We are facing a weird issue in our prod environment (self-hosted cluster in AWS) where requests containing certain multi-way synonyms are timed out with a 504 error. There are no logs available because these requests don’t reach the server. This problem only occurs with certain synonyms, not all. Deleting the problematic synonyms won’t solve the problem. It’s difficult to give an example to replicate (rather random set up), and it seems to occur at a specific time (~6am UTC). We’re running rc27. Do you have any insight into what might be causing this behaviour?