1 collectd - System information collection daemon
2 =================================================
8 collectd is a small daemon which collects system information periodically
9 and provides mechanisms to store and monitor the values in a variety of
16 * collectd is able to collect the following data:
19 Apache server utilization: Number of bytes transfered, number of
20 requests handled and detailed scoreboard statistics
23 APC UPS Daemon: UPS charge, load, input/output/battery voltage, etc.
26 Sensors in Macs running Mac OS X / Darwin: Temperature, fanspeed and
30 Statistics about Ascent, a free server for the game `World of Warcraft'.
33 Batterycharge, -current and voltage of ACPI and PMU based laptop
37 Parse statistics from websites using regular expressions.
40 Name server and resolver statistics from the `statistics-channel'
41 interface of BIND 9.5, 9,6 and later.
44 Number of nf_conntrack entries.
47 Retrieves JSON data via cURL and parses it according to user
51 CPU utilization: Time spent in the system, user, nice, idle, and related
55 CPU frequency (For laptops with speed step or a similar technology)
58 Executes SQL statements on various databases and interprets the returned
62 Mountpoint usage (Basically the values `df(1)' delivers)
65 Disk utilization: Sectors read/written, number of read/write actions,
66 average time an IO-operation took to complete.
69 DNS traffic: Query types, response codes, opcodes and traffic/octets
73 Email statistics: Count, traffic, spam scores and checks.
74 See collectd-email(5).
77 Amount of entropy available to the system.
80 Values gathered by a custom program or script.
84 Count the number of files in directories.
87 Linux file-system based caching framework statistics.
90 Receive multicast traffic from Ganglia instances.
93 Harddisk temperatures using hddtempd.
96 Interface traffic: Number of octets, packets and errors for each
100 Iptables' counters: Number of bytes that were matched by a certain
104 IPMI (Intelligent Platform Management Interface) sensors information.
107 IPVS connection statistics (number of connections, octets and packets
108 for each service and destination).
109 See http://www.linuxvirtualserver.org/software/index.html.
112 IRQ counters: Frequency in which certain interrupts occur.
115 Integrates a `Java Virtual Machine' (JVM) to execute plugins in Java
116 bytecode. See “Configuring with libjvm” below.
119 System load average over the last 1, 5 and 15 minutes.
122 CPU, disk and network I/O statistics from virtual machines.
125 Queries very detailed usage statistics from wireless LAN adapters and
126 interfaces that use the Atheros chipset and the MadWifi driver.
129 Motherboard sensors: temperature, fanspeed and voltage information,
133 Query and parse data from a memcache daemon (memcached).
136 Statistics of the memcached distributed caching system.
137 <http://www.danga.com/memcached/>
140 Memory utilization: Memory occupied by running processes, page cache,
141 buffer cache and free.
144 Information provided by serial multimeters, such as the `Metex
148 MySQL server statistics: Commands issued, handlers triggered, thread
149 usage, query cache utilization and traffic/octets sent and received.
152 Plugin to query performance values from a NetApp storage system using the
153 “Manage ONTAP” SDK provided by NetApp.
156 Very detailed Linux network interface and routing statistics. You can get
157 (detailed) information on interfaces, qdiscs, classes, and, if you can
158 make use of it, filters.
161 Receive values that were collected by other hosts. Large setups will
162 want to collect the data on one dedicated machine, and this is the
163 plugin of choice for that.
166 NFS Procedures: Which NFS command were called how often. Only NFSv2 and
170 Collects statistics from `nginx' (speak: engine X), a HTTP and mail
174 NTP daemon statistics: Local clock drift, offset to peers, etc.
177 Network UPS tools: UPS current, voltage, power, charge, utilisation,
178 temperature, etc. See upsd(8).
181 Queries routing information from the “Optimized Link State Routing”
184 - onewire (EXPERIMENTAL!)
