Azure Data Factory
The Azure Data Factory module watches your data pipelines: failed pipeline and trigger runs, daily pipelines that miss their deadline, runs that take longer than usual and the self-hosted integration runtimes they depend on. Each factory also gets a card on the Overview. Its values come from the Azure Monitor source, read with a read-only role.
What it monitors
| Area | Checks | Why it matters |
|---|---|---|
| Pipeline runs | Failed runs in the last 24 hours | A failed load leaves reports and downstream systems with stale data |
| Daily deadline | Pipelines annotated schedule:daily that have not succeeded by 06:00 | The business expects the data at the start of the day, whatever the cause of the delay |
| Duration | Each pipeline's duration against its own history | A run that suddenly takes much longer is often the first sign of a problem |
| Triggers | Trigger runs that failed or were skipped | A trigger that does not fire means a pipeline that never runs |
| Integration runtimes | Self-hosted integration runtime nodes online, and their version against the latest available | An offline node stops every pipeline that reaches on-premises data |
The daily deadline uses the check's parameters: deadline_local (06:00), timezone (America/New_York by default) and expected_pipelines_tag (schedule:daily). Change them on the check for your factory. Every check, with its rule, cause and action, is in the reference for Azure Data Factory checks.
The Data Factory card
Each factory gets a card on the Overview with today's runs (succeeded, failed, running, cancelled), the failed and running runs, its integration runtimes with their state, nodes, version and the latest available, and the next trigger runs computed from the schedule triggers, tumbling windows included. Event triggers have no next run. Today starts at midnight in the source's Time zone.
Before you begin
- The Azure Monitor source with Data factories among its resource types and Data Factory runs selected. See Connect Azure Monitor.
- Reader (
acdd72a7-3385-48ef-bd42-f606fba81ae7) on the subscriptions, for the factories themselves. - The custom role Data Factory run reader on the factories or their resource groups. Data Factory exposes run history as actions that Reader does not include, and the built-in Data Factory Contributor would add write access. The role definition and the commands to assign it are in A read-only role for Data Factory.
- The module is on. The feature
monitoring.azurecovers this module and the Azure module.
Add a resource
You do not add factories by hand. The Azure Monitor source lists them and adds them under its entity, with their checks and their card. Without the custom role, the factories still appear, and the source's Test lists Data Factory as an optional permission that failed.