Update heapster version to v0.10.0

Changed merge policy

Fix typo.

Automatically upgrade gcloud in Jenkins (reenable)

Check to prevent split brain api-servers/api-clients which can happen if...

fix e2e/kubectl for gke

Skip cleaning docker cache prior to build

This seems to be more trouble than its worth, and much of
the remaining build time (except the push itself). We can
cache the cross-build/golang images between builds - it's
not going to kill us, and it'll save the occasional weird
flake of not finding an image when it tries to rm one of
these.

Check to prevent split brain api-servers/api-clients which can happen if etcd dies between rerun of local-up-cluster.sh

Updates for gcloud 0.9.54

Remove validation for Capacity. Add default for ExternalID

Add an alternative TokenSource to the GCE CloudProvider.

Parallelize architectures in both the building and packaging phases of `make release`

Also fix breakage from ExternalID validation: Default ExternalID to node name when not specified.

Fix pod that's looked for on kube up.

Don't ever give up on retrying downloads in the GCE download-or-bust function

Small cleanup on api/types.go

Increase the time we wait for the kubelet to be healthy.

On my desktop, this took the KUBE_RELEASE_RUN_TESTS=n
"make release" down from 172s to 115s

Update issues.md with modified priority definitions

Merge release 0.14.1 to master

Issue queries in parallel for serve_hostnames soak test

Using gcr.io/google_containers/pause by default

Fedora Guide: Missing yum installation of etcd

Fix off-by-one crash in pkg/apiserver/authn.go

Convert kubernetes/serve_hostname to gcr.io/google_containers/server_hos...