Sentry on-premise lose stats


I have an issue with stats displaying in 9.1.2 on-prem version.
I have configured persistent storage in redis via “–save 500 10” and after some mysterious error (because I don’t see any in logs) sentry web stops receiving new events as well as workers stops processing (looks like redis connection error, but don’t see that in logs)

After restarting all stats disappears.

If I restarted whole system via normal way without that stuck problem (in my case it is docker-compose restart or even docker-compose down && docker-compose up -d --scale workers=30, for example) statistics is’t disappear. I see in redis logs, that dump file is successfully loaded from disk.
But I don’t understand why my system stops working and loses all statistics every day.

Please, help me to investigate where is a problem and how can I fix it?

P.S. I know, that you have just released new version (my congratulations :tada: ), but currently I cannot rapidly migrate to new one :frowning:

part of redis conf:

127) "dir"
128) "/data"
129) "save"
130) "300 10"