diagnose

Debug silent data pipeline failures by classifying the incident and tracing row counts across bronze, silver, gold, and mart layers to pinpoint where data diverges. This observation-first triage identifies the root cause before attempting inline fixes or escalating backend issues.

Trace dbt schema mismatches by performing schema discovery via information_schema queries to verify structures. Inspecting soria CLI outputs and tracing data across layers isolates whether the mismatch originates from extraction, transformation, or warehouse publishing.

Triage broken dbt marts and scraper pipelines by classifying the failure into Silent Failure, Data Trace, Schema Mismatch, or Infrastructure categories. Following a disciplined checklist, inspect CLI outputs and trace row counts to determine the exact failure point.

You can apply inline fixes for pipeline issues like env switches, dbt runs, manifest updates, or re-extraction attempts. If the issue is backend infrastructure, the triage process routes the problem by invoking the ticket workflow and producing a diagnostic artifact.

Dive manifest load failures during successful CLI runs often indicate silent data failures or schema mismatches upstream. Investigate by verifying schemas, tracing data across pipeline layers, and inspecting CLI outputs to find where the data becomes absent or inconsistent.