If you are creating a new function, see the Console Quickstart on Cloud Run.

Configure Direct VPC egress for 2nd gen functions

Direct VPC egress lets you route traffic from your Cloud Run functions (2nd gen) function directly to your VPC network.

Limitations

Direct VPC egress is not available for 1st gen functions.
You can't use Direct VPC egress and Serverless VPC Access connectors at the same time. See Compare Direct VPC egress and VPC connectors for details.

Before you begin

Enable the Cloud Functions API.
Install the Google Cloud CLI, then initialize it by running gcloud init .
Update gcloud components to version 545.0.0 or later:
```
gcloud components update
```
If you don't already have a VPC network in your project, create one .
Optional: If your function needs to access Google APIs and services using their internal IP addresses, enable Private Google Access on the subnet that you use for Direct VPC egress.

Set up IAM permissions

To authorize Direct VPC egress, ask your administrator to grant the Cloud Run Invoker ( roles/run.invoker ) role to your function's service account.

Ensure that Cloud Run has access to the VPC network by using one of the following methods:

Cloud Run Service Agent role: By default, the Cloud Run service agent has the Cloud Run Service Agent role ( roles/run.serviceAgent ) that contains the necessary permissions.
Custom permissions: For more granular control, grant the Cloud Run service agent with the following additional permissions on the project:
- compute.networks.get
- compute.subnetworks.get
- compute.subnetworks.use on the project or the specific subnet
- compute.addresses.get
- compute.addresses.list
- compute.addresses.create (required only for dual-stack subnets with external IPv6 )
- compute.addresses.delete (required only for dual-stack subnets with external IPv6)
- compute.addresses.createInternal
- compute.addresses.deleteInternal
- compute.regionOperations.get
Compute Network User role: If you don't use the default Cloud Run Service Agent role or the custom permissions, grant the Compute Network User role ( roles/compute.networkUser ) on the Cloud Run Service Agent service account. Subnets with external IPv6 also require the Compute Public IP Admin role ( roles/compute.publicIpAdmin ) .

For example, to grant the Compute Network User role, run the following command:
```
gcloud  
projects  
add-iam-policy-binding  
 PROJECT_ID 
  
 \ 
--member  
 "serviceAccount:service- PROJECT_NUMBER 
@serverless-robot-prod.iam.gserviceaccount.com" 
  
 \ 
--role  
 "roles/compute.networkUser" 
```
Replace the following:
- PROJECT_ID : the ID of your project.
- PROJECT_NUMBER : the project number where you deploy your Cloud Run function.

Configure Direct VPC egress

Configure Direct VPC egress for new or existing 2nd gen functions.

gcloud

To configure Direct VPC egress when you deploy a function, use the gcloud beta functions deploy command with flags for your network settings.
```
 gcloud  
beta  
functions  
deploy  
 FUNCTION_NAME 
  
 \ 
  
--source  
.  
 \ 
  
--runtime  
 RUNTIME 
  
 \ 
  
--trigger-http  
 \ 
  
--region  
 REGION 
  
 \ 
  
--network = 
 NETWORK 
  
 \ 
  
--subnet = 
 SUBNET 
  
 \ 
  
--network-tags = 
 NETWORK_TAG_NAMES 
  
 \ 
  
--direct-vpc-egress = 
 EGRESS_SETTING 
 
```
Replace the following:
- FUNCTION_NAME : the name of your function.
- RUNTIME : the runtime for your function, for example, nodejs20 .
- REGION : the region where you deploy your function.
- Optional: NETWORK with the name of your VPC network. Specify either a VPC network or a subnet, or both. If you specify only a network, the subnet uses the same name as the network.
- Optional: SUBNET with the name of your subnet. Specify either a VPC network or a subnet, or both. If you specify only a network, the subnet uses the same name as the network. You can deploy or execute multiple functions on the same subnet.
- Optional: NETWORK_TAG_NAMES with the comma-separated names of the network tags you want to associate with a function. Each function can have different network tags, such as network-tag-2 .
- EGRESS_SETTING with an egress setting value :
  - all : Default. Sends all outbound traffic through the VPC network.
  - private-ranges-only : Sends only traffic to internal addresses through the VPC network.
Optional: To remove all Direct VPC egress settings from a function, redeploy the function with the --clear-network and --clear-network-tags flags.

Terraform

To learn how to apply or remove a Terraform configuration, see Basic Terraform commands .

To configure Direct VPC egress, use the google-beta Terraform provider.

Using the Cloud Run functions (2nd gen) Direct VPC egress example as a starting point, update the following fields:

service_config.network : the name of your VPC network.
service_config.subnetwork : the name of your VPC subnetwork.
service_config.direct_vpc_egress : which traffic to send to the VPC network. VPC_EGRESS_ALL_TRAFFIC sends all outbound traffic through the VPC network. VPC_EGRESS_PRIVATE_RANGES_ONLY only sends traffic to private IP address ranges to the VPC network.

Example: Call an internal service from a function

This example shows how to create an internal Cloud Run service and then call it from a Cloud Run functions (2nd gen) function that uses Direct VPC egress.

