Build efficient and scalable applications using azure functions

Editor's Notes

• #2: Hello friends and welcome to this webinar: Build efficient and scalable applications using Azure Functions.
• #3: But before we starts, let me introduce myself. My name is Mayur Tendulkar and I’m a Technology Solutions Professional at Microsoft in Singapore. Before this role, I was a Program manager and Developer Evangelist with Xamarin in India. I’ve been awarded as Microsoft Most Valuable Professional in different categories since 2013. I frequently write blog posts at mayurtendulkar.com and blog.xamarin.com i.e. official xamarin blog. I’m active on social media including Twitter and you can find my twitter handle on the screen. If you’ve any questions, you can ask those in the Q&A window or you can reach out to me on Twitter. So, let’s get started with this webinar: build efficient and scalable applications using Azure Functions.
• #4: Before we jump into the discussion of Azure Functions, lets talk about the evolution of application platforms. Let’s discuss how the application platform has evolved over the years.
• #5: When I started building applications it was all on-premise. And before even starting to write the code, we had to answer many questions. For example how many users will be hitting my servers? Accordingly how many servers I’ll need? Who is going to monitor, patch, update these servers and maintain the software installed on it. At the end of the day, we had to answer lot many questions related to technology rather than focusing on the business problems.
• #6: And then came Infrastructure as a Service. In this case we didn’t have to bother about buying the server or software. Many of the questions were answered. Yet, there were still few questions that needed to be answered like: what will be the size of my virtual machines, how many of those I’ll need and how the infrastructure will scale. In this model, still most the control and management was our headache.
• #7: With the advancement in cloud, now most of the part is managed by platform vendor. This is also called as Platform as a Service which further simplified our life. PaaS allows you to just worry about your app, and let the cloud vendor (in this case Microsoft Azure) take care of not only the physical hardware, but also the ongoing management of the operating environment like Operating Systems, Virtual Machines or Containers, as well as the various application runtimes, whether it is Java Runtime, .NET, Node.js, and so on. You are not completely free from thinking about servers however, since you still have to worry about how to scale your app and accordingly specify the number and size of the various compute resources your apps would need at various times.
• #8: The latest in this series of evolution is Serverless – which further abstracts the underlying infrastructure from you, and allows you to focus only on a single question – how are you going to build your app to achieve the best results for your business, or whatever it is that you are trying to do with your apps. Whether it is the code to delight your customers through a great experience, or the logic to communicate securely with a business partner, that is where most organization want to spend their creative energies on. And serverless allows you to do just that, focus on your code and forget about the infrastructure. This is why we believe it is the platform for next generation of apps.
• #9: So. What is serverless? Let’s talk about it.
• #10: The advantage being there is no infrastructure management. For example, teams will not need to care about sizing or how much servers the app will need to execute. Because resources are provisioned as and when required. That means the app will auto scale based on the workload. This further helps in reducing cost as there are no wasted resources and you pay only for the resources which are being used.
• #11: This gives us lot of benefits. First of all developers can focus on solving business problems and they don’t need to worry about the technology problems like how many servers are required, what should be the capacity etc. It helps them to increase their efficiency and helps to build apps faster taking lesser time to market And finally it gives the flexibility to easily change the functionalities if required and do experiments.
• #12: In this case Azure helps to build serverless applications by providing tools, framework and support for serverless computing Azure provides innovative programing model and services to integrate with. Developers can build the apps within Azure portal or use Visual Studio IDE family. Lastly, Azure also gives flexibility in terms of pricing. There is consumption based as well as dedicated plans available .
• #13: We provide not just a single service but a full platform that you can use to construct cloud native applications.  At the core are three main services: - Functions: Event-driven serverless compute experience - Event Grid: Intelligent event routing service that bring both 1st party and 3rd party events to serverless compute to act upon. - Logic Apps: Ability to create workflows and orchestrate processes in a rich designer experience. Along side these core services, we provide a host of other managed building block services – like Database (Cosmos DB), Storage, Cognitive Services, IoT services etc.  These services can be utilized by developers to build fully serverless applications which don’t require any infrastructure management
• #14: Now let’s talk about the use cases where serverless makes sense.
• #15: Serverless computing finds use in many scenarios.  This is not a comprehensive list but does list some of the key themes. Let’s talk about these scenarios
• #16: Serverless can be used in application backened. An example could be a web app adding incoming orders to a queue and then Functions automatically picking up those orders to process and adding the output to a Cosmos DB.
• #17: In case of a mobile app When app makes an API call to a Function, which does some processing and stores the data in Cosmos DB.  This in turn again triggers another function which calls Notification Hub to send notifications out to multiple mobile devices.
• #18: IoT is an important workload for serverless because of inherently spiky workloads associated with IoT.  Processing huge amount of incoming data and doing some processing on that data in real time is a common serverless scenario
• #19: Bot processing can be a perfect example for going serverless as the bot is short lived for that particular execution. So whenever customer checks for lets say room availability at hotels, the bot will execute the function on serverless platform and return the results.
• #20: Data transofrmation in real-time for use between different applications is a common scenario.
• #21: In real-time stream processing: -----
• #22: Automating tasks that need to happen at regular intervals can be easily achieved using Functions.  Whether it is for infra ops scenarios or just application architecture maintenance, Functions are a good fit.
