collectd - System information collection daemon
=================================================
http://collectd.org/
About
-----
collectd is a small daemon which collects system information periodically
and provides mechanisms to store and monitor the values in a variety of
ways.
Features
--------
* collectd is able to collect the following data:
- apache
Apache server utilization: Number of bytes transfered, number of
requests handled and detailed scoreboard statistics
- apcups
APC UPS Daemon: UPS charge, load, input/output/battery voltage, etc.
- apple_sensors
Sensors in Macs running Mac OS X / Darwin: Temperature, fanspeed and
voltage sensors.
- battery
Batterycharge, -current and voltage of ACPI and PMU based laptop
batteries.
- cpu
CPU utilization: Time spent in the system, user, nice, idle, and related
states.
- cpufreq
CPU frequency (For laptops with speed step or a similar technology)
- df
Mountpoint usage (Basically the values `df(1)' delivers)
- disk
Disk utilization: Sectors read/written, number of read/write actions,
average time an IO-operation took to complete.
- dns
DNS traffic: Query types, response codes, opcodes and traffic/octets
transfered.
- email
Email statistics: Count, traffic, spam scores and checks.
See collectd-email(5).
- entropy
Amount of entropy available to the system.
- exec
Values gathered by a custom program or script.
See collectd-exec(5).
- hddtemp
Harddisk temperatures using hddtempd.
- interface
Interface traffic: Number of octets, packets and errors for each
interface.
- iptables
Iptables' counters: Number of bytes that were matched by a certain
iptables rule.
- ipvs
IPVS connection statistics (number of connections, octets and packets
for each service and destination).
See http://www.linuxvirtualserver.org/software/index.html.
- irq
IRQ counters: Frequency in which certain interrupts occur.
- load
System load average over the last 1, 5 and 15 minutes.
- libvirt
CPU, disk and network I/O statistics from virtual machines.
- mbmon
Motherboard sensors: temperature, fanspeed and voltage information,
using mbmon(1).
- memcached
Statistics of the memcached distributed caching system.
- memory
Memory utilization: Memory occupied by running processes, page cache,
buffer cache and free.
- multimeter
Information provided by serial multimeters, such as the `Metex
M-4650CR'.
- mysql
MySQL server statistics: Commands issued, handlers triggered, thread
usage, query cache utilization and traffic/octets sent and received.
- netlink
Very detailed Linux network interface and routing statistics. You can get
(detailed) information on interfaces, qdiscs, classes, and, if you can
make use of it, filters.
- network
Receive values that were collected by other hosts. Large setups will
want to collect the data on one dedicated machine, and this is the
plugin of choice for that.
- nfs
NFS Procedures: Which NFS command were called how often. Only NFSv2 and
NFSv3 right now.
- nginx
Collects statistics from `nginx' (speak: engine X), a HTTP and mail
server/proxy.
- ntpd
NTP daemon statistics: Local clock drift, offset to peers, etc.
- nut
Network UPS tools: UPS current, voltage, power, charge, utilisation,
temperature, etc. See upsd(8).
- perl
The perl plugin implements a Perl-interpreter into collectd. You can
write your own plugins in Perl and return arbitrary values using this
API. See collectd-perl(5).
- ping
Network latency: Time to reach the default gateway or another given
host.
- processes
Process counts: Number of running, sleeping, zombie, ... processes.
- sensors
System sensors, accessed using lm_sensors: Voltages, temperatures and
fan rotation speeds.
- serial
RX and TX of serial interfaces. Linux only; needs root privileges.
- snmp
Read values from SNMP (Simple Network Management Protocol) enabled
network devices such as switches, routers, thermometers, rack monitoring
servers, etc. See collectd-snmp(5).
- swap
Pages swapped out onto harddisk or whatever is called `swap' by the OS..
- tail
Follows (tails) logfiles, parses them by lines and submits matched
values.
- tape
Bytes and operations read and written on tape devices. Solaris only.
- tcpconns
Number of TCP connections to specific local and remote ports.
- users
Users currently logged in.
