ComputeAnalyzer™ Data Types

(2) LSF (Batch Jobs)

This data type contains the data generated from each Load Sharing Facility (LSF) batch job at its completion.

Classifications

Name	Description
Project	The name of the project where the job belongs.
User	The user name of the submitter.
Queue name	The name of the LSF queue where the job was submitted.
Job status	The exit status of the job. OK ERROR
Command	The command that the job executed.
Exit status	The exit status from the command.
Execution host	The name of the first execution host.
Domain	The name of the domain to which the first execution host belongs
Number of exec. hosts	The number of hosts where the job was submitted.
Number of exec. CPUs	The number of CPUs used by the job.

LSF (Batch Jobs) Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of started and completed processes in the report intervals.	Number
Time in queue	The duration the job waited in queue.	Hours
System time	The system time used by the batch command.	Hours
User time	The user time used by the batch command.	Hours
CPU time	The number of CPU hours used by the batch command (User time + System time).	Hours
CPU cost	The product of the CPU hours and CPU factor.	Number
Block I	The block input done by the batch command.	Blocks
Block O	The block output done by the batch command.	Blocks
Block I/O	The sum of the block input and output.	Blocks
Max RSS	The maximum resident set size (physical memory used).	Kilobytes
Max VM	The maximum virtual memory usage.	Kilobytes
Elapsed time	The elapsed time from start to finish of the job.	Hours
Total time	The sum of Time in queue and Elapsed time.	Hours

LSF (Batch Jobs) Measurements

(3) VMstat

VMstat is used to get information about the CPU and memory activities when SAR is inaccessible. This shows how computers haven been running instead of what they have been running.

NOTE

SAR is used by default since it provides a superset of VMstat's measurements. VMstat needs to be activated manually.

warning

This data type is already deprecated.

Classifications

Name	Description
Host	The host (without the domain) that received the VMstat output.
Architecture	The architecture of the computer that executed VMstat.
OS	The operating system of the computer that executed VMstat.
Domain	The domain to which the host belongs.
Prime time (deprecated)	Indicates whether the data collection time is classified as prime or non-prime. prime - data collected during prime time period. non-prime - data collected outside the prime time period. undef - the prime time itself is undefined.

VMstat Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of times the statistics were collected.	Number
Running queue	The number of processes in the running queue.	Number
Blocking queue	The number of processes in the blocked queue.	Number
User %	The user time as percentage of the total CPU time.	Percentage
Sys %	The system time as percentage of the total CPU time.	Percentage
Idle %	The idle time as percentage of the total CPU time.	Percentage

VMstat Measurements

(4) PSM-Shell

note

This data type is no longer in use.

Classifications

Name	Description
Host	The host that received the PSM-Shell output.
Architecture	The architecture of the computer that executed PSM-Shell.
OS	The operating system of the computer that executed PSM-Shell.
User account	The user account of the submitter.
User	The user name of the submitter.
UID	The owner of the file.
GID	The group identifier of the file owner.
Exit status	The exit status from the command.
Command	The command that the job executed.

PSM-Shell Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of times the statistics were collected.	Number
Elapsed time	The elapsed time from start to finish of the job.	Hours
CPU	The number of CPU hours used by the command.	Hours
Maximum RSS	The maximum resident set size (physical memory used).	Kilobytes
Integral RSS	The integral resident set size (physical memory used).	Kilobytes
Block I	The block input operations done by the job.	Blocks
Block O	The block output operations done by the job.	Blocks

PSM-Shell Measurements

(5) LSF (LSF Acct)

note

This data type is no longer in use.

Classifications

Name	Description
UID	The owner of the file.
Host	The host that received the LSF output.
Command	The command that the job executed.
Exit status	The exit status from the command.

LSF (LSF Acct) Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of times the statistics were collected.	Number
System time	The system time used by the batch command.	Hours
User time	The user time used by the batch command.	Hours
CPU	The number of CPU hours used by the command.	Hours
Elapsed time	The elapsed time from start to finish of the job.	Hours
Block I	The block input operations done by the job.	Blocks
Block O	The block output operations done by the job.	Blocks

LSF (LSF Acct) Measurements

(6) SAR

SAR contains many of the same variables as VMstat. It also has information about file systems or how heavily the file systems have been used.

