Performance parameters of AnalyticDB for PostgreSQL - AnalyticDB

Performance parameters can help you view the instance health status, connection status, storage status, connections, query concurrency, data storage, and node status in a visualized manner. This topic describes the performance parameters of AnalyticDB for PostgreSQL.

Elastic storage mode

You can call the DescribeHealthStatus operation to return the following metric parameters.

Parameter	Metric	Unit	Description
adbpg_status	Instance health status	N/A	critical: A coordinator node or compute node is unavailable. In this case, this metric is marked in red in the console. healthy: All nodes are available. In this case, this metric is marked in green in the console.
adbpg_connection_status	Instance connection status	N/A	The connection health status of the instance. Valid values: critical: The instance connection usage is greater than 95%. In this case, this metric is marked in red in the console. warning: The instance connection usage is greater than 90% and less than or equal to 95%. In this case, this metric is marked in yellow in the console. healthy: The instance connection usage is less than or equal to 90%. In this case, this metric is marked in green in the console. Note The instance connection usage is the maximum connection usage of all coordinator nodes and compute nodes.
adbpg_disk_status	Instance storage status	N/A	The storage status of the instance. Valid values: critical: The instance storage usage is greater than or equal to 90%. In this case, this metric is marked in red in the console and the instance is locked. warning: The instance storage usage is greater than or equal to 70% and less than 90%. In this case, this metric is marked in yellow in the console. healthy: The instance storage usage is less than 70%. In this case, this metric is marked in green in the console. Note The instance storage usage is the average storage usage of all compute nodes.
adbpg_disk_usage_percent	Instance storage usage	%	The status corresponding to the storage usage of the instance. Valid values: critical: The instance storage usage is greater than or equal to 90%. In this case, the instance is locked. warning: The instance storage usage is greater than or equal to 70% and less than 90%. healthy: The instance storage usage is less than 70%. Note The instance storage usage is the average storage usage of all compute nodes.
adbpg_master_disk_usage_percent_max	Maximum storage usage of coordinator nodes	%	The status corresponding to the maximum storage usage of all coordinator nodes. Valid values: critical: The storage usage of coordinator nodes is greater than or equal to 90%. In this case, the instance is locked. warning: The storage usage of coordinator nodes is greater than or equal to 70% and less than 90%. healthy: The storage usage of coordinator nodes is less than 70%.
adbgp_segment_disk_usage_percent_max	Maximum storage usage of compute nodes	%	The status corresponding to the maximum storage usage of all compute nodes. Valid values: critical: The storage usage of compute nodes is greater than or equal to 90%. In this case, the instance is locked. warning: The storage usage of compute nodes is greater than or equal to 80% and less than 90%. healthy: The storage usage of compute nodes is less than 80%.
node_master_status	Coordinator node health status	N/A	The health status of coordinator nodes. Valid values: critical: A coordinator node is unavailable. In this case, this metric is marked in red in the console. healthy: All coordinator nodes are available. In this case, this metric is marked in green in the console.
adbpg_master_status	Coordinator node health status	N/A
node_segment_disk_status	Compute node storage status	N/A	The storage status of compute nodes. Valid values: critical: The storage usage of compute nodes is greater than or equal to 90%. In this case, this metric is marked in red in the console and the instance is locked. warning: The storage usage of compute nodes is greater than or equal to 80% and less than 90%. In this case, this metric is marked in yellow in the console. healthy: The storage usage of compute nodes is less than 80%. In this case, this metric is marked in green in the console. Note The compute node storage usage is the maximum storage usage of all compute nodes.
node_master_connection_status	Coordinator node connection status	N/A	The connection status of coordinator nodes. Valid values: critical: The connection usage of coordinator nodes is greater than or equal to 95%. In this case, this metric is marked in red in the console. warning: The connection usage of coordinator nodes is greater than or equal to 90% and less than 95%. In this case, this metric is marked in yellow in the console. healthy: The connection usage of coordinator nodes is less than 90%. In this case, this metric is marked in green in the console. Note The coordinator node connection usage is the maximum connection usage of all coordinator nodes.
adbpg_segment_status	Coordinator node connection status	N/A
node_segment_connection_status	Compute node connection status	N/A	The connection status of compute nodes. Valid values: critical: The connection usage of compute nodes is greater than or equal to 95%. In this case, this metric is marked in red in the console. warning: The connection usage of compute nodes is greater than or equal to 90% and less than 95%. In this case, this metric is marked in yellow in the console. healthy: The connection usage of compute nodes is less than 90%. In this case, this metric is marked in green in the console. Note The compute node connection usage is the maximum connection usage of all compute nodes.
adbpg_instance_total_data_gb	Total storage	GB	The total volume of instance storage. Note This parameter is displayed in the console only for AnalyticDB for PostgreSQL instances of V6.3.11.3 or later.
adbpg_instance_hot_data_gb	Hot storage	GB	The total volume of hot storage. Note This parameter is displayed in the console only for AnalyticDB for PostgreSQL instances of V6.3.11.3 or later.
adbpg_instance_cold_data_gb	Cold storage	GB	The total volume of cold storage. Note This parameter is displayed in the console only for AnalyticDB for PostgreSQL instances of V6.3.11.3 or later.

