Installing the Analytics Engine for Discovery Search

You may have already installed the Analytics Engine as part of the Shinydocs™ Cognitive Toolkit implementation. If not, this article will guide you through the steps.

The Analytics Engine comprises two components:

1. The Indexer

2. The Visualizer

The Visualizer leverages an index pattern to determine the results to display in visualizations and dashboards. Visualizations are widgets that explore or explain items in the index using charts or graphs. Dashboards, imported into the Visualizer, are used to group multiple visualizations to present a specific explanatory or exploratory theme to stakeholders.

What You Will Need

Running the Analytics Engine installer

Two Windows services (the shinydocs-indexer and the shinydocs-visualizer) are set up during the installation of the Analytics Engine.

1. Right-click on the shinydocs-indexer-visualizer-[DATE].zip file and Unblock

2. Apply

3. Extract the contents to where you want to install the Visualizer

4. Run a command prompt window as Administrator

5. Change the directory to the location of the extracted zip file

6. Enter install.bat and run

7. Follow the prompts

Configuring the Indexer

Updating the heap space in jvm.options

1. Navigate to the directory where the Indexer is installed (for example, C:\ShinyDocs\indexer\config)

2. Open the jvm.options file with a text editor

3. Scroll down and locate the -Xms (initial heap space) and -Xmx (maximum heap space) options on lines 22 and 23

4. Determine the new value by dividing your server’s RAM allocation by two and then subtracting 4 GB (to leave enough room for the OS)

5. For example, if the server has 64GB of RAM, change the values to 28
   -Xms28g
-Xmx28g

6. Save the changes

Updating the elasticsearch.yml file

1. Navigate to the directory where the Indexer is installed (for example, C:\ShinyDocs\indexer\config)

2. Open the elasticsearch.yml file with a text editor

3. Update cluster.name (line 17)
   Note: This name will need to be the same on every node in the cluster.

   1. Remove the hashtag character # from the beginning of the line

   2. Replace my-application with a name for your cluster (e.g. acme-analytics)

4. Update node.name ( line 23)

   1. Remove the hashtag character # from the beginning of the line

   2. Replace node-1 with a unique name for this machine. This will be the name reported on the Monitoring page of Kibana. Using the server's name is a good choice!

5. Update bootstrap.memory_lock (line 43)

   1. Remove the hashtag character # from the beginning of the line

   2. Keep the value true

6. Update network.host (line 55)

   1. Remove the hashtag character # from the beginning of the line

   2. The easiest value to set network.host to is 0.0.0.0 meaning that it will bind on all network interfaces on the machine. For more details see network.host | Elasticsearch Guide [6.8] | Elastic.

   3. If this line is not modified, the index and the Visualizer will not be accessible outside of the local machine

7. If you are configuring a cluster of servers, please see our guide Cluster Setup for Analytics Engine (Shinydocs Indexer)

8. Save the changes

9. Open services.msc and restart the shinydocs-indexer service

Configuring the Visualizer

Updating the kibana.yml file

To reach the visualizer from a different machine, you will need to edit the kibana.yml file.

1. Navigate to the directory where the Visualizer is installed (for example, C:\ShinyDocs\visualizer\config)

2. Open the kibana.yml file with a text editor

3. Line 7 - server.host

   1. Remove the hashtag character # at the beginning of the line

   2. Set the server.host to 0.0.0.0

4. Save the changes

5. Open services.msc and restart the shinydocs-visualizer service

The Analytics Engine setup is complete!