Compute Cluster Metrics

This page documents Prometheus metrics emitted by the Compute Cluster component, which manages distributed query execution across compute nodes in a Yellowbrick deployment.

Purpose

These metrics are used to monitor the health, performance, and resource usage of compute clusters. They provide visibility into:

• Query activity and throughput
• Memory and CPU utilization
• Query compilation and execution times
• Cluster degradation or compute node loss

They are critical for performance tuning, detecting resource contention, and ensuring cluster reliability at scale.

Metrics

Name	Type	Freq	Labels	Description
yb_lime_active_queries	gauge	10s	cluster, cluster_name, pool	Number of active queries
yb_lime_cluster_degraded_status	gauge	10s	cluster, cluster_name, status, reason	Cluster degraded status
yb_lime_cluster_missing_workers	gauge	10s	cluster, cluster_name	Number of missing workers in each cluster
yb_lime_cluster_state	gauge	10s	cluster, state, reason	Cluster state
yb_lime_queries_completed_backend_total	counter	10s	cluster, cluster_name, pool, state	Total number of queries completed via backend
yb_lime_queries_completed_total	counter	10s	cluster, cluster_name, state	Total number of queries completed
yb_lime_queries_submitted_total	counter	10s	cluster, cluster_name	Total number of queries submitted
yb_lime_query_bytes_network	histogram	10s	cluster, cluster_name	Bytes network
yb_lime_query_bytes_read	histogram	10s	cluster, cluster_name	Bytes read
yb_lime_query_bytes_read_spill	histogram	10s	cluster, cluster_name	Bytes read spill
yb_lime_query_bytes_written	histogram	10s	cluster, cluster_name	Bytes written
yb_lime_query_bytes_written_spill	histogram	10s	cluster, cluster_name	Bytes written spill
yb_lime_query_cache_efficiency	histogram	10s	cluster, cluster_name	Query cache efficiency
yb_lime_query_compile_time	histogram	10s	cluster, cluster_name	Query compile time in seconds
yb_lime_query_cpu_usage	gauge	10s	cluster, cluster_name, pool	Duration weighted average of longest worker CPU query usage as percentage of allocated CPU
yb_lime_query_lock_time	histogram	10s	cluster, cluster_name	Query lock time in seconds
yb_lime_query_memory_granted	gauge	10s	cluster, cluster_name, pool	Total memory granted to queries in bytes
yb_lime_query_memory_used	gauge	10s	cluster, cluster_name, pool	Total memory used by queries in bytes
yb_lime_query_run_time	histogram	10s	cluster, cluster_name	Query run time in seconds
yb_lime_query_total_time	histogram	10s	cluster, cluster_name	Query total time in seconds
yb_lime_query_wait_time	histogram	10s	cluster, cluster_name	Query wait time in seconds