Kubernetes Cluster Networking
Feature state: stableAME Kubernetes Cluster Networking functionality
Avisi Cloud Kubernetes supports Calico CNI.
Networking Features
| Feature | Supported | |
|---|---|---|
| Network Policies | ✔️ | |
| Node Ports | ✔️ | |
| Traffic shaping | ✔️ |
Network Policies
Kubernetes Network Policies are fully supported by Avisi Cloud Kubernetes. Check out our help article installing network policies for help on how to install network policies within your Cluster.
Node Ports
It's possible to use Node Ports for routing traffic to specific nodes within your Cluster. Please be aware that nodes may get automatically replaced during upgrades or outages.
If you use a Private Cluster with a NAT gateway, you will need to reach the node by using its internal IP address.
NodePort Access
On AWS and Hetzner, the NodePort access setting controls who can reach the NodePort range (30000-32767, TCP and UDP) on nodes with a public IP address:
internal(default): only the cluster's internal network.allowlist: the internal network plus the IP addresses in the cluster's IP whitelist, the same list that restricts API server access. With an empty IP whitelist, only the internal network can reach the NodePorts.public: the entire internet.
Change the setting in the console: open the cluster, go to the Networking tab, select Firewalling and pick an option under NodePort Access. With the acloud CLI, run acloud cluster update <cluster-slug> --node-port-access allowlist. AME applies the change in the background.
The setting does not apply to Private Clusters on AWS. Their nodes are behind a NAT gateway and have no public IP address.
On AWS, this setting does not cover a Service of type LoadBalancer that uses a Network Load Balancer. Its NodePorts stay open to the Service's loadBalancerSourceRanges. Without loadBalancerSourceRanges, they stay open to the entire internet.
Traffic shaping
You can control the bandwidth of pods running on Avisi Cloud Kubernetes by using the CNI networking plugin for traffic shaping. See the Kubernetes documentation for more information.
DNS
Clusters are deployed with CoreDNS as the kube-dns solution. CoreDNS is configured to auto-scale based on the amount of resources available within your Cluster.
Node Local DNS cache
Optionally you can enable Node Local DNS cache. Use this functionality if you have a lot of DNS requests and experience issues with network connectivity.