Statistics describing a work queue. More...

#include <work_queue.h>

Data Fields
int	workers_connected
	Number of workers currently connected to the manager. More...

int	workers_init
	Number of workers connected, but that have not send their available resources report yet. More...

int	workers_idle
	Number of workers that are not running a task. More...

int	workers_busy
	Number of workers that are running at least one task. More...

int	workers_able
	Number of workers on which the largest task can run. More...

int	workers_joined
	Total number of worker connections that were established to the manager. More...

int	workers_removed
	Total number of worker connections that were released by the manager, idled-out, fast-aborted, or lost. More...

int	workers_released
	Total number of worker connections that were asked by the manager to disconnect. More...

int	workers_idled_out
	Total number of worker that disconnected for being idle. More...

int	workers_fast_aborted
	Total number of worker connections terminated for being too slow. More...

int	workers_blocked
	Total number of workers blocked by the manager. More...

int	workers_lost
	Total number of worker connections that were unexpectedly lost. More...

int	tasks_waiting
	Number of tasks waiting to be dispatched. More...

int	tasks_on_workers
	Number of tasks currently dispatched to some worker. More...

int	tasks_running
	Number of tasks currently executing at some worker. More...

int	tasks_with_results
	Number of tasks with retrieved results and waiting to be returned to user. More...

int	tasks_submitted
	Total number of tasks submitted to the queue. More...

int	tasks_dispatched
	Total number of tasks dispatch to workers. More...

int	tasks_done
	Total number of tasks completed and returned to user. More...

int	tasks_failed
	Total number of tasks completed and returned to user with result other than WQ_RESULT_SUCCESS. More...

int	tasks_cancelled
	Total number of tasks cancelled. More...

int	tasks_exhausted_attempts
	Total number of task executions that failed given resource exhaustion. More...

timestamp_t	time_when_started
	Absolute time at which the manager started. More...

timestamp_t	time_send
	Total time spent in sending tasks to workers (tasks descriptions, and input files.). More...

timestamp_t	time_receive
	Total time spent in receiving results from workers (output files.). More...

timestamp_t	time_send_good
	Total time spent in sending data to workers for tasks with result WQ_RESULT_SUCCESS. More...

timestamp_t	time_receive_good
	Total time spent in sending data to workers for tasks with result WQ_RESULT_SUCCESS. More...

timestamp_t	time_status_msgs
	Total time spent sending and receiving status messages to and from workers, including workers' standard output, new workers connections, resources updates, etc. More...

timestamp_t	time_internal
	Total time the queue spents in internal processing. More...

timestamp_t	time_polling
	Total time blocking waiting for worker communications (i.e., manager idle waiting for a worker message). More...

timestamp_t	time_application
	Total time spent outside work_queue_wait. More...

timestamp_t	time_workers_execute
	Total time workers spent executing done tasks. More...

timestamp_t	time_workers_execute_good
	Total time workers spent executing done tasks with result WQ_RESULT_SUCCESS. More...

timestamp_t	time_workers_execute_exhaustion
	Total time workers spent executing tasks that exhausted resources. More...

int64_t	bytes_sent
	Total number of file bytes (not including protocol control msg bytes) sent out to the workers by the manager. More...

int64_t	bytes_received
	Total number of file bytes (not including protocol control msg bytes) received from the workers by the manager. More...

double	bandwidth
	Average network bandwidth in MB/S observed by the manager when transferring to workers. More...

int	capacity_tasks
	The estimated number of tasks that this manager can effectively support. More...

int	capacity_cores
	The estimated number of workers' cores that this manager can effectively support. More...

int	capacity_memory
	The estimated number of workers' MB of RAM that this manager can effectively support. More...

int	capacity_disk
	The estimated number of workers' MB of disk that this manager can effectively support. More...

int	capacity_gpus
	The estimated number of workers' GPUs that this manager can effectively support. More...

int	capacity_instantaneous
	The estimated number of tasks that this manager can support considering only the most recently completed task. More...

int	capacity_weighted
	The estimated number of tasks that this manager can support placing greater weight on the most recently completed task. More...

int64_t	total_cores
	Total number of cores aggregated across the connected workers. More...

int64_t	total_memory
	Total memory in MB aggregated across the connected workers. More...

int64_t	total_disk
	Total disk space in MB aggregated across the connected workers. More...

int64_t	committed_cores
	Committed number of cores aggregated across the connected workers. More...

