How bad is this?

It sounds bad, but it's hard to know from the description whether it affects user experience, functionality or performance?

My theory is that this is caused by multiple save operations (from same user on the same draft) reaching the server at the same time, competing to update the corpora table. In this case one of them would succeed and hopefully there would not be much difference between the two requests.

Related to T140622, by any chance?

We did not pick this up in the current sprint because:

• it has been going on a while
• we have lot of spill over tasks to finish
• no user reports indicating that this would affect a lot of users

@Amire80 Did you use Sprint 5 on purpose instead of Sprint 4?

In T137275#2518661, @Nikerabbit wrote:

@Amire80 Did you use Sprint 5 on purpose instead of Sprint 4?

Yes, I expect enough work for Sprint 4 already.