Data Sync Docs: Run Statuses and Held Runs

The Data Sync docs now describe run behaviour that the dashboard and the API already report: why a run failed, why a full run was held, and which limits come from your plan.

What’s new

• Failure statuses. The run status table lists provider_auth_failed, provider_rate_limited, provider_error, timed_out, no_records_found, unexpected_record_count and invalid_config, with what each one means.
• Full runs held for their record count. A new section explains when a full run is held: it finds no records when the previous sync had some, or it is missing more than 30% of the previous sync’s records and at least 10 of them. A held run’s records are discarded and the last complete sync stays readable. If the next full run is held too, StackOne treats the drop as real and serves it.
• Limits a plan can change. The page says which limits come from your plan: the shortest time between syncs, 15 minutes by default, and storage per organization, 25 GiB by default.
• Events from runs that stop early. The Data Sync Events page now says that an incremental run that fails part way through still sends events for the pages it wrote, while a held full run sends none.

Synced records are queried through Deep Query, introduced in the Deep Query entry.