Typhoon for Azure is alpha. For production, use AWS, Google Cloud, or bare-metal. As Azure matures, check [errata](https://github.com/poseidon/typhoon/wiki/Errata) for known shortcomings.
We'll declare a Kubernetes cluster using the Typhoon Terraform module. Then apply the changes to create a resource group, virtual network, subnets, security groups, controller availability set, worker scale set, load balancer, and TLS assets.
Controller hosts are provisioned to run an `etcd-member` peer and a `kubelet` service. Worker hosts run a `kubelet` service. Controller nodes run `kube-apiserver`, `kube-scheduler`, `kube-controller-manager`, and `coredns`, while `kube-proxy` and `calico` (or `flannel`) run on every node. A generated `kubeconfig` provides `kubectl` access to the cluster.
Add the [terraform-provider-ct](https://github.com/poseidon/terraform-provider-ct) plugin binary for your system to `~/.terraform.d/plugins/`, noting the final name.
Read [concepts](/architecture/concepts/) to learn about Terraform, modules, and organizing resources. Change to your infrastructure repository (e.g. `infra`).
[Install](https://docs.microsoft.com/en-us/cli/azure/install-azure-cli?view=azure-cli-latest) the Azure `az` command line tool to [authenticate with Azure](https://www.terraform.io/docs/providers/azurerm/authenticating_via_azure_cli.html).
```
az login
```
Configure the Azure provider in a `providers.tf` file.
Reference the [variables docs](#variables) or the [variables.tf](https://github.com/poseidon/typhoon/blob/master/azure/container-linux/kubernetes/variables.tf) source.
Initial bootstrapping requires `bootstrap.service` be started on one controller node. Terraform uses `ssh-agent` to automate this step. Add your SSH private key to `ssh-agent`.
[Install kubectl](https://kubernetes.io/docs/tasks/tools/install-kubectl/) on your system. Obtain the generated cluster `kubeconfig` from module outputs (e.g. write to a local file).
On Container Linux clusters, install the `CLUO` addon to coordinate reboots and drains when nodes auto-update. Otherwise, updates may not be applied until the next reboot.
## Variables
Check the [variables.tf](https://github.com/poseidon/typhoon/blob/master/azure/container-linux/kubernetes/variables.tf) source.
### Required
| Name | Description | Example |
|:-----|:------------|:--------|
| cluster_name | Unique cluster name (prepended to dns_zone) | "ramius" |
| region | Azure region | "centralus" |
| dns_zone | Azure DNS zone | "azure.example.com" |
| dns_zone_group | Resource group where the Azure DNS zone resides | "global" |
| ssh_authorized_key | SSH public key for user 'core' | "ssh-rsa AAAAB3NZ..." |
!!! tip
Regions are shown in [docs](https://azure.microsoft.com/en-us/global-infrastructure/regions/) or with `az account list-locations --output table`.
#### DNS Zone
Clusters create a DNS A record `${cluster_name}.${dns_zone}` to resolve a load balancer backed by controller instances. This FQDN is used by workers and `kubectl` to access the apiserver(s). In this example, the cluster's apiserver would be accessible at `ramius.azure.example.com`.
You'll need a registered domain name or delegated subdomain on Azure DNS. You can set this up once and create many clusters with unique names.
If you have an existing domain name with a zone file elsewhere, just delegate a subdomain that can be managed on Azure DNS (e.g. azure.mydomain.com) and [update nameservers](https://docs.microsoft.com/en-us/azure/dns/dns-delegate-domain-azure-dns).
| asset_dir | Absolute path to a directory where generated assets should be placed (contains secrets) | "" (disabled) | "/home/user/.secrets/clusters/ramius" |
| worker_priority | Set priority to Low to use reduced cost surplus capacity, with the tradeoff that instances can be deallocated at any time | Regular | Low |
Check the list of valid [machine types](https://azure.microsoft.com/en-us/pricing/details/virtual-machines/linux/) and their [specs](https://docs.microsoft.com/en-us/azure/virtual-machines/linux/sizes-general). Use `az vm list-skus` to get the identifier.
Unlike AWS and GCP, Azure requires its *virtual* networks to have non-overlapping IPv4 CIDRs (yeah, go figure). Instead of each cluster just using `10.0.0.0/16` for instances, each Azure cluster's `host_cidr` must be non-overlapping (e.g. 10.0.0.0/20 for the 1st cluster, 10.0.16.0/20 for the 2nd cluster, etc).
Add `worker_priority=Low` to use [Low Priority](https://docs.microsoft.com/en-us/azure/virtual-machine-scale-sets/virtual-machine-scale-sets-use-low-priority) workers that run on Azure's surplus capacity at lower cost, but with the tradeoff that they can be deallocated at random. Low priority VMs are Azure's analog to AWS spot instances or GCP premptible instances.
