Skip to content

Releases: projectcalico/kube-controllers

Version reporting

22 Nov 05:36
Compare
Choose a tag to compare

#59 Policy controller now can report version information

$ docker run calico/kube-policy-controller:v0.5.1 version
v0.5.1

v0.5.0

06 Nov 03:24
Compare
Choose a tag to compare

70593e7 switched to using "default" tier instead of "k8s-network-policy"

Switching tier name to "default"

02 Nov 23:47
Compare
Choose a tag to compare
Pre-release

70593e7 switched to using "default" tier instead of "k8s-network-policy"

Update libcalico and get IP SAN support for TLS

31 Oct 23:58
Compare
Choose a tag to compare

Modified per-namespace policy

17 Aug 22:51
Compare
Choose a tag to compare

Per-namespace policy is now implemented using profiles, rather than policies. This allows users to side-load policy out-of-band from Kubernetes.

Updates libcalico to v0.16.0

v0.2.0: v1beta1 NetworkPolicy API

23 May 23:04
Compare
Choose a tag to compare
Pre-release

Note: this is a non-backwards compatible early release that relies on Kubernetes v1.3 v1beta1 API changes

This release switches Kubernetes api versions from v1alpha1 to v1beta1.

Notable changes:

  • Use the v1beta1 Kubernetes API.
  • The docker image has been renamed from k8s-policy-agent to kube-policy-controller.
  • Removed the policy tool, as v1beta1 NetworkPolicy can be configured via kubectl in Kubernetes v1.3

To use this version, you must be running Kubernetes v1.3 or later and have enabled the NetworkPolicy API in your apiserver:

--runtime-config=v1beta1/extensions=true,v1beta1/extensions/networkpolicies=true

To continue with the v1alpha1 API, use the k8s-policy-agent docker image and documentation for a v0.1.x release.

Performance, Leader Election, Deployment changes

19 Apr 17:12
Compare
Choose a tag to compare

Improved performance in conjunction with changes made to calico-cni v1.3.0. This version of the policy agent is incompatible with versions of the calico-cni prior to v1.3.0.

Added leader election capabilities.

Policy agent should now be deployed using a static manifest on each master node. See here.

#22 Don't configure /etc/hosts by default.
#21 Handle connection errors on startup
#20 Add leader election
#18 Label / endpoint caching for better performance

Fix API re-sync

05 Apr 17:46
Compare
Choose a tag to compare
  • #16 - error when stale resourceVersion used on k8s API
  • #17 - Cache pod labels to prevent unnecessary etcd writes.

Use new libcalico metadata API

23 Mar 00:31
Compare
Choose a tag to compare

Update to use new libcalico metadata API

Update watch resourceVersion

23 Mar 17:42
Compare
Choose a tag to compare

Bugfix: update watch resource versions so that we don't try to read stale objects from the API.