Job Management Partner 1/Performance Management - Agent Option for Platform Description, User's Guide and Reference

1.3.3 Definition examples

The following provides examples of definitions for the monitoring template and definitions for items not included in the monitoring template for each monitored resource. The following notes apply to reading the definition examples:

In the examples, the PFM - Web Console check boxes are shown as follows:
(selected) and (not selected)

In the examples, the PFM - Web Console radio buttons are shown as follows:
(selected) and (not selected)

In the examples, xxx, yyy, zzz, and dummy are variables that the user replaces with the character strings appropriate for the system environment. For other definition items, the values should be changed as required.

In the examples, the proper values for the frequency of occurrence settings (for example, m occurrence(s) during n interval(s)) differ depending on the system environment. Accordingly, specify the appropriate values. For example, assume that the status whose threshold has been exceeded for at least two minutes in the system environment is the high-load status. Further assume that the collection interval is 60 seconds and that the maximum for the number of times that the threshold can be exceeded is twice per five intervals. Under these conditions, an unacceptable high-load condition occurs when the threshold is exceeded at least three times per five intervals. The setting in this case is 3 occurrence(s) during 5 interval(s).

Organization of this subsection

(1) Processor

(2) Memory

(3) Disks

(4) Network

(5) Processes

(1) Processor

The following shows definition examples for the monitoring template and for items not included in the monitoring template.

(a) Monitoring template

l Processor-related monitoring template alarms

Processor-related alarms in the monitoring template are stored in the alarm table for PFM UNIX Template Alarms [CPU] 09.00.

Table 1-8 Processor-related monitoring template alarms

Monitoring template alarm Record Field Error threshold Warning threshold Description

Kernel CPU PI Kernel CPU % > 75 > 50 If the percentage of time during which the processor operates continues to be above the threshold, there might be a problem with the OS or the operation method.^#

User CPU PI User CPU % > 85 > 65 If the percentage of time during which the processor operates continues to be above the threshold, there might be a problem with a specific application.^#

Run Queue PI 5-Minute Run Queue Avg > 4 > 8 If the average number of threads is above the threshold, there might be a problem with the OS, operation method, or a specific application.^#

CPU Per Processor(K) PI_CPUP Processor ID >= 0 >= 0 If the CPU usage continues to be at or above the threshold, there might be a problem with the OS or operation method.

System % > 75 > 50

CPU Per Processor(U) PI_CPUP Processor ID >= 0 >= 0 If the CPU usage continues to be at or above the threshold, there might be a problem with a specific application.

User % > 85 > 65

#

Any processes that are excessively using the processor must be found, and appropriate action must be taken. If no such processes exist, operations that outstrip the kernel's scheduling capabilities, such as the generation and deletion of many processes in a short time, might be occurring. In such a case, because the system environment is not adequate for the processing, you might need to upgrade the processor or add processors.

If you want to perform more detailed monitoring of processor performance you can use your own alarms or reports in addition to the existing alarms. For details about how to create your own alarms or reports, see 1.3.2(1)(b) Monitoring methods.

For details about the settings for the existing alarms, see 5. Monitoring Templates.

l Processor-related monitoring template reports

Table 1-9 Processor monitoring template reports

Report name Displayed information

CPU Per Processor Status Real-time information about the processor status

CPU Per Processor Usage Real-time information about the CPU usage for a processor

CPU Trend Daily CPU usage for a host for the last month

CPU Usage Summary A history of the CPU usage on a minute-by-minute basis for the last hour

CPU Status(Multi-Agent) An hourly history of the CPU usage by multiple hosts for the last 24 hours

CPU Status Real-time information about the CPU usage

CPU Trend(Multi-Agent) A daily history of the CPU usage by multiple hosts for the last month

For details about settings for existing reports, see 5. Monitoring Templates.

