Site icon TheWindowsUpdate.com

Microsoft Azure Data Services Integration into ServiceNow

This post has been republished via RSS; it originally appeared at: New blog articles in Microsoft Tech Community.

ServiceNow is a SaaS based application that provides service management software via multiple offerings such as:  IT services management (ITSM), IT operations management (ITOM) and IT business management (ITBM).  ServiceNow is a leader within the Gartner Magic Quadrant and holds a position in the top right of the quadrant as a visionary.  ServiceNow ITSM, provides the ability to track/monitor tickets created and resolve quickly and effectively.  For users who want to analyze incidents and track progress, ServiceNow does provide out of the box reporting and analytics capabilities. In our use scenario, our customer wanted the ability to leverage an analytics platform like PowerBI in order to slice and dice and visualize the data in a variety of ways.  More specifically, there was a need to understand cross functional ITSM impacts on other areas of the organization to develop a comprehensive view and explore potential metrics and KPIs through a single integrated data view. Hence, extracting data from ServiceNow into Azure and/or PowerBI integration into ServiceNow was established and a defined requirement.

 

In addition, our customer wanted to create a POC to apply ML/AI on ITSM operations to understand root cause analysis and directly impact cost.  There is a finite and measurable cost for each ticket created.  The sooner major incidents can be identified and resolved, the more they would save on costs.  ML/AI leveraged models that could identify root cause, minimize duplication of tickets, reduce time to resolution and further reduce the operational complexity of managing a large number of open tickets all resulting in a simplified, better managed process with a measurable cost savings.

 
This post will demonstrate the integration options between Azure and ServiceNow as well as leveraging Azure to apply AI/ML on some of the scenarios described earlier.  We will share some of the learnings we had as we went through this journey.   Various integration points between Azure and ServiceNow:

 

  1. Azure Data Factory to ServiceNow Integration
  2. Connect directly to ServiceNow with PowerBI using the SIMBA driver
  3. Leverage PowerBI Premium ML in order to execute models for use cases described above leveraging ServiceNow data

 

At a high level, the solution will look as follows:

 

 

Azure Data Factory ServiceNow Connector Integration

With Azure Data Factory, there are two integration options into ServiceNow:

  1. ServiceNow Connector out of the box or
  2. REST API Connector

ServiceNow Connector

In you Azure Data Factory, create a new connection and search for ServiceNow as shown below

Configure the ServiceNow connectivity:

Key takeaways from the ServiceNow connectivity option:

REST API Connector

In Azure Data Factory, create a new connection and search for REST as shown below

Configure the REST API to your ServiceNow instance. 

Key takeaways from the REST API connector option:

Login to ServiceNow and Search for “REST”

As a side note, we found the Tables/Schemas documentation for ServiceNow that was relevant for the areas we were interested in like Problems leading to one or many Incidents.  The ServiceNow models and schemas provided were a great way to understand what was available to us and what table elements we would need.  This documentation can be found in the ServiceNow docs.

 

 

Here is a sample of the Incidents table and its associated relationships:

 

Once you’ve established the linked service, you can even filter using the REST APIs as shown below.

Key takeaways from configuring the REST API connector:

PowerBI Desktop to ServiceNow via SIMBA driver

We discovered that PowerBI provided a ServiceNow app that provided an out of the box dashboard into ServiceNow however after we discovered it, it was subsequently removed as an option and hence we had to look for alternative connectivity options like the SIMBA driver

 

The SIMBA is a 3rd party vendor that driver provides JDBC/ODBC connectivity to ServiceNow.  We installed the SIMBA driver and was able to connect relatively easily to ServiceNow using an JDBC/ODBC driver. The SIMBA driver was also able to read the user defined tables and columns as the Azure Data Factory REST API described earlier

Leveraging the SIMBA driver, we were able to directly query ServiceNow for our targeted tables like Incidents and Problems as shown below:

 

 

PowerBI Premium ML

 

We used a subset of Microsoft’s support ticket data; our subset contained only a description of the ticket and the category assigned. We wanted to train our machine learning model to effectively predict the category given the support ticket description in plain text. PowerBI Premium provided AutoML (Automated Machine Learning) capability that let users build ML models without writing any code. Behind the scenes AutoML iteratively trains several models optimizing model accuracy to return the best model for a given dataset.

 

After creating a PBI workspace for the project, we created a dataflow that contains all the entities (data connection, datasets and ML models).

To link our dataset, stored as an Azure SQL table, we added an entity to our dataflow setting data source as an Azure SQL DB

Specify the connection settings and set query to make sure that data is in the format you want for your model

To use AutoML to train our ML model, click ML icon in actions for our entity.

 

Pick Category field as an outcome (what we want to predict) and classification as ML problem, select Description field as data to study, provide a name for your model and hit Save and train. We may also pick how long we want training to continue.

 

Once the model is trained, training report shows various metrics on model efficiency and the top predictors as well as training details including how many models were trained to find the best model

 

 

 

 

We can also apply model on new data periodically (e.g., every hour) to get the predicted category given support ticket description which can help IT teams to effectively direct expert resources

 

Key takeaways from PBI Machine Learning:

Wrap Up

In summary, in this blog we were able to demonstrate the following:

 

About the authors:

 

 

 

 

Exit mobile version