Handling errors
Errors during apply
Frigg applies resources one at a time. It is therefore possible that one resource will be applied successfully, while another resource can fail to apply.
Some resources need to be applied in multiple stages. It’s possible a resource can only be partially applied.
The GitLab pipeline will indicate a failure in case Frigg fails to apply a resource. Any errors will be printed to the console. Typically, they can be resolved either by adjusting the configuration you’re attempting to apply, or by adjusting conflicting configuration directly on the platform in question.
When unable to resolve the issue yourself support should be contacted if your plan allows for it. In case you found a bug, yet your plan does not provide direct support, please still contact Frigg support, so that the issue can be looked into and resolved as soon as possible.
Inconsequential apply
It’s possible Frigg to not have detected a failure to apply a resource. This could occur because the API Frigg makes use of false reports a successful operation, but reverts said operation directly afterward. It’s been noted that the GitLab API does this in certain cases, normally as a result of conflicting configuration.
Errors when token expires
Frigg uses 2 access tokens within GitLab to properly work. These are:
- The access token to the Frigg Public container registry within the DOCKER_AUTH_CONFIG parameter
- The access token which allows Frigg to interact with your GitLab Group/Instance
Tokens can expire, which will result in Frigg no longer being able to work correctly. In these cases, action must be taken. I
DOCKER_AUTH_CONFIG token
If this token expires, your Frigg instance will no longer be able to access the required Docker images. Support will give you a heads up a week in advance and ensure you get a new token. But if this problem occurs and you haven’t received your new token yet, contact support in order to receive a new token and DOCKER_AUTH_CONFIG to continue using Frigg.
Frigg Access Token
If this token expires, Frigg will no longer be able to interact with your GitLab instance and top level group. Whenever this happens, you will have to create a new access token with the correct permissions.
- Create an access token on the top level group that Frigg is managing with full API access
- Update the GITLAB_TOKEN CI/CD variable with your new token
- Due to Frigg now interacting as a new user, the old state issue will be unusable
- In order to have Frigg create a new state issue, it’s recommended to run a new export of your GitLab instance