Load Monitoring

Load Monitoring

1. Log in to the GaussDB(DWS) management console.
2. On the Clusters page, locate the target cluster.
3. In the Operation column of the target cluster, click Monitoring Panel.
4. In the navigation pane on the left, choose Monitoring > Load Monitoring.

   On the Load Monitoring page, you can view the real-time and historical resource consumption of workload queues.

Workload Queues

The DMS displays the user-defined workload queue name, real-time and historical resource consumption, and workload queue resource quotas.

• Workload Queues: name of a workload queue
• Monitoring: You can click the monitoring icon to display the historical consumption trends of resources such as the CPU, memory, and disk.
• CPU Usage (%): real-time CPU usage of a workload queue
• CPU Resource (%): CPU usage quotas of a workload queue
• Real-Time Concurrent Short Queries: number of concurrent simple queries in a workload queue. Concurrent simple queries are not controlled by the workload queue.
• Concurrent Short Queries: simple concurrency quotas of the workload queue
• Real-Time Concurrent Queries: number of concurrent complex queries in a workload queue. Concurrent complex queries are controlled by the workload queue.
• Query Concurrency: complex concurrency quotas of the workload queue
• Operation

Exception Handling Rules

You can click the drop-down list of any workload queue to view the exception handling rule configured for the workload queue.

• Name: type of the supported exception handling rule, including:
  • blocktime: query blocking time. The unit is seconds.
  • elapsedtime: execution duration of a query. The unit is seconds.
  • allcputime: total CPU time spent in executing a query on all DNs. The unit is seconds.
  • cpuskewpercent: CPU time skew of a query executed on DNs. The value depends on the setting of qualificationtime.
  • qualificationtime: interval for checking the CPU skew. The unit is seconds. This parameter must be set together with cpuskewpercent.
  • spillsize: amount of query data written to disks on DNs. The unit is MB.
  • broadcastsize: size of broadcast operators of a query on DNs. The unit is MB.
  • mem_limit: maximum memory used by a query on a single instance. The unit is MB.
• Type: supported exception handling rule type. The options can be abort and penalty.
• Value: rule value, which ranges from 0 to UINT_MAX.

Waiting Queries

You can view the waiting queries in a workload queue in real time to identify the service pressure.

• User: user name of a query statement
• Application: application name of a query statement
• Database: name of the database to which a query statement is connected
• Queuing Status: queuing status of a query statement in a workload queue
• Wait Time: waiting time before a query statement is executed, in ms
• Workload Queue: workload queue to which a query statement belongs
• Query Statement: details of a query statement submitted by a user

Circuit Breaking Queries

You can view the status of a triggered circuit breaking query in a workload queue.

• Query ID: ID of a circuit breaking query
• Query Statement: circuit breaking query statement
• Blocking Time (ms): blocking time of a circuit breaking statement, in ms
• Execution Time (ms): execution time of a circuit breaking statement, in ms
• CPU Time (ms): CPU time consumed by a circuit breaking statement, in ms
• CPU Skew (%): CPU skew of a circuit breaking statement on each DN
• Exception Handling: exception handling method of a circuit breaking statement
• Status: real-time status of a circuit breaking statement