Monitors

A monitor is a recurring or manually triggered check Decube runs against a data asset.

Monitors

A monitor is a recurring, or manually triggered, check that Decube runs against a data asset: a table, or in the case of Custom SQL, the result of a query you write. Without one in place, you usually learn about a data problem when a dashboard looks wrong or a report fails to add up, days after the underlying issue started.

What a monitor does

Every monitor follows the same pattern, whatever it checks:

1. It runs on a schedule, or you trigger it manually.

2. It collects a metric relevant to what it checks, such as the age of the newest row, the number of rows, or the number of values that break a rule.

3. It evaluates that metric against the Threshold or Tolerance you configured.

4. If the evaluation fails, it raises an incident, and Decube files that incident under a case.

A Freshness monitor, for example, might run every hour, measure the age of the newest row in a table, and compare that age against the maximum you set. Once the data is older than that, the monitor raises an incident, and your alert rules decide who hears about it.

Read Building blocks first

A handful of settings reappear across almost every monitor type: schedule, detection method, scan method, Threshold, Tolerance, custom attributes, and what counts as a destructive change. Building blocks explains each one setting once which makes every monitor type page faster to read.

Building Blocks

Then work through this section

For what each check does and what you configure on it, see Monitor types. To apply one configuration across many tables at once, see:

Create monitors in bulk with Templates