What Eru reads from Snowflake

Eru connects to Snowflake with a read-only user and explores your warehouse schema. It discovers databases, schemas, tables, and views—including dbt-created models if you use them.

Tables and views

All accessible tables and views with column metadata. Eru scans schema information and samples data to understand what each table represents and how tables relate to each other.

dbt models

If you use dbt, Eru discovers your model layer. Staging models, intermediate transformations, and final marts all become part of the Truth Map, with lineage relationships preserved.

Raw and transformed data

Both raw ingested data (from Fivetran, Airbyte, etc.) and transformed analytics tables. Eru maps how raw data flows through transformations to final business entities.

Usage and query patterns

Optional access to query history helps Eru understand which tables are actually used, by whom, and for what purposes. This informs mapping confidence.

Warehouse + SaaS reconciliation

Your warehouse often contains data that also lives in SaaS tools. This creates reconciliation challenges and opportunities.

Source of truth questions

Is the warehouse or the SaaS system authoritative? Often it depends on the use case. Eru helps you define and document these decisions explicitly.

Freshness comparison

Warehouse data might lag behind live SaaS data by hours or days. Eru tracks freshness and factors it into reconciliation checks.

Transformation verification

When you transform data in dbt, you're making assumptions. Eru can verify that transformed data matches source system logic, catching when transformations drift.

Common use cases

Warehouse-to-source reconciliation

Verify that data flowing into Snowflake matches source systems. Catch ingestion failures, schema changes, and transformation bugs.

dbt model validation

Ensure dbt transformations produce expected results. Compare model outputs against source system logic.

Cross-system entity mapping

Link warehouse customer tables to CRM accounts, billing customers, and product users. Build a unified entity layer across all data.

Metric consistency

Verify that metrics calculated in the warehouse match metrics from source systems. MRR in Snowflake should equal MRR in Stripe.

Setup

1. Create a read-only user

Create a Snowflake user with SELECT access to relevant databases and schemas. Eru never needs write permissions.

-- Create Eru role and user CREATE ROLE eru_reader; GRANT USAGE ON WAREHOUSE compute_wh TO ROLE eru_reader; GRANT USAGE ON DATABASE analytics TO ROLE eru_reader; GRANT USAGE ON ALL SCHEMAS IN DATABASE analytics TO ROLE eru_reader; GRANT SELECT ON ALL TABLES IN DATABASE analytics TO ROLE eru_reader; GRANT SELECT ON FUTURE TABLES IN DATABASE analytics TO ROLE eru_reader; CREATE USER eru_user PASSWORD = '...' DEFAULT_ROLE = eru_reader; GRANT ROLE eru_reader TO USER eru_user;

2. Connect in Eru

Enter connection details through Eru's secure credential form. Account identifier, warehouse, database, and credentials are stored encrypted.

3. Eru explores your schema

The agent scans available tables, samples data, and builds a model of your warehouse structure. It identifies potential entity tables and relationship patterns.

4. Map to business entities

Eru proposes mappings between warehouse tables and business concepts. Your dim_customers table maps to the Customer entity, linked to CRM and billing representations.

Also works with

Eru supports other SQL-based data warehouses and databases:

BigQuery Redshift PostgreSQL MySQL Databricks SQL

Churn prediction guide

If you use Snowflake with dbt and want to set up cross-system churn prediction, see our detailed technical guide covering authentication, schema mapping, and dbt model compatibility.

Other integrations

Connect Snowflake to Eru

Unify your warehouse with live SaaS data.