Welcome to our blog post on integrating system tables with Overwatch! In this article, we'll delve into the exciting world of leveraging system tables to enhance the utility and insight provided by Overwatch. For an introduction to Overwatch, please refer to my initial blog post Overwatch: The Observability Tool for Databricks.
System Tables represent a repository of analytical data specific to your account, housed within the Databricks system catalog. They facilitate historical observability across your account, offering insights into usage, operational history, and other facets of your Databricks environment. Notable system tables include Audit logs, Table lineage, and Predictive optimization. Audit logs comprehensively document all audit events occurring within your Databricks account.
Overwatch, on the other hand, gets data from 3 different sources:
The audit log serves as a crucial component from which Overwatch retrieves data. It encompasses various events tied to operations within Databricks, spanning DBFS operations, notebook actions, account activities, workspace actions, cluster operations, and login activities. This audit log can be stored in an S3 container for AWS workspaces, a GCS container for GCP workspaces, or an Event Hub for Azure workspaces.
In Unity Catalog (UC) enabled Workspaces, the audit log data is accessible through the system.access.audit table within the "System" catalog. This table aggregates audit data from multiple Workspaces. Therefore, data pertaining to all Workspaces associated with the account can be accessed within the system.access.audit table.
Beginning with version 0800, Overwatch supports System Tables, making it easy to automatically retrieve audit log data. Overwatch also supports a cross-account integration with System Tables. There are several benefits for using the system table as an audit log source:
The alignment between System Tables and Overwatch ensures a seamless integration in the following ways.
Enabling system tables as a data source for Overwatch is straightforward. Simply use the keyword "system" in the auditlogprefix_source_path parameter within the Overwatch config file:
%sql
update <overwatch_config> set auditlogprefix_source_path = "system" where workspace_id = "<workspace_id>"
For further details, please refer to the provided SystemTableConfig.
System Tables serve as a repository of analytical data, including audit logs, which document various events crucial for monitoring account activities. By utilizing them as an alternative to container-based audit logs, users can streamline setup, eliminate manual intervention, and mitigate access issues. Overwatch seamlessly integrates with System Tables, offering simplified setup, effortless migration, and extended data retention, particularly beneficial for Azure deployments. Moreover, system table integration facilitates hassle-free cross-account integration, empowering users with a unified platform for comprehensive monitoring and analysis. Overall, incorporating system tables into Overwatch provides a robust solution for optimizing observability and enhancing operational efficiency within Databricks workspaces.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.