- vmem
Virtual memory statistics, e. g. the number of page-ins/-outs or the
number of pagefaults.
- vserver
System resources used by Linux VServers.
See .
- wireless
Link quality of wireless cards. Linux only.
- xmms
Bitrate and frequency of music played with XMMS.
* Output can be written or send to various destinations by the following
plugins:
- csv
Write to comma separated values (CSV) files. This needs lots of
diskspace but is extremely portable and can be analysed with almost
every program that can analyse anything. Even Microsoft's Excel..
- network
Send the data to a remote host to save the data somehow. This is useful
for large setups where the data should be saved by a dedicated machine.
- perl
Of course the values are propagated to plugins written in Perl, too, so
you can easily do weird stuff with the plugins we didn't dare think of
;) See collectd-perl(5).
- rrdtool
Output to round-robin-database (RRD) files using librrd. See rrdtool(1).
This is likely the most popular destination for such values. Since
updates to RRD-files are somewhat expensive this plugin can cache
updates to the files and write a bunch of updates at once, which lessens
system load a lot.
- unixsock
One can query the values from the unixsock plugin whenever they're
needed. Please read collectd-unixsock(5) for a description on how that's
done.
* Logging is, as everything in collectd, provided by plugins. The following
plugins keep up informed about what's going on:
- logfile
Writes logmessages to a file or STDOUT/STDERR.
- perl
Log messages are propagated to plugins written in Perl as well.
See collectd-perl(5).
- syslog
Logs to the standard UNIX logging mechanism, syslog.
* Notifications can be handled by the following plugins:
- exec
Execute a program or script to handle the notification.
See collectd-exec(5).
- logfile
Writes the notification message to a file or STDOUT/STDERR.
- network
Send the notification to a remote host to handle it somehow.
- perl
Notifications are propagated to plugins written in Perl as well.
See collectd-perl(5).
* Miscellaneous plugins:
- uuid
Sets the hostname to an unique identifier. This is meant for setups
where each client may migrate to another physical host, possibly going
through one or more name changes in the process.
* Performance: Since collectd is running as a daemon it doesn't spend much
time starting up again and again. With the exception of the exec plugin no
processes are forked. Caching in output plugins, such as the rrdtool and
network plugins, makes sure your resources are used efficiently. Also,
since collectd is programmed multithreaded it benefits from hyperthreading
and multicore processors and makes sure that the daemon isn't idle if only
one plugins waits for an IO-operation to complete.
* Once set up, hardly any maintenance is necessary. Setup is kept as easy
as possible and the default values should be okay for most users.
Operation
---------
* collectd's configuration file can be found at `sysconfdir'/collectd.conf.
Run `collectd -h' for a list of builtin defaults. See `collectd.conf(5)'
for a list of options and a syntax description.
* When the `csv' or `rrdtool' plugins are loaded they'll write the values to
files. The usual place for these files is beneath `/var/lib/collectd'.
* When using some of the plugins, collectd needs to run as user root, since
only root can do certain things, such as craft ICMP packages needed to ping
other hosts. collectd should NOT be installed setuid root since it can be
used to overwrite valuable files!
* Sample scripts to generate graphs reside in `contrib/' in the source
package or somewhere near `/usr/share/doc/collectd' in most distributions.
Please be aware that those script are meant as a starting point for your
own experiments.. Some of them require the `RRDs' Perl module.
(`librrds-perl' on Debian) If you have written a more sophisticated
solution please share it with us.
* The RRAs of the automatically created RRD files depend on the `step'
and `heartbeat' settings given. If change these settings you may need to
re-create the files, losing all data. Please be aware of that when changing
the values and read the rrdtool(1) manpage thoroughly.
collectd and chkrootkit
-----------------------
If you are using the `dns' plugin chkrootkit(1) will report collectd as a
packet sniffer (": PACKET SNIFFER(/usr/sbin/collectd[])"). The
plugin captures all UDP packets on port 53 to analyze the DNS traffic. In
this case, collectd is a legitimate sniffer and the report should be
considered to be a false positive. However, you might want to check that
this really is collectd and not some other, illegitimate sniffer.
