Looker Studio charts broke after a BigQuery table is overwritten nightly (it’s appended during the day and replaced at night). How do you investigate the cause?

The most direct way to investigate issues with data sourced from BigQuery is to examine the BigQuery job itself. The BigQuery UI provides detailed job history, including start/end times, status (success/failure), and any error messages, which are crucial for understanding why the nightly overwrite might be failing or producing unexpected data. This allows you to pinpoint if the overwrite job is failing, if the data being written is malformed, or if there are permission issues. Looking at Cloud Logging for the Looker Studio report might show errors related to Looker Studio's attempt to query BigQuery, but it won't tell you why the BigQuery data itself is problematic. Snapshot Debugger is for debugging application code, not data pipeline issues. Error Reporting aggregates application errors but is unlikely to provide specific insights into BigQuery job failures unless explicitly configured to do so, and even then, BigQuery's native logging is more detailed for this purpose.