185 Read onewire sensors using the owcapu library of the owfs project.
186 Please read in collectd.conf(5) why this plugin is experimental.
189 RX and TX of each client in openvpn-status.log (status-version 2).
190 <http://openvpn.net/index.php/documentation/howto.html>
193 Query data from an Oracle database.
196 The perl plugin implements a Perl-interpreter into collectd. You can
197 write your own plugins in Perl and return arbitrary values using this
198 API. See collectd-perl(5).
201 Network latency: Time to reach the default gateway or another given
205 PostgreSQL database statistics: active server connections, transaction
206 numbers, block IO, table row manipulations.
209 PowerDNS name server statistics.
212 Process counts: Number of running, sleeping, zombie, ... processes.
215 Counts various aspects of network protocols such as IP, TCP, UDP, etc.
218 The python plugin implements a Python interpreter into collectd. This
219 makes it possible to write plugins in Python which are executed by
220 collectd without the need to start a heavy interpreter every interval.
221 See collectd-python(5) for details.
224 RRDtool caching daemon (RRDcacheD) statistics.
227 System sensors, accessed using lm_sensors: Voltages, temperatures and
231 RX and TX of serial interfaces. Linux only; needs root privileges.
234 Read values from SNMP (Simple Network Management Protocol) enabled
235 network devices such as switches, routers, thermometers, rack monitoring
236 servers, etc. See collectd-snmp(5).
239 Pages swapped out onto harddisk or whatever is called `swap' by the OS..
242 Parse table-like structured files.
245 Follows (tails) logfiles, parses them by lines and submits matched
249 Bytes and operations read and written on tape devices. Solaris only.
252 Number of TCP connections to specific local and remote ports.
255 TeamSpeak2 server statistics.
258 Plugin to read values from `The Energy Detective' (TED).
261 Linux ACPI thermal zone information.
264 Reads the number of records and file size from a running Tokyo Tyrant
268 System uptime statistics.
271 Users currently logged in.
274 Virtual memory statistics, e. g. the number of page-ins/-outs or the
275 number of pagefaults.
278 System resources used by Linux VServers.
279 See <http://linux-vserver.org/>.
282 Link quality of wireless cards. Linux only.
285 Bitrate and frequency of music played with XMMS.
288 Statistics for ZFS' “Adaptive Replacement Cache” (ARC).
290 * Output can be written or send to various destinations by the following
294 Write to comma separated values (CSV) files. This needs lots of
295 diskspace but is extremely portable and can be analysed with almost
296 every program that can analyse anything. Even Microsoft's Excel..
299 Send the data to a remote host to save the data somehow. This is useful
300 for large setups where the data should be saved by a dedicated machine.
303 Of course the values are propagated to plugins written in Perl, too, so
304 you can easily do weird stuff with the plugins we didn't dare think of
305 ;) See collectd-perl(5).
308 It's possible to implement write plugins in Python using the python
309 plugin. See collectd-python(5) for details.
312 Output to round-robin-database (RRD) files using the RRDtool caching
313 daemon (RRDcacheD) - see rrdcached(1). That daemon provides a general
314 implementation of the caching done by the `rrdtool' plugin.
317 Output to round-robin-database (RRD) files using librrd. See rrdtool(1).
318 This is likely the most popular destination for such values. Since
319 updates to RRD-files are somewhat expensive this plugin can cache
320 updates to the files and write a bunch of updates at once, which lessens
324 One can query the values from the unixsock plugin whenever they're
325 needed. Please read collectd-unixsock(5) for a description on how that's
329 Sends the values collected by collectd to a web-server using HTTP POST
330 requests. The transmitted data is either in a form understood by the
331 Exec plugin or formatted in JSON.
333 * Logging is, as everything in collectd, provided by plugins. The following
334 plugins keep up informed about what's going on:
337 Writes logmessages to a file or STDOUT/STDERR.
340 Log messages are propagated to plugins written in Perl as well.
341 See collectd-perl(5).
344 It's possible to implement log plugins in Python using the python plugin.
345 See collectd-python(5) for details.
348 Logs to the standard UNIX logging mechanism, syslog.
350 * Notifications can be handled by the following plugins:
353 Send a desktop notification to a notification daemon, as defined in
354 the Desktop Notification Specification. To actually display the
355 notifications, notification-daemon is required.
