HP 3000 Manuals HP 3000 Manuals
Global Measurements
Global performance measurements reflect the activity of the system
workload--all processes running on the system. Use global graphs to
examine resource consumption or terminal transaction volume and response
times on a system-wide basis.
You can display nine sets of global measurements:
* Global Bottlenecks.
* Global Transaction Response.
* Global CPU Utilization.
* Global System CPU Utilization.
* Global Disc Summary.
* Global Disc Detail.
* Global Physical Disc by Task.
* Global Queue Depth.
* Global Jobs & Sessions.
The following pages list variables for all global graphs.
Table A-1. Global Bottlenecks--Variables
---------------------------------------------------------------------------------------------
| |
| The Global Bottlenecks graph displays measurements of the activity of three critical |
| system resources: CPU, disc, and memory. |
| |
| Y-Axis: I/Os per second (0 to 100) for Physical Disc and Memory Manager. Percent of |
| time (0 to 100) for CPU. |
| |
| X-Axis: Choose the scale using X-Axis options. |
| |
---------------------------------------------------------------------------------------------
| | |
| CPU Util | The percentage of elapsed time that the CPU is busy executing |
| | instructions. This is an indication of CPU use. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Phys Disc | The number of transfers per second made to or from the disc drives on |
| | behalf of user programs. Does not include transfers used in managing |
| | memory. This is an indication of overall disc activity. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Mem Mgr | The number of transfers per second made to or from the disc drives on |
| | behalf of the memory manager. This is an indication of swapping or memory |
| | management overhead. |
| | |
---------------------------------------------------------------------------------------------
Table A-2. Global Transaction Response--Variables
---------------------------------------------------------------------------------------------
| |
| Use the Global Transaction Response graph to examine interactive session activity and |
| response time. |
| |
| The graph displays terminal transaction rates and terminal response time measurements. |
| |
| |
| Y-Axis: Number of seconds (up to 20 on a linear scale; up to 100 on a logarithmic |
| scale), or number of transactions per hour (up to 20,000 on a linear scale; up to |
| 100,000 on a log scale with each unit representing 1000 transactions). |
| |
| X-Axis: Choose the scale using X-Axis options. |
| |
---------------------------------------------------------------------------------------------
| | |
| First | The average time between pressing return-enter on the keyboard and the |
| Response | time the first line is written to the terminal screen. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Prompt | The average time between pressing return-enter on the keyboard and the |
| | next prompt to the user. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Rate (x1000) | The number of terminal transactions per hour. (Each unit on the graph |
| | represents 1000 transactions.) A terminal transaction is the activity |
| | that occurs between reads issued to a terminal. Interactive and batch |
| | processes (but not system processes) that read and write to a terminal can |
| | generate terminal transactions. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Think | The average time between receiving a prompt at the terminal and the next |
| | carriage return or Enter. |
| | |
- - -
---------------------------------------------------------------------------------------------
NOTE On MPE XL systems before A.40.00 (2.1), the values displayed on
this graph are approximations.
Table A-3. Global CPU Utilization--Variables
---------------------------------------------------------------------------------------------
| |
| Use the Global CPU Utilization graph to see how the CPU is used and also to see |
| measurements of the major consumers of the CPU resource. |
| |
| This is a stacked graph. Each variable's value is stacked on the previous one. The |
| top line of the graph shows aggregate CPU use (including paused for disc). |
| |
| Y-Axis: Percent of CPU utilized (up to 100). Component percentages are stacked to |
| create an overall percent. |
| |
| X-Axis: Choose the scale using X-Axis options. |
| |
---------------------------------------------------------------------------------------------
| | |
| Session | The percentage of elapsed time the CPU spent executing instructions for |
| | interactive processes (sessions). |
| | |
---------------------------------------------------------------------------------------------
| | |
| Job | The percentage of elapsed time the CPU spent executing instructions for |
| | batch processes (jobs). |
| | |
---------------------------------------------------------------------------------------------
| | |
| System | The percentage of elapsed time the CPU spent executing instructions for |
| | system processes. System processes include data communications monitors, |
| | the spooler, and DSCOPY. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Other | The percentage of elapsed time the CPU was busy on other system overhead |
| | activities. These include memory management, disc caching, and interrupt |
| | processing. These components are broken out on the Global System CPU |
| | Utilization graph. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Paused | The percentage of elapsed time that the CPU is not executing any user or |
| | system code and disc I/O is occurring. (The CPU is waiting for I/O to |
| | complete.) |
| | |
---------------------------------------------------------------------------------------------
Table A-4. Global System CPU Utilization--Variables
---------------------------------------------------------------------------------------------
| |
| The Global System CPU Utilization is a stacked graph that shows the percentage of CPU |
| used by specific system overhead activities: Mem Mgr, Cache (MPE V only), Dispatch |
| (MPE XL only), ICS, Processes, Paused. |
| |
| Y-Axis: Percent of CPU utilized (up to 100). Component percentages are stacked to |
| create overall percent. |
| |
| X-Axis: Choose the scale using X-Axis options. |
| |
---------------------------------------------------------------------------------------------
| | |
