The Perl Toolchain Summit needs more sponsors. If your company depends on Perl, please support this very important event.

NAME

eris - Eris is the Greek Goddess of Chaos

VERSION

version 0.006

SYNOPSIS

eris exists to transform unstructuted, chaotic log data into structured messages.

Born out of disappointment and regret of existing solutions like Logstash, fluentd, and their kind, eris aims to make development and debugging of parsers easy and transparent. The goal is to provide a config that be used to to index logging data into Elasticsearch while being flexible enough to work with log files on the system. This makes it friendly to approach from a maintenance perspective as we don't need to run a massive app to figure out how a log message will be restructured.

DESCRIPTION

eris is structured to allow flexibility, extensibility, and visibility in every component.

CONCEPTS

DECODER

Decoders are pluggable thanks to eris::role::pluggable and they are searched for in the the default namespace eris::log::decoder. To add other namespaces, use the search_path paramter in a config file:

    ---
    decoders:
      search_path:
        - 'my::app::decoder'

Decoders operate on the raw string and provide rudimentary key/value pairs for the other contexts to operate on. Unlike the contexts, every discovered decoder is run for every message.

SEE ALSO

eris::log::decoders

Class providing access to installed and configured decoders on the system.

eris::log::contextualizer

Class which uses the decoders to transform the raw data into structured data.

eris::role::decoder

The abstract role which implements a decoder.

eris::log::decoder::syslog, eris::log::decoder::json

Default implementations of decoders.

CONTEXT

Contexts are pluggable and are searched for in the default namespace eris::log::decoder. To add your own namespaces, use the search_path parameter in your config file:

    ---
    contexts:
      search_path:
        - 'my::app::context'

Contexts implement the interface documented in eris::role::context. There are 4 major things to consider when implementing a new context.

contextualize_message

This method is called when the context matches the event data. This is where you can implement your own parsing or analysis of the event data. To add context to an event, use the eris::log's add_context() method. That context data will be available to future contexts.

sample_messages

Return an array of sample messages. This provides future developers with some data to use in testing and enhancing your context.

field

This specifies the field or fields that a matcher will operate on. There are two special fields * and _exists_. The * is used in conjunction with a matcher of * to match all messages. The _exists_ operator is used to check for the existance of a key in the context. A sample use of this field specifier is used by the eris::log::context::GeoIP context with an regex matcher to operate on any event data with field names matching '_ip$'.

matcher

Can be *, a string, a regex ref, an array reference, or a code reference. If * and field is * means match every message. If a literal string, or array reference, the literal string is checked against the value of in the field specified above and returns 1 if they are equivalent. If a regex reference, the regex is applied to the value in the specified field and the context is applied if the regex matches. A code reference should return 1 if the event is relevant to the context and 0 if it doesn't apply.

The default field is 'program', and the default matcher is a string with the value equal to the context's name attribute. For instance, eris::log::context::sshd defaults it's name to 'sshd', and since it doesn't override the field, this context is only applied to events with a 'program' key with a value of 'sshd'.

SEE ALSO

eris::log::contexts

Class providing access to installed and configured contexts on the system.

eris::log::contextualizer

Class which uses the contexts to transform the raw data into structured data.

eris::role::context

The abstract role which implements a context.

eris::log::context::sshd, eris::log::context::GeoIP

Selected example contexts

DICTIONARY

Dictionaries are used in conjunction with schemas to filter eris::log contexts down to only the keys and values we want. This allows better control of the data headed into storage to prevent key space explossons.

SEE ALSO

eris::dictionary

Class providing access to installed and configured dictionaries on the system.

eris::role::schema

Class which uses the dictionaries to filter structured data into a document.

eris::role::dictionary

The abstract role which implements a dictionary.

eris::dictionary::cee, eris::dictionary::eris::debug

Selected example contexts

SCHEMA

Schemas perform the transformation from structured data into documents for indexing. They allow control of the structure and destination of the document being indexed.

SEE ALSO

eris::schemas

Class providing access to installed and configured schemas on the system.

eris::role::schema

The abstract role which implements a schema.

eris::schema::syslog

Selected example contexts

IMPLEMENTATIONS

The goal of eris is to provide a set of tools that can be glued together to transform unstructured logging data into structured data and then rules for taking that structured data and storing it somewhere. That sounds cool, but there's nothing useful about it unless you can start playing with it now.

This is why eris ship withs sample implementations.

Scripts

Here's a list of the scripts installed along with eris so you can start breaking things.

eris-context.pl

This script allows you to do a few useful things. To see what happens to unstructured data, you can try performing some simple transforms via the built-in sample_messages:

    eris-context.pl --sample sshd

If you'd like to see what those samples look like as ElasticSearch build requests, you can:

    eris-context.pl --sample sshd --bulk

Without the --sample argument, you can feed data to it using STDIN or a file as it'll use the Perl magic diamond to read data until an EOF is reached.

To see what the bulk output would look like from a few sources:

Via pipe:

    tail /var/log/messages | eris-context.pl -b

Via a file:

    eris-context.pl -b /var/log/messsages

Via STDIN (for you crazy copy-pastafarians):

    eris-context.pl -b

The script provides more options, pull up it's help with:

    eris-context.pl --help
eris-field-lookup.pl

This script allows you to query the eris::dictionary to see what it knows about a particular field.

    eris-field-lookup.pl src_ip
eris-es-indexer.pl

This is a sample implementation that performs indexing of data received over syslog to an ElasticSearch cluster. It will parse all messages passed to it over STDIN and send them to an ElasticSearch cluster. It's single threaded, so it won't be able to keep up with a full speed log load. See it's help output for options and details:

    eris-es-indexer.pl --help
eris-stdin-listener.pl

This is a wrapper around eris-es-indexer.pl using POE::Component::WheelRun::Pool to provide a pool of workers for processing log data at scale. To use it with syslog-ng:

    destination d_eris { program("/usr/local/bin/eris-stdin-listener.pl" keep-alive(); ); };
    log  { source(src_network); destination(d_eris); };

See it's help output for options:

    eris-stdin-listener.pl --help
eris-eris-client.pl

This is a wrapper around eris-es-indexer.pl for use in environments using the POE::Component::Server::eris syslog service. This service is a simple, stateless syslog message dispatch system used primarily for development of new syslog parser use cases. It transforms the syslog stream into a subscription service any user on the local system can tap. IF you're using that server, you can run the eris-eris-client.pl to leverage POE::Component::Client::eris to receive messages from the upstream and dispatch them to a worker pool of eris-es-indexer.pl's. For more information, see:

    eris-eris-client.pl --help

AUTHOR

Brad Lhotsky <brad@divisionbyzero.net>

COPYRIGHT AND LICENSE

This software is Copyright (c) 2015 by Brad Lhotsky.

This is free software, licensed under:

  The (three-clause) BSD License