feat: Report number of lookups to the AsyncDataCache

[minhancao]

Add runtime stats for number of lookups to the AsyncDataCache.

[netlify]

✅ Deploy Preview for meta-velox canceled.

Name	Link
🔨 Latest commit	0f3df76
🔍 Latest deploy log	https://app.netlify.com/projects/meta-velox/deploys/690e6a72e07745000862b915

[prestodb-ci]

@ethanyzhang imported this issue as lakehouse/velox #15430

[zacw7]

curious why is this needed?

[minhancao]

@zacw7 This is needed for the tracking and calculation of the hit ratio % of the AsyncDataCache.

Hit ratio = number of hits / number of lookups to the AsyncDataCache

[zacw7]

@zacw7 This is needed for the tracking and calculation of the hit ratio % of the AsyncDataCache.

Hit ratio = number of hits / number of lookups to the AsyncDataCache

memory cache hit rate can be calculated by: memory_cache_num_hits / (memory_cache_num_hits + memory_cache_num_new)
memory + ssd cache hit rate: (memory_cache_num_hits + ssd_cache_read_entries) / (memory_cache_num_hits + memory_cache_num_new)

[minhancao]

@zacw7 Thanks for the equations! I have tried those equations before and they indeed would work for calculating the hit rate.

However, I would also like to track the total number of lookups to the memory cache. (memory_cache_num_hits + memory_cache_num_new) would not accurately track this due to memory_cache_num_hits only being incremented after the first entry access.

https://github.com/facebookincubator/velox/blob/main/velox/docs/monitoring/metrics.rst?plain=1#L335
https://github.com/facebookincubator/velox/blob/main/velox/common/caching/AsyncDataCache.cpp#L192

[zacw7]

@zacw7 Thanks for the equations! I have tried those equations before and they indeed would work for calculating the hit rate.

However, I would also like to track the total number of lookups to the memory cache. (memory_cache_num_hits + memory_cache_num_new) would not accurately track this due to memory_cache_num_hits only being incremented after the first entry access.

https://github.com/facebookincubator/velox/blob/main/velox/docs/monitoring/metrics.rst?plain=1#L335 https://github.com/facebookincubator/velox/blob/main/velox/common/caching/AsyncDataCache.cpp#L192

makes sense. thanks for the clarification.

[zacw7]

LGTM. Thanks for adding this.