Press "Enter" to skip to content
Credit: Starline on Freepik

Amazon Brings Generative AI to OpenSearch

TLDR

  • AWS has integrated its AI assistant, Amazon Q Developer, into OpenSearch, allowing developers to diagnose system issues by asking plain-language questions instead of writing complex queries.
  • The integration helps developers resolve incidents faster and frees them to focus on innovation, offering what AWS calls a “double benefit” of improving both business outcomes and developer efficiency.
  • Available in the free tier, the integration is easy to use and includes safety measures like automated reasoning and query transparency to minimize AI errors and hallucinations.

AWS is integrating its AI coding assistant, Amazon Q Developer, into Amazon OpenSearch service, a search and analytics engine that helps developers quickly find out what’s happening in their data.

Launching this week, the addition will enable developers to do searches in natural language.

With the integration, “engineers and developers can resolve operational incidents much more quickly” to reduce the impact to business, said Mukul Karnik, general manager and director of search services at AWS, in an interview with The AI Innovator. At the same time, it frees them to do more innovative tasks – “a double benefit that helps the business outcome but also improves developer productivity.”

Amazon Q Developer is an AI assistant that helps developers write, fix, understand code faster and also assists in troubleshooting problems in their systems. Developers just ask Q Developer questions in plain language instead of having to dig through logs or write complex queries to get answers.

For example, developers can prompt Q Developer to “show me all 500 errors” and “visualize the traffic in the last 30 minutes” from a specific source to get to the root cause quickly, said Karnik.

Organizations often see the following problems that Q Developer in OpenSearch aims to resolve:

  • Overwhelming volume and complexity of observability data
  • Lack of specialized knowledge among front-line engineers
  • Complex query languages and interfaces that create friction in troubleshooting

The rise of AI agents will make Q Developer in OpenSearch even more compelling, and organizations will “strongly embrace” the service, AWS predicts.

Why OpenSearch?

The decision to integrate Q Developer within OpenSearch has strategic benefits. OpenSearch, a fork of the open-source Elasticsearch project, is widely used by AWS customers for log and operational analytics. According to Karnik, tens of thousands of AWS customers rely on OpenSearch daily.

While the Q Developer capability on OpenSearch is largely targeted towards developers, the service can be used by less technical staff to build visualizations and dashboards for analysis.

“You can ask questions. You don’t need to know the query language. You don’t need to know the visualization framework, and you can just build natural language-based visualizations and add it to your dashboards,” Karnik said.

Since AWS launched the OpenSearch project in 2021 – later donating it to the Linux Foundation – the company has committed to building core capabilities in the open. However, the Q Developer integration itself is not open source.

“The underlying framework and the capabilities are actually built into open source,” said Karnik. While Q Developer is proprietary, open-source users can build their own versions.

For existing AWS customers, the integration is designed to be seamless. Developers only need an OpenSearch service managed instance and an application in the OpenSearch UX service to start using Q Developer. The capability will be included in the free tier, Karnik said.

Guardrails against hallucinations

As for how to manage hallucinations, Karnik said developers can look at the underlying query that was used to generate the answer or visualization as a way to double-check the output. They can also edit the query if they believe the output is getting something consistently wrong.

AWS is taking measures to ensure safety and reliability as well. “We’re also using automated reasoning in that system to be able to make sure there are no hallucinations or to minimize that,” Karnik said.

Karnik sees strong adoption in the financial services sector and among software and tech firms. These industries are both high-volume data consumers and early tech adopters.

“We think this capability is going to help engineers and developers reduce the amount of toil or effort they put into operational issues because most engineers and developers don’t really enjoy (doing this) – and so it gives them that time back to do more productive and interesting things,” he said.

“Another big outcome is that it reduces the time for resolution of these incidents, which matters a lot for businesses because if you think about it, time is dollars for most businesses,” Karnik added.

Author

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *