Functions often need password, API keys, and connection strings to connect to other services and retrieve data. It would be great if those secrets could easily be obtained from Key Vault.365 votes
We’re pleased to announce a public preview of our Key Vault references feature, which you can learn more about here: https://azure.microsoft.com/en-us/blog/simplifying-security-for-serverless-and-web-apps-with-azure-functions-and-app-service/
There are some limitations to the initial preview, but we’re hoping to address those very soon. We’re looking forward to your feedback!
Matthew, Azure Functions team
It would be useful to have a form of DI in our C# Functions, so we can consistently inject things like our logger class, or other Autofac DI services, into every C# Function we create.
I'd like to create a SDK package which developers can import into their Function and get access to some common functionality via DI, to remove boilerplate redundancy.
Any solution for this kind of thing?296 votes
We have started work to enable this. We discuss the approach and implementation some during our webcast last month at about minute 25 here: http://aka.ms/azurefunctionslive
Please support writing TypeScript in Azure Functions when you're creating Node functions84 votes
Stay tuned! In the meantime, check out our types file:
Currently you can only configure Azure Alerts on the App Plan that is hosting your Azure Function. However, it would be great to do this on a Function level to detect failures and other metrics.17 votes
We have been working on this, and are a few months out from enabling it.
Azure Functions team
This is partially supported today. You can click through to the App Service Settings blade and click on the graph that shows up there.
This lets you set up basic alerts on failure rates/etc or GB-sec thorughput.
We’re adding Application Insights integration soon and that will provide even more options for setting up more granular alerts: github.com/Azure/azure-webjobs-sdk-script/i..
-Chris, Functions PM
Please support python based Azure functions that run in a Linux Docker container.
Linux Docker containers are already supported so we are half of the way there and Python is working/supported on windows preview.
Currently if you try to create a simple python based Azure function on Linux preview you get the error: HttpTriggerPy1: Object reference not set to an instance of an object.
The use case here is being able to run "enterprise" python 2 environments with hard dependencies on Linux. I can verify python runs correctly and can send an example Dockerfile if needed.14 votes
The feature is currently in active development / limited preview. We invite you to try the early bits and share your feedback with us. Please follow the links below for more information -
I went to set up Azure functions for a new client and had to delete and re-create my application so I could downgrade to runtime 1.0 in order to use a powershell function. Even AWS has realized the proliferation of Powershell and now allows it to be used in Lambda functions. I have probably thousands of administrative scripts in my archives that could easily be ported to Azure functions. I'd move to AWS before I'd invest the time to re-write them all.10 votes
We’re now working on PowerShell 6 support in Functions 2.0! Please tune into the repo here – https://github.com/Azure/azure-functions-powershell-worker for more details.
Server-less Functions are great for quick and dirty development and testing. Please make this available (for testing) to DreamSpark Imagine Student license.6 votes
This work has been started. Get your pencils sharpened and your notes ready!
Azure Functions Team
Function App's PowerShell is currently v4.
But i want to use PowerShell v5 in function app.
PowerShell v5 have so nice features and i use v5 in my local machine.
So i want to run a same script in local to function app.
Please update function app's powershell.2 votes
We’re now working on PowerShell 6 support in Functions 2.0! Please tune into the repo here – github.com/Azure/azure-functions-powershell.. for more details.
It's my understanding functions execute on an A1 size virtual machine, which has 1 core and 2GB memory. The customer has no way of changing this.
Our processing requires a bit more beefy machine as it's CPU and memory intensive. We understand using a dedicated App Service Plan is an option, but we really like paying for only what we consume. Our jobs run infrequently and tend to be on the "longer" end of the run time scale.
I'd love to be able to select the VM size we run on and pay accordingly with a consumption plan.1 vote
We’re planning a new offering that allows fast scaling based on events, while you can select the VMs where your functions will run. This will allow for further networking options to connect to on-prem resources.
- Don't see your idea?