int64_t	committed_memory
	Committed memory in MB aggregated across the connected workers. More...

int64_t	committed_disk
	Committed disk space in MB aggregated across the connected workers. More...

int64_t	max_cores
	The highest number of cores observed among the connected workers. More...

int64_t	max_memory
	The largest memory size in MB observed among the connected workers. More...

int64_t	max_disk
	The largest disk space in MB observed among the connected workers. More...

int64_t	min_cores
	The lowest number of cores observed among the connected workers. More...

int64_t	min_memory
	The smallest memory size in MB observed among the connected workers. More...

int64_t	min_disk
	The smallest disk space in MB observed among the connected workers. More...

double	manager_load
	In the range of [0,1]. More...

int	total_workers_connected

int	total_workers_joined

int	total_workers_removed

int	total_workers_lost

int	total_workers_idled_out

int	total_workers_fast_aborted

int	tasks_complete

int	total_tasks_dispatched

int	total_tasks_complete

int	total_tasks_failed

int	total_tasks_cancelled

int	total_exhausted_attempts

timestamp_t	start_time

timestamp_t	total_send_time

timestamp_t	total_receive_time

timestamp_t	total_good_transfer_time

timestamp_t	total_execute_time

timestamp_t	total_good_execute_time

timestamp_t	total_exhausted_execute_time

int64_t	total_bytes_sent

int64_t	total_bytes_received

double	capacity

double	efficiency

double	idle_percentage

int64_t	total_gpus

int64_t	committed_gpus

int64_t	max_gpus

int64_t	min_gpus

int	port

int	priority

int	workers_ready

int	workers_full

int	total_worker_slots

int	avg_capacity

int	workers_blacklisted

Detailed Description

Statistics describing a work queue.

Field Documentation

◆ workers_connected

int work_queue_stats::workers_connected

Number of workers currently connected to the manager.

◆ workers_init

int work_queue_stats::workers_init

Number of workers connected, but that have not send their available resources report yet.

◆ workers_idle

int work_queue_stats::workers_idle

Number of workers that are not running a task.

◆ workers_busy

int work_queue_stats::workers_busy

Number of workers that are running at least one task.

◆ workers_able

int work_queue_stats::workers_able

Number of workers on which the largest task can run.

◆ workers_joined

int work_queue_stats::workers_joined

Total number of worker connections that were established to the manager.

◆ workers_removed

int work_queue_stats::workers_removed

Total number of worker connections that were released by the manager, idled-out, fast-aborted, or lost.

◆ workers_released

int work_queue_stats::workers_released

Total number of worker connections that were asked by the manager to disconnect.

◆ workers_idled_out

int work_queue_stats::workers_idled_out

Total number of worker that disconnected for being idle.

◆ workers_fast_aborted

int work_queue_stats::workers_fast_aborted

Total number of worker connections terminated for being too slow.

(see work_queue_activate_fast_abort)

◆ workers_blocked

int work_queue_stats::workers_blocked

Total number of workers blocked by the manager.

(Includes workers_fast_aborted.)

◆ workers_lost

int work_queue_stats::workers_lost

Total number of worker connections that were unexpectedly lost.

(does not include idled-out or fast-aborted)

◆ tasks_waiting

int work_queue_stats::tasks_waiting

Number of tasks waiting to be dispatched.

◆ tasks_on_workers

int work_queue_stats::tasks_on_workers

Number of tasks currently dispatched to some worker.

◆ tasks_running

int work_queue_stats::tasks_running

Number of tasks currently executing at some worker.

◆ tasks_with_results

int work_queue_stats::tasks_with_results

Number of tasks with retrieved results and waiting to be returned to user.

◆ tasks_submitted

int work_queue_stats::tasks_submitted

Total number of tasks submitted to the queue.

◆ tasks_dispatched

int work_queue_stats::tasks_dispatched

Total number of tasks dispatch to workers.

◆ tasks_done

int work_queue_stats::tasks_done

Total number of tasks completed and returned to user.

(includes tasks_failed)

◆ tasks_failed

int work_queue_stats::tasks_failed

Total number of tasks completed and returned to user with result other than WQ_RESULT_SUCCESS.

◆ tasks_cancelled

int work_queue_stats::tasks_cancelled

Total number of tasks cancelled.

◆ tasks_exhausted_attempts

int work_queue_stats::tasks_exhausted_attempts

Total number of task executions that failed given resource exhaustion.

