NAME
collectd-perl - Documentation of collectd's "perl plugin"
SYNOPSIS
LoadPlugin perl
# ...
<Plugin perl>
IncludeDir "/path/to/perl/plugins"
BaseName "Collectd::Plugins"
EnableDebugger ""
LoadPlugin "FooBar"
<Plugin FooBar>
Foo "Bar"
</Plugin>
</Plugin>
DESCRIPTION
The "perl plugin" embeds a Perl-interpreter into collectd and provides
an interface to collectd's plugin system. This makes it possible to
write plugins for collectd in Perl. This is a lot more efficient than
executing a Perl-script every time you want to read a value with the
"exec plugin" (see collectd-exec(5)) and provides a lot more
functionality, too.
CONFIGURATION
LoadPlugin Plugin
Loads the Perl plugin Plugin. This does basically the same as use
would do in a Perl program. As a side effect, the first occurrence
of this option causes the Perl-interpreter to be initialized.
BaseName Name
Prepends Name:: to all plugin names loaded after this option. This
is provided for convenience to keep plugin names short. All Perl-
based plugins provided with the collectd distributions reside in
the "Collectd::Plugins" namespace.
<Plugin Name> block
This block may be used to pass on configuration settings to a Perl
plugin. The configuration is converted into a config-item data type
which is passed to the registered configuration callback. See below
for details about the config-item data type and how to register
callbacks.
The name identifies the callback. It is used literally and
independent of the BaseName setting.
EnableDebugger Package[=option,...]
Run collectd under the control of the Perl source debugger. If
Package is not the empty string, control is passed to the
debugging, profiling, or tracing module installed as
Devel::Package. A comma-separated list of options may be specified
after the "=" character. Please note that you may not leave out the
Package option even if you specify "". This is the same as using
the -d:Package command line option.
See perldebug for detailed documentation about debugging Perl.
This option does not prevent collectd from daemonizing, so you
should start collectd with the -f command line option. Else you
will not be able to use the command line driven interface of the
debugger.
IncludeDir Dir
Adds Dir to the @INC array. This is the same as using the -IDir
command line option or use lib Dir in the source code. Please note
that it only has effect on plugins loaded after this option.
WRITING YOUR OWN PLUGINS
Writing your own plugins is quite simple. collectd manages plugins by
means of dispatch functions which call the appropriate callback
functions registered by the plugins. Any plugin basically consists of
the implementation of these callback functions and initializing code
which registers the functions with collectd. See the section "EXAMPLES"
below for a really basic example. The following types of callback
functions are known to collectd (all of them are optional):
configuration functions
This type of functions is called during configuration if an
appropriate Plugin block has been encountered. It is called once
for each Plugin block which matches the name of the callback as
provided with the plugin_register method - see below.
init functions
This type of functions is called once after loading the module and
before any calls to the read and write functions. It should be used
to initialize the internal state of the plugin (e. g. open sockets,
...). If the return value evaluates to false, the plugin will be
disabled.
read functions
This type of function is used to collect the actual data. It is
called once per interval (see the Interval configuration option of
collectd). Usually it will call plugin_dispatch_values to dispatch
the values to collectd which will pass them on to all registered
write functions. If the return value evaluates to false the plugin
will be skipped for an increasing amount of time until it returns
true again.
write functions
This type of function is used to write the dispatched values. It is
called once for each call to plugin_dispatch_values.
flush functions
This type of function is used to flush internal caches of plugins.
It is usually triggered by the user only. Any plugin which caches
data before writing it to disk should provide this kind of callback
function.
log functions
This type of function is used to pass messages of plugins or the
daemon itself to the user.
notification function
This type of function is used to act upon notifications. In
general, a notification is a status message that may be associated
with a data instance. Usually, a notification is generated by the
daemon if a configured threshold has been exceeded (see the section
"THRESHOLD CONFIGURATION" in collectd.conf(5) for more details),
but any plugin may dispatch notifications as well.
shutdown functions
This type of function is called once before the daemon shuts down.
It should be used to clean up the plugin (e.g. close sockets, ...).
Any function (except log functions) may set the $@ variable to describe
errors in more detail. The message will be passed on to the user using
collectd's logging mechanism.
See the documentation of the plugin_register method in the section
"METHODS" below for the number and types of arguments passed to each
callback function. This section also explains how to register callback
functions with collectd.
To enable a plugin, copy it to a place where Perl can find it (i. e. a
directory listed in the @INC array) just as any other Perl plugin and
add an appropriate LoadPlugin option to the configuration file. After
restarting collectd you're done.
