All Products
Search
Document Center

PolarDB:Performance metrics

Last Updated:Mar 30, 2026

Performance monitoring parameters for PolarDB-X instances, covering compute nodes, data nodes, Global Meta Service (GMS) nodes, and change data capture (CDC) nodes.

Compute node parameters

Parameter Description Unit
Cpu_Usage CPU utilization. %
Mem_Usage Java Virtual Machine (JVM) heap memory usage. Calculated as: used JVM heap memory / total JVM heap memory x 100%. %
Logic_RT Average response time of user requests. Measured from the point the request is received to the point the result is returned. Microseconds
Physical_RT Average response time of requests sent from compute nodes to data nodes. Measured from the point the compute node sends the request to the point it receives the result from the data node. Microseconds
Logic_QPS Average number of user requests processed per second. Requests/s
Logic_TPS Average number of transactions processed per second. Transactions/s
Physical_QPS Average number of requests sent to data nodes per second. Requests/s
Network_Usage Network throughput — the sum of inbound and outbound traffic per second. byte/s
Connections Connection counts. Returns two values: active_connection (total number of connections) and connection_count (number of connections created per second).
Thread_Detail Number of active connections.
JVM_CollectorTime Average time spent on garbage collection (GC) per second. Returns two values: time spent on GC in the young generation space per second, and time spent on full GC per second. Nanoseconds
Slow_RequestCount Slow SQL counts. Returns two values: average number of slow SQL requests sent by users per second, and average number of slow SQL requests sent to data nodes per second. Requests/s

Data node and GMS node parameters

Parameter Description Unit
MySQL_NetworkTraffic Network traffic for the PolarDB-X instance. Returns two values: average inbound traffic per second and average outbound traffic per second. KB
MySQL_QPSTPS Throughput rates. Returns two values: average number of SQL statements executed per second and average number of transactions processed per second.
MySQL_Sessions Session counts. Returns two values: number of active connections and total number of connections.
MySQL_InnoDBBufferRatio InnoDB buffer pool statistics. Returns three values: read hit ratio, usage ratio, and dirty ratio of the InnoDB buffer pool.
MySQL_InnoDBDataReadWriten InnoDB data I/O rates. Returns two values: average volume of data read per second and average volume of data written per second. KB
MySQL_InnoDBLogRequests InnoDB buffer pool operation rates. Returns two values: average number of read operations per second and average number of write operations per second on the InnoDB buffer pool. Operations/s
MySQL_InnoDBLogWrites InnoDB log write rates. Returns three values: average number of log write requests per second, average number of physical writes to log files per second, and average number of fsync() operations on log files per second. Operations/s
MySQL_TempDiskTableCreates Number of temporary tables automatically created on disk during statement execution.
MySQL_COMDML DML statement execution rates. Returns average executions per second for DELETE, INSERT, INSERT_SELECT, REPLACE, REPLACE_SELECT, SELECT, and UPDATE statements. Statements/s
MySQL_RowDML InnoDB row operation rates. Returns average rows per second for: reads, updates, deletes, and inserts on InnoDB tables, plus physical writes to log files.
MySQL_MemCpuUsage CPU and memory usage of the PolarDB-X instance.
MySQL_IOPS IOPS of the PolarDB-X instance.
MySQL_DetailedSpaceUsage Storage breakdown. Returns five values: used storage, data storage, log storage, temporary storage, and system storage.
slavestat Replication delay between the primary instance and read-only instances.
MySQL_ThreadStatus Thread counts. Returns two values: number of active threads and number of connected threads.
MySQL_ReplicationDelay Replication delay between the primary instance and the secondary instance. Seconds
MySQL_ReplicationThread States of I/O threads and SQL threads used for replication. A value of 1 means the thread is running normally; a value of 0 means the thread is lost.
IOPSUsage IOPS usage. %
MySQL_Slow_Queries Average number of slow SQL requests processed per second. Requests/s

CDC node parameters

Parameter Description Unit
Cdc_DumperDelay Latency when processing binary log events. Milliseconds
Cdc_DumperThroughput Data volume of binary log events processed per second. byte/s
Cdc_CpuUsage CPU utilization. %
Cdc_DumperHeapUsage JVM heap memory usage. Calculated as: used JVM heap memory / total JVM heap memory x 100%. %
Cdc_DumperOps Binary log event processing rates. Returns three values: dumper_total_eps (average binary log events per second), dumper_dml_eps (average binary log events with DML statements per second), and dumper_tps (average transactions per second).