unresponsive server while cleaning old data, lock wait timeout

Hi ,
I have a latest traccar server running with about 5K devices and 2K active devices , with mysql database size +20GB ,
recently i noticed the server become unresponsive with 100% cpu usage every day at an specific hour for about 1~2 hours and then become normal again ,
Logs shows the following error :

2020-05-14 12:16:36  WARN: Update device status error - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:135 < DataManager:337 < DeviceManager:268 < ...)
2020-05-14 12:16:36  WARN: Event save error - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:140 < DataManager:456 < NotificationManager:65 < ...)
2020-05-14 12:16:45  WARN: Update device status error - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:135 < DataManager:337 < DeviceManager:268 < ...)
2020-05-14 12:16:45  WARN: Update device status error - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:135 < DataManager:337 < DeviceManager:268 < ...)
2020-05-14 12:16:45  WARN: Update device status error - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:135 < DataManager:337 < DeviceManager:268 < ...)
2020-05-14 12:16:46  WARN: Failed to store position - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:140 < DataManager:456 < DefaultDataHandler:40 < ...)
2020-05-14 12:16:46  WARN: Failed to store position - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:140 < DataManager:456 < DefaultDataHandler:40 < ...)
2020-05-14 12:16:46  WARN: Event save error - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:140 < DataManager:456 < NotificationManager:65 < ...)
2020-05-14 12:16:46  WARN: Failed to update device - HikariPool-1 - Connection is not available, request timed out after 30000ms. - SQLTransientConnectionException (... < QueryBuilder:59 < *:135 < DataManager:351 < DeviceManager:296 < ...)

I have a HP G7 DL360 server with dual 2.66 GHz/4-core and 32GB RAM ,
I tracked down the problem and found out at this time the cleanup old data (<entry key='database.historyDays'>90</entry>) process is being done because of this query is being processed : DELETE FROM tc_positions WHERE serverTime < '2020-02-14 10:51:37.654' AND id NOT IN (SELECT position
the CPU usage become 100% at this time by mysql , traccar can not store some data in database,
i also did the steps on https://www.traccar.org/optimization/ but it didnt help.
is there anyway to smooth the clean-up process so we dont lose data everyday at cleanup time ?
i had the same problem with backing up database everyday with mysqldump , that i could fix it by using "mysqldump -u USER -p --single-transaction --quick --lock-tables=false" while backing up database .

You must really be at the limit what the server can take. And there is some limitations to mysql also

You should do cleanup outside of Traccar. Just write a script that would do it in reasonable batches.