NAME

YATG - Fast SNMP data poller daemon, with storage and graphing

VERSION

version 5.20002

DESCRIPTION

YATG is a daemon (background process) which at intervals wakes up and polls network devices for SNMP data, and then stores or prints that data. In this distribution are also included examples for presenting simple CGI web pages with graphs.

YATG is flexible, efficient and powerful. It can poll a large number of devices with thousands of ports in just a few seconds. The configuration is very simple, and the defaults sane (it's designed for sysadmins, after all).

You can use YATG both for historical logging, such as traffic counters on ports, as well as short-term monitoring which might feed into, say, Nagios. Wherever possible, data is translated to human-friendly formats for storage, such as using Leaf Names instead of OIDs, translated values (up, down, etc) and device port names rather than SNMP Interface Indexes.

How Does It Work?

At startup, yatg_updater loads its configuration from local files and a database, performs some basic SNMP connections to build a cache about device capabilities and so on, and then goes to sleep.

Periodically, as determined by the configuration, yatg_updater wakes up and polls all devices, then stores results, again according to instructions in the configuration.

If you have only the essential dependencies installed (see below) then you can only output results to STDOUT. With other modules, you have more options such as local or remote disk, or memcached based storage.

yatg_updater will re-load all its configuration if given a HUP signal. If you run the daemon persistently (for example with daemontools) then a cron job once a day is a good way to refresh the configuration. There is reference to this in one of the bundled example files.

What's in this distribution

yatg_updater

This is the main application, designed to be run persistently. It does not accept any input and only produces output when in debugging mode. It is a smart wrapper for the SNMP::Effective module.

YATG::Store family of modules

These are modules which take the SNMP poll results and store them to either local Disk, a Memcached server, the disk on a remote networked server, or Nagios via NSCA.

YATG::Retrieve family of modules

These are modules which read stored results back to you, for a given time window. The data can be retrieved from local Disk, a Memcached server, or the disk on a remote networked server.

yatg_trim

If using the Disk Store backend for results, eventually you'll want to save space by deleting old data. This script understands the backend file format and, given a duration, removes that amount of historical data from the file.

RPC::Serialized handlers

If storing and/or retrieving on a remote networked server, it should run an instance of RPC::Serialized, and these are the RPC Handlers for that server (see that module's documentation for further details).

CGI

For the special case of viewing graphs of disk-based poll results for switch port traffic counters, there is are two CGI scripts. One is a wrapper which presents an HTML page embedded with PNG images created from the other script.

Examples

The examples/ folder includes a copy of each of the files you should need for a complete deployment of YATG. Obviously some of them contain dummy data.

Where to go from here

To begin with, you probably want to see how to configure yatg_updater in YATG::Config.

Alongside that, there are examples of all the files you should need to install, in the examples/ folder of this distribution.

Each of the Store and Retrieve modules might have additional Perl module dependencies (i.e. from CPAN) - see the relevant docs for more details.

LOGGING and TESTING

This module uses "Log::Dispatch::Syslog" for logging, and by default will log timing data to your system's syslog service. More information is provided in the YATG::Config documentation.

To run in debug mode, where timing data is output to standard out rather than syslog, set the environment variable YATG_DEBUG to a true value.

To run the poller just once, set the YATG_SINGLE_RUN environment variable to a true value. This is great for development. It makes yatg_updater load its configuration, generate the device hints cache, sleep and then run just one poll cycle before exiting.

To override the interval between polling runs, set the YATG_INTERVAL environment variable to a number of seconds.

For example:

YATG_DEBUG=1 YATG_SINGLE_RUN=1 /usr/bin/yatg_updater /etc/yatg.yml

SEE ALSO

SNMP::Effective

This system uses SNMP::Effective at its core do the polling.

RPC::Serialized

Store polled data on another server using RPC::Serialized.

AUTHOR

Oliver Gorwits <oliver@cpan.org>

COPYRIGHT AND LICENSE

This software is copyright (c) 2015 by University of Oxford.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.