QueryContainerImages roadmap

[dekkagaijin]

Is there a publicly available roadmap re: the QueryContainerImages API? Features, GA, etc.

/ping @nickyringland

[nickyringland]

Hi @dekkagaijin,
That feature is in our Alpha API - we haven't received much feedback on it, so would be curious on what features you're interested in!

[dekkagaijin]

Thanks for the reply :)

Beyond the repositories containing the image, any combination of the following would be super helpful, roughly in order of guesstimated Impact-per-unit-effort:

• manifest (and/or manifest list) digest
  • a really useful breadcrumb that allows users to obtain a lot of information about the base image with one or two registry API calls. This would help avoid a BFS of the images and configs in a repo in order to identify the discrete image associated with a given chainID (with the caveat that there are potentially, albeit rarely, many-to-one relationships between sha(manifest)-sha(digest) and sha(digest)-chainID)
  • This'd be the huge win
• config digest
  • potentially allows for easy association with images in contexts outside the registry, e.g. pulled images and .tars
• tag(s) currently (and historically) associated with the chainID
  • potentially helpful, assuming the tag itself is meaningful to the user. This feature would probably be the least useful and most fraught in terms of design, though.

[dekkagaijin]

Also, a batch API analogous to the others would let callers query all possible chainIDs to identify the longest base image match in a single call