Disk Cache Metrics

This page documents Prometheus metrics related to disk-based caches within the Yellowbrick platform, which help accelerate query execution by reducing I/O latency for frequently accessed data.

Purpose

These metrics provide visibility into the behavior and efficiency of various internal caches used for temporary storage on disk. They can be used to:

• Monitor cache size and usage patterns
• Track hit/miss ratios for performance analysis
• Identify areas under- or over-utilizing cache space
• Inform tuning of caching policies for different subsystems

Each cache is identified by its area label, allowing for targeted troubleshooting and monitoring. The area label distinguishes between the types of internal caches. It can have one of the following values:

• ast – The Abstract Syntax Tree (AST) cache
• source – The source code cache
• object – The compiled object code cache

This labeling makes it easy to track the size and health of each cache independently in Prometheus and Grafana dashboards.

Metrics

Name	Type	Freq	Labels	Description
yb_lime_cache_entries	gauge	1m	area	Number of entries in the corresponding cache
yb_lime_cache_hits	gauge	1m	area	Hit count of the corresponding cache
yb_lime_cache_misses	gauge	1m	area	Miss count of the corresponding cache
yb_lime_cache_size	gauge	1m	area	Size in bytes of the corresponding cache