Tuesday, October 18, 2022

RangerMSP Knowledge Graph requirements

This post covers the pre-requisite packages you will need to build your Knowledge Graph for RangerMSP.  For some background see the previous post discussing MSP Knowledge Graphs

 Before we begin - you SHOULD be aware recently there have begun to develop some commercial offerings that can provide functionality like this.  They have full web user interfaces, and much prettier reporting features, and MANY more integrations than I've built.  An example of this is https://www.liongard.com/  When I started this project these products were just getting their start and were NOT very mature.  Also the platform I've developed is built entirely using open source tools, can be completely customized and self-hosted which lets you avoid the pay-per-endpoint-per-month service model.  So there is that.  Ok - enough with the disclaimer - let's get to it!

Let's talk about the tools you will need:

• RangerMSP SQL edition

This system was built using the FULL Advantage SQL Server version of RangerMSP.  The other tools used  will not work without this because the JDBC data loader is not available for the non-SQL local database.

• You must SHARE your RangerMSP (CommitCRM) folder.  You will need to grant both FILE and SHARE permissions to whichever account you run the data ingestion (HOP) workflows from.

The REAL magic of this system is the next 2 tools.  Neo4j Graph database and HOP orchestration Platform:

Neo4j is the graph database that unleashes the power of the relationships between your data.  This will house the knowledge graph database.

• For the Neo4j Database - You have a multiple options:  

Cloud hosted (AuraDB):  Free edition is limited to 200k nodes, 400k relationships: 

Download here Self hosted: (community Edition is free)   

You could also use the desktop edition if you want full enterprise functionality (note this cannot act as a shared multi-user server though - use Aura or community edition for that)

If you are installing a self-hosted Neo4j database I have instructions here

Apache HOP is the modern evolution of the popular ETL (Extract, Transform and Load) tool Pentaho Data Integration (Kettle).  This allows us to build workflows and perform API calls that import, massage, and deliver the data and relationships that will then be stored within Neo4j.

Download this and install it on a pc or server (requires JAVA).  You can install it on the same computer as your neo4j database if you like.

Additional Requirements:

• Neo4j requires the APOC libraries (automatically included in auraDB)

Advantage Database JDBC Driver  (this gets installed in HOP)

Getting your data into this system is only the first step.  Making the data usable, and to elicit ACTION is the real key!  If there's a problem, you want a ticket created with the proper details.  If there is interest I will follow-up with further parts to this series:

• REST API for RangerMSP to allow automation of ticket/charge/asset/contact/opportunity creation

• An easy-to-use DASHBOARD report viewer that works with Neo4j to present visual and tabular reporting

• Additional data integration workflows for HOP 

• A ticket-queue building workflow based on outcomes of above integrations

...Next → Step-by-Step Installation & Setup

No comments:

Post a Comment

Have a comment? Would love to hear it!