This is a Work in Progress, documenting approximately how we have been operating up to this point.
The Kubernetes community adheres to the following principles:
- Open: Kubernetes is open source. See repository guidelines and CLA, below.
- Welcoming and respectful: See Code of Conduct, below.
- Transparent and accessible: Work and collaboration should be done in public. See SIG governance, below.
- Merit: Ideas and contributions are accepted according to their technical merit and alignment with project objectives, scope, and design principles.
The Kubernetes community abides by the CNCF code of conduct. Here is an excerpt:
As contributors and maintainers of this project, and in the interest of fostering an open and welcoming community, we pledge to respect all people who contribute through reporting issues, posting feature requests, updating documentation, submitting pull requests or patches, and other activities.
As a member of the Kubernetes project, you represent the project and your fellow contributors. We value our community tremendously and we'd like to keep cultivating a friendly and collaborative environment for our contributors and users. We want everyone in the community to have positive experiences.
The project is comprised of the following types of subgroups:
- Special Interest Groups, SIGs
- Subprojects
- Working Groups, WGs
- Committees
- User Groups
The Kubernetes project is organized primarily into Special Interest Groups, or SIGs. Each SIG is comprised of members from multiple companies and organizations, with a common purpose of advancing the project with respect to a specific topic, such as Networking or Documentation. Our goal is to enable a distributed decision structure and code ownership, as well as providing focused forums for getting work done, making decisions, and onboarding new contributors. Every identifiable subpart of the project (e.g., github org, repository, subdirectory, API, test, issue, PR) is intended to be owned by some SIG.
Areas covered by SIGs may be vertically focused on particular components or functions, cross-cutting/horizontal, spanning many/all functional areas of the project, or in support of the project itself. Examples:
- Vertical: Network, Storage, Node, Scheduling, Big Data
- Horizontal: Scalability, Architecture
- Project: Testing, Release, Docs, PM, Contributor Experience
SIGs must have at least one and ideally two SIG chairs at any given time. SIG chairs are intended to be organizers and facilitators, responsible for the operation of the SIG and for communication and coordination with the other SIGs, the Steering Committee, and the broader community.
Each SIG must have a charter that specifies its scope (topics, subsystems, code repos and directories), responsibilities, areas of authority, how members and roles of authority/leadership are selected/granted, how decisions are made, and how conflicts are resolved. See the SIG charter process for details on how charters are managed. A short template for intra-SIG governance has been developed in order to simplify SIG creation, and additional templates are being developed, but SIGs should be relatively free to customize or change how they operate, within some broad guidelines and constraints imposed by cross-SIG processes (e.g., the release process) and assets (e.g., the kubernetes repo).
A primary reason that SIGs exist is as forums for collaboration. Much work in a SIG should stay local within that SIG. However, SIGs must communicate in the open, ensure other SIGs and community members can find notes of meetings, discussions, designs, and decisions, and periodically communicate a high-level summary of the SIG's work to the community.
See sig governance for more details about current SIG operating mechanics, such as mailing lists, meeting times, etc.
Specific work efforts within SIGs are divided into subprojects. Every part of the Kubernetes code and documentation must be owned by some subproject. Some SIGs may have a single subproject, but many SIGs have multiple significant subprojects with distinct (though sometimes overlapping) sets of contributors and owners, who act as subproject’s technical leaders: responsible for vision and direction and overall design, choose/approve change proposal (KEP) approvers, field technical escalations, etc.
Example subprojects for a few SIGs:
- SIG Network: pod networking (CNI, etc.), Service (incl. kube-proxy), Ingress, DNS, and Network policy
- SIG Apps: workload APIs, Helm, Kompose, ...
- SIG Cluster Lifecycle: kubeadm, kops, kubespray, minikube, ...
Subprojects for each SIG are documented in sigs.yaml.
We need community rallying points to facilitate discussions/work regarding topics that are short-lived or that span multiple SIGs.
Working groups are primarily used to facilitate topics of discussion that are in scope for Kubernetes but that cross SIG lines. If a set of folks in the community want to get together and discuss a topic, they can do so without forming a Working Group.
See working group governance for more details about forming and disbanding Working Groups.
Working groups are documented in sigs.yaml.
Some topics, such as Security or Code of Conduct, require discretion. Whereas SIGs are voluntary groups which operate in the open and anyone can join, Committees do not have open membership and do not always operate in the open. The steering committee can form committees as needed, for bounded or unbounded duration. Membership of a committee is decided by the steering committee. Like a SIG, a committee has a charter and a chair, and will report to the steering committee periodically, and to the community as makes sense, given the charter.
Some topics have long term relevance to large groups of Kubernetes users, but do not have clear deliverables or ownership of parts of the Kubernetes code base. As such they are neither good fits for SIGs or Working Groups. An example of such a topic might be continuous delivery to Kubernetes.
Though their central goal is not a a deliverable piece of work, as contributing members of the community working groups are expected to work with SIGs to either identify friction or usability issues that need to be addressed, or to provide or improve documentation in their area of expertise. However these activities are covered under general code contributions to the relevant SIGs (e.g. SIG Docs) rather than as part of the user group. These contributions are expected to be more incremental and ad-hoc versus the more targeted output of a working group.
User groups function as a centralized resource to facilitate communication and discovery of information related to the topic of the user group. User groups should not undertake to produce any deliverable, instead they should form working groups under the auspices of some SIG for such work. Likewise they shouldn't take ownership of anything in the Kubernetes process, as that is a role for SIGs.
To facilitate discoverability and engagement, user groups are documented in sigs.yaml
While most work shouldn’t require expensive coordination with other SIGs, there will be efforts (features, refactoring, etc.) that cross SIG boundaries. In this case, it is expected that the SIGs coordinate with each other and come to mutually agreed solutions. In some cases, it may make sense to form a Working Group for joint work. Cross-SIG coordination will naturally require more time and implies a certain amount of overhead. This is intentional to encourage changes to be well encapsulated whenever possible.
On the other hand, several SIGs do have project-wide impact, for example Release, Testing, and API Machinery. Even those that do not may sometimes need to make changes or impose new processes or conventions that affect other SIGs. In these cases, project-wide communication processes will need to be followed. For example, proposals with project-wide impact will need to be announced more broadly, with the opportunity for members of other SIGs to provide feedback and guidance. However, the SIG that owns the area, according to its charter, will own the decision. In the case of extended debate or deadlock, decisions may be escalated to the Steering Committee, which is expected to be uncommon.
The [KEP process] is being developed as a way to facilitate definition, agreement and communication of efforts that cross SIG boundaries. SIGs are encouraged to use this process for larger efforts. This process is also available for smaller efforts within a SIG.
All new repositories under Kubernetes github orgs should follow the process outlined in the kubernetes repository guidelines.
Note that "Kubernetes incubator" process has been deprecated in favor of the new guidelines.
All contributors must sign the CNCF CLA, as described here.