In the current provisioning setup we can only set a single github config/token. Garm is able to manage pools of workers for multiple repos and we want to use it for e.g. the kbs repo
How?
In the revised code tf will serialize the entries to json and set it as an env. In a jinja2 template it will be parsed into a dict and templated.
From my testing the update should be non-destructive, we don't need to discard an existing garm configuration, we just need to give the existing token the name "garm", since the garm schema references the github credentials by name.
Why?
In the current provisioning setup we can only set a single github config/token. Garm is able to manage pools of workers for multiple repos and we want to use it for e.g. the kbs repo
How?
In the revised code tf will serialize the entries to json and set it as an env. In a jinja2 template it will be parsed into a dict and templated.
From my testing the update should be non-destructive, we don't need to discard an existing garm configuration, we just need to give the existing token the name "garm", since the garm schema references the github credentials by name.