Skip to content

Roocs for CWT - Questions and decisions #214

Description

@agstephens

We need to decide the following:

  • how would an ESGF user discover the service?
    • do we use Intake catalogs / or just open up all the data?
      • links added to metadata records by the publisher?
  • do we enable the fixing capability?
    • switch off completely?
    • need to double-check issues around xarray assumptions/features:
      • the _FillValue issue
      • the coordinates issue
  • how do we track/monitor/report usage?
  • which projects/activities should be available through the WPS?
    • CMIP6/CMIP5 - and any others that are open data?
      • would need more tests in the codebase
  • what should the URL be: https://esgf-compute.ceda.ac.uk ?
  • build support for access through an API KEY into the first release:
    • can we re-use other middleware tools to manage load/access/allocation?
  • can we enable GET requests run by rook:
    • from a remote ESGF Data Node?
    • from a remote ESGF Compute Node?
      • might just return URLs that are passed to the user
    • would need a register of services
    • would need limits on how much can be cached/requested
  • should we support URIs that include the service/host ID?

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions