Hello, We have a 5 node cluster, each with a 5 disk JBOD storage running Kafka 0.8.1. We are interested in the possibility of using the OffsetRequest API taking a time to get offsets from a given time. Because that API can only "find" times where the log segments roll, we would need to set log.roll.hours =1 in order to have a useful granularity for that API.
I am wondering what would be the drawbacks to this? I know it would create many more files, but what affect would that have on the system(s)? Is there a general rule of thumb for what would be "too many" files?