(b) Definition examples other than for monitoring templates

l Real-time report for checking processes whose processor usage is high

Table 1-10 Definition example

Item Explanation

Name and Type Report name PD_PDI - Memory

Product UNIX (6.0)

Report type Real-time (single agent) (Select)

Historical (single agent) --

Historical (multiple agents) --

Field Record PD_PDI

Selected fields Program
PDI
CPU %
System CPU
User CPU

Filter Conditional expression: (Select Simple or Complex.)
PID <> "0"

Specify when displayed (Clear)

Indication settings Specify when displayed (Select)

Indicate delta value (Clear)

Refresh interval Do not refresh automatically (Clear)

Initial value 30

Minimum value 30

Display by ranking Field CPU %

Display number 10^#

In descending order (Clear)

Components Table All fields

List --

Graph System CPU
User CPU

Display key Field (None)

In descending order --

Graph Graph type Stacked bar graph

Series direction Row

Axis labels X-axis Process name (process ID)

Y-axis CPU usage time

Data label Data label 1 Process name

Data label 2 Process ID

Drilldown Report drilldown Arbitrary

Field drilldown Arbitrary

Legend:

--: Do not specify this item.

#

Specify a value appropriate for the circumstances.

Monitoring template alarm	Record	Field	Error threshold	Warning threshold	Description
Kernel CPU	PI	Kernel CPU %	> 75	> 50	If the percentage of time during which the processor operates continues to be above the threshold, there might be a problem with the OS or the operation method.^#
User CPU	PI	User CPU %	> 85	> 65	If the percentage of time during which the processor operates continues to be above the threshold, there might be a problem with a specific application.^#
Run Queue	PI	5-Minute Run Queue Avg	> 4	> 8	If the average number of threads is above the threshold, there might be a problem with the OS, operation method, or a specific application.^#
CPU Per Processor(K)	PI_CPUP	Processor ID	>= 0	>= 0	If the CPU usage continues to be at or above the threshold, there might be a problem with the OS or operation method.
System %	> 75	> 50
CPU Per Processor(U)	PI_CPUP	Processor ID	>= 0	>= 0	If the CPU usage continues to be at or above the threshold, there might be a problem with a specific application.
User %	> 85	> 65

Report name	Displayed information
CPU Per Processor Status	Real-time information about the processor status
CPU Per Processor Usage	Real-time information about the CPU usage for a processor
CPU Trend	Daily CPU usage for a host for the last month
CPU Usage Summary	A history of the CPU usage on a minute-by-minute basis for the last hour
CPU Status(Multi-Agent)	An hourly history of the CPU usage by multiple hosts for the last 24 hours
CPU Status	Real-time information about the CPU usage
CPU Trend(Multi-Agent)	A daily history of the CPU usage by multiple hosts for the last month

Item	Explanation
Name and Type	Report name	PD_PDI - Memory
Product	UNIX (6.0)
Report type	Real-time (single agent)	(Select)
Historical (single agent)	--
Historical (multiple agents)	--
Field	Record	PD_PDI
Selected fields	Program PDI CPU % System CPU User CPU
Filter	Conditional expression:	(Select Simple or Complex.) PID <> "0"
Specify when displayed	(Clear)
Indication settings	Specify when displayed	(Select)
Indicate delta value	(Clear)
Refresh interval	Do not refresh automatically	(Clear)
Initial value	30
Minimum value	30
Display by ranking	Field	CPU %
Display number	10^#
In descending order	(Clear)
Components	Table	All fields
List	--
Graph	System CPU User CPU
Display key	Field	(None)
In descending order	--
Graph	Graph type	Stacked bar graph
Series direction	Row
Axis labels	X-axis	Process name (process ID)
Y-axis	CPU usage time
Data label	Data label 1	Process name
Data label 2	Process ID
Drilldown	Report drilldown	Arbitrary
Field drilldown	Arbitrary

(2) Memory

The following shows definition examples for the monitoring template and for items not included in the monitoring template.

(a) Monitoring template

l Memory-related monitoring template alarms

Memory-related alarms in the monitoring template are stored in the alarm table for PFM UNIX Template Alarms [MEM] 09.00.

