There are two notable issues with the Provenance validator:
- The validator is loaded synchronously, causing the first load of the dataset page after deployment to take an extra ~8 seconds
- When the validator loads the schema, it attempts to connect to the ID url in the schema for the additional information needed by the $ref tags https://json-schema.org/understanding-json-schema/structuring.html#the-id-property
These cause dataverse to fail loading the dataset page without a network, and when there is a network the first load of the provenance validator is still slow.