Snowflake data freshness monitor with automated Slack alerts
Build a synthetic reporting data-freshness monitor for Snowflake. The goal is to catch stale reporting tables before business users notice, without having to monitor every pipeline component end-to-end. Trigger it on a schedule (e.g. hourly or daily). Main steps, at a high level: Query dbt Elementary data or Snowflake metadata tables for last-updated timestamps on the reporting tables/models that matter. Flag any table that hasn't refreshed within a tunable threshold (e.g. 24 hours). Alert the data engineering team in Slack when a table is stale, focused on the reporting layer where business impact happens. Tools used: Snowflake, dbt Elementary, Slack
What this prompt builds
This workflow monitors Snowflake data freshness by running synthetic checks on reporting tables and data models. It queries dbt Elementary data or Snowflake metadata to detect tables that haven't refreshed in over 24 hours, then alerts the data engineering team via Slack. Teams using complex data stacks benefit from automated monitoring that focuses on end-user data freshness rather than inspecting every individual pipeline component.
The problem
Data teams running complex stacks struggle to manually monitor whether reporting data in Snowflake is up-to-date across all the tables and models that business users rely on. Instead of checking each pipeline component individually, teams need a synthetic check that focuses on what end users actually care about: whether the data models feeding their dashboards, reports, and reverse ETL processes are fresh. This workflow queries Snowflake metadata or dbt Elementary data to detect any mart or reporting schema tables that haven't refreshed in over 24 hours, then sends Slack alerts to the data engineering team so they can investigate root causes only when needed.
Solution and impact
The workflow automates data freshness monitoring by focusing on the reporting layer where business impact happens, giving data teams peace of mind without requiring end-to-end stack surveillance. Engineers receive targeted Slack alerts only when tables fall behind the 24-hour threshold, allowing them to dig into specific failures rather than performing daily manual checks. Business users maintain confidence in their dashboards and operational workflows, while the data team reduces monitoring overhead and responds faster to actual data quality issues.