What’s Fixed
Connecting a Rollout integration on an account with a large history could import
only a fraction of the available records and then stop with a “Sync failed”
message on the credential card, citing database timeouts across contacts, notes,
events, calls and text messages. Retrying restarted the same doomed import
rather than continuing, so the gap never closed.
The import now reads each batch through an index that the Rollout database
actually has, instead of forcing it to re-scan and re-sort the entire account’s
history for every batch of 200 records. On the largest affected account —
6.2 million records across all record types — batches that previously timed out
after 60 seconds now return in well under a second.
What changed
- The initial backlog import completes for accounts of any size, rather than
stopping partway through contacts. - Notes, events, calls and text messages import instead of failing immediately
with a timeout. - An interrupted import (worker restart, deploy, timeout) resumes from the last
record it finished rather than starting over. - One account with an oversized backlog can no longer consume the whole
five-minute sync window, which was leaving other accounts’ data stale.