• #23: And finally Azure Functions provides a very natural way of responding to webhook requests which makes it an excellent choice for extending any SaaS applications providing webhook extensibility. For example here: whenever issue is created on Github a webhook is triggers which is internally processed by Azure Function and issue details are posted on Slack or Teams channels where developers are focusing on.
• #25: These scenarios we just saw takes us to Functions As A Service which is at the center of serverless. This programming model emphasize on: Single responsibility i.e. functions have single purpose in their life and they just process the input and return the output. These functions are short lived i.e. they are not good candidates for continues background processing. They die as soon as the execution is over. This also means, that functions are stateless i.e. they cannot hold any data or state after execution is over. And finally functions are great because they can be executed depending on events or triggers and they are scalable.
• #26: Building on top of the FaaS programming model, Azure Functions keep all the mentioned features and extend your possibilities with additional capabilities that help to reduce your development time and boost productivity, while using best in class tools. Azure functions provides integrated programming model with built-in triggers and bindings. That means they execute whenever there is http request or a blob is added to the storage. They Have enhanced dev experience providing support for your favorite code editor. You can code/test and debug your apps locally. And finally you can deploy these functions as you want depending on your requirement.
• #27: Additional information Triggers and bindings – https://docs.microsoft.com/en-us/azure/azure-functions/functions-triggers-bindings Monitoring—https://azure.microsoft.com/en-us/services/application-insights/ Local debugging—https://docs.microsoft.com/en-us/azure/azure-functions/functions-run-local#run-functions-locally CI/CD—https://docs.microsoft.com/en-us/azure/azure-functions/functions-continuous-deployment Run locally—https://docs.microsoft.com/en-us/azure/azure-functions/functions-run-local Proxies—https://docs.microsoft.com/en-us/azure/azure-functions/functions-proxies
• #28: Additional information Languages—https://docs.microsoft.com/en-us/azure/azure-functions/supported-languages Dev options— Azure Functions portal—https://docs.microsoft.com/en-us/azure/azure-functions/functions-create-first-azure-function Visual Studio—https://docs.microsoft.com/en-us/azure/azure-functions/functions-develop-vs Visual Studio Code—https://code.visualstudio.com/docs CLI—https://docs.microsoft.com/en-us/azure/azure-functions/functions-create-first-azure-function-azure-cli Java/Maven—https://docs.microsoft.com/en-us/azure/azure-functions/functions-create-first-java-maven Hosting options—https://azure.microsoft.com/en-us/pricing/details/functions/ Durable Functions—https://docs.microsoft.com/en-us/azure/azure-functions/durable-functions-overview
• #30: Azure Functions spans in about 10 separate projects on GitHub About 60 contributors from the community across hundreds of commits All issues and sprints are triaged and tracked fully in the open, anyone can see exactly what code is running in the cloud and what is being worked on next More info in - https://github.com/Azure/Azure-Functions
• #33: As Azure Serverless platform makes it easy to write serverless applications using Azure Logic Apps and Azure Functions, it can be tricky to write scalable efficient apps. So, let me share some best practices with you. First of all, avoid long running functions. Try to write short and sweet functions with single responsibility this will allow you to avoid unnecessary timeouts to execute the function. Also when you’ve long running functions, there is a chance that it may have too many dependencies. Look for dependencies and try to minimize as many of them as possible. Because whenever the function starts for the 1st time, all those dependencies are downloaded and lot of time can be wasted to download these dependencies. Next, if you’ve multiple functions communicating with each other or you want to store state between function calls, prefer Logic Apps or Durable functions. These are specifically built for such scenarios. If it is not possible to use either of them, in that case use storage ques for communication. Also, functions are meant to be stateless and idempotent. If it is must to store the state, in that case store that state with the data itself. As it is advised, if you have function that needs to be executed once in a day, write it so that every time it executes, it gives same results. One of the programming rules applies here as well. Write your code defensively. That means consider that your function may throw exception anywhere in the code it should log the exception and continue from that fail point. Get the big picture of your architecture and write defensive code at each level to make sure function executes without any issues. And finally, choose the right hosting plan. There are two hosting plans available at this moment: consumption plan and dedicated plan. Consumption plan will execute your function only when required and resources will be allocated as and when required. It will be auto scaled. If you know the usage of your application, go ahead and choose dedicated plan as it will give you a performance increase because resources are pre-allocated for your function app. I’ve created this short link for you from where you can find all the best practices required which are required for writing scalable and efficient serverless apps using azure functions.
• #34: Before concluding this webinar, let me share some resources where you can find more information about Azure Serverless Platform. First of all, our Azure Serverless Computing page is a great resource with wealth of information about documentation, learning videos and events near you. Visit the link for more information. There is also a great video about Serverless platform from Build 2018 titled ‘Serverless in Action’, note the short bit link to watch the video. And finally, our friend Jeremy has written a fantastic ebook on Serverless app which is free and available on Microsoft documentation site. Visit the short link mentioned here to download this ebook or browse through it.
• #35: Finally, I would like to Thank you all for attending this webinar build efficient and scalable applications using Azure Functions. It’s time for Questions and Answers. Feel free to ask questions in the window.