--- name: "procmeter" section: "1" platform: "linux" package: "procmeter" version: "2.5.1-11" release: "Debian 3.1 (Sarge)" platforms: ["linux"] url: "https://new.carta.tech/procmeter" --- # procmeter(1) A simple /proc based meter (like xmeter). ## NAME procmeter - A simple /proc based meter (like xmeter). ## SYNOPSIS **procmeter** **\[-h\]** **\[-update** u **\[lu\]\]** **\[-horizontal\|-vertical\]** **\[-label** \\] **\[(-log\|-logonly)** (\\|syslog\|-)\] **\[-data** \\] **\[-force-spkt\]** **\[-force-fpkt\]** **\[cpu\]** **\[cpu-user\]** **\[cpu-nice\]** **\[cpu-sys\]** **\[cpu-idle\]** **\[load\]** **\[proc\]** **\[context\]** **\[swap\]** **\[swap-in\]** **\[swap-out\]** **\[page\]** **\[page-in\]** **\[page-out\]** **\[disk\]** **\[disk-rd\]** **\[disk-wr\]** **\[intr\]** **\[mem-free\]** **\[mem-used\]** **\[mem-buff\]** **\[mem-cache\]** **\[mem-swap\]** **\[lpkt\]** **\[fpkt\]** **\[fpkt-rx\]** **\[fpkt-tx\]** **\[fpkt-coll\]** **\[spkt\]** **\[spkt-rx\]** **\[spkt-tx\]** **\[batt-life\]** **\[batt-lvl\]** ## DESCRIPTION *ProcMeter* Displays one or meter stripchart style meters showing how the system is performing. The idea for the program came from using *xmeter(1)* and *perfmeter(1)* and wanting a **small** and **simple** version for running on a Linux box. The reason for not using xmeter is the overhead interms of the rstatd and portmap that would be required and also the availability of extra statistics in /proc for Linux. The number of meters that are displayed can be changed while the program is running. ## OPTIONS The command line options available for the program are described below. - **-h** A help message is printed giving a brief description of the usage of the program. - **-update** u **\[lu\]** Sets the update interval in seconds (default 1 second). The second argument is optional and if set is multiplied by the interval to give the time between log file updates. - **-horizontal\|-vertical** Select either the horizontal or the vertical arrangement of meters. - **-label** \ Use the specified label instead of the hostname in the window and logfile. - **-log\|-logonly** \\|syslog\|- The **-log** option with a filename will open the named file for writing and append to it a log of the current values displayed on the graph. If the filename given is **syslog** then the system log is used and if the filename is - then stdout is used. The **-logonly** option does not display anything on the screen, but does log to the log file like the -log option. - **-data** \ This option allows a named pipe to be used to pass data into the program. The filename must be a named pipe that is always readable or another file that is periodically updated. A definition file with '.def' appended (instead of '.dat' if needed) contains the information that controls how the data is displayed on the graph. Each line in the defintion file matches one line in the data file (comments are allowed in the definition file only marked by a '#' in column 0). The line must contain the name of the statistic (for the menu and command line), the units and interval for the graph. - **-force-spkt** Force the program to offer the spkt, spkt-rx, spkt-tx options even if they are not detected as network devices. (Linux kernels that use modules may not have all of the network devices loaded all of the time, and thus would not be auto-detected.) - **-force-fpkt** Force the program to offer the fpkt, fpkt-rx, fpkt-tx options even if they are not detected as network devices. - **cpu** **cpu-user** **cpu-nice** **cpu-sys** **cpu-idle** The amount of CPU usage by each of the categories (user, system, nice and idle) can be displayed individualy or a total used can be shown, 20% per division. - **load** The system load averaged over one minute. - **proc** The number of processes on the system, 10 per division. (The number of processes is only available on Linux kernels 1.2? and higher). - **context** The number of context switches, 100 per division. - **swap** **swap-in** **swap-out** The amount of swapping that is performed can be displayed for swapping in or out or a total, 50 per division. - **page** **page-in** **page-out** The amount of paging that is performed can be displayed for paging in or out or a total, 100 per division. - **disk** **disk-rd** **disk-wr** The total number of disk accesses and the number of those that were reads and writes, 25 per division. - **intr** The total number of interrupts, 