Table 1-11 Memory monitoring template alarms

Monitoring template alarm Record Field Error threshold Warning threshold Description

Pagescans PI Page Scans/sec > 150 > 100 If the number of page scans that occurred is greater than the threshold, memory might be insufficient.

Swap Outs PI Swapped-Out Pages/sec > 200 > 100 If the number of pages is greater than the threshold, memory might be insufficient.

Alloc Mem Mbytes PI Alloc Swap Mbytes >= 1800 >= 1024 If the amount of used memory continues to be at or above the threshold (the Total Physical Mem Mbytes field value of the PI record), physical memory might be insufficient.

For details about the settings for the existing alarms, see 5. Monitoring Templates.

l Memory-related monitoring template reports

Table 1-12 Memory monitoring template reports

Report name Displayed information

Memory Paging A history of memory use on a minute-by-minute basis for the last hour^#

Memory Paging Status Real-time information about memory usage and the paging status^#

Memory Paging Status (Multi-Agent) An hourly history of memory usage by multiple hosts for the last 24 hours^#

Paging Trend (Multi-Agent) A daily history of page scans performed by multiple hosts for the last month^#

#

This report cannot be used in Linux.

l System-related monitoring template reports (for memory)

Table 1-13 System monitoring template reports

Report name Displayed information

I/O Overview A history of the number of I/Os on a minute-by-minute basis for the last hour. This report cannot be used in Linux.

Process Trend A daily history of the number of operated processes for the last month

System Overview (real-time report indicating the system operating status) Real-time information about the system operating status

System Overview (historical report indicating the system operating status) A history of the system operating status on a minute-by-minute basis for the last hour

System Utilization Status Real-time information about the system operating status

Workload Status Real-time information about the system workload

Workload Status (Multi-Agent) An hourly history of the workload for multiple hosts for the last 24 hours

For details about the settings for the existing reports, see 5. Monitoring Templates.

(b) Definition example other than for monitoring templates

l Real-time report for checking the memory usage of a process

Table 1-14 Definition example

Item Explanation

Name and Type Report name PD_PDI - Memory

Product UNIX (6.0)

Report type Real-time (single agent) (Select)

Historical (single agent) --

Historical (multiple agents) --

Field Record PD_PDI

Selected fields Select all fields.

Filter Conditional expression: (Select Simple or Complex.)
PID <> "0"

Specify when displayed (Clear)

Indication settings Specify when displayed (Select)

Indicate delta value (Clear)

Refresh interval Do not refresh automatically (Clear)

Initial value 30

Minimum value 30

Display by ranking Field Virtual Mem Kbytes^#1

Display number 30^#2

In descending order (Select)

Components Table Program
PID
Real Mem Kbytes
Virtual Mem Kbytes
Major Faults
Swaps
Context Switches
CPU %

List --

Graph Virtual Mem Kbytes
Real Mem Kbytes

Display name --

Display key Field (None)

In descending order --

Graph Graph type Line graph

Series direction Row

Axis labels X-axis Time

Y-axis Memory usage

Data label Data label 1 (None)

Data label 2 (None)

Drilldown Report drilldown Arbitrary

Field drilldown Arbitrary

Legend:

--: Do not specify this item.

#1

Set the fields that you want to monitor.

#2

Specify a value appropriate for the circumstances.

Monitoring template alarm	Record	Field	Error threshold	Warning threshold	Description
Pagescans	PI	Page Scans/sec	> 150	> 100	If the number of page scans that occurred is greater than the threshold, memory might be insufficient.
Swap Outs	PI	Swapped-Out Pages/sec	> 200	> 100	If the number of pages is greater than the threshold, memory might be insufficient.
Alloc Mem Mbytes	PI	Alloc Swap Mbytes	>= 1800	>= 1024	If the amount of used memory continues to be at or above the threshold (the Total Physical Mem Mbytes field value of the PI record), physical memory might be insufficient.

