gcp-service-broker

README

Open Service Broker for Google Cloud Platform

This is a service broker built to be used with Cloud Foundry and Kubernetes. It adheres to the Open Service Broker API v2.13.

Service brokers provide a consistent way to create resources and accounts that can access those resources across a variety of different services.

The GCP Service Broker provides support for:

• BigQuery
• Bigtable
• Cloud SQL
• Cloud Storage
• Dataflow (preview)
• Dataproc (preview)
• Datastore
• Dialogflow (preview)
• Firestore (preview)
• Memorystore for Redis (preview)
• ML APIs
• PubSub
• Spanner
• Stackdriver Debugger
• Stackdriver Monitoring (preview)
• Stackdriver Trace
• Stackdriver Profiler

Installation

This application can be installed as either a PCF Ops Man Tile or deployed as a PCF application. See the installation instructions for a more detailed walkthrough.

Upgrading

If you're upgrading, check the upgrade guide.

Usage

For operators: see docs/customization.md for details about configuring the service broker.

For developers: see docs/use.md for information about creating and binding specific GCP services with the broker. Complete Spring Boot sample applications which use services can be found in the service-broker-samples repository.

You can get documentation specific to your install from the /docs endpoint of your deployment.

Commands

The service broker can be run as both a server (the service broker) and as a general purpose command line utility. It supports the following sub-commands:

• client - A CLI client for the service broker.
• config - Show and merge configuration options together.
• generate - Generate documentation and tiles.
• help - Help about any command.
• migrate - Upgrade your database (you generally won't need this because the databases auto-upgrade).
• plan-info - Dump plan information from the database.
• serve - Start the service broker.
• show - Show info about the provisioned resources.

Testing

Production testing for the GCP Service Broker is administered via a private Concourse pipeline.

To run tests locally, use Ginkgo.

Integration tests require the ROOT_SERVICE_ACCOUNT_JSON environment variable to be set.

Note: Integration tests create and destroy real project resources and therefore have associated costs to run

Support

File a GitHub issue for functional issues or feature requests.

Subscribe to the gcp-service-broker Google group for discussions and updates.

Contributing

See the contributing file for more information.

This is not an officially supported Google product.