A new version of the Microsoft whitepaper “Planning a Power BI Enterprise Deployment” is now available. Once again, Melissa Coates (b|t) and Chris Webb (b|t) are the authors. I was lucky enough to be the tech editor again on this version, so I’m excited to see the new information be released to the public.

There were quite a few updates this time. Here are some of the highlights:

• Section 3, “Power BI Architectural Choices”, has updated information on dataflows and Power BI Premium. It also includes a nice section clarifying the options available for embedding Power BI content.
• Section 4, “Power BI Licensing and User Management”, has been updating to include information on self-service purchasing.
• Section 5, “Power BI Source Data Considerations” now includes information on dataflows.
• Section 6, “Power BI Dataset Storage Options” now contains information about Automatic Page Refresh and large models.
• Section 7, “Power BI Data Refresh and Data Gateway” now mentions the Power Platform Admin Center. It also discusses dataflow refreshes in addition to dataset refreshes. And more information has been added regarding the use of gateway clusters for load balancing and high availability.
• Section 8, “Power BI Dataset and Report Development Considerations” contains new information on shared datasets and .pbids (Power BI Data Source) files. It also has a new section providing guidance on information design and accessibility. And it provides updated information on the use of custom visuals.
• Section 9, “Power BI Collaboration, Sharing and Distribution”, has been updated to reflect the new workspace experience. It also discusses shared and certified datasets and the new deployment pipelines. It also contains a nice decision tree to help you determine whether to use apps, workspaces, or sharing.
• Section 10, “Power BI Administration”, has new recommendations for tenant settings. It also discusses protection metrics, custom help menus, custom branding as well as providing new information on managing workspaces and dataflows. And it discusses the new activity log and related PowerShell modules.
• Section 11, “Power BI Security and Data Protection”, now discusses the roles in the new workspace experience as well as sensitivity labels and Microsoft Information Protection.
• An updated list of deprecated items can be found in section 12, “Power BI Deprecated Items”.
• Section 13, “Support, Learning, and Third-Party Tools” contains a great list of helpful resources for the Power BI practitioner.

I hope you’ll take a glance through the updated whitepaper and catch up on all the new information. Happy reading!

Next week I’m speaking at at the Dynamic Communities Power Up event titled “Exploring the Power BI Ecosystem“. It takes place on May 27 & 28, 2020. This exciting 2-Day virtual event is designed to ensure attendees have a complete view of the Power BI product and surrounding ecosystem, provide expanded knowledge of the core components and showcase the possibilities for continued exploration and innovation.

Sessions during the event are 2.5 hours long, to really give you time to get into a topic. There are healthy 45-minute breaks between sessions to give you time to attend to personal matters. And the sessions are recorded to give you a chance to catch anything you miss. Some sessions, including mine, offer a take-home exercise to help solidify concepts discussed during the session.

I’m presenting Data Visualization and Storytelling on May 28 at 9am EST/1pm UTC. In this session, you will learn how to build eye-catching Power BI reports to support decision making. You will also see the importance and a realistic approach to data storytelling.

The following topics will be showcased through practical examples:

• Creating beautiful reports: prioritizing your KPIs, playing with colors, grid
• Choosing the best chart to illustrate your point
• Introduction to the concept of Data Storytelling
• Implementing quality checks on your report design
• Implementing navigation in your report: bookmarks, drill-through, page-report tooltips, interactive Q&A

This training is a paid event, but it’s just $399 for the full 2 days. This training is great if you are a beginner-to-intermediate Power BI user trying to round out your skills across the many areas of the Power BI suite. You can head over to the website to register. I hope to see you there!

I had the privilege of working with Tessa Hurr (PM on the Power BI team) on a presentation for the 2020 Microsoft Business Applications Summit (MBAS) about five features in Power BI that increase report accessibility. This 23-minute presentation is almost entirely demos, and only a few slides. While we talk about some features such as alt text and tab order that are primarily used for accessibility purposes, we also talk about how chart titles, header tooltips, and report themes can be used to make your report more accessible.

The conference was entirely online this year, and you can catch the sessions on demand now. I hope you’ll take some time to watch my session as well as the other great content that came from the conference. You can watch my session on the MBAS website.

Data Factory allows parameterization in many parts of our solutions. We can parameterize things such as connection information in linked services as well as blob storage containers and files in datasets. We can also parameterize certain properties in activities. For instance, we can write an expression to determine the stored procedure to be executed in a Stored Procedure Activity or the filename in the sink (destination) of a Copy Activity.

But we cannot parameterize the invoked pipeline in an Execute Pipeline Activity. This means we need to find workarounds in order to have a metadata-driven execution framework. What I mean by metadata-driven execution framework is that data is stored in a data store (in my case, a SQL Database) and used to determine what pipelines and activities get executed. With this type of framework, if I don’t want a specific pipeline to execute, I would just update my data in the datastore rather than delete the pipeline execution from the parent pipeline. We’ve been doing this type of development in SSIS for years, and Biml has played a big part in that. But SSIS allows us to parameterize the Execute Package Task.

Since we can’t implement this parameterized execution of pipelines natively, we need to look for something that Data Factory can call to accomplish the task. Paul Andrew has a nice framework that uses Azure Functions. I was working on a Data Factory solution for a client who doesn’t have C# or PowerShell developers on hand to help with the ELT process, so we needed to explore a low-code solution.

While there is no Logic App activity in Data Factory we can use a Web Activity to call the Logic App. I might have a pipeline that looks something like what is pictured below.

Within the ForEach loop is a single Web Activity.

I used some variables and parameters in an expression to populate the URL so it would be dynamic. I used a GET method in the call.

My initial version of my Logic App is shown below.

I added path parameters in my HTTP request trigger to allow me to capture the information I need to execute the appropriate pipeline. For me this included the pipeline name, a data source ID, and a country. Your parameters would vary according to your requirements.

Logic apps has an action called “Create a pipeline run”. You tell it which data factory, which pipeline, and any parameter values needed for the pipeline execution.

At this point in the workflow, our pipeline would be executing. But now we need to know when it has finished. That’s what the Initialize Variable and Until Loop actions are handling. I created a string variable called Pipeline Status and set the default value to “InProgress”. My Until loop action checks my pipeline execution status. If it’s still running, it waits 5 seconds, gets the new status, and assigns that status to the variable. This repeats until the pipeline execution is no longer in progress.

Here’s the expression I used to check whether the pipeline execution is still running:

@and(not(equals(variables('PipelineStatus'), 'InProgress')),not(equals(variables('PipelineStatus'), 'Queued')))

Once the pipeline execution is complete, an HTTP response with the pipeline status is sent back to the caller.

This is all great until you find out that Logic Apps will experience an HTTP timeout if the request takes more than 2 minutes.

Do you have any pipelines that take longer than two minutes to execute? If so, you need to change your solution to handle this. Note that you would have the same issue with Azure Functions, although it would give you 230 seconds instead of 120 seconds before it timed out. We need to switch to an asynchronous call to support long running pipelines. Paul has already done this in his framework using Azure Functions. In Logic Apps, we can change our response to an asynchronous response and then implement a polling pattern to check the status. We could alternatively implement a webhook action. I’ll write about updating the solution to handle long running pipelines in a future post.