All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
2.27.0 - 2025-01-21
- Tags defined in subnet spec should be applied to the subnet resource on AWS.
2.26.0 - 2024-12-05
- Support EC2 lifecycle hooks for machine pools
- Build: ensure using the correct variant of yq
2.25.0 - 2024-11-15
- Bump CAPA version
v2.7.1-gs-*
. This bumps CAPA tov2.7.1
and applies the customizations we already had in our fork.
- Filter out AWS internal tags from subnet tags
2.24.0 - 2024-10-17
- Expose the
maxHealthyPercentage
property to allow setting the maximum percentage of healthy machines in the Auto Scaling Group during upgrades.
2.23.0 - 2024-09-16
- Bump CAPA version v2.5.2-gs-11cb106a4. This bumps CAPA to
v2.5.2
and applies the customizations we already had in our fork.
2.22.0 - 2024-09-02
- Bump CAPA version v2.4.2-gs-70b4664c1. This bumps CAPA to
v2.4.2
and apply the customizations we already had in our fork.
2.21.0 - 2024-07-24
- Reduce sync-period to 2 minutes to improve scale-up taint removal.
2.20.0 - 2024-07-16
- Cancel instance refresh on any relevant change to ASG instead of blocking until previous one is finished (which may have led to failing nodes due to outdated join token)
- Try deleting machine pool user data file from S3 when pruning an old launch template version
2.19.1 - 2024-06-25
- Fix for the new feature of writing user data to S3 bucket for
AWSMachinePool
: on cluster deletion, delete machine pool user data files that did not get deleted yet by the S3 bucket lifecycle policy. Otherwise, CAPA would have left the S3 bucket behind on deletion since it was not empty.
2.19.0 - 2024-06-06
- Support for writing user data to S3 bucket for
AWSMachinePool
2.18.1 - 2024-05-24
- Make ServiceMonitor optional through
serviceMonitor.enabled
helm value (useful in mc-bootstrap).
2.18.0 - 2024-05-02
- Support adding custom secondary VPC CIDR blocks in
AWSCluster
(backport)
2.17.0 - 2024-04-25
- Bump CAPA version v2.3.0-gs-16d1f6ed4. Introduces the
NonRootVolumes
field to AWSMachinePools and AWSManagedMachinePools.
2.16.0 - 2024-04-17
- Bump CAPA version v2.3.0-gs-784b17920. This fixes controller not adding ID to secondary subnet when updating AWSCluster.
- Add toleration for
node.cluster.x-k8s.io/uninitialized
taint. - Remove toleration for old
node-role.kubernetes.io/master
taint.
2.15.2 - 2024-04-02
- Fix CRD installation job's settings so it can get admitted by Kyverno policies
2.15.1 - 2024-04-02
- Add ServiceMonitor for monitoring.
- Change container image registry values name to use values from
config
repo.
2.15.0 - 2024-02-20
- Bump version to
v2.3.0-gs-e9c5ab62c
. This version prevents EKS nodes from landing in the CNI subnets.
2.14.0 - 2024-02-07
- Use
app-build-suite
to build the app - Bump version to
v2.3.0-gs-3e18e8bef
. This version adds theDefaultInstanceWarmup
field to AWSMachinePools.
2.13.0 - 2024-01-23
- Backport feature: expose new field to secure VPC default security group)[kubernetes-sigs/cluster-api-provider-aws#4707].
2.12.0 - 2024-01-18
- Delete aws-node resources even when they have
Helm
labels whenAWSManagedControlPlane.spec.VpcCni.disabled
is set totrue
.
2.11.0 - 2024-01-15
- Update CRDs to make the subnet
id
field required again
2.10.1 - 2024-01-10
- Switch to image hosted on gsoci.azurecr.io
2.10.0 - 2024-01-10
-
Backported fixes and features for CAPA v2.3.x
- Use go 1.21.5, fix kubectl version detection after
--short
parameter was removed - Make VPC creation idempotent to avoid indefinite creation of new VPCs if storage of the ID fails
- Log full ARN in GC error messages
- Fix deregistering of deleted CAPI Machines
- ASG: do not set desired value for machinepool which have externally managed replicas
- Use go 1.21.5, fix kubectl version detection after
2.9.0 - 2023-12-21
-
Backported fixes and features for CAPA v2.3.x
- Enable transit encryption to S3 bucket
- Trigger machine pool instance refresh (node rollout) if bootstrap config reference changes
- Skip instance refresh attempt if ASG does not yet exist
2.8.1 - 2023-12-14
- Revert CRDs upgrade since CAPA creates unlimited VPCs for EKS clusters because the subnet
id
field cannot be set (bug)
2.8.0 - 2023-12-13
- Upgrade CAPA CRDs to v2.3.0 including required subnet
id
field
2.7.1 - 2023-12-07
- Configure
gsoci.azurecr.io
as the default container image registry.