Report name	Displayed information
Memory Paging	A history of memory use on a minute-by-minute basis for the last hour^#
Memory Paging Status	Real-time information about memory usage and the paging status^#
Memory Paging Status (Multi-Agent)	An hourly history of memory usage by multiple hosts for the last 24 hours^#
Paging Trend (Multi-Agent)	A daily history of page scans performed by multiple hosts for the last month^#

Report name	Displayed information
I/O Overview	A history of the number of I/Os on a minute-by-minute basis for the last hour. This report cannot be used in Linux.
Process Trend	A daily history of the number of operated processes for the last month
System Overview (real-time report indicating the system operating status)	Real-time information about the system operating status
System Overview (historical report indicating the system operating status)	A history of the system operating status on a minute-by-minute basis for the last hour
System Utilization Status	Real-time information about the system operating status
Workload Status	Real-time information about the system workload
Workload Status (Multi-Agent)	An hourly history of the workload for multiple hosts for the last 24 hours

Item	Explanation
Name and Type	Report name	PD_PDI - Memory
Product	UNIX (6.0)
Report type	Real-time (single agent)	(Select)
Historical (single agent)	--
Historical (multiple agents)	--
Field	Record	PD_PDI
Selected fields	Select all fields.
Filter	Conditional expression:	(Select Simple or Complex.) PID <> "0"
Specify when displayed	(Clear)
Indication settings	Specify when displayed	(Select)
Indicate delta value	(Clear)
Refresh interval	Do not refresh automatically	(Clear)
Initial value	30
Minimum value	30
Display by ranking	Field	Virtual Mem Kbytes^#1
Display number	30^#2
In descending order	(Select)
Components	Table	Program PID Real Mem Kbytes Virtual Mem Kbytes Major Faults Swaps Context Switches CPU %
List	--
Graph	Virtual Mem Kbytes Real Mem Kbytes
Display name	--
Display key	Field	(None)
In descending order	--
Graph	Graph type	Line graph
Series direction	Row
Axis labels	X-axis	Time
Y-axis	Memory usage
Data label	Data label 1	(None)
Data label 2	(None)
Drilldown	Report drilldown	Arbitrary
Field drilldown	Arbitrary

(3) Disks

The following shows definition examples for the monitoring template.

(a) Monitoring template

l Disk-related monitoring template alarms

Disk-related alarms in the monitoring template are stored in the alarm table for PFM UNIX Template Alarms [DSK] 09.00.

Table 1-15 Disk monitoring template alarms

Monitoring template alarms Record Field Error threshold Warning threshold Description

I/O Wait Time PI Wait % > 80 > 60 If the percentage of time during which the processor is waiting for I/O is greater than the threshold, I/O operations, such as updating a database, might be delayed.

Disk Service Time PI_DEVD Avg Service Time > 0.1 > 0.06 If the average operating time is greater than the threshold, the amount of information being input or output might be very large.

File System Free(L) PD_FSL File System <> dummy <> dummy If there is little unused area, disk space is insufficient.

Mbytes Free < 5120 < 10240

File System Free(R) PD_FSR File System <> dummy <> dummy If there is little unused area, disk space is insufficient.

Mbytes Free < 5120 < 10240

Disk Busy % PI_DEVD Device Name <> dummy <> dummy If the disk busy rate continues to be at or above the threshold, I/O operations might be concentrated on a specific disk.

Busy % >= 90 >= 80

Disk Queue PI_DEVD Device Name <> dummy <> dummy If the queue length continues to be at or above the threshold, the device is congested.

