As part of our regular upgrade cycle, the following Kubernetes cluster components have been updated and have been rolled out to all our managed clusters. Highlight of this update is Loki v3.0, which brings a lot of new features and performance improvements, one of which is the enabling of query results and chunk caching. If you haven’t done yet, please also make sure to verify the actions to take from our previous changelog regarding Grafana deprecations!
Most other updates focus on resolve CVEs, bug fixes and general improvements. Click each link to go to the full upstream changelog.
- amazon-eks-ami v20240506
- aws-ebs-csi-driver v1.30.0-eksbuild.1
- Add retry manager to reduce RateLimitExceeded errors
- aws-efs-csi-driver v2.0.2
- aws-vpc-cni v1.18.1-eksbuild.1
- cert-manager v1.14.5
- cluster-autoscaler v1.29.2
- coredns v1.11.1-eksbuild.9
- dex v2.39.1
- external-dns v0.14.1
- fluent-bit v3.0.3
- grafana-agent v0.40.4
- jaeger-operator v1.52.0
- karpenter v0.36.1
- keda v2.14.0
- kube-prometheus-stack chart v58.5.1
- This brings updates to alertmanager, prometheus-operator, prometheus, grafana, etc.
- grafana v10.4.1
- AngularJS plugin warnings in dashboards, see Actions to take section in previous changelog!
- Many data visulalization quality of life improvements
- Simplified alert notification routing
- prometheus v2.52.0
- prometheus-operator v0.73.2
- kube-proxy v1.29.3-eksbuild.2
- loki v3.0.0
- Major release!
- We are now including chunks and query result caching (via memcached) by default
- Pattern match filter operators
- Docs improvements
- Query acceleration with Bloom filters (experimental, not yet enabled)
- nvidia-k8s-device-plugin v0.15.0
- prometheus-blackbox-exporter
- secrets-store-csi-driver v1.4.3
- vertical-pod-autoscaler v1.1.1
Removals and deprecations:
- We have removed the optional kubernetes-dashboard feature from our clusters as it was no longer used by our customers, while we recommend client-side tools like k9s.