Create the internal backend service

Create a new directory for the backend service and change into it:
```
 mkdir  
backend-service cd 
  
backend-service 
```

Create a package.json file with the following content:

  { 
  
 "name" 
 : 
  
 "backend-service" 
 , 
  
 "version" 
 : 
  
 "1.0.0" 
 , 
  
 "description" 
 : 
  
 "" 
 , 
  
 "scripts" 
 : 
  
 { 
  
 "start" 
 : 
  
 "node index.js" 
  
 }, 
  
 "dependencies" 
 : 
  
 { 
  
 "express" 
 : 
  
 "^4.18.1" 
  
 } 
 }

Create an index.js file with the following content:

  const 
  
 express 
  
 = 
  
 require 
 ( 
 'express' 
 ); 
 const 
  
 app 
  
 = 
  
 express 
 (); 
 app 
 . 
 get 
 ( 
 '/' 
 , 
  
 ( 
 req 
 , 
  
 res 
 ) 
  
 = 
>  
 { 
  
 res 
 . 
 send 
 ( 
 "hello world" 
 ); 
 }); 
 const 
  
 port 
  
 = 
  
 parseInt 
 ( 
 process 
 . 
 env 
 . 
 PORT 
 ) 
  
 || 
  
 8080 
 ; 
 app 
 . 
 listen 
 ( 
 port 
 , 
  
 () 
  
 = 
>  
 { 
  
 console 
 . 
 log 
 ( 
 `helloworld: listening on port 
 ${ 
 port 
 } 
 ` 
 ); 
 });

Deploy the service to Cloud Run with internal ingress:

 gcloud  
run  
deploy  
backend  
 \ 
  
--source  
.  
 \ 
  
--no-allow-unauthenticated  
 \ 
  
--region = 
 REGION 
  
 \ 
  
--ingress  
internal

Replace REGION with your region, for example, us-west1 .

Save the URL of the new service. You need it in the next section.

Create and deploy the function

Create a new directory for the function and change into it:

  cd 
  
..
mkdir  
dvpc-function cd 
  
dvpc-function

Create a package.json file with the following content:

  { 
  
 "name" 
 : 
  
 "sample-http" 
 , 
  
 "version" 
 : 
  
 "0.0.1" 
 , 
  
 "dependencies" 
 : 
  
 { 
  
 "axios" 
 : 
  
 "0.21.1" 
 , 
  
 "@google-cloud/functions-framework" 
 : 
  
 "^3.0.0" 
  
 } 
 }

Create an index.js file with the following content. This code makes an authenticated request to the internal backend service.

  const 
  
 axios 
  
 = 
  
 require 
 ( 
 'axios' 
 ); 
 const 
  
 functions 
  
 = 
  
 require 
 ( 
 '@google-cloud/functions-framework' 
 ); 
 const 
  
 callVPCService 
  
 = 
  
 async 
  
 ( 
 req 
 , 
  
 res 
 ) 
  
 = 
>  
 { 
  
 const 
  
 backendUrl 
  
 = 
  
 process 
 . 
 env 
 . 
 BACKEND_URL 
 ; 
  
 if 
  
 ( 
 ! 
 backendUrl 
 ) 
  
 { 
  
 console 
 . 
 error 
 ( 
 'BACKEND_URL environment variable not set.' 
 ); 
  
 res 
 . 
 status 
 ( 
 500 
 ). 
 send 
 ( 
 'BACKEND_URL not configured.' 
 ); 
  
 return 
 ; 
  
 } 
  
 try 
  
 { 
  
 const 
  
 metadataServerURL 
  
 = 
  
 'http://metadata.google.internal/computeMetadata/v1/instance/service-accounts/default/identity?audience=' 
 ; 
  
 const 
  
 tokenUrl 
  
 = 
  
 metadataServerURL 
  
 + 
  
 backendUrl 
 ; 
  
 const 
  
 tokenResponse 
  
 = 
  
 await 
  
 axios 
 . 
 get 
 ( 
 tokenUrl 
 , 
  
 { 
  
 headers 
 : 
  
 { 
  
 'Metadata-Flavor' 
 : 
  
 'Google' 
 , 
  
 }, 
  
 }); 
  
 const 
  
 token 
  
 = 
  
 tokenResponse 
 . 
 data 
 ; 
  
 const 
  
 response 
  
 = 
  
 await 
  
 axios 
 . 
 get 
 ( 
 backendUrl 
 , 
  
 { 
  
 headers 
 : 
  
 { 
  
 Authorization 
 : 
  
 `bearer 
 ${ 
 token 
 } 
 ` 
 , 
  
 }, 
 }); 
 res 
 . 
 status 
 ( 
 200 
 ). 
 send 
 ( 
 `Response from backend: 
 ${ 
 response 
 . 
 data 
 } 
 ` 
 ); 
  
 } 
  
 catch 
  
 ( 
 error 
 ) 
  
 { 
  
 console 
 . 
 error 
 ( 
 `Error calling backend service: 
 ${ 
 error 
 . 
 message 
 } 
 ` 
 ); 
  
 res 
 . 
 status 
 ( 
 500 
 ). 
 send 
 ( 
 `Error calling backend: 
 ${ 
 error 
 . 
 message 
 } 
 ` 
 ); 
  
 } 
 }; 
 functions 
 . 
 http 
 ( 
 'callVPCService' 
 , 
  
 callVPCService 
 );

Deploy the function with Direct VPC egress configured to route all traffic to your default VPC network:

 gcloud  
beta  
functions  
deploy  
my-2ndgen-function  
 \ 
  
--source  
.  
 \ 
  
--runtime  
nodejs20  
 \ 
  
--trigger-http  
 \ 
  
--entry-point  
callVPCService  
 \ 
  
--network = 
default  
 \ 
  
--subnet = 
default  
 \ 
  
--direct-vpc-egress = 
all  
 \ 
  
--region = 
 REGION 
  
 \ 
  
--allow-unauthenticated  
 \ 
  
--set-env-vars  
 BACKEND_URL 
 = 
 BACKEND_URL

Replace the following:

REGION : the region where you deployed the backend service.
BACKEND_URL : the URL of the backend service you created.

After the function deploys, invoke it by visiting its URL. The function calls the internal backend service and returns its response.

What's next

Learn about Direct VPC egress and IP allocation .
See Troubleshooting tips .
Compare egress methods in Connecting to a VPC network .
See the gcloud beta functions deploy command reference.