dockapps/wmsupermon/README
2012-02-12 22:50:31 +00:00

163 lines
5.9 KiB
Plaintext

wmsupermon
==========
Author: Sergei Golubchik <sergii@pisem.net>
Download: http://www.dockapps.org/file.php/id/320
Universal Monitoring Dockapp.
Can be used to monitor
1. CPU usage
2. disk i/o
3. memory
4. swap
5. filesystems - space utilization
6. network traffic
7. wireless link quality
8. cpu frequency
9. CPU temperature, fan speed, voltages (no lm_sensors required!)
10. battery status (with actual - not guessed - discharge rate!)
11. traffic from your router
12. anything else :)
The dockapp is configured via ~/.wmsupermon file,
which specifies what to monitor and how to present the results.
The syntax of a config file is:
=========================
[mem]
Source = /proc/meminfo
Regex = {MemTotal: *([0-9]+).*MemFree: *([0-9]+).*Buffers: *([0-9]+) kB.Cached: *([0-9]+)}[(\1 - \2 - \3 - \4)/\1]s
[i/o]
Source = /proc/diskstats
#major minor name rio rmerge rsect ruse wio wmerge wsect wuse running use aveq
# note that "name" part is limited to letters!
# thus we count, e.g. "sda" line, but not "sda1", "sda2", etc.
Regex = /^ +[0-9]+ +[0-9]+ [a-z]+ ([0-9]+) [0-9]+ [0-9]+ [0-9]+ ([0-9]+) /DIFF(SUM(\1+\2))/
[eml]
Source = !/usr/local/bin/check_email
Interval = 60
[cpu]
Source = /proc/stat
Regex = {cpu ([0-9]+) ([0-9]+) ([0-9]+) ([0-9]+)}{DIFF(\1+\2+\3+0.01)/DIFF(\1+\2+\3+\4+0.01)}
[Tem]
Source = /sys/bus/i2c/drivers/w83627hf/9191-0290/temp2_input
Scale = 1000
Range = 37 .. 50
[[wmsupermon]]
eml=number -label
i/o=bar -label
mem=graph -small
cpu=percent -label
tem=bar
cpu=graph -medium
=========================
Empty lines are allowed, if the line starts with #, it is ignored.
First there are sources of readings. Every source starts with a label
(e.g. [mem]) which can be one to three characters long.
After the label, the following keywords are allowed:
Source = specifies where to take the data from. It's either a
filename or an executable/script (if the value starts from the
exclamation sign).
Regex = defines how to parse the source (the file or the output of
an executable). It consists of two parts delimited by any non-blank
character (e.g. regex= !...!...! or /..../..../) or two pairs of
matching brackets (e.g. [ ...]{...} or (...)<....>).
The first part is an extended regex, the second - an expression to
evaluate. This expression can use numeric constants (floating
point), \1 .. \9 to refer to parts of regex, operators + - * / > < ?:,
parentheses, and two functions SUM() and DIFF().
The DIFF function (which means differentiate) returns the difference
between the value of its argument and the value of its argument from
the previous evaluation. See [cpu] source above to for an example -
DIFF is used to convert the constantly growing number of raw "cpu
ticks" into "cpu usage".
The SUM() function is explained below.
After the expression one can put flags. Supported are
i - case insensitive match
s - single line (see below)
d - debug (the result of the expression is printed to stdout).
Normally the source is read line by line, until the regex matches
for the first time. The result of the expression is the reading of
a source.
If SUM() function is present anywhere in the expression, all lines
of the source are read, and all matches are accumulated. The result
of the SUM() is the sum of values of its argument for all matches.
See [i/o] source above for an example - SUM is used to calculate i/o
for all block devices.
If 's' (single line) flag is specified, regex is used to match
file's content as a whole, not linewise. See [mem] source above for
an example.
If no regex is specified, the source is expected to contain just one
number (many files on sysfs do).
Other recognized keywords are:
Interval = N, the reading of a source is taken every N seconds. 0
means "as fast as possible" (currently - four times a second).
Scale = N, the reading calculated from Source+Regex is divided by N.
That's most useful when there's no regex, see [Tem] example above.
Range = N1 .. N2, used by bar and graph widgets. Without explicit
range, bar expect readings to be in 0..1 range, and graph
autoscales. As with scale, this is most useful when no regex is
specified, as in [Tem] example above.
Action = shell command to be executed on left mouse click on the
widget.
Besides sources, config file defines windows and panes.
Every window definiftion starts from [[name]], where "name" is the
window title. Wmsupermon may open many windows, with different
layouts and values to monitor.
Window definition consists of widgets, which are placed on panes. A
pane has four rows, and a widget can take up to all the four,
depending on the type and options, see below. A pane is thus defined
by one to four lines, specifying widgets for this pane. Empty
lines are used to separate panes. One widget line looks like
name=widget [options]
where <name> is the name of a source, defined above, and <widget> is
the widget type. Supported widget types are "number", "percent",
"bar", and "graph". After the type, one can optionally put
<options>, from the following list:
-label - print a source name on the widget.
can be used with any widget
-smooth - smooth the output by averaging last 8 readings.
can be used with any widget
-float - print a number with the decimal dot, not an integer.
can be used only with "number" widget
-log - use logarithmic scale. only valid for "graph" widgets
-big - 4-rows graph, only valid for "graph" widgets
-medium - 2-rows graph, only valid for "graph" widgets
-small - 1-row graph, only valid for "graph" widgets.
by default a graph takes 3 rows.
-scaledown - if the graph autoscales, allow the scale to decrease,
by default it can only increase.
See example-wmsupermonrc