Skip to content

Latest commit

 

History

History
210 lines (180 loc) · 12.6 KB

GOVERNANCE.md

File metadata and controls

210 lines (180 loc) · 12.6 KB
Raw

Process for becoming a maintainer

Your organization is not yet a maintainer

  • Express interest to the senior maintainers that your organization is interested in becoming a maintainer. Becoming a maintainer generally means that you are going to be spending substantial time (>25%) on Envoy for the foreseeable future. You should have domain expertise and be extremely proficient in C++. Ultimately your goal is to become a senior maintainer that will represent your organization.
  • We will expect you to start contributing increasingly complicated PRs, under the guidance of the existing senior maintainers.
  • We may ask you to do some PRs from our backlog.
  • As you gain experience with the code base and our standards, we will ask you to do code reviews for incoming PRs (i.e., all maintainers are expected to shoulder a proportional share of community reviews).
  • After a period of approximately 2-3 months of working together and making sure we see eye to eye, the existing senior maintainers will confer and decide whether to grant maintainer status or not. We make no guarantees on the length of time this will take, but 2-3 months is the approximate goal.

Your organization is currently a maintainer

  • First decide whether your organization really needs more people with maintainer access. Valid reasons are "blast radius", a large organization that is working on multiple unrelated projects, etc.
  • Contact a senior maintainer for your organization and express interest.
  • Start doing PRs and code reviews under the guidance of your senior maintainer.
  • After a period of 1-2 months the existing senior maintainers will discuss granting "standard" maintainer access.
  • "Standard" maintainer access can be upgraded to "senior" maintainer access after another 1-2 months of work and another conference of the existing senior committers.

Maintainer responsibilities

  • Monitor email aliases.
  • Monitor Slack (delayed response is perfectly acceptable).
  • Triage GitHub issues and perform pull request reviews for other maintainers and the community. The areas of specialization listed in OWNERS.md can be used to help with routing an issue/question to the right person.
  • Triage build issues - file issues for known flaky builds or bugs, and either fix or find someone to fix any main build breakages.
  • During GitHub issue triage, apply all applicable labels to each new issue. Labels are extremely useful for future issue follow up. Which labels to apply is somewhat subjective so just use your best judgment. A few of the most important labels that are not self explanatory are:
    • beginner: Mark any issue that can reasonably be accomplished by a new contributor with this label.
    • help wanted: Unless it is immediately obvious that someone is going to work on an issue (and if so assign it), mark it help wanted.
    • question: If it's unclear if an issue is immediately actionable, mark it with the question label. Questions are easy to search for and close out at a later time. Questions can be promoted to other issue types once it's clear they are actionable (at which point the question label should be removed).
  • Make sure that ongoing PRs are moving forward at the right pace or closing them.
  • Participate when called upon in the security release process. Note that although this should be a rare occurrence, if a serious vulnerability is found, the process may take up to several full days of work to implement. This reality should be taken into account when discussing time commitment obligations with employers.
  • In general continue to be willing to spend at least 25% of ones time working on Envoy (~1.25 business days per week).
  • We currently maintain an "on-call" rotation within the maintainers. Each on-call is 1 week. Although all maintainers are welcome to perform all of the above tasks, it is the on-call maintainer's responsibility to triage incoming issues/questions and marshal ongoing work forward. To reiterate, it is not the responsibility of the on-call maintainer to answer all questions and do all reviews, but it is their responsibility to make sure that everything is being actively covered by someone.
  • The on-call rotation is tracked at Opsgenie. The calendar is visible here or you can subscribe to the iCal feed here

