Amazon Bedrock Managed Knowledge Base adds automatic sync scheduling

Bedrock Managed Knowledge Base is AWS's fully managed retrieval-augmented generation service, and stale data has been a persistent operational headache: without scheduling, teams had to script their own sync triggers or manually kick off ingestion whenever a source changed. The new automatic sync scheduling covers all native data source connectors and supports daily, weekly, or monthly cadences, directly targeting data-freshness and the operational overhead of keeping enterprise RAG current.

The practical impact is reduced glue code and fewer stale-answer incidents for production agents — a meaningful reliability improvement for customers running knowledge bases against frequently changing sources like ticketing systems, catalogs, and internal wikis. It also complements AWS's parallel push on AgentCore memory lifecycle management, both aimed at keeping long-running agents grounded in current, high-quality context.

Competitively, this is part of a broad Bedrock feature wave on September 4 that also included a native ServiceNow connector, user-managed 3LO auth setup, and federated MCP data-access patterns — AWS steadily removing friction from enterprise agent deployment. Rivals Azure (Cosmos-backed agent memory) and Google are pushing similar persistence and freshness capabilities the same week.

This is an incremental, plumbing-level update rather than a headline model launch, but it addresses a real recurring pain point. Readers running Bedrock RAG should watch how scheduling interacts with connector-specific rate limits and incremental-vs-full sync behavior, and whether the monthly/weekly/daily granularity is fine enough for high-velocity data sources that may need near-real-time refresh.