All notable changes to this project will be documented in this file.
- Add reference to kube-hetzner
- Make ICMP and kubernetes configurable in firewall
- Correct links in toc
- Perform dist-upgrade after initializing the cluster
- Add helm and cilium as recommended tools
- Add detailed component update instructions
- Move component versions into module to simplify examples
- Add k9s screenshot
- Add an additional installation step
- Hide virtual
*
host
- Add more figures
- Correct reference to pictures
- Add local haproxy to all nodes to get high availibility
- Make nu shell version configurable
- Install ccm earlier and set additional tolerations
- Set k8sServiceHost to load balancer
- Add infrastructure image source
- Migrate to system-upgrade-controller helm chart
- Display net and gross with a precision of 2
- Make count width in node names configurable
- Sequentially process examples to mitigate locking issues
- Add link to the example
- Name load balancing challenges when zone outages occur
- Use read only token to fetch prices
- Add multi-region example and improve default example
- Correct typos
- Quote ansible variable contents
- Add support for multi-region deployments
- Support image configuration per node pool
- Make network zone configurable
- Ignore ansible inventory
- Update version constraints
- Open all relevant kubernetes ports in the firewall
- Make kubectl exec and logs commands work
- Add ansible integration as feature
- Add examples to verify cilium and cluster configuration
- Describe how to add an ingress controller
- Add table of contents
- Add references to updating additional components
- Add links to charts
- Auto-generate ansible inventory and describe usage
- Add warning to auto-generated files
- Add anchor to the getting started guide
- Add instructions for using the example
- Rename Debugging section to Troubleshooting
- Place examples in subfolder so they're found by the registry
- Pass only the read only token to the gateway script
- Correct spelling
- Correct null error when retrieving gateway labels
- Require a second read only access token for the gateway
- Restore the life cycle ignore settings for image and location
- Remove KUBECONFIG setting from scripts
- Clarify OS upgrade instructions
- Clarify installation instructions and rework feature list
- Make resources immune to changes to cluster_name
- Add image and location information to output
- Improve SSH example configuration
- Add missing opentofu dependency
- Add s3 backend and etcd snapshot configuration
- Print changes before confirming the release
- Remove warning about hard coded gateway
- Simplify instructions for OS upgrades of the main node pool
- Replace sleep wait with cluster readiness check script
- Create a github release for every new version
- Update copyright and add license identifiers to files
- Remove last bit of s3 configuration
- Remove s3 sample configuation
- Update infrastructure picture
- Add link to github stars counter
- Cleanup links
- Correct typos and add hcloud-k3s reference
- Install metrics server via helm
- [breaking] Integrate control_plane_main into node_pools
- Update husky
- Replace husky with a plain githooks folder
- Add validation for node_pools
- Replace hard-coded default gateway with a dynamic calculation
- Reorder firewall rules in gateway
- Manage all scripts in scripts.tf
- Use terraform registry module as the default source
- Add missing k3s --cloud-init switch
- Shorten link to terraform module
- Make runcmd debugging use sh -e
- Reorder documentation
- Document load balancer annotations
- Add a to be added section
- Add changelog reference
- Use haproxy on gateway to proxy port 6443
- Show ports and protocols in architecture picture
- Add node_count to pools and total costs output and optimize label output
- Remove mandatory newline
- Change schedule to start at 1am
- Change no workload taint back to CriticalAddonsOnly
- Update description and lower suggested node size
- Syntax error in installation scirpt
- Improve k8s upgrade
- Update output descriptions
- Add badges and link to terraform registry
- List more features
- Remove duplicated section Addons included
- Add variable for passing k3s options to main server
- Ignore changes to server ssh_keys and location
- Add and document restore and os upgrades
- Apply and document security hardening
- Add system upgrade controller
- Allow inbound ICMP
- Prefix firewall with cluster name
- Remove local cloud init file
- Group all setup commands
- Make references absolute to work on registry.terraform.io
- Make version bump commit unconventional to not appear in the changelog
- Correct apiVersion of kustomization
- Reactivate flannel
- Reenable calico and correct network configuration
- Disable batch mode, it fails for proxied connections
- Correct workload scheduling variable meaning
- Disable IPv6 to make calico work
- Disable bgp
- Remove k8s firewall rules from gateway
- Add output of total monthly costs
- Replace calico with cilium
- Increase inotify limits to make log streaming work
- Remove dependency on HCLOUD_TOKEN variable
- Update documentation to include floating IPs
- Add example for running custom scripts with cloud-init
- Add link to channel list
- List related projects
- Add more debugging tools
- Increase server type
- Add warning to pod IP network configuration
- Add maintenance tasks and additional docs
- Add more tasks, add special thanks section
- Add features section
- Update node pool feature description
- Correct reference to hcloud_token variable
- Adds firewall configuration
- Add feature to create floating IPs
- Use internal network for cluster networking
- Add feature to run scripts with cloud-init
- Add more type information
- Output IPv6 addresses
- Add Nürnberg datacenter and location
- Set hcloud version
- Replace csi/ccm manifests with links to upstream
- Add support for k3s_version
- Update to version 1.32.2
- Add variable to specify version
- Replace flannel to support network policies
- Make installation optional
- Prevent deletion of servers and ignore user_data changes
- [breaking] Add support for multiple control plane nodes
- Make taint and additional initialization options configurable
- [breaking] Disable prevent destroy to improve usability
- [breaking] Change storage class' reclaim policy to Retain
- [breaking] Overhaul and simplify configuration
- Add cost calculation
- Make script wait until SSH connection is possible
- Add kured reboot daemon
- Add release task
- Update hetzner addon manifests
- Reduce modules by putting the configuration in /, similar to k-andy
- Bump hcloud version
- Correct name of pool file
- Remove last references to the load balancer
- Add labels also to nodes
- Ignore control plane user data file
- Remove unused control_plane file
- Update module source reference
- Move update command to into a local variable
- Install ccm and csi via helm
- [breaking] Change control plane taint to node-role.kubernetes.io/control-plane
- Remove unused ip_offset variable
- [breaking] Make ccm and csi mandatory, correct name of chart version vars
- Add commit linter and nix package references
- Bump version
- Mention that the script requires jq
- Replace template_file with templatefile
- Correct markdownlint errors
- Reindent configuration code