Monitoring Feature
Last updated: 2025-09-10 22:19:27
Monitoring in Console
To make it easier for you to view and stay up to date with how instances work, TencentDB for PostgreSQL provides a wide variety of performance monitoring metrics. You can log in to the TencentDB for PostgreSQL console and view them on the System Monitoring tab on the corresponding instance management page.
Monitoring metrics
Metric Name | Monitoring Metric Name | Unit | Metric Type |
CPU utilization | cpu | % | Actual CPU utilization |
used storage space | storage | GB | Instance space size |
Memory Usage | memory | MBytes | Actual memory usage. |
Memory usage | memory_rate | % | Actual memory utilization |
QPS | qps | Times/second | Number of executed SQL statements per second. |
Number of Connections | connections | Counts | Number of database connections during collection. |
Number of requests | read_write_calls | Times | Number of read/write requests in a statistical period. |
Number of read requests | read_calls | Times | Number of read requests in a statistical period. |
Number of write requests | write_calls | Times | Number of write requests in a statistical period. |
Other Requests | other_calls | Times | Number of other requests in a statistical period (requests other than SELECT, INSERT, UPDATE, DELETE). |
Buffer cache hit rate | hit_percent | % | Hit rate of all SQL statements executed in a statistical period |
Average execution latency | sql_runtime_avg | ms | Average execution latency of all SQL statements in one statistical period. |
Longest TOP10 Execution Latency | sql_runtime_max | ms | Longest TOP10 average execution latency of SQL statements in a statistical period. |
Shortest TOP10 Execution Latency | sql_runtime_min | ms | Shortest TOP10 average execution latency of SQL statements in a statistical period. |
The number of remaining XIDs | remain_xid | Counts | During database collection, display the remaining xid count of the database with the least xid. No such metric for read-only instances. |
Standby log shipping and playback position difference | xlog_diff | Byte | The size difference between logs sent from the primary database to the standby and replay completed reflects the speed of log application on the standby. This metric can help evaluate standby performance and network transmission speed. No such metric for read-only instances. |
Storage Usage | storage_rate | % | Total storage utilization, including temporary files, data files, log files, and other database files. |
Number of Slow Queries | slow_query_cnt | Counts | The number of slow queries in a collection cycle. |
log write delay | xlog_diff | Byte | The size difference between logs sent from the primary database to the standby and replay completed reflects the speed of log application on the standby. This metric can help evaluate standby performance and network transmission speed. No such metric for read-only instances. |
log write time delay | xlog_diff_time | Seconds | The time difference between logs sent from the primary database to the standby and logs received and written to disk. No such metric for read-only instances, and this metric is only available for instance versions 10.x or later. |
Primary-standby data synchronization delay | slave_apply_delay | Byte | The size difference between standby database replay LSN and current LSN of primary instance. For the primary instance, this metric reflects the RTO for fault switch. For read-only instances, it reflects data latency. |
Primary-standby data sync delay time | replay_lag | Seconds | The delay time of data sync between the primary and replica databases. A shorter delay means the primary-replica data is more consistent. During primary database failure, less data is lost after the switch. |
Number of Active Connections | active_conns | Counts | Trigger database collection during instantaneous active connections (non-idle connection). |
Number of idle connections | idle_conns | Counts | Query instantaneous idle connections in the database during collection. |
Number of SQL statements with execution duration exceeding 1 second | long_query | Counts | Number of SQL statements being executed with execution time exceeding 1 second during database collection. |
Number of transactions with execution duration exceeding 1 second | long_xact | Counts | Number of transactions with execution time exceeding 1 second during database collection. |
Number of idle transactions | idle_in_xact | Counts | Number of transactions in idle in transaction state when triggering database collection. |
Number of idle transactions exceeding 5 seconds | long_idle_in_xact | Counts | Number of transactions in idle in transaction state exceeding 5 seconds during database collection. |
Number of waiting sessions | waiting | Counts | Number of database connections in a waiting state. |
Number of sessions waiting exceeding 5 seconds | long_waiting | Counts | The number of database connections with a waiting time exceeding 5 seconds. |
Number of 2PC transactions | 2PC | Counts | Number of 2PC transactions during database collection. |
Number of 2PC transactions unsubmitted for over 5 seconds | long_2pc | Counts | Number of 2PC transactions with execution time exceeding 5 seconds during database collection. |
Number of new connections within 5 seconds | new_conn_in5s | Counts | Query all connection counts established within the last 5 seconds when triggering database collection. |
TPS | tps | Times/second | Number of successful transactions per second (including rollbacks and submissions). |
Number of transaction commits | xact_commit | Times/second | Average transactions per second in a statistical period. |
Transaction rollback count | xact_rollback | Times/second | The average number of rolled-back transactions per second in a statistical period. |
Number of record deletions per second | tup_deleted | Counts | Average number of tuples deletions per second in a statistical period. No Such Metric for read-only instances. |
Number of insertions per second | tup_inserted | Counts | Average number of tuples insertions per second in a statistical period. No Such Metric for read-only instances. |
Number of record updates per second | tup_updated | Counts | Average number of tuples updates per second in a statistical period. No Such Metric for read-only instances. |
Number of index scan records per second | tup_fetched | Counts | Average number of index scan tuples per second in a statistical period. |
Number of full-table scan records per second | tup_returned | Counts | The number of full-table scan tuples per second in a statistical period. |
Number of deadlocks | deadlocks | Counts | Number of deadlocks in a statistical period. |
log file size | log_file_size | GB | WAL log file occupied space size. |
Size of data files | data_file_size | GB | Data file occupied space size. |
Temporary file size | temp_file_size | MB | Temporary file size. |
throughput rate | throughput | KBytes/s | Disk read/write speed per instance per second. |
Read Throughput | throughput_read | KBytes/s | Instance disk read speed per second. |
Write Throughput | throughput_write | KBytes/s | Disk write speed per instance per second. |
connection utilization | conn_utilization | % | Current number of instance connections/Maximum number of connections. |
instance port inbound traffic | cluster_in_flow | KBytes/s | Inbound traffic on the database instance listening port. |
instance port outbound traffic | cluster_out_flow | KBytes/s | Outbound traffic on the database instance listening port. |
Was this page helpful?
You can also Contact Sales or Submit a Ticket for help.
Yes
No
Feedback