Connect Architecture Overview

[MattD]

Alteryx Connect has a few architectural components, first of which is the Alteryx Connect Server. This is a stand-alone server, either on your premises or your managed private cloud. The technical specifications and hardware requirements for this server are available here, and can scale vertically with additional CPU cores. There are three main application components to the Connect Server:

• The Alteryx Connect website allows data workers to find, assess, and enrich metadata stored in the system as a preparatory step before beginning analytic development. This web front-end runs on Apache Tomcat as an embedded service within a parent Java application.
• The Java application processes web requests, metadata ingestion, and indexing functions. This is the parent process for both the Tomcat web server and the persistence layer. 
• The persistence layer is an embedded H2 database holding both technical metadata derived from source systems (such as Tableau or MSSQL Server), and well as information generated from the Connect website (such as user comments, enriched associations, glossary terms, and data assessments). This database can be queried with an odbc or jdbc connection.

The second server related to Alteryx Connect is the Alteryx Server. Alteryx Server exposes Alteryx workflows via a website called the Gallery. These workflows may reference a variety of data sources.

• Alteryx Server also has a scheduling service. We’ll take advantage of this service to schedule the loading of metadata into Alteryx Connect.
• Loading metadata into Alteryx Connect is accomplished with a library of installed metadata loading Analytic Apps. These are available to install from our downloads site. Metadata loaders are configured with credentials to source systems and target Connect instance, and those parameters are stored in one or more YXWV files. They can be scheduled with a Connect Scheduler app that ships with the loader install.
• We support a series of third-party source systems as well, including MSSQL, Oracle, Tableau, Hive, S3, and Salesforce. The full list can be found here. An SDK is also available for you to build your own metadata loaders for sources beyond what is currently supported.
• One of the loaders is for the Alteryx Gallery. It calls the Gallery API to retrieve all Gallery workflows, parses their contents, and identifies inputs and outputs. These inputs form the foundation for workflow lineage views within the Connect site.
• Once metadata has been harvested from source systems, it is pushed to a REST API over http or https to Alteryx Connect. The metadata is staged on Connect, and reconciled with any existing entries; only the delta between a current and previous load is stored to avoid duplication.
• Data workers can view technical metadata, see relationships, have social interaction, and enrich content via the web interface.

Alteryx Connect is a centralized view into the world of data within your enterprise, unlocking tribal knowledge, and allowing data workers to become more efficient and effective!

[malibu0814]

Recently I upgraded Alteryx to 2019.1 and I am using Oracle 64-bit for my DB connections.  Since going to 2019.1, I am required to run my local connections to my Oracle DB using the alias connection process, that merely translates the connection to Oracle OCI, while moving to the Alteryx Server platform requires it to be defined as the Oracle OCI connection … without the Alias.

This means that I have to basically have two versions of my workflow with one for my local connections and another for the Server connections.  This is very cumbersome as to which workflow to run and if you forget one connection change the WF blows up and you fix it and have to start over.

Are there some options that need to be set to allow the same connection on either platform?  Before 2019.1 I did not have this issue at all.

Thanks,

Joe Sciarrino