OpenSearch Serverless: 5-Second Refresh Interval

Amazon OpenSearch Serverless now supports a 5-second default refresh interval

Editorial position: This change reduces latency for read-after-write scenarios without requiring user configuration.

The news, in plain English

Amazon OpenSearch Serverless now uses a 5-second default refresh interval for search and vector collections, down from 10 seconds. This adjustment lets newly ingested documents become searchable more quickly. The change applies automatically to all new collections in the 24 regions where the service is available.

Technical context

The refresh interval controls how often OpenSearch checks for new or updated documents to make them available for queries. With the previous 10-second default, applications needed to wait at least that long after writing data to guarantee it would appear in search results. The new 5-second setting halves this wait time automatically for all users of OpenSearch Serverless. This affects both standard search collections and vector collections used for similarity search.

In practice: Workloads that ingest data continuously and require near real-time retrieval will see improved latency characteristics without any changes to existing configurations. For example, applications that display user-generated content or log data in real time can show updates faster.

Where this fits in practice

This change primarily benefits applications that depend on quick visibility of recently written data, such as:

• Real-time analytics dashboards
• Chat applications with message history search
• Monitoring systems that alert based on recent events
• Applications that show user-generated content as it is created

Because the change is applied at the service level, no code or configuration modifications are needed for new collections. Existing OpenSearch Serverless collections created before this announcement retain the previous 10-second interval unless explicitly updated.

Adoption path

No action is required for most users. The new refresh interval applies automatically to:

• Newly created search collections
• Newly created vector collections

For existing collections, the refresh interval remains at the previously configured value (10 seconds by default) unless changed through the OpenSearch Dashboards interface or API. To update an existing collection, you would need to modify its settings using the OpenSearch Dashboards UI or the OpenSearch Serverless API.

Watch out: If your application depends on the exact timing of document visibility, verify the refresh interval for each collection, especially when upgrading or migrating. Assuming the default behavior may lead to unexpected latency differences between old and new collections.

Cost and operations

The announcement provides no details on cost implications or operational side effects. Since this is a default configuration change, it does not introduce new billable dimensions. However, reduced refresh intervals may increase the frequency of internal indexing operations, potentially affecting performance under heavy load. Monitor cluster metrics such as indexing latency and search throughput after the change if you operate near capacity. There is no indication that this change affects storage costs or query pricing.

Security and reliability

The source does not discuss security or reliability impacts. The change only affects how quickly data becomes searchable, not how it is stored or protected. Standard OpenSearch Serverless security features—including encryption at rest, IAM authentication, and network isolation—remain unchanged. There is no evidence that this adjustment impacts disaster recovery, availability, or resilience characteristics of the service.

Limits and trade-offs

The fact pack confirms the new default applies to all 24 regions where OpenSearch Serverless is available but provides no information about limits or trade-offs, such as:

• Maximum allowable refresh interval
• Minimum interval enforcement
• Performance implications under high ingestion rates
• Impact on indexing throughput or query latency at scale

Users should test workloads that are sensitive to refresh latency to understand any operational effects. For example, high-volume logging systems or real-time recommendation engines may need to evaluate whether the 5-second interval meets their performance requirements or if a custom setting is needed.

Who should act

• Developers and operators of applications requiring near real-time search capabilities should assess whether the new default meets their latency requirements.
• Teams managing existing OpenSearch Serverless collections should verify the current refresh setting and update it if needed.
• Performance engineers should monitor query latency and ingestion throughput after the change, especially in high-volume environments. This is particularly important for applications that approach the service's operational limits.
• Architects designing new systems that depend on immediate visibility of written data may rely on this default without additional configuration.

Bottom line

The reduction to a 5-second default refresh interval in OpenSearch Serverless improves read-after-write latency for applications that ingest data frequently. The change requires no action from users and applies automatically to new collections. Existing collections retain their previous settings unless explicitly modified. As with any configuration change, test in a non-production environment before applying it to critical workloads. The lack of published performance data or cost implications means users should monitor their specific use cases after deployment.

Sources

Amazon OpenSearch Serverless now supports a 5-second default refresh interval