[openstack-dev] [Openstack-sigs][Openstack-operators][all]Expose SIGs/WGs as single window for Users/Ops scenario

Rico Lin rico.lin.guanyu at gmail.com
Fri Sep 14 21:57:22 UTC 2018


The Idea has been raising around (from me or from Matt's ML), so I would
like to give people more update on this (in terms of what I have been
raising, what people have been feedbacks, and what init idea I can collect
or I have as actions.

*Why are we doing this?*
    The basic concept for this is to allow users/ops get a single window
for important scenario/user cases or issues (here's an example [1])into
traceable tasks in single story/place and ask developers be responsible (by
changing the mission of government policy) to co-work on that task.
SIGs/WGs are so desired to get feedbacks or use cases, so as to project
teams (not gonna speak for all projects/SIGs/WGs but we like to collect for
more idea for sure). And the project team got a central place to develop
for specific user requirements (Edge, NFV, Self-healing, K8s). One more
idea on this is that we can also use SIGs and WGs as a place for
cross-project docs and those documents can be some more general information
on how a user can plan for that area (again Edge, NFV, Self-healing, K8s).
There also needs clear information to Users/Ops about what's the dependency
cross projects which involved. Also, a potential way to expose more
projects. From this step, we can plan to cross-project gating ( in projects
gate or periodic) implementation

*So what's triggering and feedback:*

   - This idea has been raising as a topic in K8S SIG, Self-healing SIG
   session. Feedback from K8s-sig and Self-healing-sig are generally looking
   forward to this. SIGs appears are desired to get use cases and user issues
   (I didn't so through this idea to rest SIGs/WGs yet, but place leave
   feedback if you're in that group). Most because this can value up SIGs/WGs
   on what they're interesting on.
   - This idea has been raising as a topic in Ops-meetup session
   Most of ops think it will be super if actually anyone willing to handle
   their own issues. The concerns about this are that we have to make some
   structure or guidelines to avoid a crazy number of useless issues (maybe
   like setup template for issues). Another feedback from an operator is
   that he concerns about ops should just try to go through everything in
   detail by themselves and contact to teams by themselves. IMO it depends on
   teams to set template and say you must have some specific information or
   even figure out which project should be in charge of which failed.
   - This idea has been raising as a topic in TC session
   Public cloud WGs also got this idea as well (and they done a good job!),
   appears it's a very preferred way for them. What happens to them is public
   cloud WG collect bunch number of use cases, but would like to see immediate
   actions or a traceable way to keep tracing those task.
   Doug: It might be hard to push developers to SIGs/WGs, but SIGs/WGs can
   always raise the cross-project forum. Also, it's important to let people
   know that who they can talk to.
   Melvin: Make it easier for everyone, and give a visibility. How can we
   possible to make one thing done is very important.
   Thierry: Have a way to expose the top priority which is important for
   OpenStack.

   - Also, raise to some PTLs and UCs. Generally good, Amy (super cute UC
   member) do ask the concern about there are manual works to bind tasks to
   cross bug tracing platform (like if you like to create a story in
   Self-healing SIG, and said it's relative to Heat, and Neutron. you create a
   task for Heat in that story, but you need to create a launchpad bug and
   link it to that story.). That issue might in now still need to be manually
   done, but what we might able to change is to consider migrate most of the
   relative teams to a single channel in long-term. I didn't get the chance to
   reach most of PTLs but do hope this is the place PTLs can also share their
   feedbacks.
   - There are ML in Self-healing-sig [2]
   not like a lot of feedback to this ML, but generally looks good


*What are the actions we can do right away:*

   - Please give feedback to us
   - Give a forum for this topic for all to discuss this (I already add a
   brainstorm in TC etherpad, but it's across projects, UCs, TCs, WGs, SIGs).
   - Set up a cross-committee discuss for restructuring missions to make
   sure teams are responsible for hep on development, SIGs/WGs are responsible
   to trace task as story level and help to trigger cross-project discussion,
   and operators are responsible to follow the structure to send issues and
   provide valuable information.
   - We can also do an experiment on try on SIGs/WGs who and the relative
   projects are willing to join this for a while and see how the outcomes and
   adjust on them.
   - Can we set cross-projects as a goal for a group of projects instead of
   only community goal?
   - Also if this is a nice idea, we can have a guideline for SIGs/WGs to
   like suggest how they can have a cross-project gate, have a way to let
   users/ops to file story/issue in a format that is useful, or how to trigger
   the attention from other projects to join this.


These are what I got from PTG, but let's start from here together and
scratch what's done shall we!!

P.S. Sorry about the bad writing, but have to catch a flight.

[1] https://storyboard.openstack.org/#!/story/2002684
[2]
http://lists.openstack.org/pipermail/openstack-sigs/2018-July/000432.html

-- 
May The Force of OpenStack Be With You,

*Rico Lin*irc: ricolin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20180914/59dbd915/attachment.html>


More information about the OpenStack-dev mailing list