mastodon 4.7.3
The OpenSearch Data Prepper maintainers are happy to announce the release of Data Prepper 2.15. This version adds the ability to ingest data from Apache Iceberg, making it easier to keep OpenSearch in sync without custom pipelines. It also extends Prometheus support with a remote-write source and the ability to send data to open source Prometheus.
Big kudos to the community for keeping this project moving forward. Read the full blog here: https://bit.ly/4sotUhD
Hundreds of contributors just shipped something worth talking about. Launched today, OpenSearch 3.6 brings agent-powered tools to help you:
- Build production-ready search applications in minutes, not days
- Launch a full-stack observability solution with one command
- Tune search relevance through conversation instead of code
- Monitor distributed applications in real time with Application Performance Monitoring and built-in root cause analysis
A tour of OpenSearch 3.6’s new agentic features.
Join #OpenSearchAmbassador Eric Pugh, Doug Turnbull, and Mingshi Liu for a fast look at how agentic search is improving visibility, relevance, and migrations.
Apr 8, 10 AM PDT | Virtual | Free https://maven.com/p/aee87c/open-search-s-new-agentic-features
#OpenSearch #AI #AgenticAI #Search #Observability #OpenSource
#OpenSearch v3 marks a major step forward for open source search, introducing key innovations in neural search, AI-driven experiences, and performance.
At #FOSDEM 2026, Dotan Horovits (#OpenSearchAmbassador) and Aswath Srinivasan highlight the major features of the 3.x releases and their impact on modern search applications: https://www.youtube.com/watch?v=y6vYzD-_EZg&t=1555s
Feature Spotlight: Anomaly Detection in OpenSearch automatically spot unusual patterns in your logs and metrics with ML-powered insights. Stay one step ahead of potential issues and optimize your systems proactively. https://opensearch.org/platform/anomaly-detection/
ICYMI - The latest OpenSearch 3.5 release is a major step forward for anyone building agentic AI applications - from smarter agentic applications with conversation memory to improved agent performance with hook-based context management and more!
It’s never been easier to bridge the gap between your data and your LLMs.
See what’s new in 3.5 🧐 https://opensearch.org/blog/opensearch-3-5-is-live/
🚀 Optimizing FP16 Vector Search in #OpenSearch: A 310% Throughput Breakthrough
The result?
📈 Up to 310% throughput improvement
⚡ Dramatically reduced query latency
Learn more! 📚 https://opensearch.org/blog/accelerating-fp16-vector-search-performance-using-bulk-simd-in-opensearch-3-5/
Open source isn't just about the code; it’s about the freedom to innovate without lock-in.
Whether you’re submitting a PR, improving documentation, or helping a peer in the forums, you are the heartbeat of this project.
Want to get involved? Check out our ""Good First Issues"" and start building with us today https://opensearch.org/community/#:~:text=Whether%20you%20code,or%20triage%20meeting!
#OpenSearch named as a ""Leader and Fast Mover in Innovation/Platform Play quadrant"" in the 2025 GigaOm Radar for Vector Databases.
The project was recognized for 40-60% engancement in search relevance and 30-50% reduction in infrastructure costs.
Read on to learn about OpenSearch's robust vector search platform: https://opensearch.org/announcements/opensearch-named-a-leader-in-gigaom-radar-for-vector-databases-as-research-shows-hybrid-search-becomes-critical-for-ai/
Is modern search complexity stalling your development?
Introducing OpenSearch Launchpad: The AI assistant that takes you from 0 to a running search application in minutes.
Launchpad automates architecture, models, and indexing so you can skip the setup headaches and focus on building. No OpenSearch expertise required.
Ready to launch? Read the full story and get started in your IDE: https://opensearch.org/blog/introducing-opensearch-launchpad-from-requirements-to-a-running-search-application-in-minutes/
The #OpenSearch Project community is launching in Rome - March 30!
Join developers, data engineers, and platform engineers to explore how OpenSearch powers search, analytics, and observability at scale through technical talks and real-world architectures.
Clicca qui per l’agenda e i dettagli: https://opensearch.org/event/rome-user-group-march-30-2026/
AI is no longer a pilot. It is a mandate. Agentic AI is unlocking real autonomy and real risk, with attacks increasingly targeting access control. The choice is simple build walls or collaborate. As Carl Meadows, Governing Board Chair at the OpenSearch Software Foundation, highlights that trust comes from transparency, strong data foundations, and community-driven innovation. https://sdtimes.com/ai/to-create-trustworthy-agentic-ai-seek-community-driven-innovation/
Search modernization is no longer a nice to have.
At the LF Member Summit, OpenSearch Technical Steering Committee member Pallavi Priyadarshini and Marketing Outreach Committee chair James McIntyre got candid about the forces reshaping enterprise search and why open source is uniquely built for this moment.
Read on at: https://opensearch.org/blog/search-modernization-with-opensearch-from-vector-search-to-agentic-ai/
#OpenSearchCon China started off strong today with a fantastic welcome from Bianca Lewis, the Executive Director of the OpenSearch Software Foundation, followed by insightful sessions from our community members Jon Handler and Yakun LI. We can’t wait to see all the other great sessions ahead!
If you’re there, be sure to catch #OpenSearchAmbassador Kris Freedain on the ground—stop by and say hello!
At NDC London, #OpenSearchAmbassador Dotan Horovits discussed search methodologies, with special emphasis on semantic search, and what the OpenSearch Project provides, fully open source under The Linux Foundation. https://www.youtube.com/watch?v=N2kn4UybFBI&list=PLPFMHjhoDntvdVnYpMIRhi4CRgBwyIAET&index=1
Join OpenSearch and Netapp Instaclustr on March 11 for an evening of networking, great food, and deep dives into the future of artificial intelligence.
