Create Your First Monitors

Overview

This guide walks you through creating your first observability monitors in Matia, from enabling auto-generated freshness and row-count monitors to adding a custom monitor. Upon completion, you will understand monitor status (healthy, alerting, training, error) and where to view history in the changelog.

1. Open Observability

In the left sidebar, click Observability. The main page shows four widgets (auto-generated monitors, custom monitors, schema changes, observability issues) and a table of all monitors.

2. Enable Schemas for Auto-Generated Monitors

Matia can create one freshness monitor and one row count monitor per table for schemas you enable.

1. Click Configure on the Auto-generated monitors widget, or in the sidebar go to Settings → Account settings → Observability.
2. In the observability settings, turn on the schemas you want to monitor.
3. Save. Matia creates freshness and row count monitors for every table in those schemas.

The auto-generated monitors widget then shows how many tables are monitored and how many are healthy, alerting, training, or in error. Turning a schema off disables those monitors; it does not delete them.

3. Add a Custom Monitor

• On the Observability page, click Add Monitor (top right).
• Select data asset: Search for a table or column, select it, then click Next.
• Select monitor type: Choose a type (e.g. Custom SQL, Cardinality, Nullness, Column count for a table). Types depend on whether you selected a table or column and the column data type. Click Next.
• Configure monitor: Set Name, Recipient channels, Anomaly detection (Automatic with Low/Medium/High sensitivity, or Manual with single-value or range rules), Run frequency (Manual, Interval, or Cron), optional Time window, and Notification behavior override if needed. For Custom SQL, also set the threshold evaluation method and provide the SQL, then use Test to validate. Click Create Monitor.

Your new monitor appears in the monitors table and on the monitor details page.

4. Understand Status and Changelog

• Healthy: The result is within the expectation.
• Alerting: The result is outside the expectation; an issue is raised.
• Training: For monitors using automatic anomaly detection, the first 7 days after creation are training; the line appears blue and no alert is fired.
• Error: The query could not run; hover on the error label to see the warehouse error.

On a monitor’s details page, the Status tab shows the result over time (graph and run table). The Changelog tab is an audit log of all changes to that monitor (what changed, when, and by whom).

---

Next Steps:
See Enable schemas for auto-monitors and Add a custom monitor, Configure alert routing, or How monitors work and Monitor types for reference.