Nightly audit log of Airbyte sync jobs in Google Sheets

Every night at 1am, log yesterday's Airbyte sync jobs to a Google Sheet so we have an audit trail we can pivot on outside of Airbyte's UI.

Trigger: cron, daily at 1am in the workspace time zone.

At the top of the run, call Airbyte's List Connections once and keep the results in memory as a map from connection id to connection name. This is the cheap lookup for stamping a human-readable connection name on every row we append.

Then call Airbyte's List Jobs, paging through all results, filtered to jobs whose end time falls in the previous 24-hour window (the 24 hours ending at the trigger fire time). Include both sync and reset job types. Skip any job that is still pending or running; only log terminal states (succeeded, failed, cancelled, incomplete).

For each qualifying job, build a single row with these columns in this order: job id, connection id, connection name (from the cached lookup, falling back to an empty string if the connection is no longer listed), job type, status, start time (ISO 8601), end time (ISO 8601), duration in seconds (end minus start, rounded to a whole number), records extracted, records loaded, and bytes synced. Pull the record and byte counts from the job payload; if List Jobs doesn't include them, call Get Job Details for the id and read them from there.

Append every row for the night in a single Google Sheets Append Values call to the configured spreadsheet and tab. Use USER_ENTERED as the value input option so timestamps and numbers land as real sheet types rather than strings, and rely on the sheet's existing header row to define column meaning (do not re-append a header).

This is deterministic logging with no reasoning required. Just enumerate jobs and append rows in a stable schema so finance, ops, and data leadership can build their own pivot tables and charts on top of the sheet. If Airbyte returns zero qualifying jobs, the run should complete cleanly without appending anything.

Configuration inputs the workflow should expose: the Google Sheet spreadsheet id, the tab name, and (optionally) an Airbyte workspace id to scope the job list to a single workspace. Everything else is fixed by the schema above.