Rather than terminating, Collect the errors and print them per-file-per-munge.

fix verify gendocs

Better docs for downward API

Create schedule_algorithm file

Add monitoring docs

Increase the rate limiting of GCE's token source

ignore pods that are failed but controlled by a rc when start e2e tests

Update releasing.md with Kubernetes release process

Link verification

Links that don't work yet can be prefixed with "TODO:" to avoid the check.

Remove race in setting/reading kubelet.podCIDR

TestCreateCleanWithPrefix was failing if env variables existed

When releasing, update versions in title

Assigned emptyDir for etcd data dir in skydns pod

Use the exec-sidecar as a healthz probe for the syndns container

Demote Shell.*services e2e test as it cannot be run in parallel yet

Fixes #10127

Only subst things that exist

Updating the scripts to not delete v1beta3 conversions and copies

Promote Services shell out of flaky. Demote Elasticsearch

This document explains briefly the schedule algorithm of Kubernetes and can be complementary to scheduler.md.

Bump the ubuntu k8s version to 0.19.3

Updating nettest image to latest code

Addon update kill children

Jenkins: Allow fetching new tars for multi-job tests

Move code to prepare for e2e/restart transition

Enable e2e networking test for intra-pod communication for vagrant

Sidecar container capable of servicing exec style liveness probes over http