GCP Cloud Run Terraform Module
This module manages the deployment of containerised applications on Cloud Run. It includes the following functionality:
- Creation of the main Cloud Run service.
- Creation of a dedicated service account with required IAM bindings.
- Deployment of an optional load balancer configuration.
- Deployment of an optional "pre-deployment" Cloud Run job to handle tasks such as database migrations.
- Deployment of an optional static egress IP address for the Cloud Run service.
- Configuration of simple uptime checks and SSL certificate expiry alerts.
- Convenience variables to configure certain aspects of the Cloud Run service which would otherwise be complex, such as mounting Cloud SQL instances.
Cloud Run v2 API
Following release 9.0.0
, this module has been refactored to use the Cloud Run
v2
API resources exclusively (i.e. google_cloud_run_v2_service
). This means
that many input variables are now different and updating from a previous version
will require some changes to your module definition.
Examples
See the docs/examples.md page for a full list of detailed usage examples.
Pre-deploy Cloud Run Job
The 8.0.0
release introduced the enable_pre_deploy_job
variable which, when
set to true
, creates a Cloud Run job to execute a configurable command before
new Cloud Run service revisions are deployed. This is a useful way to run
database migrations and other commands which are tightly coupled to the release
cadence of the main Cloud Run service.
The pre-deploy job is configured via the pre_deploy_job_*
variables which can
be found in variables.tf
.
For more information on how the pre-deploy Cloud Run job works see the pre-deploy-job.md page.
Requirements
Name | Version |
---|---|
terraform | ~> 1.3 |
>= 4.0 | |
null | ~> 3.0 |
Inputs
Name | Description | Type | Default | Required |
---|---|---|---|---|
alerting_notification_channels | A list of notification channel IDs to send uptime alerts to. The format for the channel IDs should be "projects/[PROJECT_ID]/notificationChannels/[CHANNEL_ID]". |
list(string) |
[] |
no |
alerting_success_threshold_percent | If the percentage of successful uptime checks within the given uptime period falls below this, an alert will be triggered. Set to 100 to trigger an alert if any uptime check fails, set to a lower number to tolerate failures without alerting. Experience has taught us that uptime checks can fail semi-regularly due to transient problems outside our control, therefore we allow some leeway before triggering an alert. |
number |
75 |
no |
alerting_uptime_period | Frequency of uptime checks | string |
"300s" |
no |
alerting_uptime_timeout | Timeout for http polling. | string |
"30s" |
no |
allow_unauthenticated_invocations | If true, the webapp will allow unauthenticated invocations. If false, the webapp requires authentication as a Google user with the Cloud Run invoker permission on the deployment. |
bool |
true |
no |
containers | Configure one or more container instances for the service. See https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/cloud_run_v2_service#nested_containers for information on the available arguments. |
map(object({ |
n/a | yes |
create_ipv6_address | Allocate an IPv6 address to the load balancer if var.enable_ipv6 is true. | bool |
false |
no |
description | A description for the Cloud Run service. | string |
null |
no |
dns_names | DNS names to configure for the web application. Note that DNS records are _NOT_ created, they are used in the load balancer module to ensure the SSL certificate is generated with the required SANs. The map's keys are arbitrary and are only required to avoid errors when the DNS name is a value which Terraform does not know until after the apply operation. For example: { my_awesome_dns_name = "awesome.example.com" } |
map(string) |
{} |
no |
enable_alerting | Enable alerting policies. | bool |
true |
no |
enable_ipv6 | Whether to enable IPv6 address on the CDN load-balancer. | bool |
false |
no |
enable_load_balancer | Whether to configure a load balancer or use the default run.app generated hostname. |
bool |
false |
no |
enable_monitoring | Optional. If true, create uptime and SSL expiry checks. Note that this is different from not specifying an alerting email address. If no alerting email address is specified the uptime checks are still created, they just don't alert if they fail. |
bool |
false |
no |
enable_pre_deploy_job | Configure a Cloud Run Job to be executed before the main Cloud Run service is deployed. This is useful for running database migrations for example. |
bool |
false |
no |
enable_static_egress_ip | Whether to assign a static ip for egress from this cloud run instance. If enabled, the "vpcaccess.googleapis.com" API must also be enabled on the project. |
