From 72 minutes to 5: how one environment variable powers two dbt pipelines

This session walks through a single dbt environment variable that controls two completely different pipelines from one codebase: a "live" mode that filters to only active records and runs every five minutes for real-time dashboards, and a "historical" mode for full analytical workloads. Same model, two outputs.

The pattern is industry-agnostic. If your team needs both real-time operations and historical analytics from the same business logic, whether in e-commerce, fintech, SaaS, or logistics, this applies to you.

We built it for a health network running near real-time operational dashboards across nine hospitals on fragmented legacy systems. The results: a 93.6% runtime reduction and 95%+ less data processed per run on the same warehouse size, design beating brute force rather than a bigger bill, along with a measurable improvement in median operational response times and a single source of truth for the teams that depend on it.

You'll see the actual code, the two DAG propagation strategies, and the dbt platform job configuration that make it work.