Introduction
FreshGuard Core is an open-source, MIT-licensed data pipeline freshness monitoring engine. It detects stale data, volume anomalies, and schema changes across PostgreSQL, DuckDB, BigQuery, Snowflake, MySQL, and Redshift.
What it does
- Freshness monitoring — Alert when a table hasn't been updated within a tolerance window
- Volume anomaly detection — Alert when row counts deviate from a calculated baseline
- Schema change detection — Alert when columns are added, removed, or modified
Who it's for
- Teams running data pipelines who want monitoring without SaaS lock-in
- Self-hosters who want to run on their own infrastructure
- Developers building custom data quality tooling
Architecture
CLI / Direct API
|
Monitoring Algorithms
checkFreshness · checkVolumeAnomaly · checkSchemaChanges
|
Database Connectors
PostgreSQL · DuckDB · BigQuery · Snowflake · MySQL · Redshift
|
Metadata Storage (optional)
DuckDB (embedded) or PostgreSQL
Each monitoring function validates its inputs, queries the target database through a connector, analyzes the result, and returns a CheckResult with a status (ok, alert, or failed).
For the full architectural breakdown, see the Architecture doc.
Open Core model
This package is the open-source core. It contains all monitoring logic, connectors, and CLI tooling under the MIT license. It does not include multi-tenant features, authentication, dashboards, or billing — those are part of the separate FreshGuard Cloud product.
You can use, modify, and distribute this core freely.