You can call the DescribeDBClusterPerformance operation to return the following metric parameters.

Parameter	Metric	Unit	Description
adbpg_segment_cnt	Number of compute nodes	N/A	The number of compute nodes that are collected per hour.
adbpg_instance_disk_used_mb	Instance storage used	MB	The amount of storage space that is used by all compute nodes.
adbpg_instance_disk_usage_percent	Instance storage usage	%	The storage usage of the instance, which is calculated by using the following formula: Storage space used by all compute nodes/Reserved storage space of all compute nodes.
node_master_connection_count	Number of coordinator node connections	N/A	The number of connections to coordinator nodes.
node_segment_connection_count	Number of compute node connections	N/A	The number of connections to compute nodes.
node_segment_workfile_used_mb	Temporary disk file size	MB	The size of temporary files that are written to disks.
node_cpu_used_percent	CPU utilization	%	The CPU utilization of coordinator nodes or compute nodes.
node_mem_used_percent	Memory usage	%	The memory usage of coordinator nodes or compute nodes.
node_disk_iops_percent	I/O usage	%	The I/O usage of coordinator nodes or compute nodes.
node_disk_used_mb	Storage used	MB	The amount of storage space that is used by each compute node.
node_disk_usage_percent	Storage usage	%	The storage usage of each compute node, which is calculated by using the following formula: Storage space used by each compute node/Reserved storage space of each compute node.
adbpg_rsq_cost	Query cost for a resource queue	N/A	The estimated total cost of all queries that can be executed in the current queue. rsqCostLimit specifies the cost limit, and a value of -1 indicates no limit. rsqCostValue specifies the total cost of all queries that are being executed.
adbpg_rsq_count	Number of concurrent queries in a resource queue	N/A	The number of concurrent queries that can be executed in the current queue. If the number of concurrent queries reaches the upper limit, subsequent queries that are assigned to the query must wait for the queue. rsqCountLimit specifies the maximum number of concurrent queries that are allowed, and a value of -1 indicates no limit. rsqCountValue specifies the number of concurrent queries that are being executed.
adbpg_rsq_memory	Query memory for a resource queue	Byte	The amount of memory used by all queries that can be executed in the current queue. rsqMemoryLimit specifies the memory limit, and -1 indicates no limit. rsqMemoryValue specifies the amount of memory used by all queries that are being executed.
adbpg_rsq_waiters	Number of queries waiting for a resource queue	N/A	The number of queries that are waiting for resource queue slots in the current queue due to insufficient resources. If the number of queries that are being executed in a queue reaches the concurrency limit, subsequent queries that are assigned to the queue must wait until sufficient resources are available. These queries are in the waiting state in the pg_stat_activity view.
adbpg_rsq_holders	Number of executing queries in a resource queue	N/A	The number of queries that have obtained resource queue slots for execution in the current queue. These queries are not surely in the active state. If these queries have to wait for other resources such as row locks, they may be still in the waiting state. In this case, the resource queue slots that are occupied by these queries are not released. Note In most cases, the value of rsqCountValue is equivalent to that of rsqholders. In rare cases, the value of adbpg_rsq_holders may be greater than that of rsqCountValue. This is because resource queue locks may not be released after queries are executed due to system exceptions. If the issue of distinct differences between the values of adbpg_rsq_holders and rsqCountValue persists, contact Alibaba Cloud technical support.
adbpg_db_qps	Instance queries per second (QPS)	N/A	The number of read queries that are processed per second by the instance. The query statements include SELECT, SELECT INTO, SELECT FOR UPDATE, and FETCH.
adbpg_db_tps	Instance transactions per second (TPS)	N/A	The number of write operations that are processed per second by the instance. The write operations include INSERT, UPDATE, DELETE, and INSERT INTO SELECT.

