Managed Apps
Changed
- Reduced default resource requests to former profile
small (at least 500m of CPU and 600Mi of memory) and let HPA care about scaling.
Removed
- Support for
cluster.profile parameter. This parameter was not set on either management clusters nor workload clusters and so the default resource requests configured in controller.resources got used.
Fixed
- Fix prometheus metric path in the annotation
Added
- Set team annotation in Chart.yaml for alert routing.
Changed
- Align with and upgrade to upstream
stable-2.11.2. (#59)
Changed
- Downgrade fluent-bit to 1.7.9.
Fixed
- Fix missing leading slash on s3 bucket prefix.
Fixed
- Scrape only the alertmanager http port and not the gossip protocol.
Fixed
- Scrape only the alertmanager http port and not the gossip protocol.
Changed
- Upgrade vertical-pod-autoscaler to 0.10.0
API changes:
Added support for alternative recommenders.
Added support for per-VPA Object MinReplicas.
Other notable changes:
Added support for running VPA out of cluster.
Use v1 API for storage instead of v1beta2.
Allow configuring default update threshold.
Use v1 API to register admission webhook.
Bug fixes:
Use correct timestamp for checkpoints.
Issues with setting limits.
Deploying VPA in different namespaces.
Loading history.
Fixed
- Fix admission-controller webhook-service name.
- Fix webhook name in generated secret certificate.
- Prefix generated secret certificate with release-name.
Added
- Add support for
DNSEndpoint CRs. See README for further information.
Added
- Add giantswarm monitoring label to alertmanager.
Added
- Add giantswarm monitoring label to alertmanager.
