The error you see in the coordinator log (2021-07-07T13:19:30,263 ERROR [Coordinator-Exec–0] org.apache.druid.server.coordinator.rules.LoadRule - Tier[_default_tier] has no servers!) - indicates that your Historical is down.
Based on what you have described about historicals , it seems that you are hitting a similar issue being discussed recently here -
Please read that thread and try the suggestions, I think that should help in your case as well.
Most probably you may need to adjust -
Linux system limits - (No of open files,max_map_count) and maybe a historical heap.
Further, Historical logs could provide you more details for historical startup issues. On the other side , It seems that you have a lot of segments to load and I am guessing there are possibly huge no of tiny segments (few Kbs to Few MBs in size), If that’s the case, You may also want to compact them to get decent-sized segments ( ~500 MB / 5 Million rows per segment) which can drastically reduce the total segment to load and also improve overall query performance and historical stability.