Nine Kubernetes Engine

Nine Kubernetes Engine or NKE is a fully managed Kubernetes offering running in Nine-operated data centers in Switzerland. Supplemented by additional services, letting you completely focus on your application development.

Getting started

To get started with NKE you need a login to access our Cockpit. Then simply select Managed Kubernetes from our products overview and use Add Cluster to create your first cluster. The initial creation can take a few minutes until all services are deployed. To access the cluster, have a look at our cluster login article.

Locations

NKE is available in the following locations:

Name	API Name
NTT Zurich 1	nine-es34
ColoZüri 4.1	nine-cz41

Subnets in each Location

All Nine Kubernetes Engine cluster nodes will have public IPs by default. Each Pod running on a node will use the public IP of the node it is running when initiating connections to cluster external systems (Kubernetes default networking model). The IP subnets which will be used depend on the location of the Nine Kubernetes Engine cluster. You can find a current list of subnets in JSON format by using our insight service:

Location	API Endpoint
All Subnets	https://insight.nineapis.ch/subnets
nine-es34	https://insight.nineapis.ch/subnets/nine-es34
nine-cz41	https://insight.nineapis.ch/subnets/nine-cz41

Please note that new subnets might be added at any point.

Example

You can use the curl command line tool to request a current list:

curl https://insight.nineapis.ch/subnets

{"data":{"nine-cz41":["94.230.211.128/26"],"nine-es34":["178.209.59.64/26","185.88.237.64/26","5.148.184.0/26"]}}

Node Pools

Like many Kubernetes offerings, NKE manages nodes in pools. A pool is defined by a name, machine type, disk size and an amount of nodes. Node pools simplify isolating different workloads on a node level from each other. Also, they can be individually scaled up and down.

NKE uses three nine-standard-2 as control plane nodes. Depending on the services configured and the specific usage of the cluster, it may be necessary to adjust the machine type of these nodes. If such a change is required, we will reach out to you proactively.

Machine Types

NKE offers the following machine types:

Name	vCPU	RAM (GiB)
nine-standard-1	1	4
nine-standard-2	2	8
nine-standard-4	4	16
nine-highmem-2	2	16
nine-highmem-4	4	32
nine-highcpu-2	2	4
nine-highcpu-4	4	8
nine-highcpu-8	8	16

If you require a type that is not yet defined here, don't hesitate to contact us and we will consider adding it to the list.

Load Balancing

NKE supports services with the type load balancer out of the box. After creating your service it will be assigned a public IPv4 address for the lifetime of the service.

Simply follow the Kubernetes documentation to create a new load balancer.

Though for most use cases an Ingress might be the simpler alternative, if you just need to expose HTTP(S) traffic. Head over to our ingress documentation for that.

Persistent Storage

Persistent Storage can be requested by creating a PersistentVolumeClaim. Our controller will take care of creating a PersistentVolume and mounting it to your pod.

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: example
  namespace: example
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 5Gi
  storageClassName: standard

Our controller supports online expansion of a PVC. Simply edit the PVC in question and increase the storage requests to your liking.

Read Write Many (RWX)

⚠ Note: RWX is currently only available in the location nine-es34.

If you have a use case for RWX storage, you can create such volumes with the storage class files. This will create a NFS-backed volume that can be mounted in multiple pods at once. In most cases we recommend against using RWX and instead make use of object storage where possible.

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: example
  namespace: example
spec:
  accessModes:
    - ReadWriteMany
  resources:
    requests:
      storage: 5Gi
  storageClassName: files

Backups

All NKE clusters are backed up by default. For more details see our article on backups.

Maintenance Windows

We have three windows a week: Tuesday, Wednesday and Thursday starting at 00:00 UTC to 04:00 UTC. This allows us to upgrade many nodes in different clusters in a rolling fashion.

Additional Services

To supplement NKE, we offer a number of services to make a fully featured application platform. You can see all available services in Cockpit, pricing on our website and more details in other support articles.

Further Information

For further information or sales please contact info@nine.ch and for support contact us on our support portal.