NAME
    rrdpoller - Retrieve RRD file data and apply some threshold algorithm

SYNOPSIS
    rrdpoller [OPTIONS] list|ls *filename*

    rrdpoller [OPTIONS] get *filename* *datasource* [--offset *time*] [--cf
    *func-name*]

    rrdpoller [OPTIONS] exact *filename* *datasource* *value*

    rrdpoller [OPTIONS] boundaries *filename* *datasource* [--min *number*]
    [--max *number*]

    rrdpoller [OPTIONS] relation *filename* *datasource*
    [*<*|*>*]*threshold*[*%*] [--target *filename*] [--compare-ds
    *datasource*] [--offset *time*]

    rrdpoller [OPTIONS] quotient *filename* *datasource*
    [*<*|*>*]*threshold*[*%*] [--target *filename*] [--target-ds
    *datasource*] [--offset *time*]

    rrdpoller [OPTIONS] hunt *filename* *datasource* *roll* [--parent
    *filename*] [--parent-ds *datasource*]

DESCRIPTION
    rrdpoller allows you to poll RRD files and get current value of a given
    datasource. Additionally it implement some advanced checks able to use
    past values to decide if the current one is out of threshold or not.

    The main purpose of this tool is to do the interface between an existing
    monitoring system like BigBrother, Nagios or Mon and a set of RRD graphs
    generated by some specialized data collectors like rrdcollect, Cricket
    and such. Thus you can build a very modular monitoring/trend
    architecture without the need of double polling.

EXAMPLES
    Those examples are fictive, they doesn't represent real numbers. It's
    just to let you figure out what you can do with each algorithm.

    rrdpoller ls host1-net.rrd
        Returns the whole list of datasources contained in the file.

    rrdpoller get host1-load.rrd 1min_avg
        Get the current value of the *1min_avg* datasource of the
        *host1-load.rrd* file.

    rrdpoller --proxy http://xmlrpc.monserver.com:8080/ get host1-load.rrd
    1min_avg
        Same as precedent but querying a remove daemon (see rrdpollerd).

    rrdpoller exact host1-service.rrd http 1
        Get the current value of the *http* datasource of the
        *host1-load.rrd* file and return false if not equal to 1.

    rrdpoller boundaries host1-load.rrd 5min_avg --max 15
        Get the current value of the *5min_avg* datasource of the
        *host1-load.rrd* file and return false if greater than *15*.

    rrdpoller relation host1-disk.rrd usage 100000 --offset 1min
        Compare the current *usage* datasource value of the *host1-disk.rrd*
        file with the same datasource 1 minute ago and return false if the
        delta is greater than 100000.

    rrdpoller quotient host1-mem.rrd usage 90% --target-ds total
        Compare the current *usage* datasource of the *host1-mem.rrd* file
        and divide it by the *total* datasource of the same file. The
        command returns false if the obtained percentage is greater than
        *90%*.

    rrdpoller hunt pop2.rrd users 40 --target pop1.rrd
        If the *users* datasource of the *pop2.rrd* file is non-zero and the
        same datasource of the *pop1.rrd* hasn't yet reach *40*, the command
        will return false.

    rrdpoller quotient host1-mem.rrd high_mem,low_mem,+ 90% --target-ds
    low_total,high_total,+
        Same as former quotient example but using RPN datasource feature.
        Here the high_mem and the low_mem datasources will be summed then
        the low_total and the high_total too before to do the quotient
        computation.

COMMON OPTIONS
  --proxy *URL*
    If you use this option, the rrdpoller command will turn into an XML-RPC
    client and will ask send all RRD queries to a remote rrdpoller daemon
    (see rrdpollerd for more information).

  --max-last *seconds*
    If this option is specified with an integer, the test will return false
    if the last updated row is older than X *seconds*. The exit code will be
    2 if this test fails instead of 1 for normal tests.

SUB-COMMANDS
    All sub-commands needs at least the too first arguments *filename* and
    *datasource* (except for list command). Except for the get and list
    commands (which doesn't perform a test), all sub-command will change the
    rrdpoller return code to false (1) if the test fails. All sub-commands
    will print the given *datasource* value of the given *filename* on the
    standard output.

    The *datasource* can be an RPN expression (see Math::RPN) as well as all
    other datasource options. Note that the Math::RPN module have to be
    installed in order to use this feature.

  list|ls *filename*
    This command list all datasources of the given file.

  get *filename* *datasource* [OPTIONS]
    This sub-command the last inserted value in the given *datasource* of
    the given RRD *filename*. You can get an older value by giving an offset
    that will be expressed as: <last_inserted_time> - <offset>.

   options:
    --offset *time*
        Makes the get command to retrieve an earlier value. It substract the
        value of *time* to the timestamp of the last value inserted in the
        *datasource*. The *time* can be expressed in second (ie *60*) or in
        RRD time reference specification format (see "TIME REFERENCE
        SPECIFICATION" in rrdfetch).

