Microsoft 365 Community Community Demos, 18 Aug 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Performance Insights - New Power Apps self-service tool for developers

Description

In this 10-minute citizen developer-focused demo, Varda Shrivastava shows a new Power Apps self-service tool for enterprise app makers that analyzes runtime user data and delivers a prioritized list (daily) of recommendations to make Model-driven or customer engagement Power Apps perform more efficiently. Evaluates usage patterns, forms performance, customizations, configurations, environment, and networking. Suggests mitigation and provides links to references. Available/Visible (Public Preview) in apps now. This PnP Community demo is taken from the Microsoft Power Apps community call – August 2021, recorded on August 18, 2021.

Demo Presenter: Varda Shrivastava (Microsoft)

Supporting documents:
• Blog - Introducing Performance Insights (preview) for Model-driven apps | https://powerapps.microsoft.com/blog/introducing-performance-insights-preview-for-model-driven-apps/
• Demo - Performance Insights For Power Apps - Power CAT Live | https://www.youtube.com/watch?v=-RWzvTZzMfA
• Feedback - Feedback thread for Performance Insights (preview) | https://powerusers.microsoft.com/t5/Community-Feedback/Feedback-thread-for-Performance-Insights-preview/m-p/1193936

Stay connected:
• Twitter https://twitter.com/microsoft365dev
• YouTube https://aka.ms/m365pnp/videos
• Blogs https://aka.ms/m365pnp/community/blog
• Call attention to your great work by using #PowerAppsCC and #PnPWeekly on Twitter.

A

Hey everyone, I'm really excited to be joining you today to introduce a new feature that we've added to really empower our makers to build performance model driven apps, so I'll be talking about performance insights, which is a new self-service tool that is now available in public preview, on the powerapps maker portal and before diving in and you know, showing the cool demo. I wanted to provide some context as to why we built this feature and really give some sense of the problem we were trying to solve.

A

So, as you know, better performance continues to be a top ask from powerapps users and while we as a product team, are investing in proof improvements to the platform. Certain performance issues, such as those caused by customizations, need to be mitigated by our makers.

A

While we have tools like solution, checker that do static code analysis, we found that people aren't necessarily aware of the runtime performance issues of their apps. So this is when the end users are facing issues while using their apps, and this is where performance insights comes in. We want to empower you to build faster apps that your end users love with this new feature. We analyze runtime and usage telemetry data to provide you with a daily set of insights and recommendations related to the performance of your power apps.

A

This includes standalone model driven apps or customer engagement. Apps such as dynamics, sales, marketing or customer service, and more and this list of recommendations is prioritized by severity to help you turn these insights into action and to get better performance to go over some key points of this new feature. Performance insights generates recommendations based on usage telemetry that has been collected over a 24-hour period, and these insights are available for up to seven days.

A

A really important point is that if your app doesn't have enough usage, then you won't have insights or recommendations since it's based on the telemetry data, so the recommendation for from our site would be to use performance insights with apps that have usage and in a environment that has users such as a production environment.

A

Also, these insights are sorted by severity, from critical being the most important ones that will have the highest performance gains to warning and then information.

A

The severity of each category varies by how it's determined, but factors such as performance of the page load time and the number of end users impacted by the event influence what the severity would be. These insights are across several categories from different areas like customizations, page performance, form, performance network and latency, and more and I'll go over some of these in the next slide.

A

But really our goal for this feature is to be a platform where several first party teams can now add their own recommendations and insights to help you gain better performance and have a richer set of insights. So these are some of the areas and categories that we have today.

A

Things like client environment, which will tell you about the minimum system requirements or the browser type. The browser version that your users are on that are affecting your performance or how different customizations are like xhr, request, type or different plugins, that you're using are impacting the overall performance of your model driven app and before I dive into the demo. Here's the quick summary of how you can use this feature the front door for this feature is going to be in the maker portal to make our power apps.

A

When you go to the left navigation pane, you can select apps and then for any model driven app. If you click on the context, menu of the command bar performance should show up in the preview, and if there aren't any questions I'll hop in the the demo uh fork, uh is there going to be a version with of canvas apps? uh For now the focus is on model driven apps, um the roadmap for canvas apps. Isn't there isn't a clear timeline for that?

