Copyright 2012 Rashid Khan
Kibana is a browser based interface for Logstash and ElasticSearch that allows you to efficiently search, visualize, analyze and otherwise make sense of your logs.
ruby >= 1.8.7 (probably?)
logstash >= 1.1.0
elasticsearch >= 0.18.0
java >= 1.6 if you want to run Kibana in JRuby
warbler if you want to create an executable standalone war file
git clone --branch=kibana-ruby https://github.com/rashidkpc/Kibana.git
gem install bundler
Set your elasticsearch server in KibanaConfig.rb:
Elasticsearch = "elasticsearch:9200"
Point your browser at http://localhost:5601
To run Kibana with JRuby, e.g. if you have to run in on a windows machine, you can create a (executable) WAR archive.
git clone --branch=kibana-ruby https://github.com/rashidkpc/Kibana.git cd Kibana jruby -S gem install bundler jruby -S bundle install
Configure your environment (see above).
jruby -S rake war
jruby -S warble executable war
if you want to include a webserver (default: jetty).
java [-Djetty.port=5601] -jar Kibana.war
Todo: Externalize the configuration. Any help would be appreciated.
Q: Why is there no last button?
A: ElasticSearch isn't so hot at going to the last result of a many million result query.
Q: Why is this Ruby instead of PHP now?
Q: Why do I have to set a limit on events to analyze? A: Big result sets take a long time to retrieve from elasticsearch and parse out
Q: Well then why don't you use the Elastic Search terms facet?
A: I've found the terms facet to cause out of memory crashes with large result sets. I don't know a way to limit the amount of memory a facet may use. Until there's a way to run a facet and know for sure it won't crash Elastic Search, I'm going to keep analysis features implemented in Ruby. I'm open to other suggestions though. I suggest you be careful with the Statistics mode, its more stable than terms, and I try to detect when it might be dangerous but can still bite you.
Q: Why do some results not show up when I search for a string I know is in
the elasticsearch indexes?
A: If you are searching analyzed fields, which is the default in ES for string fields, remember that they are broken down into terms. For instance, a search for "test" will match records containing firstname.lastname@example.org, since @ is a term boundary and is broken down into "test" and "bleh.com". However, this will NOT match records containing email@example.com because "test.com" is the full term and you are searching for an exact match. You would need to use test to match both of these records. Note you may also want to configure the ES analyze behavior for certain fields if this is not the desired behavior. Helpful References:
Q: How do I run Kibana under Apache? A: There's a few samples in the sample/apache directory on how to do this.
Q: Kibana is great, but I want to make it so users have to authenticate in order to access it. How do I do that? A: This can be handled a number of ways. The best way is to run Kibana with Passenger and Apache or Nginx. There's sample configurations in the sample directory. You can then handle your preferred authentication mechanism with Apache or Nginx.
Q: Where can I get some help with this?
A: Find me on Freenode - rashidkpc in #logstash