flume-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wolfgang Hoschek <whosc...@cloudera.com>
Subject Re: Flume Jambalaya - A Flume Plugin with Multiple Components
Date Fri, 02 May 2014 22:00:48 GMT
My sense is that a) is interesting if it evolves into a capable true native tailer, whereas
b) is already available in flume and c) and d) are already available in flume via the MorphlineInterceptor


On May 3, 2014, at 12:18 AM, Israel Ekpo <israel@aicer.org> wrote:

> Flume Community,
> I created a Flume Plugin with multiple components that complements the current version
of Apache Flume.
> This was necessary as part of a personal project as I working on.
> It is code named - Flume Jambalaya
> Jambalaya is a standalone Apache Flume plugin that contains a variety of sources, interceptors,
channels, sinks, serializers and other components designed to extend the Flume architecture.
It has been released under the Apache License version 2.0
> https://github.com/aicer/flume-jambalaya
> It currently contains:
> (a) File Source - This source lets you ingest data by tailing files from a specific path
> (b) ElasticSearch HTTP Sink - This sink sends events to an ElasticSearch cluster via
HTTP with no dependency on the ElasticSearch versions between Flume and the Server cluster.
> (c) DateInterceptor - The date interceptor is used for parsing dates from fields and
using that date or timestamp as the timestamp for the Flume event.
> (d) Grok Interceptor - allows you to extract structured data from unstructured text and
inject them as headers into the event
> Sample configuration files are available here
> https://github.com/aicer/flume-jambalaya/tree/master/sample-configuration-files
> I did not realize that the Flume trunk already has a HTTP Sink for ElasticSearch so you
can decide whether or not to use the sink that comes with it
> I am still testing and integrating the various components.
> Please check it out when you get a chance and send me some feedback
> Thanks.

View raw message