-
What work did the SIG do this year that should be highlighted?
- Storage layer test unification (kubernetes/kubernetes#109831)
- CEL for admission control to alpha1 (kubernetes/enhancements#3488)
- Aggregated Discovery to alpha (kubernetes/enhancements#3352)
- StorageVersions revived interest from sig-auth (kubernetes/enhancements#2339)
- APIServer identity, revived interest from sig-auth (kubernetes/enhancements#1965)
- CRD CEL validation to beta (kubernetes/enhancements#2876)
- Efficient watch resumption to stable (kubernetes/enhancements#1904)
-
What initiatives are you working on that aren't being tracked in KEPs?
- Contacting individuals to widen the reviewer pool (including, not limited to kubernetes/kubernetes#113904, kubernetes/kubernetes#113959 as examples)
- Twice a week issue/PR triage meetings.
- Storage layer test unification (kubernetes/kubernetes#109831)
-
KEP work in 2022 (v1.24, v1.25, v1.26):
- alpha:
- 3488 - CEL for Admission Control - v1.26
- 3352 - Aggregated Discovery - v1.26
- 3156 - HTTP3 - v1.24
- beta:
- 1965 - kube-apiserver identity - v1.26
- 2876 - CRD Validation Expression Language - v1.25
- 2885 - Server Side Unknown Field Validation - v1.25
- 2896 - OpenAPI V3 - 1.24
- stable:
-
What areas and/or subprojects does your group need the most help with? Any areas with 2 or fewer OWNERs? (link to more details)
- There are some packages that API Machinery owns and come out usually in our triage meetings, and that we most likely don't know much about: this happens often when Kubernetes is upgrading libraries for example - dependency updates span multiple SIGs due to which the entire PR also comes under the purview of API Machinery.
- Technical support for triages could have improvements, for example sometimes we remove API Machinery in PRs, and every update to the issue / PR get's us re-tagged again.
- The ecosystem of the different Kubernetes Clients that we own grows more or less organically. Client-go and Python-client are probably the bigger ones.
- The subprojects that have a total reviewer + approver count <= 2 across all its OWNERS files are as follows:
- https://github.com/kubernetes-client/c (count: 2)
- https://github.com/kubernetes-client/go-base (count: 2)
- https://github.com/kubernetes-client/perl (count: 1)
- There also exist parts of a subproject that come under review/approver crunch. One example is the
cacher
component. Components like this come under a subproject with enough OWNERS but the particular component itself might be dependent on the expertise of a small subset of these OWNERS.
-
What metrics/community health stats does your group care about and/or measure?
On the technical health of the SIG, we look at the following metrics tracked in Devstats:
- Ratio of open/close PRs
- Ratio of open/close issues
- Overall age of open issues
- Number of active contributors to the SIG
On the inclusion health of the SIG, we look at:
- Representation of diversity and of multiple companies in the SIG participants
-
Does your CONTRIBUTING.md help new contributors engage with your group specifically by pointing to activities or programs that provide useful context or allow easy participation?
- No, one can be created. The README points contributors to SIG meetings and triage meetings that happens twice a week.
-
If your group has special training, requirements for reviewers/approvers, or processes beyond the general contributor guide, does your CONTRIBUTING.md document those to help existing contributors grow throughout the contributor ladder?
- No, it can be improved.
-
Does the group have contributors from multiple companies/affiliations?
- Yes, there are contributors from multiple companies. We see all sorts of contributions, varying from issues, to comments, to PRs, to designs, to sig meeting participation, and user-survey data.
-
Are there ways end users/companies can contribute that they currently are not? If one of those ways is more full time support, what would they work on and why?
- There currently isn't a charted course for end users/companies to get more involved.
- The most common, albeit ad-hoc, method has been showing up in the SIG meeting and starting/continuing discussions on the slack channel.
- There is plenty of areas were the SIG could use help in reviewing and maintaining the code, but it's a complicated code base to ramp up. If people don't stick around long enough, the effort does not make sense.
- Primary slack channel member count: 3912
- Primary mailing list member count: 771
- Primary meeting attendee count (estimated, if needed): 30
- Primary meeting participant count (estimated, if needed): 10
- Unique reviewers for SIG-owned packages: 111
- Unique approvers for SIG-owned packages: 94
Include any other ways you measure group membership
Continuing:
- component-base
- control-plane-features
- idl-schema-client-pipeline
- json
- kubernetes-clients
- server-api-aggregation
- server-binaries
- server-crd
- server-frameworks
- server-sdk
- universal-machinery
- yaml
Continuing:
- API Expression
- Multitenancy
- Structured Logging
Operational tasks in sig-governance.md:
- README.md reviewed for accuracy and updated if needed
- CONTRIBUTING.md reviewed for accuracy and updated if needed (or created if missing and your contributor steps and experience are different or more in-depth than the documentation listed in the general contributor guide and devel folder.)
- Subprojects list and linked OWNERS files in sigs.yaml reviewed for accuracy and updated if needed
- SIG leaders (chairs, tech leads, and subproject owners) in sigs.yaml are accurate and active, and updated if needed
- Meeting notes and recordings for 2022 are linked from README.md and updated/uploaded if needed
- Did you have community-wide updates in 2022 (e.g. community meetings, kubecon, or kubernetes-dev@ emails)? Links to email, slides, or recordings: - KubeCon + CloudNativeCon NA 2022 Maintainer Track