Skip to main content

Version History

Bedrock Chat has evolved through several major versions, each introducing significant improvements and architectural changes. Understanding the migration path between versions is crucial for maintaining your deployment.

Current Version: V3

V3 introduces fine-grained permission control, Bot Store functionality, and enhanced administrative features. This version requires significant schema changes in DynamoDB. Key Features:
  • Fine-grained permission control with user group-based access
  • Bot Store marketplace for sharing and discovering bots
  • Enhanced administrative features for API management and bot analytics
  • Essential bot marking capability

Previous Versions

Version 2

V2 replaced pgvector on Aurora Serverless with Amazon Bedrock Knowledge Bases, improving RAG accuracy and reducing maintenance complexity.

Version 1

V1 introduced bot personalization and RAG capabilities using pgvector on Aurora PostgreSQL.

Version 0

Initial release with basic chat functionality powered by Amazon Bedrock.

Migration Paths

To ensure a smooth migration with minimal data loss, follow the recommended upgrade path:
V0 → V1 → V1.4 → V2 → V2.10.0 → V3
Important: Skipping versions during migration may result in data loss or compatibility issues. Always upgrade through intermediate versions as recommended.

Migration Guide Selection

Choose the appropriate migration guide based on your current version:

V2 to V3 Migration

Migrate from V2 to V3 with schema changes and Bot Store

V1 to V2 Migration

Migrate from pgvector to Bedrock Knowledge Bases

V0 to V1 Migration

Upgrade from basic chat to bot personalization

Pre-Migration Checklist

Before starting any migration:
1

Backup Your Data

Create backups of all DynamoDB tables and S3 buckets containing documents
# Backup DynamoDB table
aws dynamodb create-backup \
  --backup-name "BedrockChatBackup-$(date +%Y%m%d)" \
  --table-name YOUR_TABLE_NAME
2

Document Bot Configurations

Record all important bot settings, instructions, and knowledge sources in case manual recreation is needed
3

Review Breaking Changes

Read the specific migration guide for your version to understand all breaking changes
4

Plan Downtime

Schedule a maintenance window as migrations typically cause service interruption
5

Verify Prerequisites

Ensure all prerequisites for the target version are met (user groups, permissions, etc.)

Common Migration Considerations

Service Disruption

Most major version migrations require downtime:
  • V2 to V3: ~60 minutes (depends on data volume)
  • V1 to V2: Varies based on bot recreation time
  • V0 to V1: Minimal downtime for database encryption

Data Preservation

  • Conversation history
  • Bot configurations (with schema updates)
  • User accounts and authentication
  • Document uploads (may require migration)
  • Incompatible features (e.g., pgvector bots in V2+)
  • Published APIs (must be republished)
  • Custom configurations not supported in new versions

Regional Considerations

Ensure your deployment region supports all required services:
  • Amazon Bedrock
  • Amazon OpenSearch Serverless
  • Amazon Bedrock Knowledge Bases
See Supported Regions for detailed information.

Post-Migration Verification

After completing migration:
  1. Verify User Access: Confirm all users can log in and access their data
  2. Test Bot Functionality: Validate that bots respond correctly with RAG content
  3. Check Permissions: Ensure permission controls work as expected (V3)
  4. Review Analytics: Confirm administrative features are functioning
  5. Monitor Costs: Check that new services are within expected cost ranges

Need Help?

If you encounter issues during migration:

Version Comparison Matrix

| Feature | V0 | V1 | V2 | V3 | |---------|----|----|----|----|| | Basic Chat | ✓ | ✓ | ✓ | ✓ | | Bot Personalization | ✗ | ✓ | ✓ | ✓ | | RAG (pgvector) | ✗ | ✓ | ✗ | ✗ | | Bedrock Knowledge Bases | ✗ | ✗ | ✓ | ✓ | | Bot Sharing | ✗ | Limited | Limited | ✓ | | Fine-grained Permissions | ✗ | ✗ | ✗ | ✓ | | Bot Store | ✗ | ✗ | ✗ | ✓ | | API Publishing | ✗ | ✓ | ✓ | ✓ | | Administrative Dashboard | ✗ | ✗ | Limited | ✓ |

Build docs developers (and LLMs) love