Core needs for Glacier Connector

Implement Glacier storage:

Low-level platform tasks:

• changes to S3 blob provider to apply storage class to blob (Glacier mode for the S3 blob provider)
• APIs to request a blob retrieve
• APIs to request a blob's status
• mock for file storage / inmemory storage in order to be able to unit test higher-level tasks

High-level platform tasks:

• define ColdStorage facet with:
  • coldstorage:content for the blob dispatched to a Glacier blob provider
  • coldstorage:contentStorageClass for the storage class (glacier / deep archive / ...)
  • coldstorage:beingRetrieved: boolean for blobs for which a retrieve is in progress
• operations:
  • moveToColdStorage: request blob to be moved to Glacier and be replaced with a given preview
    file:content -> coldstorage:content
    preview (tbd) -> file:content
    (postponed to other milestone: dealing with versions)
    (postponed to other milestone: delete original blob from main blob store to avoid costs)
  • get blob status (regular / unavailable / being retrieved / available until 2020-01-01) (actually a DownloadService REST API)
  • retrieveBlobFromColdStorage: request a retrieve of a blob for n days
• scheduled job to check if blobs being retrieved are finally available, sending an event
• listener sending a mail when a blob becomes available

UI: postponed to annother milestone.