◆ time_when_started

timestamp_t work_queue_stats::time_when_started

Absolute time at which the manager started.

◆ time_send

timestamp_t work_queue_stats::time_send

Total time spent in sending tasks to workers (tasks descriptions, and input files.).

◆ time_receive

timestamp_t work_queue_stats::time_receive

Total time spent in receiving results from workers (output files.).

◆ time_send_good

timestamp_t work_queue_stats::time_send_good

Total time spent in sending data to workers for tasks with result WQ_RESULT_SUCCESS.

◆ time_receive_good

timestamp_t work_queue_stats::time_receive_good

Total time spent in sending data to workers for tasks with result WQ_RESULT_SUCCESS.

◆ time_status_msgs

timestamp_t work_queue_stats::time_status_msgs

Total time spent sending and receiving status messages to and from workers, including workers' standard output, new workers connections, resources updates, etc.

◆ time_internal

timestamp_t work_queue_stats::time_internal

Total time the queue spents in internal processing.

◆ time_polling

timestamp_t work_queue_stats::time_polling

Total time blocking waiting for worker communications (i.e., manager idle waiting for a worker message).

◆ time_application

timestamp_t work_queue_stats::time_application

Total time spent outside work_queue_wait.

◆ time_workers_execute

timestamp_t work_queue_stats::time_workers_execute

Total time workers spent executing done tasks.

◆ time_workers_execute_good

timestamp_t work_queue_stats::time_workers_execute_good

Total time workers spent executing done tasks with result WQ_RESULT_SUCCESS.

◆ time_workers_execute_exhaustion

timestamp_t work_queue_stats::time_workers_execute_exhaustion

Total time workers spent executing tasks that exhausted resources.

◆ bytes_sent

int64_t work_queue_stats::bytes_sent

Total number of file bytes (not including protocol control msg bytes) sent out to the workers by the manager.

◆ bytes_received

int64_t work_queue_stats::bytes_received

Total number of file bytes (not including protocol control msg bytes) received from the workers by the manager.

◆ bandwidth

double work_queue_stats::bandwidth

Average network bandwidth in MB/S observed by the manager when transferring to workers.

◆ capacity_tasks

int work_queue_stats::capacity_tasks

The estimated number of tasks that this manager can effectively support.

◆ capacity_cores

int work_queue_stats::capacity_cores

The estimated number of workers' cores that this manager can effectively support.

◆ capacity_memory

int work_queue_stats::capacity_memory

The estimated number of workers' MB of RAM that this manager can effectively support.

◆ capacity_disk

int work_queue_stats::capacity_disk

The estimated number of workers' MB of disk that this manager can effectively support.

◆ capacity_gpus

int work_queue_stats::capacity_gpus

The estimated number of workers' GPUs that this manager can effectively support.

◆ capacity_instantaneous

int work_queue_stats::capacity_instantaneous

The estimated number of tasks that this manager can support considering only the most recently completed task.

◆ capacity_weighted

int work_queue_stats::capacity_weighted

The estimated number of tasks that this manager can support placing greater weight on the most recently completed task.

◆ total_cores

int64_t work_queue_stats::total_cores

Total number of cores aggregated across the connected workers.

◆ total_memory

int64_t work_queue_stats::total_memory

Total memory in MB aggregated across the connected workers.

◆ total_disk

int64_t work_queue_stats::total_disk

Total disk space in MB aggregated across the connected workers.

◆ committed_cores

int64_t work_queue_stats::committed_cores

Committed number of cores aggregated across the connected workers.

◆ committed_memory

int64_t work_queue_stats::committed_memory

Committed memory in MB aggregated across the connected workers.

◆ committed_disk

int64_t work_queue_stats::committed_disk

Committed disk space in MB aggregated across the connected workers.

◆ max_cores

int64_t work_queue_stats::max_cores

The highest number of cores observed among the connected workers.

◆ max_memory

int64_t work_queue_stats::max_memory

The largest memory size in MB observed among the connected workers.

◆ max_disk

int64_t work_queue_stats::max_disk

The largest disk space in MB observed among the connected workers.

◆ min_cores

int64_t work_queue_stats::min_cores

The lowest number of cores observed among the connected workers.

◆ min_memory

int64_t work_queue_stats::min_memory

The smallest memory size in MB observed among the connected workers.

◆ min_disk

int64_t work_queue_stats::min_disk

The smallest disk space in MB observed among the connected workers.

