Outage on any endpoint in catalog blocks usage of other endpoints

I was trying to create a template using the ComputeService backed by the nova API (I was connecting to OVH, which does not have a jClouds provider). However, one of their regions was down for maintenance. This caused the following error:

SEVERE: Cannot retry after server error, command has exceeded retry limit 5: [method=org.jclouds.openstack.nova.v2_0.features.ImageApi.public abstract org.jclouds.collect.PagedIterable org.jclouds.openstack.nova.v2_0.features.ImageApi.listInDetail()[], request=GET https://compute.bhs3.cloud.ovh.net/v2/[tenant]/images/detail HTTP/1.1

I feel that jClouds shold handle this case gracefully and simply skip the offending API, rather than failing with NullPointerExceptions and making it impossible to create templates as long as any endpoint is down for whatever reason.

This is the reply one gets when querying the given URL directly:

$ curl https://compute.bhs3.cloud.ovh.net/v2/[tenant]/images/detail
curl: (52) Empty reply from server

I verified on OVH's site that this region was indeed down for maintenance: http://travaux.ovh.net/?do=details&id=32795&PHPSESSID=b0747f889c42d0b97fedeb393cc16b0e