deploy-cloud-functions

This action deploys your function source code to Cloud Functions and makes the URL available to later build steps via outputs.

Caution

This README corresponds to the "v3" GitHub Action, which is currently in beta. If you are using "v2", see the documentation for google-github-actions/deploy-cloud-functions@v2.

This is not an officially supported Google product, and it is not covered by a Google Cloud support contract. To report bugs or request features in a Google Cloud product, please contact Google Cloud support.

Prerequisites

• This action requires Google Cloud credentials that are authorized to access the secrets being requested. See Authorization for more information.

• This action runs using Node 20. If you are using self-hosted GitHub Actions runners, you must use a version of the GitHub Actions runner that supports Node 20 or higher.

Usage

jobs:
  job_id:
    runs-on: 'ubuntu-latest'
    permissions:
      contents: 'read'
      id-token: 'write'

    steps:
    - uses: 'actions/checkout@v4'

    - id: 'auth'
      uses: 'google-github-actions/auth@v2'
      with:
        project_id: 'my-project'
        workload_identity_provider: 'projects/123456789/locations/global/workloadIdentityPools/my-pool/providers/my-provider'

    - id: 'deploy'
      uses: 'google-github-actions/deploy-cloud-functions@v3'
      timeout-minutes: 10
      with:
        name: 'my-function'
        runtime: 'nodejs22'

    # Example of using the output
    - id: 'test'
      run: 'curl "${{ steps.deploy.outputs.url }}"'

Inputs

Important

In addition to these inputs, we highly recommend setting job and step-level timeouts, which can be used to control total deployment time.

• project_id: (Optional) ID of the Google Cloud project in which to deploy the service. The default value is computed from the environment.

• region: (Optional, default: us-central1) Region in which the function should be deployed.

• universe: (Optional, default: googleapis.com) The Google Cloud universe to use for constructing API endpoints. Trusted Partner Cloud and Google Distributed Hosted Cloud should set this to their universe address.

  You can also override individual API endpoints by setting the environment variable GHA_ENDPOINT_OVERRIDE_<endpoint> where <endpoint> is the API endpoint to override. For example:

  env:
    GHA_ENDPOINT_OVERRIDE_oauth2: 'https://oauth2.myapi.endpoint/v1'
  

  For more information about universes, see the Google Cloud documentation.

• name: (Required) Name of the Cloud Function.

• description: (Optional) Human-friendly description of the Cloud Function.

• environment: (Optional, default: GEN_2) Runtime environment for the Cloud Function. Allowed values are "GEN_1" and "GEN_2", but this GitHub Action only provides support for "GEN_2".

• kms_key_name: (Optional) Resource name of a Google Cloud KMS crypto key used to encrypt/decrypt function resources. If specified, you must also provide an artifact registry repository using the 'docker_repository' field that was created with the same key.

• labels: (Optional) List of labels that should be set on the function. These are comma-separated or newline-separated KEY=VALUE. Keys or values that contain separators must be escaped with a backslash (e.g. \, or \\n) unless quoted. Any leading or trailing whitespace is trimmed unless values are quoted.

  labels: |-
    labela=my-label
    labelb=my-other-label
  

  This value will only be set if the input is a non-empty value. If a non-empty value is given, the field values will be overwritten (not merged). To remove all values, set the value to the literal string {}.

  Google Cloud restricts the allowed values and length for labels. Please see the Google Cloud documentation for labels for more information.

• source_dir: (Optional, default: ./) Path on disk to the root of the the function's source code. Defaults to current directory. This does NOT follow symlinks to directories or files when generating the upload artifact.

  NOTE: The function source code must exist on the GitHub Actions filesystem. This means you must have use: actions/checkout@v4 before the deployment step!.

• runtime: (Required) Runtime for the function, such as "nodejs20". For a list of all available runtimes, run:

  $ gcloud functions runtimes list
  

  The available runtimes change over time.

• build_environment_variables: (Optional) List of environment variables that should be set in the build environment. These are comma-separated or newline-separated KEY=VALUE. Keys or values that contain separators must be escaped with a backslash (e.g. \, or \\n) unless quoted. Any leading or trailing whitespace is trimmed unless values are quoted.

  build_environment_variables: |-
    FRUIT=apple
    SENTENCE=" this will retain leading and trailing spaces "
  

  This value will only be set if the input is a non-empty value. If a non-empty value is given, the field values will be overwritten (not merged). To remove all values, set the value to the literal string {}.