DATA TYPES
The following complex types are used to pass values between the Perl
plugin and collectd:
Config-Item
A config-item is one structure which keeps the information provided
in the configuration file. The array of children keeps one entry
for each configuration option. Each such entry is another config-
item structure, which may nest further if nested blocks are used.
{
key => key,
values => [ val1, val2, ... ],
children => [ { ... }, { ... }, ... ]
}
Data-Set
A data-set is a list of one or more data-sources. Each data-source
defines a name, type, min- and max-value and the data-set wraps
them up into one structure. The general layout looks like this:
[{
name => 'data_source_name',
type => DS_TYPE_COUNTER || DS_TYPE_GAUGE || DS_TYPE_DERIVE || DS_TYPE_ABSOLUTE,
min => value || undef,
max => value || undef
}, ...]
Value-List
A value-list is one structure which features an array of values and
fields to identify the values, i. e. time and host, plugin name and
plugin-instance as well as a type and type-instance. Since the
"type" is not included in the value-list but is passed as an extra
argument, the general layout looks like this:
{
values => [123, 0.5],
time => time (),
interval => $interval_g,
host => $hostname_g,
plugin => 'myplugin',
type => 'myplugin',
plugin_instance => '',
type_instance => ''
}
Notification
A notification is one structure defining the severity, time and
message of the status message as well as an identification of a
data instance. Also, it includes an optional list of user-defined
meta information represented as (name, value) pairs:
{
severity => NOTIF_FAILURE || NOTIF_WARNING || NOTIF_OKAY,
time => time (),
message => 'status message',
host => $hostname_g,
plugin => 'myplugin',
type => 'mytype',
plugin_instance => '',
type_instance => '',
meta => [ { name => <name>, value => <value> }, ... ]
}
Match-Proc
A match-proc is one structure storing the callbacks of a "match" of
the filter chain infrastructure. The general layout looks like
this:
{
create => 'my_create',
destroy => 'my_destroy',
match => 'my_match'
}
Target-Proc
A target-proc is one structure storing the callbacks of a "target"
of the filter chain infrastructure. The general layout looks like
this:
{
create => 'my_create',
destroy => 'my_destroy',
invoke => 'my_invoke'
}
METHODS
The following functions provide the C-interface to Perl-modules. They
are exported by the ":plugin" export tag (see the section "EXPORTS"
below).
plugin_register (type, name, data)
Registers a callback-function or data-set.
type can be one of:
TYPE_CONFIG
TYPE_INIT
TYPE_READ
TYPE_WRITE
TYPE_FLUSH
TYPE_LOG
TYPE_NOTIF
TYPE_SHUTDOWN
TYPE_DATASET
name is the name of the callback-function or the type of the data-
set, depending on the value of type. (Please note that the type of
the data-set is the value passed as name here and has nothing to do
with the type argument which simply tells plugin_register what is
being registered.)
The last argument, data, is either a function name or an array-
reference. If type is TYPE_DATASET, then the data argument must be
an array-reference which points to an array of hashes. Each hash
describes one data-set. For the exact layout see Data-Set above.
Please note that there is a large number of predefined data-sets
available in the types.db file which are automatically registered
with collectd - see types.db(5) for a description of the format of
this file.
Note: Using plugin_register to register a data-set is deprecated.
Add the new type to a custom types.db(5) file instead. This
functionality might be removed in a future version of collectd.
If the type argument is any of the other types (TYPE_INIT,
TYPE_READ, ...) then data is expected to be a function name. If the
name is not prefixed with the plugin's package name collectd will
add it automatically. The interface slightly differs from the C
interface (which expects a function pointer instead) because Perl
does not support to share references to subroutines between
threads.
These functions are called in the various stages of the daemon (see
the section "WRITING YOUR OWN PLUGINS" above) and are passed the
following arguments:
TYPE_CONFIG
The only argument passed is config-item. See above for the
layout of this data type.
TYPE_INIT
TYPE_READ
TYPE_SHUTDOWN
No arguments are passed.
TYPE_WRITE
The arguments passed are type, data-set, and value-list. type
is a string. For the layout of data-set and value-list see
above.
TYPE_FLUSH
The arguments passed are timeout and identifier. timeout
indicates that only data older than timeout seconds is to be
flushed. identifier specifies which values are to be flushed.
TYPE_LOG
The arguments are log-level and message. The log level is small
for important messages and high for less important messages.
The least important level is LOG_DEBUG, the most important
level is LOG_ERR. In between there are (from least to most
important): LOG_INFO, LOG_NOTICE, and LOG_WARNING. message is
simply a string without a newline at the end.