Serverless mode

You can call the DescribeHealthStatus operation to return the following metric parameters.

Parameter	Metric	Unit	Description
adbpg_status	Instance health status	N/A	critical: A coordinator node or compute node is unavailable. In this case, this metric is marked in red in the console. healthy: All nodes are available. In this case, this metric is marked in green in the console.
adbpg_connection_status	Instance connection status	N/A	The connection health status of the instance. Valid values: critical: The instance connection usage is greater than 95%. In this case, this metric is marked in red in the console. warning: The instance connection usage is greater than 90% and less than or equal to 95%. In this case, this metric is marked in yellow in the console. healthy: The instance connection usage is less than or equal to 90%. In this case, this metric is marked in green in the console. Note The instance connection usage is the maximum connection usage of all coordinator nodes and compute nodes.
node_master_status	Coordinator node health status	N/A	The health status of coordinator nodes. Valid values: critical: A coordinator node is unavailable. In this case, this metric is marked in red in the console. healthy: All coordinator nodes are available. In this case, this metric is marked in green in the console.
node_master_connection_status	Coordinator node connection status	N/A	The connection status of coordinator nodes. Valid values: critical: The connection usage of coordinator nodes is greater than or equal to 95%. In this case, this metric is marked in red in the console. warning: The connection usage of coordinator nodes is greater than or equal to 90% and less than 95%. In this case, this metric is marked in yellow in the console. healthy: The connection usage of coordinator nodes is less than 90%. In this case, this metric is marked in green in the console. Note The coordinator node connection usage is the maximum connection usage of all coordinator nodes.
node_segment_connection_status	Compute node connection status	N/A	The connection status of compute nodes. Valid values: critical: The connection usage of compute nodes is greater than or equal to 95%. In this case, this metric is marked in red in the console. warning: The connection usage of compute nodes is greater than or equal to 90% and less than 95%. In this case, this metric is marked in yellow in the console. healthy: The connection usage of compute nodes is less than 90%. In this case, this metric is marked in green in the console. Note The compute node connection usage is the maximum connection usage of all compute nodes.
adbpg_master_disk_usage_percent_max	Maximum storage usage of coordinator nodes	%	The status corresponding to the maximum storage usage of all coordinator nodes. Valid values: critical: The storage usage of coordinator nodes is greater than or equal to 90%. In this case, the instance is locked. warning: The storage usage of coordinator nodes is greater than or equal to 70% and less than 90%. healthy: The storage usage of coordinator nodes is less than 70%.

You can call the DescribeDBClusterPerformance operation to return the following metric parameters.

Parameter	Metric	Unit	Description
adbpg_acu_used	ACU usage	AnalyticDB compute unit (ACU)	The amount of computing resources that are used by the instance.
adbpg_segment_cnt	Number of compute nodes	N/A	The number of compute nodes that are collected per hour.
adbpg_instance_disk_used_mb	Instance storage used	MB	The amount of Object Storage Service (OSS) storage space that is used by the instance.
node_master_connection_count	Number of coordinator node connections	N/A	The number of connections to coordinator nodes.
node_segment_connection_count	Number of compute node connections	N/A	The number of connections to compute nodes.
node_segment_workfile_used_mb	Temporary disk file size	MB	The size of temporary files that are written to disks.
node_cpu_used_percent	CPU utilization	%	The CPU utilization of coordinator nodes or compute nodes.
node_mem_used_percent	Memory usage	%	The memory usage of coordinator nodes or compute nodes.