Databricks' Native Monitoring Resources
Datadog, Friday, October 2nd, 2026
Datadog explains how to collect Databricks and Spark telemetry using system tables, the Spark UI, job notifications and the REST API.
The second part of Datadog's Databricks series covers how to gather telemetry with native tools.
System tables are the main source for performance, cost and usage analysis, from audit logs to SQL warehouse events, but their latency varies so they are not suited for real-time alerting.
For real-time needs, the post recommends job notifications, Jobs API queries and host-based telemetry, and covers the Databricks and Spark UIs for compute, job, query, data quality, lineage and Model Serving visibility.