NetApp's Carlos Rolo will be breaking down "Agentic Conversation Memory in OpenSearch 3.5" and showing us how to simplify this complex but game-changing technology.
Register now: https://opensearch.org/event/lisbon-user-group-march-11-2026/
With #OpenSearch Security Analytics, the security log data from your critical infrastructure can provide insight into potential risks to your systems, users, confidential data, and applications.
Learn how: https://opensearch.org/platform/security-analytics/
And that’s a wrap for #OpenSearchCon China—what an incredible experience from start to finish. Grateful for the insightful conversations, inspiring talks, and the chance to connect with such a passionate community. Looking forward to what’s next! #OpenSearch
Why choose between keyword search and semantic understanding? With Hybrid Search in #OpenSearch, you get the best of both worlds. 🧠
Improve your user experience by making your search ""smarter"" without sacrificing performance. 🔗 https://opensearch.org/platform/vector-engine
📊 Data Prepper 2.14: OpenTelemetry-powered APM service maps and production-ready Prometheus support
The #OpenSearch Data Prepper maintainers announced the release of Data Prepper 2.14! Check out the blog to learn how this version expands support for observability uses cases with new application performance monitoring service map and improved Prometheus support.
Vector search for modern applications.
OpenSearch keeps advancing performance and reliability like the agentic memory support for self-learning applications shown in the recent 3.5 release. Learn more about vector search in OpenSearch: https://opensearch.org/platform/vector-engine/
#OpenSearchAmbassador Dotan Horovits highlights key features from OpenSearch 3.5: smarter AI agents, Prometheus support in Dashboards, new PPL commands, and LLM-powered relevance judgments.
https://www.youtube.com/watch?v=siXu7UNKCgM
Which of these highlights stands out to you? Comment below.
OpenSearch Agent Health is an open source observability and evaluation solution for AI agents. It’s available as a zero-installation NPX tool that delivers three core capabilities in a lightweight (~4 MB) package. https://opensearch.org/blog/opensearch-agent-health-open-source-observability-and-evaluation-for-ai-agents/
OpenSearch believes open collaboration drives measurable innovation ✨
To better understand the business value behind open source, explore the latest ROI report from Hilary Carter, SVP of Research at the Linux Foundation 📈:
https://www.linuxfoundation.org/hubfs/Research%20Reports/2025%20Open%20Source%20ROI%20Survey%202.23.26.pdf?hsLang=en
What if data could answer questions in natural language?
Agentic search in #OpenSearch delivers relevant results and accurate queries, showing how conversational search can simplify complex information retrieval across both structured and unstructured data. https://opensearch.org/blog/evaluating-agentic-search-in-opensearch/
Your AI agent can fail silently—calling the wrong API, hallucinating reasoning, running unnecessary tool calls, and incurring unexpected costs, only to tell the user: “I can’t help with that.”
Dotan Horovits, #OpenSearchAmbassador, walks through OpenSearch Agent Health in this video, showing how it lets you trace every decision, evaluate performance automatically, and integrate with any agent framework. One NPX command gets you started in under a minute. https://youtu.be/MU3tTv4lKtc
The OpenSearch Software Foundation announced that BigData Boutique, OpenSource Connections and Resolve Technology have joined as General Members.
The companies will contribute technical expertise and participate in working groups to support enterprise search, observability and vector database innovation.
Learn how the additions reflect growing global demand for open source search and analytics as organizations adopt hybrid search and AI-driven architectures: https://opensearch.org/announcements/opensearch-software-foundation-welcomes-bigdata-boutique-opensource-connections-and-resolve-technology-as-general-members/
Meet our new #OpenSearchAmbassadors—community leaders dedicated to advancing #OpenSearch, sharing knowledge, and helping others build and innovate together.
Learn more about our ambassadors here: https://opensearch.org/blog/introducing-the-2026-2027-opensearch-ambassadors
Meet the new #OpenSearch code contributors for February! Huge thanks to our community for powering the project—through code, docs, talks, and user groups.
https://forum.opensearch.org/t/new-contributors-february-2026/27925
Discover the power of OpenSearch Dashboards! Visualize your data like never before with intuitive charts, real-time analytics, and customizable layouts. From log analysis to user behavior insights, see what truly matters in your data. https://opensearch.org/platform/opensearch-dashboards/
REST APIs are great. Until they are not.
OpenSearch now supports native gRPC endpoints with Protocol Buffer schemas, giving teams a high-performance alternative to JSON-based REST without having to give up REST entirely.
Both protocols run in parallel. Adopt at your own pace. https://opensearch.org/blog/advancing-opensearch-with-grpc-and-protocol-buffers/
Upcoming #OpenSearch User Groups this week:
- Lisbon - https://opensearch.org/event/lisbon-user-group-march-11-2026/
- Washington DC (First-ever meeting!) - https://opensearch.org/event/washington-dc-user-group-launch-march-12-2026/
- Cape Town (First-ever meeting!) - https://opensearch.org/event/cape-town-user-group-march-14-2026/
For all events, see here: https://opensearch.org/events/
Where is OpenSearch headed in 2026? Our technical roadmap is live, focused on building the industry’s most robust AI-native search and analytics platform.
Search modernization
Observability and analytics
Scalability and resiliency
Community and platform
https://opensearch.org/blog/the-2026-opensearch-roadmap-four-pillars-for-ai-native-innovation/
#OpenSearch #AI #VectorSearch #Observability
#Atlassian migrated critical workloads to #OpenSearch as demand for artificial intelligence (AI) and contextual search increased. Check out how #Atlassian is achieving 99% availability across 300+ clusters: https://bit.ly/46SWekn
