Overview of AI Cluster Monitoring Metrics

This topic describes the new AI cluster monitoring metrics added in OCP 4.5.0, including Logservice monitoring metrics and OceanBase Shared Storage (SS) mode monitoring metrics.

Metric Types

AI cluster monitoring metrics are divided into the following two categories:

Logservice Monitoring Metrics

Logservice monitoring focuses on the independently deployed oblogservice process, with the monitoring dimension being LOG_SERVICE.

Logservice monitoring metrics include the following categories:

• Log Reception and Writing
• Log Latency
• Log Reading
• FetchLog Operations
• RPC Requests
• HTTP Requests
• GC Collection
• Object Storage Upload
• Thread Pool
• Process Resources
• Latency and Status

OceanBase Shared Storage (SS) Mode Monitoring Metrics

The following monitoring groups apply to OceanBase Shared Storage (SS) mode, requiring an OB cluster version of ≥ 4.6.2.0.

Shared storage monitoring metrics include the following categories:

• SS Macroblock Hit Rate
• SS Macroblock Cache Size
• Shared Data
• Local Cache Warm-up