DatavineDatavine/Docs
Data Features

Anomaly Detection

Create monitors on BigQuery tables to automatically detect data quality issues: stale data, volume drops, null rate spikes, schema changes, and more.

Check types

FreshnessDetects when a table hasn't been updated within the expected window.
VolumeFlags unusual row count changes — too many or too few rows compared to historical baseline.
Null RateAlerts when null percentage in a column spikes above normal.
SchemaDetects added, removed, or changed columns.
DistributionFlags when value distributions shift significantly from historical patterns.
DuplicatesDetects duplicate rows based on key columns.

Creating a monitor

  1. Go to Anomaly Detection
  2. Click Add Monitor
  3. Select dataset and table
  4. Choose which checks to enable
  5. Set sensitivity: Relaxed, Normal, or Strict
  6. Click Create

Sensitivity levels

Relaxed
Only flags severe deviations. Good for noisy tables.
Normal
Standard thresholds. Recommended for most tables.
Strict
Flags small deviations. For critical tables with tight SLAs.

AI root cause analysis

When an anomaly is detected, click Analyze with AI to get a Gemini-powered explanation including: likely root cause, business impact, step-by-step fix instructions, and prevention tips.