Manage multiple deployment profiles for Google Cloud Run.
Cloud Run's cost model fits very well with multiple staging environments. For example, you might want to deploy multiple services with slightly different environmental variables (e.g. different Cloud SQL connection). Although Google offers a friendly UI, it's not very helpful when you need to manage multiple profiles within a CI/CD pipeline.
Here is an example of running two environments.
{
"name": "hello",
"image": "gcr.io/cloudrun/hello",
"platform": "managed",
"concurrency": 10,
"max-instances": 1,
"memory": "512Mi",
"region": "asia-northeast1",
"env": {
"MYSQL_USER": "local",
"MYSQL_PASSWORD": "berglas://my-bucket/mysql-password-dev"
},
"cloudsql-instances": [
"staging"
]
}When you want to override some, but not all parameters, run-env can automatically read from other profiles using the include key.
{
"include": ["dev.json"],
"max-instances": 10,
"allow-unauthenticated": true,
"env": {
"MYSQL_USER": "production",
"MYSQL_PASSWORD": "berglas://my-bucket/mysql-password-prod"
},
"cloudsql-instances": [
"production"
]
}Several feature that might be extra helpful:
- You can include multiple files.
- You can nest multiple files.
- When used with tools like berglas, the sensitive information can be safely encrypted and stored, outside of your git repository.
- Text fields supports shell-like interpolation so you can dynamically configure service name (e.g.
prefix-$CIRCLE_BRANCH).
MIT