OpenSearch Project Blog

Learn how community feedback transformed OpenSearchCon's Search track into a dedicated home for Apache Lucene practitioners. Discover the new Search & Apache Lucene track at OpenSearchCon North America 2026, September 22–24 in San Jose.

Learn how the memory retention policy in OpenSearch automatically manages the lifecycle of agentic memory, controlling storage growth while preserving specific memories.

The OpenSearch Observability TAG's first year: 10+ orgs shipped OTLP ingestion, Prometheus metrics, an open-source Observability Stack, and AI agent observability.

AI agents generate 10x the telemetry, and proprietary per-GB pricing punishes you for it. See how the open-source OpenSearch Observability Stack unifies logs, traces, and agent telemetry without vendor lock-in.

At OpenSearchCon Europe 2026, Bobby Mohammed (AWS) and Daniel Wrigley (OpenSource Connections) introduced Agentic Relevance Tuning (ART), a framework that uses specialized LLM agents to fully automate the search relevance lifecycle in OpenSearch.

Connect OpenSearch to ML models hosted on private infrastructure. Configure ML Commons connectors for VPC-hosted endpoints, private SageMaker models, and internal inference servers without exposing services to the public internet

At OpenSearchCon Europe 2026, Unnati Mishra and Akshat Khanna (Angel One) argued that privacy in search must be handled during the ingest phase rather than retroactively through access controls.

Scale Kubernetes workloads using OpenSearch query results as KEDA signals. Count 429 errors or compute p95 latency directly from your logs and traces. No extra pipeline needed.

At OpenSearchCon Europe 2026, Amazon's Shubham Chaudhary shared how bypassing Java's JVM limitations to implement native SIMD architecture achieved a 60% reduction in Lucene vector search latency.

The OpenSearch Agent Skills Hackathon is officially underway! Running from July 15 to August 17, 2026, this event offers U.S.-based developers a chance to build production-ready agentic skills, compete for a $5,000 prize pool, and potentially showcase their work at OpenSearchCon North America. Join the developer community today to start building and get your code in front of the OpenSearch maintainers.

Celebrating five years of the OpenSearch Project, true open source governance under the Linux Foundation has proven to be the most sustainable model for enterprise data infrastructure. Backed by over 2 billion downloads, OpenSearch delivers superior predictability, security, and native AI vector capabilities compared to restrictive "source-available" alternatives.

As vector search workloads scale, retrieving vectors from search results becomes a bottleneck. Vector retrieval using doc values in OpenSearch 3.7 bypasses the expensive _source vector reconstruction path, delivering 5.5x faster end-to-end retrieval and 14.7x faster server-side performance.

The OpenSearch Ambassador Program now accepts applications year-round. Apply today to join our next cohort—reviews begin August 7, 2026.

At OpenSearchCon Europe 2026, IBM's Dom Couldwell argued that truly relevant enterprise search requires design based on user intent (query, journey, and product intent) rather than just the literal keywords typed

Deploy the OpenSearch Observability Stack—a single OpenTelemetry-native platform unifying metrics, traces, and logs—and follow a step-by-step root cause analysis from error spike to failing log line, without switching tools.

Data Prepper 2.16 adds end-to-end metrics with a pull-based Prometheus source and OpenSearch-TSDB, experimental OpenSearch pull-based ingestion, other improvements to the OpenSearch sink, and more.

The rise of agentic AI has sparked a debate on whether open-source infrastructure is becoming obsolete. However, as David Nalley of AWS highlighted at OpenSearchCon India 2026, the economics of token consumption make recreating established infrastructure from scratch a waste of resources. Rather than reinventing foundational software, AI agents should utilize proven, open-source building blocks like OpenSearch. Open source provides essential "known good" design patterns and fosters "communities of problems," where human collaboration continues to drive innovation in ways that raw compute power cannot replace.

At OpenSearchCon Europe 2026, Fernando Rejon Barrera from Zeta Alpha challenged standard vector search benchmarks by evaluating Lucene’s HNSW, Faiss, and jVector under realistic production workloads.

Learn how vector search in OpenSearch eliminates zero-result pages in e-commerce by acting as a semantic fallback when lexical query relaxing falls short.