Last year on November 10, 2020; .Net 5 was released. And Azure Functions, one of the biggest features for Azure, was left out of the release. It wasn’t until March 10, 2021, that a new version was released to support .Net 5.
But with this new release, we found a new approach to the execution model. Here is where we now can see that Azure Functions are executed in an Isolated Process. But what is about this Isolated Process execution?
Since their inception the .Net and .Net Core function applications have run in the same host process, bringing with them unique advantages such as having a set of bindings and injections to the SDK. But on the other hand, the execution of functions for other languages is executed in an “out-of-process” model (outside the process) that allows executing the functionality in a separate process. That is, under the execution of a process that in turn is in charge of the execution of the functions.
And this is where the Azure Functions development and support team decides to adopt this model for .Net 5 support. Thus, if we want to work on this version, we will find a new way of working.
The first change that we are going to notice is that when we create a new project, let’s say in Visual Studio, we will find the presence of the Program.cs class. Which is the entry point of the execution of the functions, that is, the process in charge of the execution.
Previously in Azure Functions for .Net Core 3.1, we could create a class that inherits from FunctionsStartup so that it is executed on launch. But in this aproach, we already have this functionality already in our projects.
Another change that we are going to find is that the base Nuget packages that we are going to use will be,
and the link extension packages will be,
Where * represents the trigger package (s) that we will be using in our functions. For instance,
Here is a table with the changes or differences with Azure Functions for .Net 3.1