bool |
false |
no |
encryption_key | The ID of a customer managed encryption key (CMEK) to use to encrypt this container image. |
string |
null |
no |
execution_environment | The sandbox environment to host this revision. Possible values are EXECUTION_ENVIRONMENT_GEN1, and EXECUTION_ENVIRONMENT_GEN2. |
string |
"EXECUTION_ENVIRONMENT_GEN1" |
no |
grant_sql_client_role_to_webapp_sa | When set to true the roles/cloudsql.client role will be granted to the webapp service account at the project level to allow it to connect to Cloud SQL. |
bool |
false |
no |
ingress | The ingress setting for the Cloud Run service. Possible values are INGRESS_TRAFFIC_ALL, INGRESS_TRAFFIC_INTERNAL_ONLY, and INGRESS_TRAFFIC_INTERNAL_LOAD_BALANCER. If var.use_load_balancer == true, the provided var.ingress will be ignored and the ingress will be set automatically to "INGRESS_TRAFFIC_INTERNAL_LOAD_BALANCER". |
string |
"INGRESS_TRAFFIC_ALL" |
no |
launch_stage | The launch stage for the Cloud Run service. Possible values are UNIMPLEMENTED, PRELAUNCH, EARLY_ACCESS, ALPHA, BETA, GA, and DEPRECATED. |
string |
"GA" |
no |
max_instance_request_concurrency | Sets the maximum number of requests that each serving instance can receive. | number |
null |
no |
min_ports_per_vm | When using Cloud NAT to provide an egress route, Cloud NAT's minimum ports per VM can be configured to determine how many concurrent connections can be established to the same destination IP address and port. |
number |
64 |
no |
monitoring_path | Path component of url to be monitored. | string |
"/" |
no |
monitoring_scoping_project | The ID of a Cloud Monitoring scoping project to create monitoring resources in. If omitted, var.project will be used instead. |
string |
null |
no |
mount_cloudsql_instance | Mount a single CloudSQL instance in each container for the service. This value should be the Cloud SQL instance connection name, for example "example-devel-e662dd2b:europe-west2:sql-6e1dd60b". This is a convenience variable to simplify mounting a single Cloud SQL instance. If you require more control you can define one or more Cloud SQL mounts directly using var.containers and var.volumes instead. |
string |
null |
no |
name | Prefix used to form resource names. | string |
"webapp" |
no |
pre_deploy_job_annotations | Map of key/value pairs containing annotations to assign to the pre-deploy Cloud Run job. |
map(string) |
null |
no |
pre_deploy_job_container | Configure the container instance for the pre-deploy job. See https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/cloud_run_v2_job#nested_containers for more information on these options. |
object({ |
null |
no |
pre_deploy_job_encryption_key | The ID of a customer managed encryption key (CMEK) to use to encrypt this container image. |
string |
null |
no |
pre_deploy_job_execution_environment | The execution environment to host this task. Possible values are EXECUTION_ENVIRONMENT_GEN1, and EXECUTION_ENVIRONMENT_GEN2 |
string |
"EXECUTION_ENVIRONMENT_GEN2" |
no |
pre_deploy_job_force | When true, and only when used in addition to var.pre_deploy_job_trigger, the pre-deploy Cloud Run job is executed at every terraform apply, regardless of # the status of var.pre_deploy_job_container.image. |
bool |
false |
no |
pre_deploy_job_labels | Map of key/value pairs containing labels to assign to the pre-deploy Cloud Run job. |
map(string) |
null |
no |
pre_deploy_job_launch_stage | The launch stage for the pre-deploy Cloud Run job. Possible values are UNIMPLEMENTED, PRELAUNCH, EARLY_ACCESS, ALPHA, BETA, GA, and DEPRECATED. |
string |
"GA" |
no |
pre_deploy_job_max_retries | Configure the maximum number of retries for the pre-deploy job. | number |
null |
no |
pre_deploy_job_mount_cloudsql_instance | Mount a CloudSQL instance in the pre-deploy job container. This is a convenience variable to simplify mounting a Cloud SQL instance. However, if you require more control over this you should define it directly in var.pre_deploy_job_container instead. |
string |
null |
no |
pre_deploy_job_parallelism | Specifies the maximum desired number of tasks the execution should run at given time. |
number |
null |
no |
pre_deploy_job_task_count | Specifies the desired number of tasks the execution should run. | number |
null |
no |
pre_deploy_job_timeout | Configure a timeout, in seconds, for the pre-deploy job. | string |
null |
no |
pre_deploy_job_trigger | When true, the pre-deploy Cloud Run job is executed via a null_resource-triggered gcloud command whenever Terraform detects that var.pre_deploy_job_container.image has changed. |
bool |
true |
no |
pre_deploy_job_volumes | Configure one or more volumes for the pre-deploy job. See https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/cloud_run_v2_job#nested_volumes for more information on these options. |
