feat: add Bedrock Knowledge Base retriever to amazon-bedrock integration

[PVidyadhar]

feat: add Bedrock Knowledge Base retriever to amazon-bedrock integration

Related Issues

N/A — new feature adding Knowledge Base retrieval to the existing Amazon Bedrock integration.

Proposed Changes:

Adds BedrockKnowledgeBaseRetriever component to the amazon-bedrock-haystack integration, enabling document retrieval from Amazon Bedrock Managed Knowledge Bases.

• Created BedrockKnowledgeBaseRetriever as a Haystack @component with run(query) method
• Returns List[Document] with content, metadata (source URI, score)
• Supports both Managed (recommended) and Vector knowledge base types
• Agentic retrieval (query decomposition + managed reranking) with automatic fallback to standard Retrieve API
• Includes to_dict/from_dict serialization for pipeline persistence
• Required IAM permissions: bedrock:Retrieve, bedrock:AgenticRetrieve
• SDK requirement: boto3 >= 1.43

How did you test it?

• Unit tests included (mocked boto3 calls)
• Manual E2E: Created Haystack Pipeline with retriever component → pipe.run() returned 3 Documents with scores. Tested against managed KB in us-west-2.
• Serialization verified via to_dict() round-trip

Notes for the reviewer

• This extends the existing amazon-bedrock integration (which already has Embedder, Generator, Ranker) with a Retriever component
• Follows the same pattern as existing components in integrations/amazon_bedrock/
• managedSearchConfiguration requires boto3 >= 1.43 (released April 2026)

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: fix:, feat:, build:, chore:, ci:, docs:, style:, refactor:, perf:, test:.

[CLAassistant]

All committers have signed the CLA.

[github-actions]

Heads-up for maintainers

This PR is from a fork and touches integrations whose integration tests require API keys.
Those tests are skipped in CI because fork PRs don't have access to repo secrets for security reasons.

Affected integrations:

• amazon_bedrock

Please run the integration tests locally (hatch run test:integration inside each folder) before approving.

[github-actions]

Coverage report (amazon_bedrock)

Click to see where and how coverage changed

File	Statements	Missing	Coverage	Coverage (new stmts)	Lines missing
integrations/amazon_bedrock/src/haystack_integrations/components/retrievers/amazon_bedrock
knowledge_base_retriever.py					27-38, 132-137, 174-188, 190-192
Project Total

This report was generated by python-coverage-comment-action

[davidsbatista]

Hi @PVidyadhar - thanks for this PR.

I noticed that there are no integrations/live tests. In any case, did you actually test this Retriever against a live Amazon Bedrock Knowledge Base?

[PVidyadhar]

Hi @PVidyadhar - thanks for this PR.

I noticed that there are no integrations/live tests. In any case, did you actually test this Retriever against a live Amazon Bedrock Knowledge Base?

Hi @davidsbatista — yes, fully tested against a live Managed KB (ZDYIO8E7GK) in us-west-2:

• Standard Retrieve: ✅ 5 documents returned
• Agentic Retrieve: ✅ 5 documents returned
• User-agent confirmed on service-side logs: haystack/bedrock-kb

I've also added integration tests (test_knowledge_base_retriever_integration.py) marked with @pytest.mark.integration - they run with KNOWLEDGE_BASE_ID env var set. Push is up.

[davidsbatista]

Thanks for your contribution, I've done some adjustments + adding the tests to the workflow