re-Invent virtual conference.
AWS states that when DevOps Guru detects abnormal app behavior that could lead to service disruptions or outages, such as approaching resource limits, code and config changes, it alerts developers with details like the issue details, a timeline of events, and recommendations for remediation. It does this through the Amazon Simple Notification Service, (SNS), and partner integrations with companies such as Atlassian’s Opsgenie or PagerDuty.
Amazon DevOps Guru also “spotlights” things like database I/O overutilization, under-provisioned compute capacities, and memory leaks.
Swami Sivasubramanian (head of the AWS Machine Learning) stated that customers have asked him to continue adding services in areas where he can use his own expertise to improve application availability and draw on the years of experience that he has gained running Amazon.com. “With Amazon DevOps Guru we have taken our experience to build specialized machine learning models that help customers identify, troubleshoot and prevent operational problems while providing intelligent recommendations when they do occur.” This allows teams to immediately benefit the operational best practices that Amazon.com has learned, saving customers time and effort in configuring and managing multiple monitoring systems.
According to AWS, the Amazon DevOps Guru service analyzes system and app data in order to identify anomalies. It also groups this data into “operational insight” which includes anomalous metrics, visualizations and recommendations for remediation. The service also correlates and groups related metrics such as Web app latency spikes or running out of disk space, bad deployments, and memory leaks.
This results in fewer redundant alarms and more help for users who are dealing with high-severity issues. The Amazon DevOps Guru console allows users to view configuration change history and deployment events. This can be combined with system and user activity to create a prioritized list that may indicate the possibility of an operational problem.
This service also provides intelligent recommendations and integration with AWS Systems Manager runbook and collaboration tooling. It allows users to manage their infrastructure and applications more efficiently.
The company stated that Amazon DevOps Guru is a partner with Amazon CodeGuru which provides intelligent recommendations and identifies the most costly lines of code in an application. It also provides machine learning-powered developer tools that provide intelligent recommendations and identify the most expensive lines of code in an application. This tool, powered by machine learning, provides users with the automated benefits that machine learning can bring to their operational data so that developers can improve the availability of their applications.

Preview: Machine Learning-Based Amazon DevOps Guru
Scroll to top