lots of 'MANIFEST_UNKNOWN' errors in fluxd logs

[rade]

In a k8s cluster on GKE I am seeing many of the following errors in the fluxd logs:

2017-11-13T17:19:40.515106619Z ts=2017-11-13T17:19:40.514852143Z caller=warming.go:148 component=warmer err="requestingmanifests: getting remote manifest: Get https://eu.gcr.io/v2/google_containers/kubernetes-dashboard-amd64/manifests/v1.4.1: http: non-successful response (status=404 body=\"{\\\"errors\\\":[{\\\"code\\\":\\\"MANIFEST_UNKNOWN\\\",\\\"message\\\":\\\"Manifest with tag 'v1.4.1' has media type 'application/vnd.docker.distribution.manifest.v1+prettyjws', but client accepts 'application/vnd.docker.distribution.manifest.v1+json'.\\\"}]}\")"

#790 looks related, but it's not clear whether it would actually fix this particular v1+prettyjws instead of v1+json condition.

[squaremo]

#790 looks related, but it's not clear whether it would actually fix this particular v1+prettyjws instead of v1+json condition.

The thrust of #790 is to accommodate all the content types, including signed manifests (v1+prettyjws). So it is very much intended to fix this problem.