Fix wrong config used for node-exporter container limits

*: Fix AlertmanagerConfigInconsistent many-to-many matching errors

ignore swp files

If there is more than one prometheus-operator pod, which happens briefly
when we delete the prometheus-operator pod, we can see the errors of
many-to-many matching, this whitelists the labels matching, and excluded
the pod.

Use local dependency for kube-prometheus jsonnet

there's never a need to commit swap files, so just ignore them

Add mixin to strip cpu/memory limits (issue #72)

manifests: Re-generate

set kubeStateMetrics version to 1.7.2

Signed-off-by: Guus van Weelden <guus.vanweelden@moia.io>

Fix io time yaxe

Fix incorrect KubeVersionMismatch alert when `{job=coredns}`

Increase ntp clockskewdetected accuracy to 5ms

adding a note about ServiceMonitor definition

Bump kube-state-metrics to v1.7.0

The resource is called storageclasses not storageclass.

In the **Adding additional namespaces to monitor** section I appended a note showing the need for ServiceMonitor when adding additional namespaces... 

see: https://github.com/coreos/prometheus-operator/issues/2557#issuecomment-498996568

update README, change minikube k8s-version to 1.14.4