This topic describes the metrics for PolarDB for MySQL.
When you call an API operation provided by CloudMonitor, set the Namespace and Period parameters.
- Set the Namespace parameter to acs_polardb.
- Set the Period parameter to an integral multiple of 60. Default value: 60. Unit: seconds.
The following table describes the valid values of the MetricName and Dimensions parameters.
Metric | Unit | Metric | Dimensions | Statistics |
---|---|---|---|---|
ActiveSessions | Count | cluster_active_sessions | userId, clusterId, and nodeId | Average |
ConnectionUtilization | % | cluster_connection_utilization | userId, clusterId, and nodeId | Average |
CPUUtilization | % | cluster_cpu_utilization | userId, clusterId, and nodeId | Average |
Input And Output | KB | cluster_data_io | userId, clusterId, and nodeId | Average |
Innodb Input/Output Per Second | CountSecond | cluster_data_iops | userId, clusterId, and nodeId | Average |
Input/Output Per Second | CountSecond | cluster_iops | userId, clusterId, and nodeId | Average |
Memory Buffer Hit Ratio | % | cluster_mem_hit_ratio | userId, clusterId, and nodeId | Average |
MemoryUtilization | % | cluster_memory_utilization | userId, clusterId, and nodeId | Average |
QueriesPerSecond | Count | cluster_qps | userId, clusterId, and nodeId | Average |
Slow Queries Per Second | CountS | cluster_slow_queries_ps | userId, clusterId, and nodeId | Average |
TransactionsPerSecond | CountS | cluster_tps | userId, clusterId, and nodeId | Average |
cluster_replica_lag | Second | cluster_replica_lag | userId, clusterId, and instanceId | Average, Minimum, and Maximum |
TotalSessions | Count | cluster_total_session | userId, clusterId, and nodeId | Average, Maximum, and Minimum |
OutputTraffic | KB/s | cluster_output_traffic | userId, clusterId, and nodeId | Average, Maximum, and Minimum |
ManipulationsPerSecond | CountSecond | cluster_mps | userId, clusterId, and instanceId | Average, Maximum, and Minimum |
IOPSUsage | % | cluster_iops_usage | userId, clusterId, and nodeId | Average, Maximum, and Minimum |
InputTraffic | KB/s | cluster_input_traffic | userId, clusterId, and nodeId | Average, Maximum, and Minimum |
Note The following information describes how the Slow Queries Per Second metric is measured
and the precautions that you must take when you set the threshold for this metric:
- The value of the Slow Queries Per Second metric is the number of slow SQL queries within a minute divided by 60s. For example, if a slow SQL query occurs within a minute, the value of the Slow Queries Per Second metric is calculated by using the following formula: 1/60 = 0.016.
- During the routine maintenance for ApsaraDB RDS instances, slow SQL queries may occur. Therefore, we recommend that you do not set the threshold of the Slow Queries Per Second metric to a value that is too small, for example, 0. Instead, you can set the threshold to 1.