◆ manager_load

double work_queue_stats::manager_load

In the range of [0,1].

If close to 1, then the manager is at full load and spends most of its time sending and receiving taks, and thus cannot accept connections from new workers. If close to 0, the manager is spending most of its time waiting for something to happen. deprecated fields:

◆ total_workers_connected

int work_queue_stats::total_workers_connected

Deprecated:: Use workers_connected instead.

◆ total_workers_joined

int work_queue_stats::total_workers_joined

Deprecated:: Use workers_joined instead.

◆ total_workers_removed

int work_queue_stats::total_workers_removed

Deprecated:: Use workers_removed instead.

◆ total_workers_lost

int work_queue_stats::total_workers_lost

Deprecated:: Use workers_lost instead.

◆ total_workers_idled_out

int work_queue_stats::total_workers_idled_out

Deprecated:: Use workers_idled_out instead.

◆ total_workers_fast_aborted

int work_queue_stats::total_workers_fast_aborted

Deprecated:: Use workers_fast_aborted instead.

◆ tasks_complete

int work_queue_stats::tasks_complete

Deprecated:: Use tasks_with_results.

◆ total_tasks_dispatched

int work_queue_stats::total_tasks_dispatched

Deprecated:: Use tasks_dispatched instead.

◆ total_tasks_complete

int work_queue_stats::total_tasks_complete

Deprecated:: Use tasks_done instead.

◆ total_tasks_failed

int work_queue_stats::total_tasks_failed

Deprecated:: Use tasks_failed instead.

◆ total_tasks_cancelled

int work_queue_stats::total_tasks_cancelled

Deprecated:: Use tasks_cancelled instead.

◆ total_exhausted_attempts

int work_queue_stats::total_exhausted_attempts

Deprecated:: Use tasks_exhausted_attempts instead.

◆ start_time

timestamp_t work_queue_stats::start_time

Deprecated:: Use time_when_started.

◆ total_send_time

timestamp_t work_queue_stats::total_send_time

Deprecated:: Use time_send.

◆ total_receive_time

timestamp_t work_queue_stats::total_receive_time

Deprecated:: Use time_receive.

◆ total_good_transfer_time

timestamp_t work_queue_stats::total_good_transfer_time

Deprecated:: Use time_send_good + time_receive_good.

◆ total_execute_time

timestamp_t work_queue_stats::total_execute_time

Deprecated:: Use time_workers_execute.

◆ total_good_execute_time

timestamp_t work_queue_stats::total_good_execute_time

Deprecated:: Use time_workers_execute_good.

◆ total_exhausted_execute_time

timestamp_t work_queue_stats::total_exhausted_execute_time

Deprecated:: Use time_workers_execute_exhaustion.

◆ total_bytes_sent

int64_t work_queue_stats::total_bytes_sent

Deprecated:: Use bytes_sent.

◆ total_bytes_received

int64_t work_queue_stats::total_bytes_received

Deprecated:: Use bytes_received.

◆ capacity

double work_queue_stats::capacity

Deprecated:: Use capacity_cores.

◆ efficiency

double work_queue_stats::efficiency

Deprecated:: broken.

◆ idle_percentage

double work_queue_stats::idle_percentage

Deprecated:

◆ total_gpus

int64_t work_queue_stats::total_gpus

Deprecated:: : broken.

◆ committed_gpus

int64_t work_queue_stats::committed_gpus

Deprecated:: : broken.

◆ max_gpus

int64_t work_queue_stats::max_gpus

Deprecated:: : broken.

◆ min_gpus

int64_t work_queue_stats::min_gpus

Deprecated:: : broken.

◆ port

int work_queue_stats::port

Deprecated:: Use work_queue_port Port of the queue.

◆ priority

int work_queue_stats::priority

Deprecated:: Not used.

◆ workers_ready

int work_queue_stats::workers_ready

Deprecated:: Use workers_idle instead.

◆ workers_full

int work_queue_stats::workers_full

Deprecated:: Use workers_busy insead.

◆ total_worker_slots

int work_queue_stats::total_worker_slots

Deprecated:: Use tasks_running instead.

◆ avg_capacity

int work_queue_stats::avg_capacity

Deprecated:: Use capacity_cores instead.

◆ workers_blacklisted

int work_queue_stats::workers_blacklisted

Deprecated:: Use workers_blocked instead.

The documentation for this struct was generated from the following file:

work_queue.h

Data Fields