The Banana project was forked from Kibana, and works with all kinds of time series (and non-time series) data stored in Apache Solr. It uses Kibana's powerful dashboard configuration capabilities, ports key panels to work with Solr, and provides significant additional capabilities, including new panels that leverage D3.js.

The goal is to create a rich and flexible UI, enabling users to rapidly develop end-to-end applications that leverage the power of Apache Solr. Data can be ingested into Solr through a variety of ways, including LogStash, Flume and other connectors.

Pull the repo from the "release" branch; version 1.4 will be tagged as banana-1.4.

Banana 1.4 contains many new features, new panels, enhancements and bug fixes to improve the overall user experience and stability. Thank you to our growing community for your suggestions and contributions! Please continue sending us your feedbacks, so that we can further extend and improve Banana!

This release includes the following key new features and improvements:

1. Banana 1.4 provides much improved performance - by better utilizing Solr's caches in Timepicker Module. This improvement is discernable in the Relative and Since time modes.
2. A new Full Text Search panel provides a more traditional search interface to view textual data.
3. Enhancements to the Table panel improve performance and user experience:
   • Sorting option can now be turned off in order to speed up the search results returned from Solr.
   • A particular column - corresponding to a URI field - can be set as the hyperlink column. When set, this value will become clickable and linked to a specific URI.
   • You can now display images inside a table column.
4. In the Range Facet panel, users now have the ability to set the chart's precision automatically or manually.
5. In the Terms panel, chart colors can be customized by changing the default color template or by using field values as colors.
6. It is now possible to load and save a dashboard to Gist.
7. We have fixed the Solr server location for the banana-int collection, the internal collection that stores dashboards. Now banana-int should be located on the same Solr server as specified in Solr Settings in the Dashboard configuration. You do not have to manually edit config.js file anymore.
8. The dashboard contains a new button that enables users to quickly create a new dashboard from templates. Currently, we provide two default templates: a time-series dashboard template and a non time-series one.
9. The Histogram panel no longer requires you to set a time field; as logically expected, it will get the time field from the Timepicker panel.
10. We have enhanced the usability by allowing you to specify custom help messages inside each panel. You no longer need to use a separate Text panel for this purpose. Instead, you can now embed information and instructions within each panel in the dashboard to better communicate with your users.

Banana 1.3 improves on its already powerful capability to visualize and interpret generalized time series data (banana is not only used to search log files, but also visualize social media streams, call center logs, medical records, and etc.). It starts leveraging the power of D3.js (data-driven documents) and provides new panels and enhancements, while also allowing visualization of non-time series data. Key new features include:

1. Stats and aggregations are now available in the Terms and Map panels. In addition to count mode, you can now visualize stats such as mean, max, min, sum, etc.
2. A new Range Facet panel allows you to visualize and graphically explore distributions on numeric fields, with selections being reflected across the entire dashboard.
3. A new Heatmap panel provides for visualization of the powerful pivot faceting capability of Solr.
4. A new Ticker panel provides a stock ticker like representation of trends in your time series data.
5. The Export functionality in the the Table Module has been optimized for vastly improved performance and now allows you to export only a subset of the fields in the returned documents.
6. Previous versions required a Timepicker and time fields set in all panels for them to work. We have cleaned up the code so that it will now work without a Timepicker and a time filter, which will help visualize non-time series data. The time field provided in the Timepicker is used by all panels.
7. General improvements in the UI and in-product help documentation makes Banana 1.3 easier to use.
8. The directory structure is now cleaned up and legacy files have been removed. Instructions for enabling CORS in Solr and for setting the schema/config for banana's internal collections are now contained in the resources directory.

Following release 1.1, we have addressed a number of user requests, including:

1. This release provides panels for representing geo-spatial data—a map module that provides a heat map-style representation based on two-letter country codes or US state codes, and a bettermap module that provides a clustered representation of location (LatLonType) data.
2. The Table Module now has a Save button that enables you to save to csv, JSON or XML formats so that you can use other tools like MS Excel for further analysis. The number of rows downloaded will be equal to number of “pageable” hits configured in the Paging tab within the Table Panel Configuration Menu (accessed by clicking on the cog wheel icon near the top right of the table panel).
3. You can now control whether a dashboard can be saved and/or edited from the Editable checkbox in the General tab, and the Controls tab, both within the Dashboard Configurator (accessed from the cog-wheel icon to very top and right of dashboard).
4. We have added a hits panel that provides you with the number of matching results returned while using the global query parameters. This is useful if you want to make the number prominent or if you are not using the histogram panel prominently.
5. You can now provide additional Global Query Parameters that apply to all panels of the dashboard from the Solr tab in the Dashboard Configurator. Among other uses, this feature is invaluable for:
   • Specifying a custom query parser (Solr query parameter: &defType) or search handler (&qt)
   • Specifying a user type for use in custom business rules at the Solr server.
   • Specifying default search fields (&df)
6. We fixed a bug in the values mode within the histogram module, where missing values were previously assumed to be zero. This led to jagged graphs when the “group by” option was used. We no longer set them to zero but rather have the individual lines skip the missing values.
7. In the Absolute Time and Since modes, the timepicker used to skip back one day if your browser time was behind UTC. This issue has now been fixed.
8. Banana 1.1 hardcoded certain default search fields (df's) to work with our LogStash output writer. Specifically, it hardcoded a df=message. This means that your old dashboards may not be fetching query results with Banana 1.2, though they were doing so with 1.1. To fix this, add a Global Query Parameter &df=message (or whatever field you want to search on) within the Dashboard Configurator. Alternately, you can set the default search field in your solrconfig (recommended).