| Mem Mgr CPU | The percentage of elapsed time the CPU spent managing main memory. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Cache | (MPE V only.) The percentage of elapsed time the CPU spent on disc |
| | caching activity. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Dispatch | (MPE XL only.) The percentage of elapsed time the CPU spent on |
| | dispatching processes. |
| | |
---------------------------------------------------------------------------------------------
| | |
| ICS | The percentage of elapsed time the CPU spent on other activities, such as |
| | using the interrupt control stack (ICS). The principal activity is |
| | handling interrupts. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Processes | The percentage of elapsed time the CPU spent executing code on behalf of |
| | user sessions and jobs and system processes. These components are |
| | displayed individually on the Global CPU Utilization graph. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Paused | The percentage of elapsed time that the CPU is not executing any user or |
| | system code and disc I/O is occurring. (The CPU is waiting for I/O to |
| | complete.) |
| | |
---------------------------------------------------------------------------------------------
Table A-5. Global Disc Summary--Variables
---------------------------------------------------------------------------------------------
| |
| The Global Disc Summary graph displays the read-write rates for all disc device |
| input-output activity and the peak utilization of the busiest disc. |
| |
| Three read-write rates are displayed: Logical, Physical, Mem Mgr. Peak disc |
| utilization is represented by Peak Disc Util. |
| |
| Y-Axis: I/Os per second for logical, physical, and memory manager (up to 100 on a |
| linear scale and up to 1000 on a logarithmic scale). Percent of time busy for peak |
| disc utilization (up to 100). |
| |
| X-Axis: Choose the scale using X-Axis options. |
| |
---------------------------------------------------------------------------------------------
| | |
| Logical (all | MPE V: The number of transfers per second to or from all discs that would |
| discs) | have occurred if MPE software disc caching was not on. Does not include |
| | transfers due to memory management. MPE XL: The number of read or write |
| | attempts; roughly the number of times the file system read or write |
| | routines are called. Note: MPE XL mapped files can perform physical |
| | transfers without calling file system routines, so they would not generate |
| | corresponding logical transfers. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Phys Disc | The number of transfers per second made to or from the disc drives on |
| | behalf of user programs. Does not include transfers used in managing |
| | memory. This is an indication of overall disc activity. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Mem Mgr | The number of transfers per second made to or from the disc drives on |
| | behalf of the memory manager. This is an indication of swapping or memory |
| | management overhead. |
| | |
---------------------------------------------------------------------------------------------
Table A-5. Global Disc SummaryVariables (cont.)
---------------------------------------------------------------------------------------------
| Peak Disc | The percent of time when transfers to or from the busiest disc for the |
| Util | interval were in progress. This includes the seek and physical transfer |
| | times. This metric emphasizes disc bottlenecks and can be thought of as a |
| | high watermark for disc utilization. |
| | |
| | * 5-minute detail data. |
| | |
| | Each 5-minute data point represents the disc utilization value for |
| | the disc that was busiest during the 5-minute interval. If there |
| | are multiple discs on the system and the load is well-balanced |
| | among them, it is likely that different discs are responsible for |
| | the series of peaks generated. |
| | |
| | * Hourly/daily summary data. |
| | |
| | Each hourly or daily data point represents the average of the |
| | 5-minute peak disc utilization values over the hourly or daily |
| | interval. |
| | |
---------------------------------------------------------------------------------------------
See chapter 3 for a discussion of how we compute peak disc utilization.
Table A-6. Global Disc Detail
------------------------------------------------------------------------------------------
| |
| Use the Global Disc Detail graphs to display the utilization percentage and read-write |
| rates for individual discs. |
| |
| The two Disc Detail graphs are: |
| * Busiest Discs. This graph displays first when you select Disc Detail. |
| * Individual Discs. This graph displays when you scroll down on the vertical |
| scroll bar. |
| |
------------------------------------------------------------------------------------------
Table A-7. Global Disc Detail Graph 1
---------------------------------------------------------------------------------------------
| |
| The initial display when Disc Detail is selected shows ranked utilization for the five |
| most highly used discs in the display window. |
| |
| Y-Axis: Percent of disc utilized (up to 100). |
| |
| X-Axis: Choose the scale using X-Axis options. The ldev labels on the graph are |
| ranked (left to right) in decreasing order of utilization percentage for the period. |
| Use the vertical scroll bar to see I/Os per second and utilization for up to 32 |
| individual discs. |
| |
---------------------------------------------------------------------------------------------
| | |
| Busiest | Disc with the highest average utilization for the period displayed. The |
| Logical | measurement reflected in the graph is disc utilization. |
| Device | |
| | |
---------------------------------------------------------------------------------------------
| | |
| Second | Disc with the second highest average utilization for the period displayed. |
| Busiest | The measurement reflected in the graph is disc utilization. |
| Device | |
| | |
---------------------------------------------------------------------------------------------
| | |
| Third | Disc with the third highest average utilization for the period displayed. |
| Busiest | The measurement reflected in the graph is disc utilization. |
| Device | |
| | |
---------------------------------------------------------------------------------------------
| | |
| Fourth | Disc with the fourth highest average utilization for the period displayed. |
| Busiest | The measurement reflected in the graph is disc utilization. |
| Device | |
