single server, 32GB, 8 cores, HDD (Lets say i can't use SSD for now)

Total connect worker : 3

Total broker : 3

Schema registry : 1
Kafbat UI : 1

resource reserve :
connect worker : 4GB (min/max) each

broker : 3GB (min/max) each

Schema registry : 1 (min/max) GB
Kafbat UI : 1 (min/max) GB

Total connectors : 279 connectors

most of connectors are JDBC and there are some debezium classes too

I don't set poll,batch rate at all for every connectors (using default)

Default partition : 3 per topic

My broker crash sometimes and my worker freeze or down sometimes also

How do i improve the survivalbility ?

I want it to just stay alive and don't down or crash no more

or should i use 2 workers with more reserve ram ? like 2 workers with 6 GB each

Please help :(