356 See http://www.galago-project.org/specs/notification/.
359 Send an E-mail with the notification message to the configured
363 Execute a program or script to handle the notification.
364 See collectd-exec(5).
367 Writes the notification message to a file or STDOUT/STDERR.
370 Send the notification to a remote host to handle it somehow.
373 Notifications are propagated to plugins written in Perl as well.
374 See collectd-perl(5).
377 It's possible to implement notification plugins in Python using the
378 python plugin. See collectd-python(5) for details.
380 * Value processing can be controlled using the "filter chain" infrastructure
381 and "matches" and "targets". The following plugins are available:
383 - match_empty_counter
384 Match counter values which are currently zero.
387 Match values by their identifier based on regular expressions.
390 Match values with an invalid timestamp.
393 Select values by their data sources' values.
395 - target_notification
396 Create and dispatch a notification.
399 Replace parts of an identifier using regular expressions.
402 Set (overwrite) entire parts of an identifier.
404 * Miscellaneous plugins:
407 Sets the hostname to an unique identifier. This is meant for setups
408 where each client may migrate to another physical host, possibly going
409 through one or more name changes in the process.
411 * Performance: Since collectd is running as a daemon it doesn't spend much
412 time starting up again and again. With the exception of the exec plugin no
413 processes are forked. Caching in output plugins, such as the rrdtool and
414 network plugins, makes sure your resources are used efficiently. Also,
415 since collectd is programmed multithreaded it benefits from hyperthreading
416 and multicore processors and makes sure that the daemon isn't idle if only
417 one plugins waits for an IO-operation to complete.
419 * Once set up, hardly any maintenance is necessary. Setup is kept as easy
420 as possible and the default values should be okay for most users.
426 * collectd's configuration file can be found at `sysconfdir'/collectd.conf.
427 Run `collectd -h' for a list of builtin defaults. See `collectd.conf(5)'
428 for a list of options and a syntax description.
430 * When the `csv' or `rrdtool' plugins are loaded they'll write the values to
431 files. The usual place for these files is beneath `/var/lib/collectd'.
433 * When using some of the plugins, collectd needs to run as user root, since
434 only root can do certain things, such as craft ICMP packages needed to ping
435 other hosts. collectd should NOT be installed setuid root since it can be
436 used to overwrite valuable files!
438 * Sample scripts to generate graphs reside in `contrib/' in the source
439 package or somewhere near `/usr/share/doc/collectd' in most distributions.
440 Please be aware that those script are meant as a starting point for your
441 own experiments.. Some of them require the `RRDs' Perl module.
442 (`librrds-perl' on Debian) If you have written a more sophisticated
443 solution please share it with us.
445 * The RRAs of the automatically created RRD files depend on the `step'
446 and `heartbeat' settings given. If change these settings you may need to
447 re-create the files, losing all data. Please be aware of that when changing
448 the values and read the rrdtool(1) manpage thoroughly.
451 collectd and chkrootkit
452 -----------------------
454 If you are using the `dns' plugin chkrootkit(1) will report collectd as a
455 packet sniffer ("<iface>: PACKET SNIFFER(/usr/sbin/collectd[<pid>])"). The
456 plugin captures all UDP packets on port 53 to analyze the DNS traffic. In
457 this case, collectd is a legitimate sniffer and the report should be
458 considered to be a false positive. However, you might want to check that
459 this really is collectd and not some other, illegitimate sniffer.
465 To compile collectd from source you will need:
467 * Usual suspects: C compiler, linker, preprocessor, make, ...
469 * A POSIX-threads (pthread) implementation.
