Documentation

Enterprise Search Connectors

In this section, you will learn how to manage and configure the enterprise search and retrieval augmented generation connectors, provided by the RheinInsights Retrieval Suite.

The following content sources are available for crawling. The following connectors come with full support for secure search, i.e., early binding security trimming:

Atlassian Confluence Connector

Atlassian Confluence Cloud Connector

Atlassian Jira Connector

Atlassian Jira Cloud Connector

File Share Connector

GitHub Enterprise Cloud Connector

GitHub Enterprise Server Connector

Google Drive Connector

Google Mail (GMail) Connector

JDBC and SQL Server Connector - see the security concept described at the connector page

Microsoft SharePoint Online Connector

Microsoft OneDrive Connector

OData via REST Connector

SAP Success Factors (documentation follows soon)

Slack Connector

Trello Connector

Web Page Connector

The following connector does not come with security trimming as of now

Git Connector - due to the general Git protocol

Managing Connectors

A key feature of our Retrieval Suite is that you can configure any combination of content source and search engine. Also each connector can be configured to use distinct content processing and principal processing pipelines.

  1. Please first add the connectors, you like to use at Managing Connectors .

  2. Afterwards, configure

    1. the content source connection, cf. Sources .

    2. the search engine connection, cf. Search Engines .

    3. content transformation, cf. Content Transformation .

    4. principal transformation, cf. Security Transformation

  3. If you successfully validated all configurations, you can start or schedule crawls for your new connector.

Starting and Managing Crawls

For crawl modes and crawling, please refer to Crawl Modes .

Multi-Node Environments and Load Balancing

As written at Deployment and Base Configuration , the RheinInsights Retrieval Suite is not just multi-threaded but can run in multi-node environments.

All connector will be configured centrally for all nodes. The configuration is valid for all instances of the RheinInsights Retrieval Suite which are connected to the database instance at Database Settings . When starting a crawl, the crawl is picked up by the first node of the RheinInsights Retrieval Suite, which has sufficient capacity. This is where the crawl will be performed.

A node can execute up to four crawls in parallel. Where a crawl can be a principal synchronization, an incremental, a checksum or a full crawl.