  Previous versions of this GitHub Action also included a separate input for sourcing values from a value, but this is no longer supported. Use a community action or script to read the file in a separate step and import the contents as an output.

• build_service_account: (Optional) Service account to be used for building the container.

• build_worker_pool: (Optional) Name of the Cloud Build Custom Worker Pool that should be used to build the function. The format of this field is:

  projects/<project>/locations/<region>/workerPools/<workerPool>
  

  where <project> and <region> are the project id and region respectively where the worker pool is defined and <workerPool> is the short name of the worker pool.

  If the project ID is not the same as the function, then the Cloud Functions Service Agent must be granted the role Cloud Build Custom Workers Builder in the project.

• docker_repository: (Optional) Repository in Artifact Registry to which the function docker image will be pushed after it is built by Cloud Build. If unspecified, Cloud Functions will create and use a repository named 'gcf-artifacts' for every deployed region.

  The value must match the pattern:

  projects/<project>/locations/<location>/repositories/<repository>.
  

  Cross-project repositories are not supported. Cross-location repositories are not supported. Repository format must be 'DOCKER'.

• entry_point: (Optional) Name of a Google Cloud Function (as defined in source code) that will be executed. Defaults to the resource name suffix (ID of the function), if not specified.

• all_traffic_on_latest_revision: (Optional, default: true) If true, the latest function revision will be served all traffic.

• available_cpu: (Optional) The number of available CPUs to set (e.g. 0.5, 2, 2000m). By default, a new function's available CPUs is determined based on its memory value.

• available_memory: (Optional) The amount of memory available for the function to use. Allowed values are of the format: with allowed units of "k", "M", "G", "Ki", "Mi", "Gi" (e.g 128M, 10Mb, 1024Gib).

  For all generations, the default value is 256MB of memory.

• environment_variables: (Optional) List of environment variables that should be set in the runtime environment. These are comma-separated or newline-separated KEY=VALUE. Keys or values that contain separators must be escaped with a backslash (e.g. \, or \\n) unless quoted. Any leading or trailing whitespace is trimmed unless values are quoted.

  environment_variables: |-
    FRUIT=apple
    SENTENCE=" this will retain leading and trailing spaces "
  

  This value will only be set if the input is a non-empty value. If a non-empty value is given, the field values will be overwritten (not merged). To remove all values, set the value to the literal string {}.

  Previous versions of this GitHub Action also included a separate input for sourcing values from a value, but this is no longer supported. Use a community action or script to read the file in a separate step and import the contents as an output.

• ingress_settings: (Optional, default: ALLOW_ALL) Ingress settings controls what traffic can reach the function. Valid values are "ALLOW_ALL", "ALLOW_INTERNAL_ONLY", and "ALLOW_INTERNAL_AND_GCLB".

• max_instance_count: (Optional) Sets the maximum number of instances for the function. A function execution that would exceed max-instances times out.

• max_instance_request_concurrency: (Optional) Sets the maximum number of concurrent requests allowed per container instance.

• min_instance_count: (Optional) Sets the minimum number of instances for the function. This is helpful for reducing cold start times.

• secrets: (Optional) List of KEY=VALUE pairs to use as secrets. These are comma-separated or newline-separated KEY=VALUE. Keys or values that contain separators must be escaped with a backslash (e.g. \, or \\n) unless quoted. Any leading or trailing whitespace is trimmed unless values are quoted.

  These can either be injected as environment variables or mounted as volumes. Keys starting with a forward slash '/' are mount paths. All other keys correspond to environment variables:

  with:
    secrets: |-
      # As an environment variable:
      KEY1=secret-key-1:latest
  
      # As a volume mount:
      /secrets/api/key=secret-key-2:latest
  

  This value will only be set if the input is a non-empty value. If a non-empty value is given, the field values will be overwritten (not merged). To remove all values, set the value to the literal string {}.

