Relevant issues:
ClickHouse/ClickHouse#2261
ClickHouse/ClickHouse#1531

Clickhouse do not respect limits for cpu and memory and gets restarted by Kubernetes constantly.

Workaround 1: use the whole node for clickhouse — fix cpu and memory problems

Workaround 2: to limit the memory use these settings:
max_memory_usage_for_all_queries
use_uncompressed_cache
merge_tree_max_rows_to_use_cache

There is no settings for number of cpu cores until ClickHouse/ClickHouse#2261
is fixed.

Need a note about these settings usage in README, NOTES.txt and values.yaml.