Prerequisites
-------------
To compile collectd from source you will need:
* Usual suspects: C compiler, linker, preprocessor, make, ...
* A POSIX-threads (pthread) implementation.
Since gathering some statistics is slow (network connections, slow devices,
etc) the collectd is parallelized. The POSIX threads interface is being
used and should be found in various implementations for hopefully all
platforms.
* libcurl (optional)
If you want to use the `apache' and/or `nginx' plugins.
* libhal (optional)
If present, the uuid plugin will check for UUID from HAL.
* libiptc (optional)
For querying iptables counters.
* libmysqlclient (optional)
Unsurprisingly used by the `mysql' plugin.
* libnetlink (optional)
Used, obviously, for the `netlink' plugin.
* libnetsnmp (optional)
For the `snmp' plugin.
* liboping (optional, if not found a version shipped with this distribution
can be used)
Used by the `ping' plugin to send and receive ICMP packets.
* libpcap (optional)
Used to capture packets by the `dns' plugin.
* libperl (optional)
Obviously used by the `perl' plugin. The library has to be compiled with
ithread support (introduced in Perl 5.6.0).
* librrd (optional; headers and library; rrdtool 1.0 and 1.2 both work fine)
If built without `librrd' the resulting binary will be `client only', i.e.
will send its values via multicast and not create any RRD files itself.
Alternatively you can chose to write CSV-files (Comma Separated Values)
instead.
* libsensors (optional)
To read from `lm_sensors', see the `sensors' plugin.
* libstatgrab may be used to collect statistics on systems other than Linux
and/or Solaris. Note that CPU- and disk-statistics, while being provided
by this library, are not supported in collectd right now..
* libupsclient/nut (optional)
For the `nut' plugin which queries nut's `upsd'.
* libxmms (optional)
* librt, libsocket, libkstat, libdevinfo
Various standard Solaris libraries which provide system functions.
* CoreFoundation.framework and IOKit.framework
For compiling on Darwin in general and the `apple_sensors' plugin in
particular.
* libvirt (optional)
Collect statistics from virtual machines.
* libxml2 (optional)
Parse XML data provided by libvirt.
Configuring / Compiling / Installing
------------------------------------
To configure, build and install collectd with the default settings, run
`./configure && make && make install'. For detailed, generic instructions
see INSTALL. For a complete list of configure options and their description,
run `./configure --help'.
By default, the configure script will check for all build dependencies and
disable all plugins whose requirements cannot be fulfilled (any other plugin
will be enabled). To enable a plugin, install missing dependencies (see
section `Prerequisites' above) and rerun `configure'. If you specify the
`--enable-' configure option, you can force the plugin to be built.
This will most likely fail though unless you're working in a very unusual
setup and you really know what you're doing.
By default, collectd will be installed into `/opt/collectd'. You can adjust
this setting by specifying the `--prefix' configure option - see INSTALL for
details. If you pass DESTDIR= to `make install', will be
prefixed to all installation directories. This might be useful when creating
packages for collectd.
Crosscompiling
--------------
To compile correctly collectd needs to be able to initialize static
variables to NAN (Not A Number). Some C libraries, especially the GNU
libc, have a problem with that.
Luckily, with GCC it's possible to work around that problem: One can define
NAN as being (0.0 / 0.0) and `isnan' as `f != f'. However, to test this
``implementation'' the configure script needs to compile and run a short
test program. Obviously running a test program when doing a cross-
compilation is, well, challenging.
If you run into this problem, you can use the `--with-nan-emulation'
configure option to force the use of this implementation. We can't promise
that the compiled binary actually behaves as it should, but since NANs
are likely never passed to the libm you have a good chance to be lucky.
Contact
-------
For questions, bugreports, development information and basically all other
concerns please send an email to collectd's mailinglist at
.
For live discussion and more personal contact visit us in IRC, we're in
channel #collectd on freenode.
Author
------
Florian octo Forster ,
Sebastian tokkee Harl ,
and many contributors (see `AUTHORS').
Please send bugreports and patches to the mailinglist, see `Contact' above.