list(object({ |
[] |
no |
pre_deploy_job_vpc_access | Configure VPC access for the pre-deploy job. See https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/cloud_run_v2_job#nested_vpc_access for more information on these options. |
object({ |
null |
no |
project | Project containing the webapp. | string |
n/a | yes |
region | Location used to create Cloud Run service and other resources. | string |
n/a | yes |
revision | The unique name for the revision. If this field is omitted, it will be automatically generated based on the Service name. |
string |
null |
no |
scaling | The minimum number of auto-scaled instances defaults to 0, thus, the container will stop if it doesn't receive requests for a period of time and the following request will make the container start from cold. This should be carefully considered for containers that take a significant amount of time starting from cold. For a container with N-workers, the maximum number of auto-scaled instances should be less than 1/N of the maximum connection count for the Cloud SQL instance. |
object({ |
null |
no |
service_account_display_name | If non-empty, override the default display name of the webapp service account. | string |
"" |
no |
service_account_id | A service account is always created for the web application. If non-empty this variable overrides the default service account id. The default id is formed from the "name" variable value with "-run" appended. |
string |
"" |
no |
service_annotations | Map containing additional annotations to be added to the Cloud Run service itself. |
map(string) |
{} |
no |
service_labels | A set of key/value label pairs to assign to the Cloud Run service. | map(string) |
{} |
no |
session_affinity | Enables session affinity. For more information, go to https://cloud.google.com/run/docs/configuring/session-affinity. |
bool |
null |
no |
sql_instance_project | Project containing SQL instance. Defaults to var.project. | string |
null |
no |
ssl_certificates | A list of self-links to any custom TLS certificates to add to the load balancer. Requires that var.use_load_balancer be "true". The self-link is available as the "self_link" attribute of "google_compute_ssl_certificate" resources. |
list(any) |
[] |
no |
ssl_policy | By default, the google_compute_ssl_policy.default SSL policy is applied to the load balancer in load_balancer.tf. This sets the SSL profile to MODERN and restricts TLS to >= 1.2. If a different SSL policy is required, it should be created outside of this module and its ID passed through using this variable. |
string |
null |
no |
static_egress_ip_cidr_range | The cidr range used to create a subnet that this cloud run will use if assigned a static ip |
string |
"10.124.0.0/28" |
no |
static_egress_ip_subnetwork_id | When using an existing VPC Access Connector with the static egress IP configuration an existing subnetwork must be provided. |
string |
null |
no |
template_annotations | Map containing additional annotations to be added to the Cloud Run service template. |
map(string) |
{} |
no |
template_labels | A set of key/value label pairs to assign to the Cloud Run service revision. | map(string) |
{} |
no |
timeout_seconds | The maximum duration, in seconds, the instance is allowed for responding to a request. Maximum is 900s. |
string |
"300s" |
no |
traffic | Configure traffic allocation between one or more service revisions. | list(object({ |
[] |
no |
use_ssl_certificates | Whether to use the custom TLS certs in var.ssl_certificates for the load balancer or the Google-managed certs for the specified var.dns_names. |
bool |
false |
no |
volumes | Configure one or more volumes for the service. See https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/cloud_run_v2_service#nested_volumes for more information on these options. |
list(object({ |
[] |
no |
vpc_access | Configure VPC access for the Cloud Run service. For more information on these options see https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/cloud_run_v2_service#nested_vpc_access |
object({ |
null |
no |
vpc_access_connector_max_throughput | Optional. The maximum throughput of the connector in megabytes per second. Defaults to 300. |
number |
300 |
no |
Outputs
Name | Description |
---|---|
load_balancer | Load balancer for the webapp |
network_endpoint_group | Network endpoint group for the load balancer. |
service | Webapp Cloud Run service resource |
service_account | Service account which service runs as |
ssl_policy | The ssl_policy object, if one is being created. |
static_egress_ip | The static egress IP assigned to this cloud run instance. Only populated if the variable enable_static_egress_ip is true. |