A

Yet, okay, thanks awesome, all right, so I have uh my makeup: our apps open and I've selected, an environment which does have some usage, I'm on the apps apples side and I'm going to go ahead and select an app that I want to view in size. For so, as you can see, it's in the command bar or in the context, menu you're going to see a performance preview label and when I click on that it takes me to this grid, which has several categories listed that are ranked by severity.

A

So the way I would navigate this is if I want to deep dive into any specific one.

A

You click on that and then the details pane is populated with different information such as the severity, the area, the category, but also the motivation of that specific category. The insight and the recommendation as to what you can do to mitigate this issue.

A

We also link to our documentation that, in some cases, provides uh deeper details into how you can mitigate the issue, as well as providing certain data that can help supplement it. So in in some categories, uh we also provide additional data that you can download to to get a deep dive more into the category, and this is the date picker, where you can pick a date up to seven previous days for this app.

A

Since I wasn't usage uh in the last few days, the the most recently available data that I bought was from 13 but yeah. This is the feature and I'm happy to answer any questions.

B

I have a question right off the bat, um so how much? How much telemetry do you usually recommend you need to gather before you really see results in there? I know you said you had to have a minimum amount for it to work, but it do you find that within one week or you know two weeks of usage or you need a month to shake it all out.

A

So using an app.

B

How does that usually go in flow.

A

Yeah, so the cool thing about this is that it's looking at daily, it's going to look at the telemetry data over a 24 hour period.

A

So what that really entails is that we're gonna analyze telemetry data, that's collected over 24 hour period in the utc time zone. So, depending on when you view it, it's going to be a different date. So if I'm, for example, if I open it up today on august 18th, the latest data that I would have for my app could be collected over august 17th in the utc time zone. So it's over a day period, but within that, what we recommend is that there was usage on that specific date.

A

So it's not really aggregating information over a month or data over a week, but really for that daily time period um and the recommendation to get to use apps that have users just more from you know which environment you're using it from as well like if you use it from a test environment versus from from a production environment.

B

That makes sense. I I see a couple questions here in chat. um One person is asking: can performance preview be exported to power bi using power, automation.

A

Yeah, we don't have that capability right now, but that is a good suggestion that I will note down.

B

Right on um and right now this only works for model driven apps, so I'm sure everybody is hasn't, asked it yet. But uh do you have plans for campus apps for this too.

A

Yeah, so really for now, the focus has been on monitoring apps, because we want to empower our enterprise level, app makers to solve this problem in model driven apps. That's what we we're finding has the greatest performance issues. um I don't want to comment on the timeline yet because we haven't necessarily committed to a timeline for canvas apps yet, but for now the focus is going to be on model driven apps and with conversion. Apps coming out is what that's going to be. Our next focus too.

B

Oh the best of both worlds that way yeah.

A

That's where the future is yep.

B

Super and so um for this feature right now, is there anything special anybody needs to do to turn it on, or is it just going to be there in every tenant in every environment.

A

Yeah, no, the really cool thing is that it's already available for all your model driven apps, there's no special need to turn on since it's in public preview, it should be available. The main caveat is that your app needs to have users. So if, if there isn't usage, then these insights wouldn't be generated, so you wouldn't have access to the grid.

B

That makes sense cool. Let's see, if there's any other questions in here, a lot of people are saying they think this would fit great into the center of excellence as a module when they're uh yeah.

B

I think I think that's it for the open questions that I've seen in there. I'm sure if somebody else has one they'll, they'll pop it in there and that's great, I can't wait to go. Try this out. I I know there's a couple of model driven apps that we are developing right now that I want to go run this scan on just to see if there's room for improvement.

A

Yeah, we would love for everyone to try out this feature and- and I think dorad in some of the slides we've linked a forum as well, that you can leave your feedback on, but always feel free to reach out to me, as well with any feedback that you have and looking forward to questions and feedback from folks.

B

Fantastic awesome. Thank you.

A

Very much thank you.