| | |
---------------------------------------------------------------------------------------------
| | |
| Fifth | Disc with the fifth highest average utilization for the period displayed. |
| Busiest | The measurement reflected in the graph is disc utilization. |
| Device | |
| | |
---------------------------------------------------------------------------------------------
Table A-8. Global Disc Detail Graph 2
---------------------------------------------------------------------------------------------
| |
| The second Global Disc Detail graph displays individual disc activity for up to 32 |
| discs. |
| |
| Individual disc metrics: Logical, Physical, Mem Mgr, Util. |
| |
| You can select the disc to display by using the vertical scroll bar. |
| |
| Y-Axis: I/Os per second (up to 100) for logical, physical, and memory manager. |
| Percent of time (up to 100) for utilization. |
| |
| X-Axis: Choose the scale using X-Axis options. |
| |
---------------------------------------------------------------------------------------------
| | |
| Device | Indicates which logical device's data is being displayed in the graph. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Logical | MPE V: The number of transfers per second to or from the disc that would |
| | have occurred if MPE software disc caching was not on. Does not include |
| | transfers due to memory management. |
| | |
| | MPE XL: On MPE XL systems, logical disc I/Os for individual devices are |
| | not available at this time. Therefore, these values are set to zero on |
| | Global Disc Detail graphs. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Physical | The number of transfers per second made to or from the disc drives on |
| | behalf of user programs. Does not include transfers used in managing |
| | memory. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Mem Mgr | The number of transfers per second made to or from the disc drives on |
| | behalf of the memory manager. This is an indication of swapping or memory |
| | management overhead. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Util (%) | The percent of time when transfers to or from the disc are in progress. |
| | |
---------------------------------------------------------------------------------------------
Table A-9. Global Physical Disc by Task--Variables
---------------------------------------------------------------------------------------------
| |
| Use the Global Physical Disc by Task graph to examine overall disc I/O activity |
| attributable to four types of processes: Sessions, Jobs, System Processes, Memory |
| Manager. This is a stacked graph. The value of each variable is stacked on top of |
| the previous variable's value. The top line on the graph represents total transfers |
| per second to or from all discs. |
| |
| Y-Axis: I/Os per second (0 to 100 on a linear scale; up to 1000 on a logarithmic |
| scale). |
| |
| X-Axis: Choose the scale using X-Axis options. |
| |
---------------------------------------------------------------------------------------------
| | |
| Session | The number of transfers per second to and from all discs attributable to |
| | interactive user processes (sessions). |
| | |
---------------------------------------------------------------------------------------------
| | |
| Job | The number of transfers per second to and from all discs attributable to |
| | batch user processes (jobs). |
| | |
---------------------------------------------------------------------------------------------
| | |
| System | The number of transfers per second to and from all discs attributable to |
| | any system process. Examples of system processes include data |
| | communications monitors, the spooler, and DSCOPY. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Mem Mgr | The number of transfers per second made to or from the disc drives on |
| | behalf of the memory manager. This is an indication of swapping or memory |
| | management overhead. |
| | |
---------------------------------------------------------------------------------------------
Table A-10. Global Queue Depth--Variables
---------------------------------------------------------------------------------------------
| |
| The Global Queue Depth graph displays the average number of processes using or waiting |
| on resources: CPU, Disc, Memory, Impede. |
| |
| Y-Axis: Average number of processes (0 to 10) waiting on the resource. (Logarithmic |
| scaling depicts a range from 0.1 to 100.) |
| |
---------------------------------------------------------------------------------------------
| | |
| CPU | The average number of processes waiting on or using the CPU. This is an |
| | indication of system CPU load. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Disc | The average number of processes blocked, waiting on a disc resource. This |
| | is an indication of disc I/O load. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Memory | The average number of processes blocked, waiting on memory. This is an |
| | indication of memory load. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Impede | The average number of processes blocked, waiting on impede; for example, |
| | software locks or semaphores. This is an indication of impede load. |
| | |
---------------------------------------------------------------------------------------------
Table A-11. Global Jobs & Sessions--Variables
---------------------------------------------------------------------------------------------
| |
| The Global Jobs & Sessions graph shows the number of jobs and sessions logged on and |
| active. Variables are: Sessions, Busy Sess, Jobs, Busy Jobs. |
| |
| Y-Axis: Average number of jobs or sessions. Scaling is automatic. |
| |
---------------------------------------------------------------------------------------------
| | |
| Sessions | The average number of interactive sessions logged on during the time |
| | interval. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Busy Sess | The average number of interactive sessions that were active (consumed CPU) |
| | during the time interval. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Jobs | The average number of batch jobs logged on during the time interval. |
| | |
---------------------------------------------------------------------------------------------
| | |
| Busy Jobs | The average number of batch jobs that were active (consumed CPU) during |
| | the time interval. |
| | |
---------------------------------------------------------------------------------------------