100 per division. This includes the timer interrupt which always (on x86 at least) runs at 100 per second. - **mem-free** **mem-used** **mem-buff** **mem-cache** **mem-swap** The memory in the system can be displayed in terms of that used (not including buffers or cached), the amount free, the amount for buffers, the amount cached and the amount of swap used can be shown, the number of MB per division depends on the total amount of RAM. (Memory cached is only available on Linux kernels 2.0? and higher). - **lpkt** The number of network packets on local (loopback or dummy) devices, 100 packets per division. - **fpkt** **fpkt-rx** **fpkt-tx** **fpkt-coll** The number of network packets total, received and transmitted, and collisions on fast (ethernet and all others not in spkt) devices, 50 packets per division. - **spkt** **spkt-rx** **spkt-tx** The number of network packets total, received and transmitted on slow (slip, ppp, plip, flip, isdn, ippp) devices, 5 packets per division. - **batt-life** **batt-lvl** The remaining battery life on systems that have apm enabled, as a percentage and the battery level as one of four states high, low, critical and unknown. If no statistics are selected then just the menu button is displayed. ## USAGE When started the program will display all of the requested statistics in a set of vertically aligned meters. At the top (or bottom) of each of the meters is a label that indicates the statistic that is being displayed and the units. At the very top (left when horizontal) of the window is a menu buton labeled with the hostname. By selecting items from this menu, meters are toggled, if the meter already exists then it is removed, if not then a new meter is added to the bottom. In either case, the meters will all resize to allow room and make all of the meters the same size. The log file contains the a header line consisting of a '#' character followed by a list of the statistics in the file separated by spaces. The data itself is written into the file in the same order as listed in the header comment following the current time in seconds since the UNIX epoch. When the displayed statistics change, a new header line is written into the file. The data values are written to the file in absolute units, and do not use the same scaling as used for the graphs themselves. ## X RESOURCES The name of the class used is ProcMeter, the widget hierarchy used is (Widget-Class:name): ProcMeter\ Paned:pane\ MenuButton:menu\ SimpleMenu:menushell\ smeBSBObject:{statistic}\ ...\ ProcMeter:{statistic}\ ... The standard X resources that can usefully be applied to the widgets are: - **background** The colour to use to draw the background of the Widget, specified as an X colour name. - **foreground** The colour to use to draw the foreground of the Widget, specified as an X colour name. - **font** The font to use for the text, specified as an X font name. The extra resources available on each of the meters are the following: - **solid** A boolean value that specifies if the area under the graph is to be filled in or just the line drawn. - **grid** The colour that is to be used to draw the grid lines onto the meter. - **gridMin** The minimum number of grid lines that are drawn on the meter, if the data requires more grid lines, then they are added. If this number is negative then no grid lines are drawn, but the meter is scaled the same way as for a positive number. If the number of grid lines needed to cope with the data reaches 1/3 of the height of the window in pixles then only one grid line is drawn. This single grid line is drawn at the level that corresponds to the height of the window when the 1/3 limit is reached. - **labelPosition** The location of the label that is drawn with each of the meters. This can take the values -1 for a label at the bottom, 0 for no label and 1 for a label at the top. To set the foreground of all of the menu items to blue, use cpu menu item to red use '\*menushell.cpu.foreground: red'. (Note, that the background of the individual menu items cannot be set.) The resources can be set on the command line as well using the Xt standard method. The command line arguments -xrm '\*menushell.cpu.foreground: red' will do the same as putting the line in the application defaults file. (The single quatation marks are important here.) ## SEE ALSO xmeter(1) perfmeter(1) ## AUTHOR Andrew M. Bishop 1994,1995,1996,1997 (amb@gedanken.demon.co.uk)