As part of our regular upgrade cycle, the following Kubernetes cluster components have been updated. We’ve already rolled these out to all staging clusters. Production clusters will follow in the coming days.
Most notable changes include bug fixes and minor improvements.
Core functionality
- aws-node-termination-handler v1.7.0 -> v1.11.0
- Preview support for NTH Queue Processing Mode as alternative to IMDS mode. Note, we still use IMDS
- Support for EC2 Instance Rebalance Recommendations
- Reduced NTH startup time across all modes via lowering IMDS retry rate
- Several bug fixes and small improvements
- aws-vpc-cni v1.6.3 -> v1.7.5
- Adds support for Pod Security Groups
- #972 Persist IPAM state to local file and use across restarts
- #955 Add init container to copy
portmapandloopbackbinaries and set uprp_filter
- calico v3.13.4 -> v3.15.4
- kube-proxy v1.17.9 -> v1.17.12
- Many bug and/or regression fixes
- metrics-server v0.3.7 -> v0.4.1
Autoscaling
Monitoring and Logging
- loki v1.6.1 -> v2.0.0. This is a major new release of Loki, it contains multiple performance improvements and optimizations under-the-hood, plus some nice new features, among them:
- Significant enhancements to the LogQL query language
- Generate prometheus alerts directly from your logs! Until now, we needed to configure a Grafana dashboard to trigger alerts from log patterns, with this release, Loki itself can be configured to ship alerts directly to Alertmanager. Stay tunned as we’ll adapt our reference solution so you can configure those alerts yourself.
- grafana v7.1.5 -> v7.3.4
- prometheus v2.20.1 -> v2.22.2
- prometheus-operator v0.40.0 -> v0.43.2
- There’s a new CRD introduced in a recent version:
AlertmanagerConfig, which allows to split the Alertmanager configuration in different objects, much like how Grafana dashboards are managed, or Prometheus rules and targets. This will allow us to streamline our reference solution a bit more, while enabling our customers to better adjust Alertmanager configuration.
- There’s a new CRD introduced in a recent version:
- prometheus-blackbox-exporter v0.17.0 -> v0.18.0
- prometheus-cloudwatch-exporter Helm chart 0.8.0 -> 0.12.0
- prometheus-elasticsearch-exporter Helm chart 2.1.1 -> 4.0.0
- metrics-server 0.3.7 -> 0.4.1
Others
- dex v0.24.0 -> v0.26.0
- Add team groups support to bitbucket connector
- Support username, email and groups claim in OIDC connector
- cert-manager v0.16.1 -> v1.0.4
- v1 API
kubectl cert-manager statuscommand to help with investigating issues- Using new and stable Kubernetes APIs
- Improved logging
- ACME improvements
- external-dns v0.7.3 -> v0.7.4
- Add service annotation to set public/private iface for NodePort
- kubernetes-dashboard v2.0.3 -> v2.0.4
- Add network policy support
- Add role support
- Add role bindings and cluster role bindings support
- nginx-ingress v0.34.1 -> v0.41.2
- Following the Ingress
extensions/v1beta1deprecation, please usenetworking.k8s.io/v1beta1for new Ingress definitions
- Following the Ingress
- oauth2_proxy v6.0.0 -> v6.1.1
- velero v1.4.2 -> v1.5.2
- Auto Volume Backup Using Restic with
--default-volumes-to-resticflag - DeleteItemAction plugins
- Code modernization
- Restore Hooks: InitContianer Restore Hooks and Exec Restore Hooks
- Auto Volume Backup Using Restic with