Open StekPerepolnen opened 8 months ago
We do not monitor restarts of root system tablets. For Hive, SS, BSC, we could see tablets in the check if they are unresponsive. But root kesus, coordinators, mediators, tx-allocators can suffer in any way, and we do not monitor this directly
Check that tablets restarting in the lbs-sas cluster domain are reported to the HC API HC over lbk* clusters will not detect faulty tablets in /Root. I reproduced it like this: I banned the launch of tablets on storage nodes and dropped one DS. HC is silent on the database. Even in database monitoring.