Posted On: Apr 4, 2022

You can now run OpenSearch and OpenSearch Dashboards version 1.2 on Amazon OpenSearch Service. This version includes a new observability interface in OpenSearch Dashboards, improvements to several other features such as anomaly detection, k-NN, and SQL/PPL.

Amazon OpenSearch Service lets you run and scale OpenSearch and open-source Elasticsearch (up to version 7.10) clusters with ease. OpenSearch is a community-driven, open source search and analytics suite, originally derived from Apache 2.0 licensed Elasticsearch 7.10.2 & Kibana 7.10.2. It consists of a search engine, OpenSearch, and visualization capabilities powered by OpenSearch Dashboards.

We launched support for OpenSearch 1.0 on Amazon OpenSearch Service in September 2021, and for OpenSearch 1.1 in January 2022. The support included features such as transforms, data streams, notebooks, cross-cluster replication, and improvements to anomaly detection and alerting. We are now adding support for OpenSearch 1.2 with the following new features:

  • New observability interface: The new observability capabilities allows developers and devops engineers to more easily analyze trace data and log data in a single interface. A Piped Processing Language (PPL) based event explorer helps developers interactively explore log data and visualize the results in simple to configure charts. Developers can save their PPL-based visualization and view multiple saved visualizations on a custom operational panel. Also integrated into the new observability interface is OpenSearch Notebooks. With notebooks, developers can interactively and collaboratively develop rich reports that combine markdown, SQL/PPL queries, and visualizations with support for multi-timelines and live data so that users can easily tell a story. For more information about this feature, please see the documentation.
  • Feature Attribution in Anomaly Detection: You can now provide an attribution ratio for each input feature to help you understand how they contributed to an anomaly. With this data, you can quickly identify the cause of the anomaly. For more information, please see documentation here
  • More efficient k-NN dense vectors: k-NN now has support for Faiss library, allowing you to expand the size of feature vectors. The Faiss library brings efficient similarity search and clustering of dense vectors and allows for search in data sets of vectors in sizes larger than what fits in memory. For more information, please see documentation here.
  • ‘Match’ query support in SQL and PPL: The match query type returns documents that match a provided text, number, date, or Boolean value for a specified field. The Match search query type is now supported in both SQL and PPL query languages. For more information, please see documentation here for PPL and here for SQL.

To learn more about Amazon OpenSearch Service, please visit the product page.