Configure JXM and Enable Hive connector storage caching

[rohank07]

Configure storage caching following: https://trino.io/docs/current/connector/hive-caching.html

To verify caching is configured correctly - setting up JMX connector can help to query the metrics of the caching system

[rohank07]

Configured caching on the trino coordinator - got cache hits from running Select * queries on dummy tables. Still trying to figure how how to populate the cache on the worker pods.

[rohank07]

SELECT * FROM jmx.current."metrics:name=rubix.bookkeeper.gauge.cache_hit_rate";

[rohank07]

Configuration for Hive cache:
hive.cache.enabled=true
hive.cache.location=/opt/hive-cache
hive.metastore-cache-ttl=1440s
hive.cache.disk-usage-percentage=80
hive.cache.start-server-on-coordinator=true

hive.cache.start-server-on-coordinator property only works when enabled on the coordinator and does not write to worker pod directory.

[rohank07]

Initial configuration is complete. But when querying SELECT * FROM jmx.current."metrics:name=rubix.bookkeeper.gauge.cache_hit_rate";
It only updates the cache hit rate on the unclassified catalog and not the lsddp one.

Also, when runnin SELECT avg(cache_hit) FROM jmx.current."rubix:catalog=unclassified,name=stats" WHERE NOT is_nan(cache_hit); the hit rate value returns NaN

Caching has been configured for the unclassifed catalog. When configuring for the lsddp catalog cache records are being recorded but jxm is unable to query the cache hit rate

[rohank07]

Closing - will create a seperate issue for multiple catalogs