The taskcluster authentication server manages permissions and credentials in the taskcluster eco-system. Identifiers, credentials and authorized scopes will be stored in azure table storage, and various components will be granted read-only access in-order to authorize requests.
On the client side, an authorized client must have a CLIENT_ID and an
ACCESS_TOKEN to be used with hawk for making requests.
On the server side, CLIENT_IDs will resolve to ACCESS_TOKEN for HMAC
validation and a set of scopes, which will be used to determine what resources
the client is authorized to access.
If you deploy a copy of this application that fails to start, all of TaskCluster will come to a screeching halt immediately. Don't do that.
This app auto-deploys from Github to the staging environment, which uses a different Azure backend.
Happily, there's a nice, automated way of testing that staging environment
before promoting it to production. First, if you haven't already, run npm run checkStagingSetup and copy-paste the results into your user-config.yml.
Note that you will need Heroku app access to do this! Once your config is set
up, just run npm run checkStaging to check the staging site for
functionality.