TYPE_NOTIF
The only argument passed is notification. See above for the
layout of this data type.
plugin_unregister (type, plugin)
Removes a callback or data-set from collectd's internal list of
functions / datasets.
plugin_dispatch_values (value-list)
Submits a value-list to the daemon. If the data-set identified by
value-list->{type} is found (and the number of values matches the
number of data-sources) then the type, data-set and value-list is
passed to all write-callbacks that are registered with the daemon.
Note: Prior to version 4.4 of collectd, the data-set type used to
be passed as the first argument to plugin_register. This syntax is
still supported for backwards compatibility but has been deprecated
and will be removed in some future version of collectd.
plugin_write ([plugins => ...][, datasets => ...], valuelists => ...)
Calls the write function of the given plugins with the provided
data sets and value lists. In contrast to plugin_dispatch_values,
it does not update collectd's internal cache and bypasses the
filter mechanism (see collectd.conf(5) for details). If the plugins
argument has been omitted, the values will be dispatched to all
registered write plugins. If the datasets argument has been
omitted, the required data sets are looked up according to the
"type" member in the appropriate value list. The value of all three
arguments may either be a single scalar or a reference to an array.
If the datasets argument has been specified, the number of data
sets has to equal the number of specified value lists.
plugin_flush ([timeout => timeout][, plugins => ...][, identifiers =>
...])
Flush one or more plugins. timeout and the specified identifiers
are passed on to the registered flush-callbacks. If omitted, the
timeout defaults to "-1". The identifier defaults to the undefined
value. If the plugins argument has been specified, only named
plugins will be flushed. The value of the plugins and identifiers
arguments may either be a string or a reference to an array of
strings.
plugin_flush_one (timeout, plugin)
This is identical to using "plugin_flush (timeout => timeout,
plugins => plugin".
Note: Starting with version 4.5 of collectd, plugin_flush_one has
been deprecated and will be removed in some future version of
collectd. Use plugin_flush instead.
plugin_flush_all (timeout)
This is identical to using "plugin_flush (timeout => timeout)".
Note: Starting with version 4.5 of collectd, plugin_flush_all has
been deprecated and will be removed in some future version of
collectd. Use plugin_flush instead.
plugin_dispatch_notification (notification)
Submits a notification to the daemon which will then pass it to all
notification-callbacks that are registered.
plugin_log (log-level, message)
Submits a message of level log-level to collectd's logging
mechanism. The message is passed to all log-callbacks that are
registered with collectd.
ERROR, WARNING, NOTICE, INFO, DEBUG (message)
Wrappers around plugin_log, using LOG_ERR, LOG_WARNING, LOG_NOTICE,
LOG_INFO and LOG_DEBUG respectively as log-level.
The following function provides the filter chain C-interface to Perl-
modules. It is exported by the ":filter_chain" export tag (see the
section "EXPORTS" below).
fc_register (type, name, proc)
Registers filter chain callbacks with collectd.
type may be any of:
FC_MATCH
FC_TARGET
name is the name of the match or target. By this name, the
callbacks are identified in the configuration file when specifying
a Match or Target block (see collectd.conf(5) for details).
proc is a hash reference. The hash includes up to three callbacks:
an optional constructor (create) and destructor (destroy) and a
mandatory match or invoke callback. match is called whenever
processing an appropriate match, while invoke is called whenever
processing an appropriate target (see the section "FILTER
CONFIGURATION" in collectd.conf(5) for details). Just like any
other callbacks, filter chain callbacks are identified by the
function name rather than a function pointer because Perl does not
support to share references to subroutines between threads. The
following arguments are passed to the callbacks:
create
The arguments passed are config-item and user-data. See above
for the layout of the config-item data-type. user-data is a
reference to a scalar value that may be used to store any
information specific to this particular instance. The daemon
does not care about this information at all. It's for the
plugin's use only.
destroy
The only argument passed is user-data which is a reference to
the user data initialized in the create callback. This callback
may be used to cleanup instance-specific information and
settings.
match, invoke
The arguments passed are data-set, value-list, meta and user-
data. See above for the layout of the data-set and value-list
data-types. meta is a pointer to an array of meta information,
just like the meta member of the notification data-type (see
above). user-data is a reference to the user data initialized
in the create callback.
GLOBAL VARIABLES
$hostname_g
As the name suggests this variable keeps the hostname of the system
collectd is running on. The value might be influenced by the
Hostname or FQDNLookup configuration options (see collectd.conf(5)
for details).
$interval_g
This variable keeps the interval in seconds in which the read
functions are queried (see the Interval configuration option).
Any changes to these variables will be globally visible in collectd.