470 Since gathering some statistics is slow (network connections, slow devices,
471 etc) the collectd is parallelized. The POSIX threads interface is being
472 used and should be found in various implementations for hopefully all
475 * CoreFoundation.framework and IOKit.framework (optional)
476 For compiling on Darwin in general and the `apple_sensors' plugin in
478 <http://developer.apple.com/corefoundation/>
480 * libclntsh (optional)
481 Used by the `oracle' plugin.
484 If you want to use the `apache', `ascent', `curl', `nginx', or `write_http'
486 <http://curl.haxx.se/>
489 Used by the `dbi' plugin to connect to various databases.
490 <http://libdbi.sourceforge.net/>
492 * libesmtp (optional)
493 For the `notify_email' plugin.
494 <http://www.stafford.uklinux.net/libesmtp/>
496 * libganglia (optional)
497 Used by the `gmond' plugin to process data received from Ganglia.
499 * libgcrypt (optional)
500 Used by the `network' plugin for encryption and authentication.
503 If present, the uuid plugin will check for UUID from HAL.
504 <http://hal.freedesktop.org/>
507 For querying iptables counters.
508 <http://netfilter.org/>
510 If not found on the system, a version shipped with this distribution can
511 be used. It requires some Linux headers in /usr/include/linux. You can
512 force the build system to use the shipped version by specifying
513 --with-libiptc=shipped
514 when running the configure script.
517 Library that encapsulates the `Java Virtual Machine' (JVM). This library is
518 used by the Java plugin to execute Java bytecode. See “Configuring with
521 * libmemcached (optional)
522 Used by the `memcachec' plugin to connect to a memcache daemon.
524 * libmysqlclient (optional)
525 Unsurprisingly used by the `mysql' plugin.
526 <http://dev.mysql.com/>
528 * libnatapp (optional)
529 Required for the “netapp” plugin.
530 This library is part of the “Manage ONTAP SDK” published by NetApp.
532 * libnetlink (optional)
533 Used, obviously, for the `netlink' plugin.
534 <http://www.linuxfoundation.org/en/Net:Iproute2>
536 * libnetsnmp (optional)
537 For the `snmp' plugin.
538 <http://www.net-snmp.org/>
540 * libnotify (optional)
541 For the `notify_desktop' plugin.
542 <http://www.galago-project.org/>
544 * liboping (optional)
545 Used by the `ping' plugin to send and receive ICMP packets.
546 <http://verplant.org/liboping/>
548 * libowcapi (optional)
549 Used by the `onewire' plugin to read values from onewire sensors (or the
551 <http://www.owfs.org/>
554 Used to capture packets by the `dns' plugin.
555 <http://www.tcpdump.org/>
558 Obviously used by the `perl' plugin. The library has to be compiled with
559 ithread support (introduced in Perl 5.6.0).
560 <http://www.perl.org/>
563 The PostgreSQL C client library used by the `postgresql' plugin.
564 <http://www.postgresql.org/>
566 * libpython (optional)
567 Used by the `python' plugin. Currently, only 2.3 ≦ Python < 3 is supported.
568 <http://www.python.org/>
571 Used by the `rrdtool' and `rrdcached' plugins. The latter requires RRDtool
572 client support which was added after version 1.3 of RRDtool. Versions 1.0,
573 1.2 and 1.3 are known to work with the `rrdtool' plugin.
574 <http://oss.oetiker.ch/rrdtool/>
576 * librt, libsocket, libkstat, libdevinfo (optional)
577 Various standard Solaris libraries which provide system functions.
578 <http://developers.sun.com/solaris/>
580 * libsensors (optional)
581 To read from `lm_sensors', see the `sensors' plugin.
582 <http://www.lm-sensors.org/>
584 * libstatgrab (optional)
585 Used by various plugins to collect statistics on systems other than Linux
587 <http://www.i-scream.org/libstatgrab/>
589 * libupsclient/nut (optional)
590 For the `nut' plugin which queries nut's `upsd'.
591 <http://networkupstools.org/>
594 Collect statistics from virtual machines.
595 <http://libvirt.org/>
598 Parse XML data. This is needed for the `ascent' and `libvirt' plugins.
599 <http://xmlsoft.org/>
602 <http://www.xmms.org/>
605 Parse JSON data. This is needed for the `curl_json' plugin.
606 <http://www.lloydforge.org/projects/yajl/>
608 Configuring / Compiling / Installing
609 ------------------------------------
611 To configure, build and install collectd with the default settings, run
612 `./configure && make && make install'. For detailed, generic instructions
613 see INSTALL. For a complete list of configure options and their description,
614 run `./configure --help'.
