Model + test failures
Names, statuses, and duration — the short list of what to fix first.
Technology
dbt Core & Cloud
Failed models and tests, grouped — not twenty Slack pings overnight.
The problem
Overnight dbt fails. CI is red. Models skip. Tests fail. Everyone digs logs while leadership still sees yesterday’s numbers.
The outcome
Upload a run artifact or connect a webhook. VectorData surfaces failed models, failed tests, diffs vs the last good run, and one calm alert for the blast radius.
run_results.json
In-browser analyzer available
Diffs
What broke vs last good run
1 alert
Per upstream root cause
Capabilities
Run history
Pass / fail / skip counts you can scan in seconds.
Pipeline context
See where the failure sits so you don’t chase a child model.
Same command center
dbt sits beside freshness and cost — one morning check.
- ✓Live use case: /vectordata/dbt-failures
- ✓Works from a pasted artifact before BigQuery credentials exist
- ✓Pro and above include dbt run & test failure monitoring
- ✓Demo: watch a broken run in /app
Related
Google BigQuery
Freshness and cost signals from BigQuery metadata — never your result rows.
Read →Slack alerts
One calm alert per root cause — routed where your team already works.
Read →Snowflake
Freshness SLAs on Snowflake marts — same calm Watchdog queue as BigQuery + dbt.
Read →Looker
Watch the BigQuery marts behind Looker before explores lie.
Read →