What excites me most about Orchestra is how it brings orchestration and observability together in one place. In my experience, debugging pipeline failures across disconnected tools is one of the biggest time sinks for data teams. Orchestra's end-to-end lineage and unified metadata mean you can see exactly what failed and why — without switching between systems. As someone who works closely with dbt and Snowflake, I also love that Orchestra is built for the modern data stack, making it feel like a natural fit for the way lean data teams actually work today.

Visual orchestration pipeline and the numerous connectors to key tools in the data ecosystem.

What I like best is that Orchestra stays focused on orchestration and observability without trying to take over my whole stack. We run dbt on BigQuery with dlt and Cloud Run, and Orchestra gives me a single control plane to schedule, monitor, and debug all of it. The visibility into failures is the biggest win. When something breaks, I can immediately see where it happened and why, so we no longer have customers spotting data freshness issues before we do. Debugging that used to eat hours now takes minutes. It's also refreshingly simple to adopt. I rebuilt our entire orchestration in about a week. The serverless environment for running dlt and dbt core, plus the large connector library, means I'm scaling the business instead of writing boilerplate orchestration code. I now run business-critical pipelines monthly without babysitting them.