A user reported that we are hammering the S3 endpoint quite excessively if there is an error to fetch the latest.json file. The respective message can be found here: https://discord.com/channels/1128210118216007792/1391447848964591696/1391535161434312755. Part of the problem seems to be that all partitions trying to fetch the file from S3 at the same time (and are also retrying).