This folder contains the configuration files and data mappings for Amazon Kendra, an intelligent search service used to enable advanced search functionality within the application.
The Kendra setup is organized as follows:
- /data-sources: Configurations for connecting to external data sources like S3 buckets, DynamoDB tables, and RDS instances.
- /indexes: Definitions of Kendra indexes, including field mappings and metadata settings.
- /query-suggestions: Custom configurations for query auto-completion and suggestions.
The AWS CDK manages all Kendra resources, simplifying their deployment and configuration. Key components include:
- Indexes: Centralized storage for searchable data, with customizable relevance tuning.
- Data Source Connectors: Automatically syncs data from various AWS and third-party services.
- Access Policies: Configures fine-grained access controls to ensure data security.
- Natural Language Processing (NLP): Supports queries in natural language, making search more intuitive for users.
- Custom Relevance Tuning: Allows fine-tuning of search results to prioritize certain data fields or sources.
- Multilingual Support: Processes and understands queries in multiple languages.
- Incremental Data Syncing: Periodically syncs only the changes from data sources, improving efficiency.
- Ensure that all data sources are properly indexed before enabling search functionality.
- Relevance tuning is an iterative process; update configurations based on user feedback.
- Indexes are updated asynchronously, so allow sufficient time for changes to propagate.
- Verify data source connectivity and permissions if indexing fails.
- Check query logs in CloudWatch for insights into search performance and user behavior.
- Test search queries across various scenarios to ensure accuracy and relevance.