616 By default, the configure script will check for all build dependencies and
617 disable all plugins whose requirements cannot be fulfilled (any other plugin
618 will be enabled). To enable a plugin, install missing dependencies (see
619 section `Prerequisites' above) and rerun `configure'. If you specify the
620 `--enable-<plugin>' configure option, the script will fail if the depen-
621 dencies for the specified plugin are not met. In that case you can force the
622 plugin to be built using the `--enable-<plugin>=force' configure option.
623 This will most likely fail though unless you're working in a very unusual
624 setup and you really know what you're doing. If you specify the
625 `--disable-<plugin>' configure option, the plugin will not be built. If you
626 specify the `--enable-all-plugins' or `--disable-all-plugins' configure
627 options, all plugins will be enabled or disabled respectively by default.
628 Explicitly enabling or disabling a plugin overwrites the default for the
629 specified plugin. These options are meant for package maintainers and should
630 not be used in everyday situations.
632 By default, collectd will be installed into `/opt/collectd'. You can adjust
633 this setting by specifying the `--prefix' configure option - see INSTALL for
634 details. If you pass DESTDIR=<path> to `make install', <path> will be
635 prefixed to all installation directories. This might be useful when creating
636 packages for collectd.
638 Configuring with libjvm
639 -----------------------
641 To determine the location of the required files of a Java installation is not
642 an easy task, because the locations vary with your kernel (Linux, SunOS, …)
643 and with your architecture (x86, SPARC, …) and there is no ‘java-config’
644 script we could use. Configuration of the JVM library is therefore a bit
647 The easiest way to use the `--with-java=$JAVA_HOME' option, where
648 `$JAVA_HOME' is usually something like:
649 /usr/lib/jvm/java-1.5.0-sun-1.5.0.14
651 The configure script will then use find(1) to look for the following files:
657 If found, appropriate CPP-flags and LD-flags are set and the following
658 library checks succeed.
660 If this doesn't work for you, you have the possibility to specify CPP-flags,
661 C-flags and LD-flags for the ‘Java’ plugin by hand, using the following three
662 (environment) variables:
668 For example (shortened for demonstration purposes):
670 ./configure JAVA_CPPFLAGS="-I$JAVA_HOME/include -I$JAVA_HOME/include/linux"
672 Adding "-ljvm" to the JAVA_LDFLAGS is done automatically, you don't have to
678 To compile correctly collectd needs to be able to initialize static
679 variables to NAN (Not A Number). Some C libraries, especially the GNU
680 libc, have a problem with that.
682 Luckily, with GCC it's possible to work around that problem: One can define
683 NAN as being (0.0 / 0.0) and `isnan' as `f != f'. However, to test this
684 ``implementation'' the configure script needs to compile and run a short
685 test program. Obviously running a test program when doing a cross-
686 compilation is, well, challenging.
688 If you run into this problem, you can use the `--with-nan-emulation'
689 configure option to force the use of this implementation. We can't promise
690 that the compiled binary actually behaves as it should, but since NANs
691 are likely never passed to the libm you have a good chance to be lucky.
693 Likewise, collectd needs to know the layout of doubles in memory, in order
694 to craft uniform network packets over different architectures. For this, it
695 needs to know how to convert doubles into the memory layout used by x86. The
696 configure script tries to figure this out by compiling and running a few
697 small test programs. This is of course not possible when cross-compiling.
698 You can use the `--with-fp-layout' option to tell the configure script which
699 conversion method to assume. Valid arguments are:
701 * `nothing' (12345678 -> 12345678)
702 * `endianflip' (12345678 -> 87654321)
703 * `intswap' (12345678 -> 56781234)
709 For questions, bug reports, development information and basically all other
710 concerns please send an email to collectd's mailing list at
711 <collectd at verplant.org>.
713 For live discussion and more personal contact visit us in IRC, we're in
714 channel #collectd on freenode.
720 Florian octo Forster <octo at verplant.org>,
721 Sebastian tokkee Harl <sh at tokkee.org>,
722 and many contributors (see `AUTHORS').
724 Please send bug reports and patches to the mailing list, see `Contact'