Sdílet prostřednictvím


sys.dm_resource_governor_workload_groups (Transact-SQL)

Returns workload group statistics and the current in-memory configuration of the workload group. This view can be joined with sys.dm_resource_governor_resource_pools to get the resource pool name.

Column name

Data type

Description

group_id

int

ID of the workload group. Is not nullable.

name

sysname

Name of the workload group. Is not nullable.

pool_id

int

ID of the resource pool. Is not nullable.

statistics_start_time

datetime

Time that statistics collection was reset for the workload group. Is not nullable.

total_request_count

bigint

Cumulative count of completed requests in the workload group. Is not nullable.

total_queued_request_count

bigint

Cumulative count of requests queued after the GROUP_MAX_REQUESTS limit was reached. Is not nullable.

active_request_count

int

Current request count. Is not nullable.

queued_request_count

int

Current queued request count. Is not nullable.

total_cpu_limit_violation_count

bigint

Cumulative count of requests exceeding the CPU limit. Is not nullable.

total_cpu_usage_ms

bigint

Cumulative CPU usage, in milliseconds, by this workload group. Is not nullable.

max_request_cpu_time_ms

bigint

Maximum CPU usage, in milliseconds, for a single request. Is not nullable.

NoteNote
This is a measured value, unlike request_max_cpu_time_sec, which is a configurable setting. For more information, see CPU Threshold Exceeded Event Class.

blocked_task_count

int

Current count of blocked tasks. Is not nullable.

total_lock_wait_count

bigint

Cumulative count of lock waits that occurred. Is not nullable.

total_lock_wait_time_ms

bigint

Cumulative sum of elapsed time, in milliseconds, a lock is held. Is not nullable.

total_query_optimization_count

bigint

Cumulative count of query optimizations in this workload group. Is not nullable.

total_suboptimal_plan_generation_count

bigint

Cumulative count of suboptimal plan generations that occurred in this workload group due to memory pressure. Is not nullable.

total_reduced_memgrant_count

bigint

Cumulative count of memory grants that reached the maximum query size limit. Is not nullable.

max_request_grant_memory_kb

bigint

Maximum memory grant size, in kilobytes, of a single request since the statistics were reset. Is not nullable.

active_parallel_thread_count

bigint

Current count of parallel thread usage. Is not nullable.

importance

sysname

Current configuration value for the relative importance of a request in this workload group. Importance is one of the following, with Medium being the default:

  • Low

  • Medium

  • High

Is not nullable.

request_max_memory_grant_percent

int

Current setting for the maximum memory grant, as a percentage, for a single request. Is not nullable.

request_max_cpu_time_sec

int

Current setting for maximum CPU use limit, in seconds, for a single request. Is not nullable.

request_memory_grant_timeout_sec

int

Current setting for memory grant time-out, in seconds, for a single request. Is not nullable.

group_max_requests

int

Current setting for the maximum number of concurrent requests. Is not nullable.

max_dop

int

Maximum degree of parallelism for the workload group. The default value, 0, uses global settings. Is not nullable.

Remarks

This dynamic management view shows the in-memory configuration. To see the stored configuration metadata, use the sys.resource_governor_workload_groups catalog view.

When ALTER RESOURCE GOVERNOR RESET STATISTICS is successfully executed, the following counters are reset: statistics_start_time, total_request_count, total_queued_request_count, total_cpu_limit_violation_count, total_cpu_usage_ms, max_request_cpu_time_ms, total_lock_wait_count, total_lock_wait_time_ms, total_query_optimization_count, total_suboptimal_plan_generation_count, total_reduced_memgrant_count, and max_request_grant_memory_kb. statistics_start_time is set to the current system date and time, the other counters are set to zero (0).

Permissions

Requires VIEW SERVER STATE permission.