EXPORTS
By default no symbols are exported. However, the following export tags
are available (:all will export all of them):
:plugin
plugin_register ()
plugin_unregister ()
plugin_dispatch_values ()
plugin_flush ()
plugin_flush_one ()
plugin_flush_all ()
plugin_dispatch_notification ()
plugin_log ()
:types
TYPE_CONFIG
TYPE_INIT
TYPE_READ
TYPE_WRITE
TYPE_FLUSH
TYPE_SHUTDOWN
TYPE_LOG
TYPE_DATASET
:ds_types
DS_TYPE_COUNTER
DS_TYPE_GAUGE
DS_TYPE_DERIVE
DS_TYPE_ABSOLUTE
:log
ERROR ()
WARNING ()
NOTICE ()
INFO ()
DEBUG ()
LOG_ERR
LOG_WARNING
LOG_NOTICE
LOG_INFO
LOG_DEBUG
:filter_chain
fc_register
FC_MATCH_NO_MATCH
FC_MATCH_MATCHES
FC_TARGET_CONTINUE
FC_TARGET_STOP
FC_TARGET_RETURN
:fc_types
FC_MATCH
FC_TARGET
:notif
NOTIF_FAILURE
NOTIF_WARNING
NOTIF_OKAY
:globals
$hostname_g
$interval_g
EXAMPLES
Any Perl plugin will start similar to:
package Collectd::Plugins::FooBar;
use strict;
use warnings;
use Collectd qw( :all );
A very simple read function might look like:
sub foobar_read
{
my $vl = { plugin => 'foobar' };
$vl->{'values'} = [ rand(42) ];
plugin_dispatch_values ('gauge', $vl);
return 1;
}
A very simple write function might look like:
sub foobar_write
{
my ($type, $ds, $vl) = @_;
for (my $i = 0; $i < scalar (@$ds); ++$i) {
print "$vl->{'plugin'} ($vl->{'type'}): $vl->{'values'}->[$i]\n";
}
return 1;
}
A very simple match callback might look like:
sub foobar_match
{
my ($ds, $vl, $meta, $user_data) = @_;
if (matches($ds, $vl)) {
return FC_MATCH_MATCHES;
} else {
return FC_MATCH_NO_MATCH;
}
}
To register those functions with collectd:
plugin_register (TYPE_READ, "foobar", "foobar_read");
plugin_register (TYPE_WRITE, "foobar", "foobar_write");
fc_register (FC_MATCH, "foobar", "foobar_match");
See the section "DATA TYPES" above for a complete documentation of the
data types used by the read, write and match functions.
NOTES
o Please feel free to send in new plugins to collectd's mailing list
at <collectd at verplant.org> for review and, possibly, inclusion
in the main distribution. In the latter case, we will take care of
keeping the plugin up to date and adapting it to new versions of
collectd.
Before submitting your plugin, please take a look at
<http://collectd.org/dev-info.shtml>.
CAVEATS
o collectd is heavily multi-threaded. Each collectd thread accessing
the perl plugin will be mapped to a Perl interpreter thread (see
threads(3perl)). Any such thread will be created and destroyed
transparently and on-the-fly.
Hence, any plugin has to be thread-safe if it provides several
entry points from collectd (i. e. if it registers more than one
callback or if a registered callback may be called more than once
in parallel). Please note that no data is shared between threads by
default. You have to use the threads::shared module to do so.
o Each function name registered with collectd has to be available
before the first thread has been created (i. e. basically at
compile time). This basically means that hacks (yes, I really
consider this to be a hack) like "*foo = \&bar; plugin_register
(TYPE_READ, "plugin", "foo");" most likely will not work. This is
due to the fact that the symbol table is not shared across
different threads.
o Each plugin is usually only loaded once and kept in memory for
performance reasons. Therefore, END blocks are only executed once
when collectd shuts down. You should not rely on END blocks anyway
- use shutdown functions instead.
o The perl plugin exports the internal API of collectd which is
considered unstable and subject to change at any time. We try hard
to not break backwards compatibility in the Perl API during the
life cycle of one major release. However, this cannot be
guaranteed at all times. Watch out for warnings dispatched by the
perl plugin after upgrades.
KNOWN BUGS
o Currently, it is not possible to flush a single Perl plugin only.
You can either flush all Perl plugins or none at all and you have
to use "perl" as plugin name when doing so.
SEE ALSO
collectd(1), collectd.conf(5), collectd-exec(5), types.db(5), perl(1),
threads(3perl), threads::shared(3perl), perldebug(1)
AUTHOR
The "perl plugin" has been written by Sebastian Harl
<sh at tokkee.org>.
This manpage has been written by Florian Forster <octo at verplant.org>
and Sebastian Harl <sh at tokkee.org>.