Integrations Overview

Entropy Data provides integrations for Data Platforms, Data Catalogs, Data Quality, Business Intelligence, Notifications, SSO, Git Providers, Automation, Access Control, AI Agents, AI Models, and Custom solutions.

Data Platforms

Data Catalogs

Data Quality

Business Intelligence

Notifications

SSO

Git Providers

Automation

Access Control

AI Agents

AI Models

Custom

Types of Integrations

Entropy Data integrations fall into several categories:

Native Integrations

Native integrations are built directly into Entropy Data and require no additional deployment. They provide seamless connectivity with minimal configuration.

A native integration has one purpose. An ingestion reads the assets of a data platform on a schedule: its databases, schemas, tables and views, and where the platform has them, its semantic models, such as the semantic views of Snowflake. A metadata writer writes metadata from Entropy Data back to the platform whenever it changes, for example the descriptions and tags of a data contract to Databricks Unity Catalog, the AWS Glue Data Catalog or Snowflake, or the semantic namespaces as business glossaries to Google Cloud Knowledge Catalog or, in preview, the business context of the Glue Data Catalog. An access writer grants access agreements in the platform, for example in Databricks Unity Catalog. Add one integration per purpose; a platform that offers several purposes asks for the purpose in the Add Integration wizard.

Connector Integrations

Connector integrations are separately deployed services that communicate with Entropy Data via APIs and webhooks. They offer more flexibility and control, and can be customized to your specific requirements. Many connectors are built using our SDK.

A connector can report its version and its health to Entropy Data. Both are shown on the integrations page: the version next to the connector type, and the health as a badge with the time it was last reported and, when the connector reports a problem, its error message. This makes a connector that has stopped working distinguishable from one that is fine. Reporting is optional — a connector that reports neither keeps working and is shown as not reported. See the connector resource in the API specification for the fields.

Link Integrations

Link integrations connect Entropy Data with external tools through webhooks and event forwarding. They enable real-time communication for notifications and event-driven workflows.

External Tools

External tools are standalone applications that interact with Entropy Data's API. They can be used independently or integrated into your CI/CD pipelines.

Scheduled Ingestion Runs

Ingestions synchronize assets on a schedule that you configure per integration. You can review the status and the last runs of an ingestion under Settings > Integrations. A writer has no schedule: a metadata writer writes whenever the metadata changes in Entropy Data, and an access writer whenever an access agreement is activated or ends. The integration page of a writer shows the tables or access agreements it wrote and what still waits in its queue.

Automatic Disabling After Repeated Failures

If the scheduled runs of an ingestion keep failing, Entropy Data disables the integration automatically. This prevents a broken integration from failing indefinitely, for example after credentials expired or a host became unreachable.

• By default, an integration is disabled after 5 consecutive failed scheduled runs. Self-hosted installations can change the threshold with APPLICATION_INGESTIONS_AUTO_DISABLE_AFTER_CONSECUTIVE_FAILURES, or set it to 0 to switch the behavior off.
• Only scheduled runs count. Manually triggered runs never disable an integration, so you can debug a failing integration without it being disabled under you.
• The failure streak resets on any successful run, and whenever someone edits the integration configuration or re-enables the integration. Each change grants a fresh set of attempts.
• When an integration is disabled automatically, the integration details page states that it was disabled because its runs kept failing. Fix the cause and re-enable the integration to resume the schedule.

Entropy Data notifies the organization when this happens: organization owners receive the "Integration disabled after repeated failures" email, and the asset owner team is notified through its notification channels.

Getting Started

Most integrations require an API Key for authentication. Visit the individual integration pages for detailed setup instructions and configuration examples.

For custom integration requirements or support, reach out to support@entropy-data.com.