    --cf *func-name*
        Default consolidation function is AVERAGE. This option allow you to
        change it to whichever you want.

  exact *filename* *datasource* *exact*
    This threshold allows you to monitor the *datasource* for an *exact*
    match. This is useful in cases where an enumerated (or boolean) SNMP
    object instruments a condition where a transition to a specific state
    requires attention. For example, a datasource might return either
    true(1) or false(2), depending on whether or not a power supply has
    failed.

  boundaries *filename* *datasource* [OPTIONS]
    This threshold takes too optional values, a *min*imum and a *max*imum
    value. If the *datasource* strays outside of this interval, the test
    fail.

   options:
    --min *number*
        If current value is lower than *number*, the test fail

    --max *number*
        If current value is greater than *number*, the test fail

  relation *filename* *datasource* *[<|>]threshold[%]* [OPTIONS]
    A relation threshold considers the difference between two *datasources*
    (possibly from different RRD *files*), or alternatively, the difference
    between two temporally distinct values for the same *datasource*. The
    difference can be expressed as absolute value, or as a percentage of the
    second *datasource* (comparison) value. This difference is compared to a
    *threshold* argument with either the greater than (>) or lesser than (<)
    operator. The criteria fails when the expression (<absolute or relative
    difference> <either greater-than or less-than> <threshold>) evaluates to
    false.

    The *threshold* number, optionally preceded by the greater than (>) or
    less than (<) symbol, and optionally followed by the symbol percent (%).
    If omitted, greater than is used by default and the expression:
    difference > threshold, is evaluated. "<10%", ">1000", "50%", and "500"
    are all examples of valid thresholds.

   options:
    --target *filename*
        The path of the comparison RRD *file*. This argument is optional and
        if omitted the first RRD *file* is also taken as the comparison
        target.

    --target-ds *datasource*
        The name of the comparison *datasource*. This *datasource* must
        belong to the comparison RRD file. This argument is optional and if
        omitted the first *datasource* name is also taken as the comparison
        datasource name. If the value is a number, the value is considered
        as a fix value and is taked for the comparison.

    --offset *time*
        The temporal *offset* to go back from the first value to fetch the
        comparison datasource value. Note that a data source value must
        exist in the RRD file for that exact offset. If This argument is
        optional and if omitted, it is set to 0. The *time* value can be
        expressed in second (ie *60*) or in RRD time reference specification
        format (see "TIME REFERENCE SPECIFICATION" in rrdfetch).

  quotient *filename* *datasource* *[<|>]threshold[%]* [OPTIONS]
    Quotient thresholds are similar to relation thresholds, except that they
    consider the quotient of two data sources, or alternatively, the same
    data source at two different time points. For a quotient monitor
    threshold, the value of the first data source is computed as a
    percentage of the second data source value (such as 10 (first
    datasource) is 50% of 20 (second datasource)). This percentage is then
    compared to a threshold argument with either the greater than (>) or
    less than (<) operator. The criteria fails when the expression
    (<percentage> <either greater-than or less-than> <threshold>) evaluates
    to false.

    The *threshold* number, optionally preceded by the greater than (>) or
    less than (<) symbol and followed by the symbol percent (%). If omitted,
    greater than is used by default and the expression: difference >
    threshold, is evaluated. "<10%" and "50%" are all examples of valid
    thresholds.

   options:
    --target *filename*
        The path of the comparison RRD *file*. This argument is optional and
        if omitted the first RRD *file* is also taken as the comparison
        target.

    --target-ds *datasource*
        The name of the comparison *datasource*. This *datasource* must
        belong to the comparison RRD file. This argument is optional and if
        omitted the first *datasource* name is also taken as the comparison
        datasource name. If the value is a number, the value is considered
        as a fix value and is taked for the comparison.

    --offset *time*
        The temporal *offset* to go back from the first value to fetch the
        comparison datasource value. Note that a data source value must
        exist in the RRD file for that exact offset. If This argument is
        optional and if omitted, it is set to 0. The *time* value can be
        expressed in second (ie *60*) or in RRD time reference specification
        format (see "TIME REFERENCE SPECIFICATION" in rrdfetch).

  hunt *roll* --parent <filename> --parent-ds <datasource> [OPTIONS]
    The hunt threshold is designed for the situation where the data source
    serves as an overflow for another data source; that is, if one data
    source (the parent) is at or near capacity, then traffic will begin to
    appear on this (the monitored) data source.�One application of hunt
    monitor thresholds is to identify premature rollover in a set of modem
    banks configured to hunt from one to the next. Specifically, the
    criteria of the hunt monitor threshold fails if the value of the
    monitored data source is non-zero and the current value of the parent
    data source falls below a specified capacity threshold.

    The *roll* argument is the threshold of the parent data source.
    Generally this should be slightly less than the maximum capacity of the
    target.

   options:
    --parent *filename*
        The path of the comparison RRD *file*. This argument is optional and
        if omitted the first RRD *file* is also taken as the comparison
        target.

    --parent-ds *datasource*
        The name of the comparison *datasource*. This *datasource* must
        belong to the comparison RRD file. This argument is optional and if
        omitted the first *datasource* name is also taken as the comparison
        datasource name. If the value is a number, the value is considered
        as a fix value and is taked for the comparison.

EXIT STATUS
    0   Successful test

    1   Performed test returned a false value

    2   The max-last was reached

    255 An exception was thrown

AUTHOR
    Olivier Poitrey <rs@rhapsodyk.net>

LICENCE
    rrdpoller retrieves RRD file data and apply some threshold algorithm.
    Copyright (C) 2004 Olivier Poitrey

    This program is free software; you can redistribute it and/or modify it
    under the terms of the GNU General Public License as published by the
    Free Software Foundation; either version 2 of the License, or (at your
    option) any later version.

    This program is distributed in the hope that it will be useful, but
    WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
    Public License for more details.

    You should have received a copy of the GNU General Public License along
    with this program; if not, write to the Free Software Foundation, Inc.,
    59 Temple Place, Suite 330, Boston, MA 02111-1307 USA

SEE ALSO
    RRD::Query, RRD::Threshold, rrdtool, RRDs, rrdpollerd