Pinned Tweet
OpenSearch 3.10 is now available, with updates across search, observability, and cluster operations.
Here’s what’s new:
- AI agent observability: Trace multi-turn conversations down to LLM and tool calls
- Simpler hybrid search: Configure score fusion directly in queries without a search pipeline
- Faster PPL queries: Skip irrelevant time-range indexes and query across clusters
- Expanded ingestion: Pull data at rest directly from object storage, including Amazon S3, Google Cloud Storage, and Azure Blob Storage
- Improved resiliency: Restore Alerting and Index State Management state from snapshots more reliably
Explore the full release to see what else is new in OpenSearch 3.10: bit.ly/4iNzsAK
Paris, let's talk search relevance and agentic AI!
Join the OpenSearch Project Paris community on November 19 from 7–10 PM at STATION F in Paris for an evening of technical deep dives, live demos, and community conversations.
On the agenda:
• Vector Query Relaxing: Reducing zero-result searches with semantic retrieval, ANN, and quantization.
• Agentic Search: Improving search personalization and performance with locally hosted SLMs, fine-tuning, hybrid search, and reranking.
Register: bit.ly/4BXTLDb
1
136
OpenSearch Data Prepper 2.17 is here, introducing Splunk HEC compatibility so existing forwarders and HEC clients can send data to OpenSearch without client modifications beyond updating their destination.
Other updates include:
- Multi-feed RSS ingestion with independent failure handling
- JDBC coordination using PostgreSQL or MySQL for multi-node scaling
- Improved gRPC connection management for better OpenTelemetry load balancing
A big thank you to all the contributors and maintainers who made this release possible!
Explore what's new in Data Prepper 2.17: bit.ly/4kaagpd
3
226
At OpenSearchCon North America, lead analyst Rob Strechay of @VentureBeat shares what stood out to him: growing community involvement, major companies contributing code, and the potential for OpenSearch in agentic AI.
1
151
At KubeCon + CloudNativeCon NA, OpenSearch Ambassador Brian Graf shared how to index Kubernetes objects in OpenSearch to make cluster data searchable and easier to explore.
A few moments from his Tip Talk, Your Cluster Is a Dataset, at the Starlight Summit Zone.
Missed it? Stop by the OpenSearch booth to continue the conversation!

ALT A person is speaking into a microphone at a conference, gesturing with one hand. Behind them is a large screen displaying a presentation slide titled "Multiple ways to get the answer." Audience members are seated in front, listening attentively. The setting is a spacious, indoor convention center.
1
1
197
Great open source maintainers do more than write great code.
OpenSearch maintainers share what the role really takes, their journeys into maintainership, and how to grow the next generation of open source leaders.
Read the blog from OpenSearch Ambassador @KrisFreedain: bit.ly/4z8zxUF

ALT The image features a promotional graphic for a blog post by Kris Freddain. It includes the OpenSearch logo and details about a panel discussion titled "Inside open source maintainership: Realities, myths, and community growth." Next to this text is a photo of four individuals sitting on stage, engaged in conversation, with a backdrop showing "OpenSearchCon North America 2026.
1
151
Since joining the Linux Foundation, OpenSearch has grown from 700 million to 2.4 billion downloads, with contributor diversity increasing 31% in the past year.
In a conversation with @TFiR_io, OpenSearch Software Foundation Executive Director Bianca Lewis discusses how vendor-neutral governance is strengthening community trust, expanding contributions, and giving enterprises greater confidence in open source infrastructure.
Read here: bit.ly/4zyGoHX
117
At KubeCon + CloudNativeCon North America, OpenSearch Ambassador Brian Graf is delivering a Tip Talk, “Your Cluster Is a Dataset: Indexing Kubernetes Objects in OpenSearch.”
📍 Tomorrow, November 10 at 1:15 PM MST | Starlight Summit Zone, Solutions Showcase (Hall B)
Brian will also be at the OpenSearch booth throughout the week. Stop by and bring your questions!
Session details: bit.ly/48IdDN3
2
165
If you're at Observability Day today in Salt Lake City, catch a live demo of agent observability with OpenSearch and OpenTelemetry's GenAI semantic conventions.
See how developers can trace agent behavior, investigate failures, and gain better visibility into complex AI workflows.
You don't need to know every line of code to become an OpenSearch maintainer.
For Srikanth Padakanti, it started with fixing a flaky test. Three and a half years later, he's an OpenSearch Data Prepper maintainer, contributing to everything from ingestion and observability to an RFC for indexing data directly from object storage without Kafka in the middle.
Srikanth shares how community feedback, code reviews, and a willingness to keep contributing shaped his journey from first-time contributor to maintainer.
1
202
OpenSearch is becoming far more than a search engine.
As the retrieval layer for enterprise AI, OpenSearch helps find the right information, enforce access permissions, and provide relevant context before an AI model responds or an agent takes action.
Hear from Stephanie Walter, Practice Leader for the AI Stack at @HyperFRAME_Res
Research: bit.ly/4A7qHXS
OpenSearch production adoption has nearly doubled. AI is a big reason why.
In a conversation with @TFiR_io, OpenSearch Software Foundation Executive Director Bianca Lewis discusses what's driving adoption, why enterprises are rethinking fragmented data infrastructure, and where OpenSearch fits.
Read here: bit.ly/4eBjOpN
1
1
194
An AI agent finding the right document isn't the same as an AI agent making the right decision. @IBM's Pushkala Pattabhiraman uses a simple customer return scenario to unpack retrieval, memory, and evidence in agentic systems.
Watch the keynote: bit.ly/48uzRC3
1
3
172
Traffic spikes are a brutal test of search infrastructure. At OpenSearchCon North America 2026, member company @Uber's Mingmin Chen shared how OpenSearch supports rides, maps, and Uber Eats, and what it takes to keep search resilient at scale. The deeper engineering story is in load shedding and improvements contributed back to the ecosystem.
Watch: bit.ly/4kcPEgb
2
211
OpenSearch 3.9 introduces a native engine for neural sparse search that moves the forward index off the JVM heap and runs scoring in C++. In benchmarks, queries ran 6–32× faster than the Lucene engine at the same recall.
Explore the benchmarks and setup: bit.ly/4BvxGvz
Kerala, your OpenSearch community starts here.
The first OpenSearch Kerala User Group meetup is happening October 31 in Kochi!
From search and observability to AI, vector search, and Kubernetes, come learn from real-world implementations, exchange ideas, and meet the builders behind the tech.
October 31 | 9 AM–1 PM IST
Find your community. Build something together.
RSVP: bit.ly/4zgyAKR
1
204
How do you make search work for one of the world's most demanding scientific environments?
Member company @CERN's Sokratis Papadopoulos takes us through ten years of search infrastructure, from the Large Hadron Collider to security and monitoring workloads. There's also a look at using an OpenSearch MCP server to interact with monitoring data.
Watch the session: bit.ly/4zZZdUd
106
From vector search to observability, OpenSearch is doing a lot more than traditional search.
At OpenSearchCon NA, we caught up with Dmitry Kan and Tony Piazza from Aiven to discuss OpenSearch's expanding use cases, capabilities, and ability to support workloads at scale.
Thanks to Aiven for sponsoring the event, delivering a keynote, and supporting the OpenSearch community.
Watch their keynote here: bit.ly/4A2eDrb
1
155
OpenSearch 3.9 is here, and Command-K is just the beginning.
OpenSearch Ambassador @dep4b walks through what’s new: smarter trace visualization, unified alerting, faster neural sparse search, improved vector compression, and adaptive concurrency controls.
Less configuration, more visibility, and better performance. See what’s new in OpenSearch 3.9: bit.ly/4iFppxn
134














