v0.5

Features

• Syncing all ArgoCD apps at the end of the cluster bootstrap process

• Added support for specifying additional users | Inferring VM specs from the specified type, for each node-group | Codebase refactoring

• Complete cluster upgrade implementation (supporting only AWS for now) | Adding HCloud support (tested)

• Successfully provisioning a cluster in Hetzner Cloud

• Add a command to spin up a dev environment

Chores

• Writing short help message for each leaf CLI command | Fixing Makefile | Small README.md updates

• Removing replicas field for a node-group

• Solved last 2 issues mentioned in issue #25

• Update AWS specific sample Kubeaid Bootstrap Script config file

Fixes :

• Removing monitoring namespace from cluster-vars.jsonnet.tmpl, so that Prometheus doesn't monitor itself

• Fix Disaster Recovery Setup : Link IAM Role and Policy

• Managing ClusterAPI components versions in KubeAid and not here

• Replacing the skip-kubeaid-config-setup flag with skip-kube-prometheus-build flag

• Checking out to default branch when repo already exists | Proper error handling while creating IAM Roles and Policies

• Creating Cilium ArgoCD App and corresponding values file when kubeaid-config gets updated post cluster provisioning

---

Corresponding changes done in KubeAid

• Limited the Infrastructure Provider component's cache to its own namespace.

• Updated ClusterAPI version.

• Fixed scale from 0 support.

• Fixed kubeaid-bootstrap-script #41

Hetzner specific updates :

• Added support for provisioning clusters in Hetzner Cloud (tested).

• Restructured the capi-cluster/hetzner subchart, keeping in mind that we want to mix Hetzner bare-metal and HCloud in the future.

• Fixed postKubeadm commands, which were causing HCloud cluster provisioning to fail.

• Updated ClusterAPI provider Hetzner version.

AWS specific updates :

• Allowing the user to specify the control-plane Loadbalancer scheme.

• Distributing a node-group across multiple Failure Domains (Availability Zones) for high-availability.

• Making it mandatory for the user to specify the min and max size explicitly for each node-group.

• Setting PUT HTTP response hop limit to 2, supporting both IMDSv1 and IMDSv2.

• Using hostNetwork for AWS CCM. Otherwise the NoCredentialsProvider error pops out after some time.

• Added support for specifying additional users, each with their corresponding SSH public key.

• Fixed Sealed Secret keys backuper container image and ServiceAccount name

• Updating part 1 of the KubeAid demonstration blog series for AWS

Ignore changes in the capi-cluster/azure subchart and hetzner-failover-script. They are untested anyways.
updating part 1 of the KubeAid demonstration blog series for AWS
updating part 1 of the KubeAid demonstration blog series for AWS