• service_account: (Optional) The email address of the IAM service account associated with the Cloud Run service for the function. The service account represents the identity of the running function, and determines what permissions the function has. If not provided, the function will use the project's default service account for Compute Engine.

  Note this differs from the service account used to deploy the Cloud Function, which is the currently-authenticated principal. However, the deploying service account must have permission to impersonate the runtime service account, which can be achieved by granting the deployment service account "roles/iam.serviceAccountUser" permission on the runtime service account.

• service_timeout: (Optional, default: 60s) The function execution timeout, specified as a time duration (e.g. "30s" for 30 seconds).

• vpc_connector: (Optional) ID of the connector or fully qualified identifier for the connector.

• vpc_connector_egress_settings: (Optional, default: PRIVATE_RANGES_ONLY) Egress settings controls what traffic is diverted through the VPC Access Connector resource. Allowed values are "PRIVATE_RANGES_ONLY" and "ALL_TRAFFIC".

• event_trigger_location: (Optional) The location of the trigger, which must be a region or multi-region where the relevant events originate.

• event_trigger_type: (Optional) Specifies which action should trigger the function. For a list of acceptable values, run:

  $ gcloud functions event-types list
  

  This usually requires the eventarc API to be enabled:

  $ gcloud services enable eventarc.googleapis.com
  

  The available trigger types may change over time.

• event_trigger_pubsub_topic: (Optional) Name of Google Cloud Pub/Sub topic. Every message published in this topic will trigger function execution with message contents passed as input data of the format:

  projects/<project_id>/topics/<topic_id>
  

  The service account must have permissions on this topic.

• event_trigger_service_account: (Optional) The email address of the IAM service account associated with the Eventarc trigger for the function. This is used for authenticated invocation.

• event_trigger_retry: (Optional, default: true) Describes whether event triggers should retry in case of function's execution failure.

• event_trigger_channel: (Optional) The name of the channel associated with the trigger in the format:

  projects/<project>/locations/<location>/channels/<channel>
  

  You must provide a channel to receive events from Eventarc SaaS partners.

Allowing unauthenticated requests

The Cloud Functions product recommendation is that CI/CD systems not set or change settings for allowing unauthenticated invocations. New deployments are automatically private services, while deploying a revision of a public (unauthenticated) service will preserve the IAM setting of public (unauthenticated). For more information, see Controlling access on an individual service.

Outputs

• name: Full resource name of the Cloud Function, of the format:

  projects/<project>/locations/<location>/functions/<function>
  

• url: The URL of your Cloud Function.

Authorization

The deployment service account must have the following IAM permissions:

• Cloud Functions Deployer (roles/cloudfunctions.deployer)

Additionally, the deployment service account must have permissions to act as (impersonate) the runtime service account, which can be achieved by granting the deployment service account "roles/iam.serviceAccountUser" permissions on the runtime service account. See the Google Cloud documentation to learn more about custom runtime service accounts and additional configuration for deployment

Via google-github-actions/auth

Use google-github-actions/auth to authenticate the action. You can use Workload Identity Federation or traditional Service Account Key JSON authentication.

Authenticating via Workload Identity Federation

jobs:
  job_id:
    permissions:
      contents: 'read'
      id-token: 'write'

    steps:
    - uses: 'actions/checkout@v4'

    - id: 'auth'
      uses: 'google-github-actions/auth@v2'
      with:
        project_id: 'my-project'
        workload_identity_provider: 'projects/123456789/locations/global/workloadIdentityPools/my-pool/providers/my-provider'

    - id: 'deploy'
      uses: 'google-github-actions/deploy-cloud-functions@v3'
      timeout-minutes: 10
      with:
        name: 'my-function'
        runtime: 'nodejs22'

Via Application Default Credentials

If you are hosting your own runners, and those runners are on Google Cloud, you can leverage the Application Default Credentials of the instance. This will authenticate requests as the service account attached to the instance. This only works using a custom runner hosted on GCP.

jobs:
  job_id:
    steps:
    - uses: 'actions/checkout@v4'

    - id: 'deploy'
      uses: 'google-github-actions/deploy-cloud-functions@v3'
      timeout-minutes: 10
      with:
        name: 'my-function'
        runtime: 'nodejs22'

The action will automatically detect and use the Application Default Credentials.