Introducing memory retention for agentic memory in OpenSearch
OpenSearch

Introducing memory retention for agentic memory in OpenSearch


Summary

OpenSearch 3.8 introduces an experimental memory retention feature for agentic memory containers, allowing users to set age-based or count-based limits to automatically delete stale sessions, long-term memory, and history. This functionality helps reduce storage costs and maintains agent performance by preventing context degradation from outdated data. Additionally, the feature supports pinning important memories to prevent deletion and allows for configuring cluster-wide defaults for easier management across multiple containers.
Read the Original Article

This article originally appeared on OpenSearch.

Read Full Article on Original Site

Related Articles

Online index migration and shard scaling in OpenSearch with the AOSC plugin
Online index migration and shard scaling in OpenSearch with the AOSC plugin

Arpit Singla Aug 19, 2026 2 shared categories

The real cost of your observability stack
The real cost of your observability stack

Shenoy Pratik Gurudatt Aug 7, 2026 2 shared categories

Scaling Kubernetes Workloads with the OpenSearch KEDA Scaler
Scaling Kubernetes Workloads with the OpenSearch KEDA Scaler

Karsten Schnitter Jul 22, 2026 2 shared categories

Retrieve vectors 5x faster with docvalue_fields in OpenSearch
Retrieve vectors 5x faster with docvalue_fields in OpenSearch

Navneet Verma Jul 15, 2026 2 shared categories

Single pane of glass for all your telemetry: The OpenSearch Observability Stack
Single pane of glass for all your telemetry: The OpenSearch Observability Stack

Shenoy Pratik Gurudatt Jul 10, 2026 2 shared categories

Popular from OpenSearch