We have added a number of exciting new features and fixed key issues, including:

1. You can now add a Filtering panel that supports global filter queries (fq's). Now, if you click on a facet in the terms panel, the results will be filtered for that particular value.
2. The terms, histogram and table modules allow you to specify a panel-specific filter query (within the Query Tab while configuring the panel) allowing greater flexibility in designing dashboards.
3. The inspector icon on these panels shows the Solr query, which is very useful for debugging dashboards.
4. The Histogram module allows you to plot values in addition to counts. It also allows you to group values by another field. This would be useful if for example you plot CPU utilization over time and want to group by hostname.
5. The sort operation in the Table module is now fixed and works correctly on single-valued fields.
6. We have refactored the code to enable easier addition of new modules and fixes to existing modules.

If you created dashboards for Banana 1.0, you did not have a global filtering panel. In some cases, these filter values can be implicitly set to defaults that may lead to strange search results. We recommend updating your old dashboards by adding a filtering panel. A good way to do it visually is to put the filtering panel on its own row and hide it when it is not needed.

• A modern web browser. The latest version of Chrome and Firefox have been tested to work. Safari also works, except for the "Export to File" feature for saving dashboards. We recommend that you use Chrome or Firefox while building dashboards.
• A webserver.
• A browser reachable Solr server. The Solr endpoint must be open, or a proxy configured to allow access to it.

Run Solr at least once to create the webapp directories

	cd $SOLR_HOME/example  
	java -jar start.jar

Copy banana folder to $SOLR_HOME/example/solr-webapp/webapp/

Browse to http://<solr_server>:<port_number>/solr/banana/src/index.html#/dashboard

If your Solr server/port is different from localhost:8983, edit banana/src/config.js and banana/src/app/dashboards/default.json to enter the hostname and port that you are using. Remember that banana runs within the client browser, so provide a fully qualified domain name (FQDN), because the hostname and port number you provide should be resolvable from the client machines.

If you have not created the data collections and ingested data into Solr, you will see an error message saying "Collection not found at .." You can use any connector to get data into Solr. If you want to use LogStash, please go to the Solr Output Plug-in for LogStash Page (https://github.com/LucidWorks/solrlogmanager) for code, documentation and examples.

LucidWorks has packaged Solr, LogStash (with a Solr Output Plug-in), and Banana (the Solr port of Kibana), along with example collections and dashboards in order to rapidly enable proof-of-concepts and initial development/testing. See http://www.lucidworks.com/lucidworks-silk/.

Pull the repo from the "release" branch; versions 1.3, 1.2 and 1.1 will be tagged as banana-1.3, banana-1.2 and banana-1.1 respectively. Run "ant" from within the banana directory to build the war file.

cd $BANANA_REPO_HOME  
ant 

The war file will be called banana-buildnumber.war and will be located in $BANANA_REPO_HOME/build

cp $BANANA_REPO_HOME/build/banana-buildnumber.war $SOLR_HOME/example/webapps/banana.war   
cp $BANANA_REPO_HOME/jetty/banana-context.xml $SOLR_HOME/example/contexts/      

Run Solr:

cd $SOLR_HOME/example/
java -jar start.jar    

Browse to http://localhost:8983/banana (or the FQDN of your solr server).

Banana is a an AngularJS app and can be run in any webserver that has access to Solr. You will need to enable CORS on the Solr instances that you query, or configure a proxy that makes requests to banana and Solr as same-origin. We typically recommend the latter approach.

If you want to save and load dashboards from Solr, create a collection using the configuration files provided in either the resources/banana-int-solr-4.4 (for Solr 4.4) directory or the resources/banana-int-solr-4.5 directory (for Solr 4.5 and above). If you are using Solr Cloud, you will need to upload the configuration into ZooKeeper and then create the collection using that configuration.

The Solr server configured in config.js will serve as the default node for each dashboard; you can configure each dashboard to point to a different Solr endpoint as long as your webserver and Solr put out the correct CORS headers. See the README file under the resources/enable-cors directory for a guide.

Q: How do I secure my solr endpoint so that users do not have access to it?
A: The simplest solution is to use a Apache or nginx reverse proxy (See for example https://groups.google.com/forum/#!topic/ajax-solr/pLtYfm83I98).

Q: Can I use banana for non-time series data?
A: Yes, from version 1.3 onwards, non-time series data are also supported.

1. LucidWorks SILK: http://www.lucidworks.com/lucidworks-silk/
2. Webinar on LucidWorks SILK: http://programs.lucidworks.com/SiLK-introduction_Register.html.
3. LogStash: http://logstash.net/
4. SILK Use Cases: https://github.com/LucidWorks/silkusecases. Provides example configuration files, schemas and dashboards required to build applications that use Solr and Banana.

Banana uses the dashboard configuration capabilities of Kibana (from which it is forked) and ports key panels to work with Solr. Moreover, it provides many additional capabilities like heatmaps, range facets, panel specific filters, global parameters, and visualization of "group-by" style queries. We are continuing to add many new panels that go well beyond what is available in Kibana, helping users build complete applications that leverage the data stored in Apache Solr, HDFS and a variety of sources in the enterprise.

If you have any questions, please contact Andrew Thanalertvisuti ([email protected]) or Ravi Krishnamurthy ([email protected]).

Kibana is a trademark of Elasticsearch BV
Logstash is a trademark of Elasticsearch BV