Cutting a release

  • We do releases every 3 months, at the end of each quarter, as described in the release schedule.
  • Take a look at open issues tagged with the current release, by searching for "is:open is:issue milestone:[current milestone]" and either hold off until they are fixed or bump them to the next milestone.
  • Begin marshalling the ongoing PR flow in this repo. Ask maintainers to hold off merging any particularly risky PRs until after the release is tagged. This is because we aim for main to be at release candidate quality at all times.
  • Do a final check of the release notes:
    • Make any needed corrections (grammar, punctuation, formatting, etc.).
    • Check to see if any security/stable version release notes are duplicated in the major version release notes. These should not be duplicated.
    • If the "Deprecated" section is empty, delete it.
    • Remove the "Pending" tags and add dates to the top of the release notes for this version.
    • Switch the VERSION from a "dev" variant to a final variant. E.g., "1.6.0-dev" to "1.6.0".
    • Update the RELEASES doc with the relevant dates. Now, or after you cut the release, please also make sure there's a stable maintainer signed up for next quarter, and the deadline for the next release is documented in the release schedule.
    • Get a review and merge.
  • Wait for tests to pass on main.
  • Create a tagged release. The release should start with "v" and be followed by the version number. E.g., "v1.6.0". This must match the VERSION.
  • From the envoy landing page use the branch drop-down to create a branch from the tagged release, e.g. "release/v1.6". It will be used for the stable releases.
  • Monitor the AZP tag build to make sure that the final docker images get pushed along with the final docs. The final documentation will end up in the envoyproxy.github.io repository.
  • Update the website (example PR) for the new release.
  • Craft a witty/uplifting email and send it to all the email aliases including envoy-announce@.
  • Make sure we tweet the new release: either have Matt do it or email [email protected] and ask them to do an Envoy account post.
  • Do a new PR to setup the next version
    • Update VERSION to the next development release. E.g., "1.7.0-dev".
    • git mv docs/root/version_history/current.rst docs/root/version_history/v1.6.0.rst, filling in the previous release version number in the filename and delete empty sections (like Incompatible Behavior Changes, Minor Bahavior Changes, etc). Add an entry for the new file in the toctree in version_history.rst.
    • Create a new "current" version history file at the release notes for the following version. E.g., "1.7.0 (pending)". Use this text as the template for the new file:
1.7.0 (Pending)
===============

Incompatible Behavior Changes
-----------------------------
*Changes that are expected to cause an incompatibility if applicable; deployment changes are likely required*

Minor Behavior Changes
----------------------
*Changes that may cause incompatibilities for some users, but should not for most*

Bug Fixes
---------
*Changes expected to improve the state of the world and are unlikely to have negative effects*

Removed Config or Runtime
-------------------------
*Normally occurs at the end of the* :ref:`deprecation period <deprecated>`

New Features
------------

Deprecated
----------
  • Run the deprecate_versions.py script (e.g. sh tools/deprecate_version/deprecate_version.sh) to file tracking issues for runtime guarded code which can be removed.
  • Check source/common/runtime/runtime_features.cc and see if any runtime guards in disabled_runtime_features should be reassessed, and ping on the relevant issues.

When does a maintainer lose maintainer status

If a maintainer is no longer interested or cannot perform the maintainer duties listed above, they should volunteer to be moved to emeritus status. In extreme cases this can also occur by a vote of the maintainers per the voting process below.

xDS API shepherds

The xDS API shepherds are responsible for approving any PR that modifies the api/ tree. They ensure that API style and versioning policies are enforced and that a consistent approach is taken towards API evolution.

The xDS API shepherds are also the xDS API maintainers; they work collaboratively with the community to drive the xDS API roadmap and review major proposed design changes. The API shepherds are intended to be representative of xDS client and control plane developers who are actively working on xDS development and evolution.

As with maintainers, an API shepherd should be spending at least 25% of their time working on xDS developments and expect to be active in this space in the near future. API shepherds are expected to take on API shepherd review load and participate in meetings. They should be active on Slack #xds and responsive to GitHub issues and PRs on which they are tagged.

The API shepherds are distinct to the UDPA working group, which aims to evolve xDS directionally towards a universal dataplane API. API shepherds are responsible for the execution of the xDS day-to-day and guiding xDS implementation changes. Proposals from UDPA-WG will be aligned with the xDS API shepherds to ensure that xDS is heading towards the UDPA goal. xDS API shepherds operate under the envoyproxy organization but are expected to keep in mind the needs of all xDS clients (currently Envoy and gRPC, but we are aware of other in-house implementations) and the goals of UDPA-WG.

If you wish to become an API shepherd and satisfy the above criteria, please contact an existing API shepherd. We will factor in PR and review history to determine if the above API shepherd requirements are met. We may ask you to shadow an existing API shepherd for a period of time to build confidence in consistent application of the API guidelines to PRs.

Extension addition policy

Adding new extensions has a dedicated policy. Please see this document for more information.

External dependency policy

Adding new external dependencies has a dedicated policy. Please see this document for more information.

Conflict resolution and voting

In general, we prefer that technical issues and maintainer membership are amicably worked out between the persons involved. If a dispute cannot be decided independently, the maintainers can be called in to decide an issue. If the maintainers themselves cannot decide an issue, the issue will be resolved by voting. The voting process is a simple majority in which each senior maintainer receives two votes and each normal maintainer receives one vote.

Adding new projects to the envoyproxy GitHub organization

New projects will be added to the envoyproxy organization via GitHub issue discussion in one of the existing projects in the organization. Once sufficient discussion has taken place (~3-5 business days but depending on the volume of conversation), the maintainers of the project where the issue was opened (since different projects in the organization may have different maintainers) will decide whether the new project should be added. See the section above on voting if the maintainers cannot easily decide.