12
Sometimes the real fix is a product change so you stop needing that join at all.
This worked in dev on sample data but fails at full volume. Details:
Context: Data vault hype, when is it actually justified?
Happy to share schema snippets or metrics if useful.
Sometimes the real fix is a product change so you stop needing that join at all.
Idempotent writes with merge keys saved us during backfills.
This matches our runbook for skewed keys.
Could you share a sketch of the salting logic?
Sign in to reply.