Sweetwater's proactive data observability playbook with Datadog
A recurring gap in one sales dashboard turned out to trace back to a mis-scheduled legacy script, and it's exactly the kind of failure Sweetwater now catches before anyone notices.
Overview
Sweetwater's analytics team struggled with stale report data and no visibility into problems until they escalated. Covers combining dbt with Datadog for proactive monitoring, how dbt Mesh gave departments more ownership, and tradeoffs in centralizing data systems.
What you'll take away
- 01
Extending Datadog, already used by DBAs to monitor source systems, upstream into the dbt pipeline itself turned a known tool into cross-system observability instead of a hard internal sell.
- 02
The team traced a recurring sales-dashboard issue, invoices missing a sales channel, back to one mis-scheduled legacy script, using dashboards it already had for Kafka inserts by topic.
- 03
Current monitoring auto-configures across entire schemas, with row-count and freshness monitors that adapt to each table's expected volume and machine-learning-driven alert thresholds.
Session speakers
Daniel Gonzalez
Sweetwater SoundDerek Andres
Sweetwater