Classifications

Name	Description
Host	The host name where the statistics were collected.
Architecture	The architecture of the host.
OS	The operating system of the host.
Domain	The domain to which the host belongs.
Prime time (deprecated)	Indicates whether the data collection time is classified as prime or non-prime. prime - data collected during prime time period. non-prime - data collected outside the prime time period. undef - the prime time itself is undefined.

SAR Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of times the statistics were collected.	Number
User %	The user time as percentage of the total CPU time.	Percentage
Sys %	The system time as percentage of the total CPU time.	Percentage
Wait IO %	The wait for IO time as percentage of the total CPU time.	Percentage
CPU %	The total time or sum of User %, Sys %, and Wait IO %.	Percentage
Idle %	The idle time as percentage of the total CPU time.	Percentage
iget/s	The file access routines.	Number
namei/s	The file access routines (path name lookups).	Number
dirblk/s	The file access routines (blocks read during directory lookups).	Number
scall/s	The number of system calls per second (all kinds).	Number
sread/s	The number of system calls per second (read system calls).	Number
swrit/s	The number of system calls per second (write system calls).	Number
fork/s	The number of forks per second.	Number
exec/s	The number of execs per second.	Number
rchar/s	The number of characters transferred by read system calls (HP-UX: block devices only).	Number
wchar/s	The number of characters transferred by write system calls (HP-UX: block devices only).	Number
bread/s	The read operations between the buffer cache and disk (or block device).	Number
bwrit/s	The write operations between the buffer cache and disk (or block device).	Number
lread/s	The average number of logical reads per second from the buffer cache.	Number
lwrit/s	The average number of logical writes per second from the buffer cache.	Number
rcache%	The cache hit ratio when reading.	Percentage
wcache%	The cache hit ratio when writing.	Percentage
pread/s	The read/s from character devices using raw I/O (physical).	Number
pwrit/s	The write/s from character devices using raw I/O (physical).	Number
rq size	The average size of run queue or processes in the memory.	Number
rq occ%	The percentage run queue occupied with processes.	Percentage
sq size	The size of swap queue or processes ready to run. This is unavailable on Solaris.	Number
sq occ%	The percentage swap queue occupied with processes.	Percentage
swap in req/s	The number of swap-in requests per second (AIX: page replacements cycles/s).	Number
swap out req/s	The number of swap-out requests per second (AIX: number of page faults/s).	Number
swap in blk/s	The number of 512 bytes units swapped per second (AIX: for Solaris, number of pages).	Number
swap out blk/s	The number of 512 bytes units swapped out per second (AIX: for Solaris, number of pages).	Number
switches/s	The number of process switches per second.	Number
free swap	The average amount of free swap pages available (HP-UX: N/A).	Number
free memory	The average amount of free memory pages available (HP-UX: N/A; AIX: N/A).	Number

SAR Measurements

NOTE

Some statistics are not reported on AIX since most access to files bypasses kernel block buffering. The variables concerned are:

• bread/s
• bwrit/s
• lread/s
• lwrit/s
• rcache%
• wcache%
• pread/s
• pwrit/s

NOTE

Swap information is unavailable under all systems and varies on how it is reported.

(27) IOstat

warning

This data type is already deprecated.

Classifications

Name	Description
Host	The host that received the IOstat output.
Architecture	The architecture of the host.
OS	The operating system of the host.
Domain	The domain to which the host belongs.
Disk

IOstat Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of times the statistics were collected.	Number
Numread/s	The number of read requests per second.	Number
Numwrite/s	The number of write requests per second	Number
Numread	The number of read requests.	Number
Numwrite	The number of write requests.	Number
Read/s	The bytes read per second.	Number
Write/s	The bytes written per second.	Number
Read	The number of megabytes read per second.	Megabytes
Write	The number of megabytes written per second.	Megabytes
Waitavg	The average number of transactions waiting for service.	Number
Actavg	The average number of transactions actively being serviced	Number
Wsvc_t	The average service time in wait queue in milliseconds.	Milliseconds
Asvc_t	The average service time active transactions in milliseconds.	Milliseconds
Wait	The percent of time transactions are waiting for service.	Percentage
Busy	The percent of time the disk is busy.	Percentage

IOstat Measurements

(36) WinSAR

This data type is similar to SAR but is specialized for Windows.

