Elastic unveils multimodal indexing, agent upgrades, faster OpenTelemetry storage and CI/CD detector

Elastic announced four core platform deliveries on June 2, 2026 that aim to tighten retrieval, telemetry and agent workflows: a multimodal Elastic Inference Service model, general availability of Elastic Agent Builder with new context features and connectors, a redesigned metrics engine tuned for OpenTelemetry, and an open‑source CI/CD pipeline detector from Elastic Security Labs. The company framed these updates as improvements to multimodal retrieval, large‑scale agent context fidelity and observability performance.

The Elastic Inference Service (EIS) now hosts jina-embeddings-v5-omni, a multimodal embedding model that places text, images, video and audio into a single Elasticsearch index. The model supports nearly 100 languages and is available in two sizes — small and nano-enabling a single retrieval layer for mixed media workloads without separate indexes for each modality. Elastic Agent Builder reached general availability and adds built‑in skills plus native Microsoft 365 connectors for SharePoint and Drive. Agent Builder’s context management includes offloading, compaction and summarization features designed to preserve accuracy across long conversations, and the release bundles purpose‑built skills for security operations and site‑reliability workflows.

On telemetry and indexing performance, Elastic says its rebuilt metrics engine stores OpenTelemetry (OTel) data at 3.75 bytes per data point and returns queries up to 160x faster than the prior Elasticsearch TSDS. The release also introduces GPU‑accelerated indexing via NVIDIA cuVS (GA), which Elastic reports can deliver a 12x indexing throughput improvement, and DiskBBQ vector indexing, which the company says reduces query latency by at least 3x for queries with restrictive filters.

Elastic positions Elasticsearch as the retrieval layer for agent workflows built on Azure AI Foundry and compatible with Azure OpenAI Service, highlighting hybrid search (BM25 plus vector), reranking and context engineering as core capabilities. For high‑cardinality embedding workloads on Azure, Elastic argues the indexing and query improvements translate into lower latency and cost in production deployments.

Developer and operational updates target real projects: a fluent, type‑safe ES|QL query builder for TypeScript and JavaScript (shipped in April 2026) aims to reduce runtime query errors, and Agent Builder includes five security‑ops and five site‑reliability skills. The CI/CD pipeline detector from Elastic Security Labs has been open‑sourced to help teams catch malicious GitHub Actions and Azure DevOps workflows before they reach production.

Taken together, these releases simplify multimodal retrieval by consolidating media into one index, keep agents accurate at scale through context management and skills, and improve observability economics and speed for OTel workloads. Builders on Azure should evaluate EIS model sizing, GPU indexing paths and Agent Builder connectors to measure latency, cost and context fidelity for their own agent and telemetry pipelines.