Read allowed_hosts from actor definitions yaml and store in database

[evantahler]

From the tech spec:

ActorDefintions are updated so that connectors can declare which hosts/IPs they intend to connect to. This is a new optional field in source_definitions.yaml and destination_definitions.yaml.

- name: HubSpot
  sourceDefinitionId: 36c891d9-4bd9-43ac-bad2-10e12756272c
  dockerRepository: airbyte/source-hubspot
  dockerImageTag: 0.3.0
  documentationUrl: https://docs.airbyte.com/integrations/sources/hubspot
  icon: hubspot.svg
  sourceType: api
  releaseStage: generally_available
  # ~~~ NEW ~~~
  allowed_hosts:
    - *.hubspot.com 

The new “allowed_hosts” property is an array. It’s values can be URL matches (per the example above, e.g. "*.instagram.com"), IP address and/or CIDR ranges, or a reference to a value that will be completed by the user completing the connector’s configuration. The reference use-case is required for those connectors who accept a dynamic hostname or IP address to connect to (e.g. a database hostname or an API sub-domain). A reference specification could look like:

allowed_hosts:
 - “${#/definitions/config/hostname}”

Finally, we can combine the approaches for when there’s a customer-specific subdomain in use with interpolation:

allowed_hosts:
- “${#/definitions/config/subdomain}.snowflake.com”

This information should be loaded by the bootloader and stored in the actor_defintions table in the database in a new column as a JSON array.

[evantahler]

Grooming:

• we should do some validation. E.g. the values from configs should exist in the SPEC. We also might want to validate the URL strings?
• Filling this info out in OSS will eventually propagate to cloud

[evantahler]

Following the work done in #18300 is a really good guide about how to add a migration and load data from the actor_definitions!