project-team-guide/doc/source/ptl.rst

7.6 KiB

Project Team Lead (PTL) Responsibilities

So you want to be a PTL? Here is a list of items that you can expect to perform. This document is meant to serve as a general guide for current and future PTLs. It is not all encompassing, nor are all sections required, it is simply a guide that you may refer to. Each OpenStack project is different and will release, plan, and delegate differently, too.

Weekly Tasks

  1. Organize the team meeting agenda. Most teams will use a wiki or etherpad.

  2. Follow the weekly [release] guidelines email to keep track of the development cycle tasks (unless you have an appointed release liaison to follow it for you). It is also very useful to subscribe into the release calendar.

  3. Attend Technical Committee and Cross-Project meetings when possible. For more information see meetings.

  4. If the team does not have an appointed bug czar, then perform a bug triage. Ensure all new bugs that have been reported are triaged. The below is a URL to view all new bugs for keystone.

    https://bugs.launchpad.net/keystone/+bugs?orderby=status&start=0
  5. If the team does not have an appointed bug czar, then remember to also tag bugs accordingly. The below is a URL to view all untagged bugs for keystone.

    https://bugs.launchpad.net/keystone/+bugs?field.tag=-*

Core member maintenance

This will vary greatly from team to team, but here is a general guide you may consider.

Criteria when considering new cores

A successful appointment can be made solely by the PTL, however it is encouraged to hear the opinions of the existing core team. This can be done through private messages on IRC or publicly through the openstack-dev mailing list.

  1. Are their review numbers (quantity and disagreements) similar to those of existing core members?
  2. Does the person follow up on reviews?
  3. Are the reviews helpful, critical, and nice to authors?
  4. Does the person help in triaging bugs? (For example: reproducing the issue)
  5. Does the person provide bug fixes?
  6. Does the person create and implement new features?
  7. Does the person review specifications?
  8. Does the person participate in weekly meetings?
  9. Does the person participate in the forum discussions?
  10. Does the person participate at the mid-cycle / project-team-gathering?
  11. Does the person know the project direction and priorities?
  12. Does the person help others on IRC?

Actions to perform when adding new cores

Once a person is ready, perform the following:

  1. Announce new core member on mailing list

  2. Give them voice on IRC (if applicable):

    /msg chanserv flags <room> <irc_handle> +V
  3. Add them to the project drivers group in launchpad.

  4. Update the core team in Gerrit.

Criteria when removing new cores

  1. Is the core reviewer reviewing enough? Use data (i.e. stackalytics) to determine if the core reviewer is among the other core members in terms of overall review count.
  2. Does the core reviewer have enough time to review? Job responsibilities may have changed.
  3. Does the core reviewer no longer have knowledge of the code base and new features?
  4. Does the core reviewer still provide useful feedback, even if infrequently?

At the beginning of a new cycle

  1. Add placeholder migrations, these should be the first things to merge once the n-1 stable branch has been created. Do not merge any migrations before this is done
  2. Appoint cross project liaisons (Docs, release, QA, Oslo, etc.).
  3. Appoint FirstContact SIG liaisons (By default, the liaison will be the PTL).
  4. Check if the meeting time works for most active contributors.
  5. If you keep release liaison responsibilities, join #openstack-release and be sure to follow [release] emails from the mailing list.
  6. If necessary, squash database migrations. This is usually not necessary, but will reduce the amount of time necessary for upgrading. A sample squash can be seen by the keystone team.
  7. Track removed and deprecated features, for example using deprecated-as-of-<series> and removed-as-of-<series> blueprints.
  8. Organize the specifications page.

During the cycle

  1. Help first time contributors, they will be struggling the most.
  2. Lack of reviews? Reach out to the core team and remind them.
  3. Release libraries as necessary but don't wait too long! Some teams will release after 4 weeks even if the changes are minor. More often is better than less often.

Before the Forum

  1. Start an etherpad to brainstorm potential session topics. For example: http://lists.openstack.org/pipermail/openstack-dev/2017-March/114123.html
  2. Based on that brainstorming, propose sessions. Create an etherpad for every session, prime the content. List these etherpads in the Wiki.

During the Forum

  1. Reach out to potential new contributors to the project, participate in project on-boarding sessions.
  2. Attend as many cross-project sessions as possible.
  3. In the discussion sessions you moderate:
    • Take notes on the etherpad (or delegate a scribe)
    • Act as a moderator rather than actively participate (or delegate a moderator)
  4. After the discussion, post a summary of the session outcome to the ML, for the benefit of those who could not be present.

At the end of the cycle

  1. Clean up release notes.
  2. Coordinate with the release management team for deliverables, unless a liaison has been appointed
  3. Expect queries from the release marketing staff to name release highlights and major features
  4. Perform a retrospective via an etherpad. Suggested sections include: What went well?, What didn't go well.
  5. Analyze how complete each new feature is. Does it have DevStack support? Horizon support? Client bindings? CLI support? Documentation? Does the install guide need to be updated?

Before the PTG

  1. Decide if your team will hold a team meeting at the PTG, and communicate with the events organizers
  2. If your team gathers at the PTG, create an etherpad to dynamically build the room agenda, and list it on the event wiki page.

During the PTG

  1. Be flexible, attend inter-project sessions as appropriate.
  2. Keep the event schedule up to date on what the current topics of discussion in your team room is.

Stable

Alternatively, the responsibilities in this section can be delegated to a local stable maintenance czar.

  1. Ensure the stable branches gates are not broken.
  2. Co-ordinate with the stable release team to ensure releases are performed when a critical fix is backported, or sufficient smaller fixes have landed.

One offs

When necessary, the following can be performed at unscheduled times.

  1. Bug smashes
  2. API sprints