Queue Length >= 5 >= 3

For details about the settings for the existing alarms, see 5. Monitoring Templates.

l Disk-related monitoring template reports

Table 1-16 Disk monitoring template reports

Report name Displayed information

Avg Service Time - Top 10 Devices Real-time information about the 10 devices with the longest average operating time

Avg Service Time Status - Top 10 Devices Real-time information about the 10 devices with the longest average operating time

Device Detail Real-time information about the selected device

Device Usage Status Real-time information about device usage

Device Usage Status (Multi-Agent) An hourly history of device usage by multiple hosts for the last 24 hours

Free Space Mbytes - Top 10 Local File Systems Real-time information about the 10 local file systems with the smallest amount of free space

Local File System Detail Real-time information about the selected local file system

Remote File System Detail Real-time information about the selected remote file system

Space Usage - Top 10 Local File Systems Real-time information about the 10 local file systems with the highest usage

Space Usage - Top 10 Remote File Systems Real-time information about the 10 remote file systems with the highest usage

NFS Activity Overview A history of the operating status of the NFS clients and server on a minute-by-minute basis for the last hour^#

NFS Load Trend A daily history of the operating status of the NFS clients and server for the last month^#

NFS Usage Status Real-time information about the operating status of the NFS clients and server^#

NFS Usage Status (Multi-Agent) An hourly history of the operating status of the NFS clients and server for the last 24 hours^#

#

This report cannot be used in Linux.

For details about the settings for the existing reports, see 5. Monitoring Templates.

Monitoring template alarms	Record	Field	Error threshold	Warning threshold	Description
I/O Wait Time	PI	Wait %	> 80	> 60	If the percentage of time during which the processor is waiting for I/O is greater than the threshold, I/O operations, such as updating a database, might be delayed.
Disk Service Time	PI_DEVD	Avg Service Time	> 0.1	> 0.06	If the average operating time is greater than the threshold, the amount of information being input or output might be very large.
File System Free(L)	PD_FSL	File System	<> dummy	<> dummy	If there is little unused area, disk space is insufficient.
Mbytes Free	< 5120	< 10240
File System Free(R)	PD_FSR	File System	<> dummy	<> dummy	If there is little unused area, disk space is insufficient.
Mbytes Free	< 5120	< 10240
Disk Busy %	PI_DEVD	Device Name	<> dummy	<> dummy	If the disk busy rate continues to be at or above the threshold, I/O operations might be concentrated on a specific disk.
Busy %	>= 90	>= 80
Disk Queue	PI_DEVD	Device Name	<> dummy	<> dummy	If the queue length continues to be at or above the threshold, the device is congested.
Queue Length	>= 5	>= 3

Report name	Displayed information
Avg Service Time - Top 10 Devices	Real-time information about the 10 devices with the longest average operating time
Avg Service Time Status - Top 10 Devices	Real-time information about the 10 devices with the longest average operating time
Device Detail	Real-time information about the selected device
Device Usage Status	Real-time information about device usage
Device Usage Status (Multi-Agent)	An hourly history of device usage by multiple hosts for the last 24 hours
Free Space Mbytes - Top 10 Local File Systems	Real-time information about the 10 local file systems with the smallest amount of free space
Local File System Detail	Real-time information about the selected local file system
Remote File System Detail	Real-time information about the selected remote file system
Space Usage - Top 10 Local File Systems	Real-time information about the 10 local file systems with the highest usage
Space Usage - Top 10 Remote File Systems	Real-time information about the 10 remote file systems with the highest usage
NFS Activity Overview	A history of the operating status of the NFS clients and server on a minute-by-minute basis for the last hour^#
NFS Load Trend	A daily history of the operating status of the NFS clients and server for the last month^#
NFS Usage Status	Real-time information about the operating status of the NFS clients and server^#
NFS Usage Status (Multi-Agent)	An hourly history of the operating status of the NFS clients and server for the last 24 hours^#

(4) Network

The following shows definition examples for the monitoring template.

(a) Monitoring template

l Network-related monitoring template alarms

Network-related alarms in the monitoring template are stored in the alarm table for PFM UNIX Template Alarms [NET] 09.00.

