Azure Functions Core Tools
The Azure Functions Core Tools provide a local development experience for creating, developing, testing, running, and debugging Azure Functions.
v1 (v1.x branch): Requires .NET 4.7.1 Windows Only
v2 (master branch): Self-contained cross-platform package
Both v1 and v2 of the runtime can be installed on Windows.
To install v1 with npm:
npm i -g azure-functions-core-tools@1
To install v1 with chocolatey:
choco install azure-functions-core-tools --version 1.0.15
To install v2 with npm:
npm i -g azure-functions-core-tools --unsafe-perm true
To install v2 with chocolatey:
choco install azure-functions-core-tools
brew tap azure/functionsbrew install azure-functions-core-tools
- Set up package feed
wget -q https://packages.microsoft.com/config/ubuntu/18.04/packages-microsoft-prod.debsudo dpkg -i packages-microsoft-prod.deb
wget -q https://packages.microsoft.com/config/ubuntu/17.10/packages-microsoft-prod.debsudo dpkg -i packages-microsoft-prod.deb
Ubuntu 16.04 / Linux Mint 18
wget -q https://packages.microsoft.com/config/ubuntu/16.04/packages-microsoft-prod.debsudo dpkg -i packages-microsoft-prod.deb
sudo apt-get updatesudo apt-get install azure-functions-core-tools
Other Linux Distributions
- Download latest release
Download the latest release for your platform from here.
- Unzip release zip
Using your preferred tool, unzip the downloaded release. To unzip into an
azure-functions-cli directory using the
unzip tool, run this command from the directory containing the downloaded release zip:
unzip -d azure-functions-cli Azure.Functions.Cli.linux-x64.*.zip
- Make the
Zip files do not maintain the executable bit on binaries. So, you'll need to make the
func binary executable. Assuming you used the instructions above to unzip:
cd azure-functions-clichmod +x func./func
- Optionally add
To execute the
func command without specifying the full path to the binary, add its directory to your
$PATH environment variable. Assuming you're still following along from above:
NOTE: npm can be used on all platforms. On unix platforms, you may need to specify
--unsafe-perm if you are running npm with sudo. That's due to npm behavior of post install script.
NOTE: If you're running the v2 on Windows, Linux, or Mac, make sure to enable the
beta runtime in function app settings, otherwise you may not see the same results as running locally.
Getting Started on Kubernetes
Using the Core Tools, you can easily run Azure Functions on 1.7+ Kubernetes clusters. The Core Tools will build and push a Docker image of the function to a given registry and create corresponding Kubernetes objects including a Deployment, Service and Horizontal Pod Autoscaler.
First, make sure you init a Docker file.
func init --docker
Deploy a function to Kubernetes
func deploy --platform kubernetes --name myfunction --registry <docker-hub-id or registry-server>
Deploy a function with a minimum of 3 instances and a maximum of 10
func deploy --platform kubernetes --name myfunction --registry <docker-hub-id or registry-server> --min 3 --max 10
Scaling out Http Trigger
Currently the solution is configured to scale out using the Horizontal Pod Autoscaler when any pod reaches a CPU of 60%.
Get function logs
func logs --name myfunction --platform kubernetes
Provide a kubeconfig file
func deploy --platform kubernetes --name myfunction --registry <docker-hub-id or registry-server> --config /mypath/config
Deploy a function to Knative
Deploying Azure Functions to knative is supported with the
--platform knative flag.
The Core Tools CLI identifies non HTTP trigger functions and annotates the knative manifest with the the
minScale annotation to opt out of scale-to-zero.
func deploy --platform knative --name myfunction --registry <docker-hub-id or registry-server>
Deploying a function to AKS using ACR
Using the configuration options an Azure Function app can also be deployed to a AKS (Azure Kubernetes Service) Kubernetes cluster and use ACR as the registry server. Do all of the following before you run the deployment command.
Create a AKS cluster
Once your AKS cluster is created make sure that you can access it using kubectl. To make kubectl run in the context of your cluster, configure a connection using the command below.
az aks get-credentials \--name FunctionsCluster \--resource-group <resource-group-name>
To verify the connection to your cluster run the following command
> kubectl get nodesNAME STATUS ROLES AGE VERSIONaks-agentpool-20257154-0 Ready agent 1d v1.11.5aks-agentpool-20257154-1 Ready agent 1d v1.11.5aks-agentpool-20257154-2 Ready agent 1d v1.11.5
Create a ACR Registry
An ACR instance can be created using the Azure Portal or the Azure CLI
Login to the ACR Registry
Before pushing and pulling container images, you must log in to the ACR instance.
az acr login --name <acrName>
Give the AKS cluster access to the ACR Registry
The AKS cluster needs access to the ACR Registry to pull the container. Azure creates a service principal to support cluster operability with other Azure resources. This can be used for authentication with an ACR registry. See here for how to grant the right access here: Authenticate with Azure Container Registry from Azure Kubernetes Service
Run the deployment
The deployment will build the docker container and upload the container image to your referenced ACR instance (Note: Specify the ACR Login Server in the --registry parameter this is usually of the form <container_registry_name>.azurecr.io) and then your AKS cluster will use that as a source to obtain the container and deploy it.
func deploy --platform kubernetes --name myfunction --registry <acr-registry-loginserver>
If the deployment is successful, you should see this:
Function deployed successfully! Function IP: 22.214.171.124
Verifying your deployment
You can verify your deployment by using the Kubernetes web dashboard. To start the Kubernetes dashboard, use the az aks browse command.
az aks browse --resource-group myResourceGroup --name myAKSCluster
In the Kubernetes dashboard look for the namespace "azure-functions" and make sure that a pod has been deployed sucessfully with your container.
Deploying Azure Functions with Virtual-Kubelet
Azure Functions running on Kubernetes can take advantage of true serverless containers model by getting deployed to different providers of Virtual Kubelet, such as Azure Container Instances.
Functions deployed to Kubernetes already contain all the tolerations needed to be schedulable to Virtual Kubelet nodes. All you need to do is to set up VKubelet on your Kubernetes cluster:
Important note: Virtual Kubelet does not currently allow for Kubernetes Services to route external traffic to pods. This means that HTTP triggered functions will not receive traffic running on a VKubelet provider (including ACI).
A good usage scenario for using functions with VKubelet would be with event triggered / time triggered functions that do not rely on external HTTP traffic.
func extensions command require the
dotnet cli to be installed and on your path. This requirement is tracked here. You can install .NET Core for your platform from https://www.microsoft.com/net/download/
CurrentDirectory: is the default directory the functions runtime looks for functions in.
%TMP%\LogFiles\Application\Functions: is the default directory for logs. It mirrors the logs directory on Azure as well.
For questions on Azure Functions or the tools, you can ask questions here:
File bugs at Azure Functions Core Tools repo on GitHub.