Classifications

Name	Description
Host	The host where the statistics were collected.
Architecture	The architecture of the host.
OS	The operating system of the host.
Domain	The domain to which the host belongs.
Prime time (deprecated)	Indicates whether the data collection time is classified as prime or non-prime. prime - data collected during prime time period. non-prime - data collected outside the prime time period. undef - the prime time itself is undefined.

WinSAR Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of times the statistics were collected.	Number
Proc DPC Time	The percentage of time the processor spent receiving and servicing deferred procedure calls (DPCs) during the sample interval.	Percentage
Proc Interrupt Time	The percentage of time the processor spent receiving and servicing hardware interrupts during the sample interval.	Percentage
Proc Privileged Time	The percentage of the non-idle processor time spent on privileged mode.	Percentage
Proc Processor Time	The percentage of time that the processor is executing an application or operating system processes other than idle.	Percentage
Proc User Time	The percentage of the non-idle processor time spent on user mode.	Percentage
Proc Interrupts	The average number of hardware interrupts the processor receives and services per second.	Number
Idle Processor Time	The percentage of time occupied by the idle process.	Percentage
Sys Exception (per second)	The rate at which the system dispatches exceptions.	Number
Sys Context Switches (per second)	The combined rate at which all processors on the computer are switched from one thread to another.	Number
Sys File Read Operations (per second)	The combined rate of file system read requests to all devices on the computer, including requests to read from the file system cache.	Number
Sys File Write Operations (per second)	The combined rate of file system write requests to all devices on the computer, including requests to write to data in the file system cache.	Number
Sys Processor Queue Length	The number of threads in the processor queue.	Number
Sys Total Interrupts	The combined rate of hardware interrupts received and processed by all processors on the computer per second.	Number
Mem Available (in Kb)	The amount of physical memory immediately available for allocation to a process or system use.	Number
Mem Cache (in Kb)	The sum of the values of System Cache Resident Bytes, System Driver Resident Bytes, System Code Resident Bytes, and Pool Paged Resident Bytes.	Number
Mem Cache Faults (per second)	The rate at which faults occur when a page sought in the file system cache is not found and must be retrieved from elsewhere in the memory (soft fault) or disk (hard fault).	Number
Mem Commit Limit (in Kb)	The allowable amount of virtual memory to commit without the need to extend the paging file(s).	Number
Mem Committed (in Kb)	The amount of committed virtual memory.	Number
Mem Page Faults (per second)	The average number of pages faulted per second.	Number
Mem Pages (per second)	The rate at which pages are read from or written to disk to resolve hard page faults.	Number
Phys Disk Read	The rate of read operations on the disk per second.	Number
Phys Disk Write	The rate of write operations on the disk per second.	Number
Logical Disk Read	The rate at which read operations are performed on the disk per second.	Number
Logical Disk Write	The rate at which write operations are performed on the disk per second.	Number
Cache Copy Read Hits	The percentage of cache copy read requests that did not require a disk read to access the page in the cache.	Percentage
Cache Copy Reads	The rate per second at which read operations from pages of the file system cache involve a copy read.	Number
Cache Data Flush Pages	The rate per second at which the file system cache has flushed to the disk in response to a request to flush or satisfy a write-through-file write request.	Number
Cache Data Flushes	The rate per second at which the file system cache has flushed its contents to the disk in response to a request to flush or satisfy a write-through-file write request.	Number
Cache Data Maps	The rate per second at which a file system, such as NTFS, maps a page of a file into the file system cache to read the page.	Number
Cache Fast Reads	The rate per second at which read operations from the file system cache bypass the installed file system and retrieve the data directly from the cache	Number
Cache Lazy Write Flushes	The rate per second at which the Lazy Writer thread writes to the disk.	Number
Cache Lazy White Pages	The rate per second at which the Lazy Writer thread has written to the disk.	Number
Cache Read Aheads	The rate per second at which read operations from the file system cache detect sequential access to a file.	Number

WinSAR Measurements

(48) Grid Engine

This data type contains data on jobs run via Sun Grid Engine.

Classifications

Name	Description
Department	The department to which the job belongs (enterprise edition only).
Project	The project to which the job belongs (enterprise edition only).
Account	The account string specified for the job.
User name	The user name of the submitter.
Queue name	The name of the Grid Engine queue where the job was dispatched.
Job name	The name of the program run or a name set by the submitter or submitting program.
Job status	The exit status from the Grid Engine and operating system.
Execution host	The name of the host where the job was executed.
Resources	The string consisting of "bat" or "int" for batch or interactive, respectively.
Number of slots	The number of job slots assigned to the job execution.

