Golden DB runs hygiene continuously (on every data event) plus scheduled sweeps (nightly). Continuous handles incoming data; scheduled handles drift and gaps. Both are on by default; advanced users can tune frequency.
Continuous hygiene
Runs on every data event: new record, source update, identity-resolution decision, or manual edit. It validates the change, runs matching/dedup if relevant, updates the record, and logs the action. Sub-second latency for most events.
Scheduled hygiene
Runs nightly. Sweeps the entire database for drift, missed duplicates, stale records, and consistency violations — surfacing issues that continuous hygiene missed (e.g. two records that became duplicates after both updated). Configure in Settings → Data Hygiene → Schedule.
What scheduled sweeps catch
- Late-discovered duplicates — two records that became duplicates after both were edited.
- Stale records — records untouched in 12+ months that might need re-enrichment.
- Validation drift — fields that no longer match their schema (e.g. an email format change).
- Cross-source consistency — fields that disagree across sources without an active conflict-resolution.
Tuning the schedule
Default schedule: nightly at 2am UTC. Adjust in Settings → Data Hygiene → Schedule. Heavier workspaces (10M+ records) might want twice-weekly to reduce job duration; lighter workspaces can run weekly.