Table 1-17 Network monitoring template alarms

Monitoring template alarm Record used Field used Abnormal condition Warning condition Meaning

Network Rcvd/sec PI_NINS Pkts Rcvd/sec >= 9 >= 8 If the number of packets is large, many packets have been received successfully.

For details about the settings for the existing alarms, see 5. Monitoring Templates.

l Network-related monitoring template reports

Table 1-18 Network monitoring template reports

Report name Displayed information

Network Interface Detail Real-time information about network usage of the selected system

Network Interface Summary (real-time report indicating the network usage) Real-time information about network usage

Network Interface Summary (historical report indicating the network usage) A history of network usage on a minute-by-minute basis for the last hour

Network Overview A history of network usage on a minute-by-minute basis for the last hour

Network Status (Multi-Agent) An hourly history of the network usage by multiple hosts for the last 24 hours

Network Status Real-time information about network usage

For details about the settings for the existing reports, see 5. Monitoring Templates.

Monitoring template alarm	Record used	Field used	Abnormal condition	Warning condition	Meaning
Network Rcvd/sec	PI_NINS	Pkts Rcvd/sec	>= 9	>= 8	If the number of packets is large, many packets have been received successfully.

Report name	Displayed information
Network Interface Detail	Real-time information about network usage of the selected system
Network Interface Summary (real-time report indicating the network usage)	Real-time information about network usage
Network Interface Summary (historical report indicating the network usage)	A history of network usage on a minute-by-minute basis for the last hour
Network Overview	A history of network usage on a minute-by-minute basis for the last hour
Network Status (Multi-Agent)	An hourly history of the network usage by multiple hosts for the last 24 hours
Network Status	Real-time information about network usage

(5) Processes

The following shows definition examples for the monitoring template.

(a) Monitoring template

l Process-related monitoring template alarms

Process-related monitoring template alarms are stored in the alarm table for PFM UNIX Template Alarms [PS] 09.00.

Table 1-19 Process monitoring template alarms

Monitoring template alarms Record used Field used Abnormal condition Warning condition Meaning

Process End PD_PDI Program = jpcsto = jpcsto If performance data is not collected, the process has stopped.

Process Alive PI_WGRP Process Count > 0 > 0 This indicates that the workgroup process is running.

Workgroup = workgroup = workgroup

For details about the settings for the existing alarms, see 5. Monitoring Templates.

l Process-related monitoring template reports

Table 1-20 Process monitoring template reports

Report name Displayed information

CPU Usage - Top 10 Processes Real-time information about the 10 processes with the highest CPU usage

I/O Activity - Top 10 Processes Real-time information about the 10 processes that processed the most I/O operations. This report cannot be used in HP-UX, AIX, and Linux.

Major Page Faults - Top 10 Processes Real-time information about the 10 processes with the most page faults causing physical I/O operations

Process Detail Real-time information about the processes on the selected host

Process Overview A history of the process operating status on a minute-by-minute basis for the last hour

Process Summary Status Real-time information about the process operating status

For details about the settings for the existing reports, see 5. Monitoring Templates.

Monitoring template alarms	Record used	Field used	Abnormal condition	Warning condition	Meaning
Process End	PD_PDI	Program	= jpcsto	= jpcsto	If performance data is not collected, the process has stopped.
Process Alive	PI_WGRP	Process Count	> 0	> 0	This indicates that the workgroup process is running.
Workgroup	= workgroup	= workgroup

Report name	Displayed information
CPU Usage - Top 10 Processes	Real-time information about the 10 processes with the highest CPU usage
I/O Activity - Top 10 Processes	Real-time information about the 10 processes that processed the most I/O operations. This report cannot be used in HP-UX, AIX, and Linux.
Major Page Faults - Top 10 Processes	Real-time information about the 10 processes with the most page faults causing physical I/O operations
Process Detail	Real-time information about the processes on the selected host
Process Overview	A history of the process operating status on a minute-by-minute basis for the last hour
Process Summary Status	Real-time information about the process operating status

[Contents][Back][Next]

[Trademarks]