8b0cd5b9c5
Automatic merge from submit-queue (batch tested with PRs 47906, 47910) Reduce scheduler CPU request to 75m On a 1 cpu master we are over budget with CPU requests. Components like npd or cluster autoscaler don't have *any* space to run. We need to reduce some requests. cc: @gmarek @mikedanese @roberthbailey @davidopp @dchen1107
This directory forms the base of the main SaltStack configuration. The
place to start with any SaltStack configuration is
top.sls. However, unless you are particularly keen on
reading Jinja templates, the following tables break down what
configurations run on what providers. (NB: The _states
directory is a special directory included by Salt for ensure blocks,
and is only used for the docker config.)
Key: M = Config applies to master, n = config applies to nodes
| Config | GCE | Vagrant | AWS | Azure |
|---|---|---|---|---|
| debian-auto-upgrades | M n | M n | M n | M n |
| docker | M n | M n | M n | M n |
| etcd | M | M | M | M |
| generate-cert | M | M | M | M |
| kube-addons | M | M | M | M |
| kube-apiserver | M | M | M | M |
| kube-controller-manager | M | M | M | M |
| kube-proxy | n | n | n | n |
| kube-scheduler | M | M | M | M |
| kubelet | M n | M n | M n | M n |
| logrotate | M n | n | M n | M n |
| supervisord | M n | M n | M n | M n |
| base | M n | M n | M n | M n |
| kube-client-tools | M | M | M | M |