Welcome to your new Function project!

This sample project contains a single function based on Spring Cloud Function: functions.CloudFunctionApplication.echo(), which returns an echo of the data passed via HTTP request.

Make sure that Java 17 SDK is installed since Spring Boot 3.0 requires Java 17.

To start server locally run ./mvnw spring-boot:run. The command starts http server and automatically watches for changes of source code. If source code changes the change will be propagated to running server. It also opens debugging port 5005 so a debugger can be attached if needed.

To run tests locally run ./mvnw test.

It's recommended to set FUNC_REGISTRY environment variable.

# replace ~/.bashrc by your shell rc file
# replace docker.io/johndoe with your registry
export FUNC_REGISTRY=docker.io/johndoe
echo "export FUNC_REGISTRY=docker.io/johndoe" >> ~/.bashrc

This command builds an OCI image for the function. By default, this will build a JVM image.

func build -v                  # build image

Note: If you want to enable the native build, you need to edit the func.yaml file and set the BP_NATIVE_IMAGE BuilderEnv variable to true:

buildEnvs:
  - name: BP_NATIVE_IMAGE
    value: "true"

Note: If you have issues with the Spring AOT processing in your build, you can turn this off by editing the func.yaml file and remove the -Pnative profile from the BP_MAVEN_BUILD_ARGUMENTS BuilderEnv variable:

buildEnvs:
  - name: BP_MAVEN_BUILD_ARGUMENTS
    value: -Pnative -Dmaven.test.skip=true --no-transfer-progress package

Removing the -Pnative profile means that you no longer will be able to build as a native image.

This command runs the func locally in a container using the image created above.

func run

This command will build and deploy the function into cluster.

func deploy -v # also triggers build

Building Spring Boot apps with Paketo Buildpacks on an ARM processor based system, like an Apple Macbook with an M1 or M2 chip, doesn't work well at the moment. To work around this you can build the image on your local system using Jib. Then, you would deploy the Jib generated image.

./mvnw compile com.google.cloud.tools:jib-maven-plugin:3.3.1:dockerBuild -Dimage=$FUNC_REGISTRY/echo
func deploy --build=false --image=$FUNC_REGISTRY/echo

For the examples below, please be sure to set the URL variable to the route of your function.

You get the route by following command.

func info

Note the value of Routes: from the output, set $URL to its value.

TIP:

If you use kn then you can set the url by:

# kn service describe <function name> and show route url
export URL=$(kn service describe $(basename $PWD) -ourl)

Using func invoke command with Path-Based routing:

func invoke --target "$URL/echo" --data "$(whoami)"

If your function class only contains one function, then you can leave out the target path:

func invoke --data "$(whoami)"

curl -v "$URL/echo" \
  -H "Content-Type:text/plain" \
  -w "\n" \
  -d "$(whoami)"

If your function class only contains one function, then you can leave out the target path:

curl -v "$URL" \
  -H "Content-Type:text/plain" \
  -w "\n" \
  -d "$(whoami)"

echo "$(whoami)" | http -v "$URL/echo"

If your function class only contains one function, then you can leave out the target path:

echo "$(whoami)" | http -v "$URL"

To clean the deployed function run:

func delete