Grid Engine Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of identical jobs.	Number
Wait time	The waiting duration before dispatching a job to a queue.	Minutes
Avg wait time	The average wait time for each job.	Minutes
System time	The system time used by the job.	Minutes
User time	The user time used by the job.	Minutes
CPU time	The total CPU time the job used (User time + System time).	Minutes
Block I	The block input operations done by the job.	Block operations
Block O	The block output operations done by the job.	Block operations
Block I/O	The sum of block input and output operations.	Block operations
Max RSS	The maximum resident set size (physical memory used).	Kilobytes
Max VM	The maximum virtual memory usage.	Kilobytes
Wall time	The elapsed time from start to finish of the job.	Minutes
Total time	The elapsed time from job submission to termination (Time in queue + Wall time).	Minutes
Integral memory usage	The integral of memory consumption over the lifespan of the job.	Gbs (Gigabytes/second)

Grid Engine Measurements

(78) LSF Pending Reason

This data type contains information on how long jobs are kept pending and why.

NOTE

A single job can have multiple reasons for pending, which can differ for different hosts. Therefore, summarizing the Elapsed time for all Pending reasons may not provide a meaningful report.

Classifications

Name	Description
Project	The project to which the job belongs.
User	The user name of the submitter.
Queue name	The name of the LSF queue where the job was submitted.
Pending reason	The reason a job is still pending.
Command	The command that the job executed.
Domain	The domain to which the first execution host belongs.
Requested resource	The string describing the resources that the pending job requires.
Number of hosts pending	The number of hosts where a given pending reason was valid.

LSF Pending Reason Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of pending reasons in a report interval.	Number
Elapsed time	The elapsed time for selected pending reasons.	Hours

LSF Pending Reason Measurements

(80) PBS Accounting

This data type contains data on jobs run via a Portable Batch System (PBS). The system collects data from Open PBS and PBS Professional.

Classifications

Name	Description
User name	The user name of the submitter.
Group name	The group name of the submitter.
Account	The account string specified for the job.
Queue name	The name of the PBS queue where the job was dispatched.
Job name	The name of the program run or a name set by the submitter or submitting program.
Exit status	The exit statuses from the PBS and operating system.
Execution host	The name of the host where the job was executed. In cases where jobs require multiple CPUs, only the host running the MPI controller job is reported.
Resources	The string consisting of the job requirements given in name-value pairs.
Number of CPUs	The number of CPUs assigned to execute the job.

PBS Accounting Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of identical jobs.	Number
Wait time	The waiting duration before dispatching a job to a queue.	Hours
Avg wait time	The average time in queue for each job.	Hours
Wall time	The elapsed time from start to finish of the job.	Hours
Total time	The elapsed time from job submission to termination (Wait time + Wall time).	Hours
CPU time	The total CPU time used by the job.	Hours
Max RSS	The maximum resident set size (physical memory used).	Kilobytes
Max VM	The maximum virtual memory usage.	Kilobytes

PBS Accounting Measurements

(133) Azure Sign-in Events

Azure Sign-in Events reports users' sign-in activity and details (including the authentication used and sign-in status) in a device, application, or service through Azure Active Directory.

Classifications

Name	Description
Time	The start time of the work period.
User	The name of the signed-in user.
Mail	The email address or principal name of the user.
Status	The sign-in status or event type of the authentication. Success Failed
Application	The name of the application that requested authentication.
Authentication	The type of authentication used.

Azure Sign-in Events Classifications

Measurements

Name	Description	Unit of Measurement
Count	The number of authentications.	Number
Events	The number of events	Number

Azure Sign-in Events Measurements

(134) Azure Resource Usage Cost

Azure Resource Usage Cost reports the workload in the cloud, which includes cost and usage data.

Classifications

Name	Description
Time	The start time of the work period.
Cost	The resource cost.
Cost USD	The resource cost in USD.
Name	The name of the subscription.
Charge Type	The charge fee type.
Publisher Type	The usage poller of the customer.
Service	The name of the Azure service.
Group	The name of the resource group.
Location	The resource location.
Currency	The currency of the Cost.

Azure Resource Usage Cost Classifications

Measurements

Name	Description	Unit of Measurement
Total Cost	The sum of cost per resource group.	Number

Azure Resource Usage Cost Measurements