/
New feature Beta

Gmail connector

Gmail connector is a Databricks data engineering capability, introduced August 2026.

A managed Lakeflow Connect connector that ingests messages, labels, drafts, filters and profile data from one Gmail mailbox, using a Google service account with domain-wide delegation.

  • Strictly look, don't touch: it is read-only with the gmail.readonly scope and never modifies the source mailbox.
  • Gmail only remembers about a week of history, so a pipeline that runs less than weekly loses its historyId cursor and falls back to a full refresh.
  • Deeply nested email is cut off: MIME parts are materialized up to 8 levels deep, and anything deeper stays unexpanded.

Limitations: In Beta. One mailbox per connection, and domain-wide delegation needs a Google Workspace super administrator. Only messages and message_labels sync incrementally, and they do not support SCD Type 2; every other table is fully re-ingested each run. Run at least weekly or the history cursor expires and forces a full refresh.

Open in REbricked →
Category
Data engineering
Introduced
August 2026
Also known as
Gmail, Lakeflow Connect for Gmail
Verified
2026-10-20

Sources

Related in Data engineering