DragonFly On-Line Manual Pages

Search: Section:  

SYMON(8)               DragonFly System Manager's Manual              SYMON(8)


NAME

     symon - system monitor


SYNOPSIS

     symon [-dtuv] [-f filename]


DESCRIPTION

     symon is a lightweight system monitor that measures cpu, load,
     filesystem, interface, disk, memory, pf, pf queues, mbuf, proc and sensor
     statistics every 5 seconds. This information is then spooled to symux(8)
     for further processing.

     symon has been designed to inflict minimal performance and security
     impact on the system it monitors.  symux(8) has performance impact
     proportional to the amount of streams it needs to manage. Ideally symux
     should live on a different system and collect data from several symon
     instances in a LAN.

     By default, symon will drop privileges and chroot(2) to home of the symon
     user.  This behaviour is not strictly needed for the cpu, mem, mbuf, disk
     debug and interface probes as these will work even when symon is started
     as nobody.

     The options:

     -d      Stop symon from becoming a daemon and show debug information on
             stdout.

     -f filename
             Read configuration from filename instead of /etc/symon.conf.

     -t      Test configuration file and exit.

     -u      By default symon will chroot(2) into _symon user home directory.
             The -u disables this behaviour.

     -v      Show version information.


CONFIGURATION

     symon obtains configuration data from /etc/symon.conf.  The configuration
     file contains monitor stanzas that define what resources should be
     monitored and to which symux(8) the information should be streamed to.

     Multiple monitor statements are allowed. Whitespace, newlines and text
     behind '#' are ignored. The format in BNF:

           monitor-rule = "monitor" "{" resources "}" [every]
                          "stream" ["from" host] ["to"] host [ port ]
           resources    = resource [ version ] ["(" argument ")"]
                          [ ","|" " resources ]
           resource     = "cpu" | "cpuiow" | "debug" | "df" | "flukso" |
                          "if" | "io" | "load" | "mbuf" | "mem" | "pf" |
                          "pfq" | "proc" | "sensor" | "smart"
           version      = number
           argument     = number | name
           every        = "every" time
           time         = "second" | number "seconds"
           host         = ip4addr | ip6addr | hostname
           port         = [ "port" | "," ] portnumber

     Note that symux(8) data files default to receiving data every 5 seconds.
     Adjusting the monitoring interval will also require adjusting the
     associated symux(8) datafile(s).

     The pf probe will return data that is collected for the loginterface set
     in /etc/pf.conf(5).

     The Linux io, df, and smart probes support device names via id, label,
     path and uuid.

     The FreeBSD io, df, and smart probes support gpt names, ufs names, ufs
     ids and paths.

     The OpenBSD io probe supports device uuids.


EXAMPLE

     Here is an example OpenBSD symon.conf that monitors cpu, memory, pf,
     interfaces xl0/de0/lo0/wi0, disks wd[0-3]/cd[0-1], debug variables debug0
     to debug19 and streams that information to localhost on port 2100.

           monitor { cpu(0),  mem, pf, if(xl0), if(de0),
                     if(lo0), if(wi0), io(wd0), io(wd1),
                     io(wd2), io(wd3), io(cd0), io(cd1),
                     io(ccd0), df(sd0a), df(sd0d), df(sd0e),
                     debug, proc(httpd) } stream to 127.0.0.1 2100


EXAMPLE

     Here is an example Linux symon.conf that monitors cpu including iowait,
     memory, load, interface eth0, io and df for a set of disks every 5
     seconds. Smart data is to be collected every 60 seconds.  Disks in the
     smart and io statements are identified using ids, filesystem volumes in
     df using labels.
           monitor { smart(ata-Hitachi_HDS722020ALA330_JK1130ABABABAB),
                     smart(ata-Hitachi_HDS722020ALA330_JK1130ACACACAC),
                   } every 60 seconds stream to 192.168.0.2 port 2100

           monitor { cpuiow(0), cpuiow(1), mem, if(eth0),
                     io(ata-Hitachi_HDS722020ALA330_JK1130ABABABAB),
                     io(ata-Hitachi_HDS722020ALA330_JK1130ACACACAC),
                     df(data_1),
                     df(data_2),
                     df(data_3),
                     df(home),
                     df(var),
                     load
                   } stream to 192.168.0.2 port 2100


SIGNALS

     SIGHUP  Causes symon to read /etc/symon.conf.  symon will keep the old
             configuration if errors occured during parsing of the
             configuration file. Note that the chroot(2) may cause resources
             to become unattainable, most notably the configuration file
             itself.


FILES

     /var/run/symon.pid
             Contains the program id of the symon daemon.

     /etc/symon.conf
             symon system wide configuration file.


BUGS

     Every monitored resource mentioned /etc/symon.conf gets queried.
     Mentioning, for example, cpu(0) twice for different muxes will result in
     two distinct cpu(0) measurement actions.

     The proc module is too simple: memory shared between two instances of the
     same process is simply counted twice.

     symon does not check whether all resources mentioned in /etc/symon.conf
     exist.


AUTHOR

     Willem Dijkstra <wpd@xs4all.nl>. Daniel Hartmeier helped to port to big-
     endian architectures. Matthew Gream helped to port symon to other BSD
     platforms.

     Port contributors: Marc Balmer, Tito Dal Canton, Matthew Gream, Daniel
     Hartmeier, Lars Kotthoff, Constantine A. Murenin, J. Martin Petersen,
     Fredrik Soderblom, Harm Schotanus and Martin van der Werff.


SEE ALSO

     symux(8)

DragonFly 6.5-DEVELOPMENT        April 4, 2012       DragonFly 6.5-DEVELOPMENT

Search: Section: