Amazon Web Services Inc. today made its application observability offering generally available, about six months after its launch in preview.

Amazon DevOps Guru is a machine learning-powered service that helps to detect operational issues for applications, generating reports and notifications and providing insights and recommendations for developers to remedy those problems.

The service is a fully managed offering that works by analyzing logs, metrics and events across 25 AWS resources in order to identify anomalous application behavior such as increased latency, error rates and resource constraints. The idea is to remedy these issues before they can cause outages or service disruptions, Amazon explained.

When an issue is identified, the service automatically sends out an alert together with a summary of related anomalies, the likely root cause and additional context about how and when it occurred. Where possible, Amazon DevOps Guru will also suggest possible ways to fix the issues it finds.

Amazon DevOps Guru was announced in December during the virtual AWS re:Invent 2020 event by AWS Chief Executive Andy Jassy. He said at the time that DevOps Guru relies on machine learning informed by years of operational data from the company’s own applications and services.

The service is designed to complement Amazon CodeGuru, which was launched in 2019 and helps developers find issues in their code before it gets deployed. DevOps Guru does much the same for applications that are already up and running.

The service is accessible through an integrated dashboard focused on an insights page that lists all of the anomalies it discovers. These reports are presented with contextual information and recommendations and can be either reactive, highlighting existing issues, or proactive, identifying issues that may occur in future.

A reactive insight would alert developers to a sudden increase in latency within a Lambda function, for example, while a proactive insight would notify teams of an expected increase in latency resulting from increased memory use of that function. Alerts can be sent through various channels, including SMS texts, Slack or a similar service to warn teams that they need to investigate something.

Amazon Machine Learning Vice President Swami Sivasubramanian said the service takes advantage of the company’s decades of experience in running the Amazon.com website to improve application availability.

“With Amazon DevOps Guru, we have taken that expertise and built specialized machine learning models to detect, troubleshoot, and prevent operational issues long before they impact customers and without dealing with cold starts each time an issue arises,” Sivasubramanian said. “Amazon DevOps Guru immediately provides customers the benefits of operational best practices we have learned running Amazon.com, and we designed Amazon DevOps Guru to be so simple that turning it on would be an easy choice for every AWS customer.”

Amazon DevOps Guru enables Amazon to compete in an emerging application observability market that’s dominated by companies such as Splunk Inc., Sumo Logic Inc. and Datadog Inc.

Constellation Research Inc. analyst Holger Mueller told SiliconANGLE that observability tools are key for enterprises because they make it possible to automate software development processes, enabling “self-driving” or “autonomous” software operations.

“Software development needs to move from archaic manual processes to an automated way of building, implementing and operating software, to enable the long-term vision of Automated Software Operations, or ASO,” Mueller said. “Today Amazon is pushing the yardstick further in the direction of ASO with the general availability of Amazon DevOps Guru. It promises to achieve higher developer velocity and get code assets into production faster and more reliably.”

Amazon said Amazon DevOps Guru integrates with third-party developer tools such as Atlassian Corp. Plc.’s Opsgenie and PagerDuty Inc.’s incident response platform. It also plays nicely with Amazon’s own Systems Manager offering that provides visibility into company’s use of Amazon’s cloud infrastructure.

Early adopters of Amazon DevOps Guru, not unexpectedly, heaped praise on the new offering. Indian information technology services provider HCL Technologies Ltd. said the service has led to a big reduction in the amount of time its teams spend on resolving operational issues.

“With the insights Amazon DevOps Guru provides, our teams can now quickly find issues without having to start from scratch trying to root cause problems,” said Anchal Gupta, HCL Technologies’ senior technical lead of DevOps. “Our IT team has significantly reduced our mean time to recovery, and they are saving hours upon hours of time resolving issues.”

Amazon said there are no monthly or service-level fees to use DevOps Guru, with customers only paying for the data the service uses. Amazon DevOps Guru is available now in the US East (N. Virginia), US East (Ohio), and US West (Oregon), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Ireland) and Europe (Stockholm) regions, with further availability to come in the months ahead.

Image: AWS