Stale read on element creation

We have another stale read on the transkribus import: a child element is immediately created on a freshly created parent

Example : https://arkindex.teklia.com/process/2bdffebc-e8d0-40e8-b5b8-8711ddb94015/0

added 1 deleted label

It's a funny one I can't reproduce locally in cluster mode

Have you tried setting a very low CPU limit on Postgres containers? If you slow the replica down enough, it might work…

mentioned in merge request !1078 (merged)

closed

mentioned in commit 5ce1e044