2.7.0 - 2023-11-30
- Upgrade CAPA to v2.3.0 but comment out CRD upgrade so existing clusters reconcile fine without having the newly-required subnet
id
field yet
2.6.1 - 2023-11-13
- Bump to CAPA version with the backported feature "Use
AdditionalTags
for S3 buckets"
2.6.0 - 2023-11-08
- Add
global.podSecurityStandards.enforced
value for PSS migration. - Bump to CAPA version with the backported feature "Tag S3 bucket as owned by cluster"
2.5.0 - 2023-10-09
- Upgraded to CAPA v2.2.4.
2.4.0 - 2023-09-19
- Add network policies (
CiliumNetworkPolicy
objects depend onciliumNetworkPolicy.enabled
)
- Upgrade to cluster-api-provider-aws v2.2.1
- Switch to using Giant Swarm cluster-api-provider-aws fork's GitHub releases to make things consistent with the cluster-api fork
2.3.0 - 2023-07-14
- Add necessary values for PSS policy warnings
- Add common labels to pods so that Hubble shows the app name
2.2.0 - 2023-04-19
- Change default registry in Helm chart from quay.io to docker.io.
2.1.0 - 2023-03-31
- Enabled Flatcar ignition bootstrap feature gate.
2.0.2 - 2023-03-22
- Added
node-role.kubernetes.io/control-plane
to crd install jobs toleration
2.0.1 - 2023-03-17
- Add psp and seccomp profile.
2.0.0 - 2023-03-02
- Change
kubectl
image source togiantswarm/kubectl
- Bump
kubectl
version tov1.24.10
- Upgrade image and CRDs to a version newer than v2.0.2 (still including our
LoadBalancerReadyCondition
patch)
1.9.2 - 2023-01-19
- Set Helm chart ownership to team hydra.
- Bumped controller to latest changes from upstream
release-1.5
branch
1.9.1 - 2022-11-30
- Generation of Helm chart will fail if we get a 404 from Github.
- Re-generate helm chart after adading feature gate.
1.9.0 - 2022-11-30
- Enabled external resource gc feature gate.
1.8.3 - 2022-11-22
- Switched to using fork-built image for
v1.5.2
. The fork contains a fix setting the LoadBalancerReadyCondition to Deleted
1.8.2 - 2022-11-21
- Switched to using upstream-built image for
v1.5.2
1.8.1 - 2022-11-11
- Upgraded image used to
v1.5.2-gs
(fork until upstream1.5.2
release is available)
1.8.0 - 2022-11-03
- Add clusterctl labels to CRDs to support
clusterctl move
.
1.7.0 - 2022-10-18
- Don't delete crd-install job when the
Job
fails so that we can inspect what happened.
1.6.0 - 2022-10-04
PodSecurityPolicy
are removed on newer k8s versions, so only apply it in thecrd-install
job if object is registered in the k8s API.
1.5.0 - 2022-08-11
- Bumped to cluster-api-provider-aws v1.5.0
1.2.3 - 2022-08-06
- Remove ssh SSO public key value from the repository as its no longer used.
1.2.2 - 2022-08-05
1.2.1 - 2022-07-07
- Add VerticalPodAutoscaler CR.
- Value to set image registry domain
- Fix CRD path on
make
target.
1.2.0 - 2022-03-09
- Installing CRDs via crd-install job.
- Generating all manifests with kustomize by using upstream manifests a base and applying custom overlays.
- Push to aws-app-collection.
0.6.8-gs10-crd - 2021-11-24
0.6.8-gs10 - 2021-11-24
- Fix selector labels for EKS webhook.
0.6.8-gs9 - 2021-11-24
- Fix match labels to be unique for all deployments and services.
0.6.8-gs8 - 2021-10-11
0.6.8-gs7 - 2021-10-11
0.6.8-gs6 - 2021-10-11
- Fix chart uniqueness.
- Enable EKS in CAPA webhook
0.6.8-gs5 - 2021-09-09
- Fix labels for all deployment and services.
0.6.8-gs4 - 2021-09-02
- Moving helm templating to new config management.
0.6.8-gs3 - 2021-08-31
- Deactivate eks flag only for core aws webhook
0.6.8-gs2 - 2021-08-25
- Activated eks feature flags
0.6.8-gs1 - 2021-08-02
- Updated CAPA to v0.6.8.
0.6.6-gs2 - 2021-07-28
- Add secret to store ssh sso public key for later use.
- Updated CAPA to version to 0.6.6
- Remove kube-rbac-proxy for the metrics endpoint.
0.6.5-gs3 - 2021-05-27
- Changed watch-filter value to
capi
.