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.
- References to certificates/v1beta1 preventing the operator from working on K8s 1.22
- Added CHANGELOG.md to track changes across releases
- More information to operator logs (e.g. reconcilerID)
- Support for taints and tolerations
- Support for custom annotations
- Mutating and validating webhooks and associated TLS configuration
- Several enhancements for local development (e.g.
make dev/up)
- Refactor templates to leverage kustomize bases/overlays in //config
- Updated validation markers for the CRD (required, minimum, etc.)
- Introduced Director to consolidate the handles logic from the actors
- Standardized on cockroachdb/errors for wrapping errors
- Made enormous-sized tests run for less long
- Propagate affinity and tolerations for version checker job
- Ensure the controller watches owned job objects
- CertificateGenerated conditional correctly set during deploys
- No longer requeueing permanent errors
- PVCs no longer missing
AdditionalLabels - Some flakiness in e2e tests
- Version checker jobs no longer pile up when pods take >2m to come up
- More information to operator logs (e.g. reconcilerID)
- Refactor templates to leverage kustomize bases/overlays in //config
- Propagate affinity and tolerations for version checker job
- Ensure the controller watches owned job objects
- CertificateGenerated conditional correctly set during deploys
- Support for taints and tolerations
- Support for custom annotations
- Mutating and validating webhooks and associated TLS configuration
- Several enhancements for local development (e.g.
make dev/up)
- Updated validation markers for the CRD (required, minimum, etc.)
- Introduced Director to consolidate the handles logic from the actors
- Standardized on cockroachdb/errors for wrapping errors
- No longer requeueing permanent errors
- PVCs no longer missing
AdditionalLabels - Some flakiness in e2e tests
- Version checker jobs no longer pile up when pods take >2m to come up
AdditionalLabelswhich are added to all managed resources- Examples for new features (addition labels, affinity rules, etc)
- e2e tests for EKS, OpenShift, and downgrading
- Skip the initContainer when running an insecure cluster
- Can now pass additional parameters to OpenShift
- Updating of crdbversions.yaml now automated via GitHub actions
- Dead code from multiple packages
- Dependency on ginko and gomega