Hello Team, Have set these parameters in Openobserve deployment to improve query performance but not all of the 4096 MB memory cache is being used. Is there a particular reason for this. Have fetched the logs for a span of 1 week . name: ZO_MEMORY_CACHE_DATAFUSION_MAX_SIZE value: "4096" name: ZO_MEMORY_CACHE_MAX_SIZE value: "4096"

This is expected, both are ceilings, not preallocated blocks. •

ZO_MEMORY_CACHE_MAX_SIZE

caches parquet files on demand as queries touch them. If your 1-week query's working set is under 4GB, usage stays below the cap. It doesn't preload. •

ZO_MEMORY_CACHE_DATAFUSION_MAX_SIZE

is used only while a query is actively running, so it drops back down at idle. You'd only see it near the ceiling under heavy/concurrent load. • If this is a multi-querier setup, each node only caches its own slice, so per-pod usage will naturally be lower than the full configured size.