►
Description
[SIG ContribEx] GitHub Administration Subproject Bi-Weekly Meeting for 20220825
A
Hi
everyone
and
welcome
to
the
github
admin
meeting-
and
this
is
we
are
august
25
and
just
want
to
remind
the
that
this
meeting
is
under
the
code
of
contract.
So
I
will
I
invite
everyone
in
this
call
to
be
external
to
each
other.
A
Okay,
the
only
thing
I
see
the
agenda,
it's
you
greg.
B
Yeah
I
wanted
to
share
a
discussion
that
I'm
having
with
country
becks
and
just
help
other
people
be
aware
of
some
things
that
are
going
on
there,
just
to
make
sure
that
we're
kind
of
aligned
and
there's
agreement
that
it's
a
sensible
thing
to
do
and
make
sure
I
capture
any
objections
or
feedback
about
the
approach
and
also
see
if
it
spurs
any
other
ideas.
B
That
might
be
interesting.
So
for
anybody
who
doesn't
know
me,
I've
been
in
the
kubernetes
community
for
a
long
time,
but
I
now
work
for
github
prior,
I
was
working
at
microsoft
in
the
kubernetes
team.
B
What
we
called
upstream
so
working
entirely
with
community
and
contributions
back,
and
I
work
now
on
a
product
called
code
spaces
and
anybody
familiar
with
codespace
is
awesome.
If
not
quick,
real
quick
introduction
is
it's
cloud-based
development
environments,
and
so,
if
you
may
have
heard
of
it
originally
as
vs
code
online,
it
was
came
out
of
that
effort.
B
On
the
microsoft
side,
it
was
handed
over
to
github
to
make
it
much
more
agnostic
and
focused
less
on
enterprise
use
cases,
and
my
mission
at
github
is
to
enable
code
spaces
for
everyone,
and
specifically,
my
role
is
enabling
open
source
communities
to
take
advantage
of
this
new
capability
in
github,
and
there
are
two
major
use
cases
that
are
kind
of
users
that
I
focus
on
benefiting
one
is
maintainers
and
the
other
are
new
contributors
and
they're
very
complementary.
B
The
story
is
that
in
a
code,
space
there's
a
default
experience
that
actually
works
for
lots
of
different
projects
or
there's
a
highly
customized
experience
that
is
tailored
perfectly
to
a
project
that
encapsulates
encoding
all
of
the
dependencies
and
startup
requirements
in
a
spec
called
the
dev
container
and
then
codespaces
reads
that
launches
a
codes.
You
know
this
dev
environment
with
a
pre-built
container
that
can
be
a
pre-built
container
or
it
builds
on
the
flyout
container.
B
Obviously,
pre-built
go
much
faster,
but
basically
everything
that's
in
the
readme
is
done
automatically
and
you
get
a
virtual
dev
environment
with
either
bs
code
or
you
can
use
vi
or
emacs
or
we're
also
adding
support
for
other
ides
as
well,
and
you
don't
have
to
spend
hours
days
weeks
figuring
out
how
to
install
all
the
dependencies
and
make
sure
all
the
little
tools
are
working.
B
That's
already
done
for
you,
and
so
as
a
new
contributor.
Instead
of
being
all
fiddly,
you
can
run
a
test
and
make
a
change
and
see
the
output
of
that
change
immediately
and
then
work
out
from
there.
So
it's
sort
of
encourages
new
contributors
by
giving
them
a
very
fast
time
to
joy
and
then
on
the
maintainer
side.
B
One
of
the
benefits
is,
you
can
encode
all
of
the
dependencies
in
the
configuration
of
code
spaces
so
that
new
contributor,
you
don't
spend
all
your
time
helping
people
debug
their
local
setup
problems,
but
even
more
importantly,
because
you
have
these
ephemeral,
many
different
dev
environments,
you
can
work
in
many
different
projects
simultaneously
without
having
to
worry
about
conflicts
on
your
local
machine.
B
You
can
it's
great
for
branches,
it's
great
for
pr,
so
you
can
start
a
codespace
on
a
pr
and
it
automatically
clones
all
the
code
and
make
sure
the
pr
is
in
place.
So
it's
perfect
for
testing
prs
and
it
has
a
great
kind
of
side
benefit
of
just
making
collaboration
across
maintainers
and
with
other
new
contributors,
which
straight
more
easy
and
so
I'll
pause
for
a
second.
Before
I
go
into
what
I've
been
talking
to
contributes
about
see
if
there
any
questions
about
code
spaces
as
a
product.
B
B
We're
going
to
be
doing
a
big
marketing
announcement
about
that
later
this
year.
That
will
make
it
easy
for
everyone
to
use
that.
I
can't
say
more
than
that
at
this
point,
especially
since
I'm
being
recorded,
and
this
will
be
on
youtube-
and
I
don't
want
to
get
myself
in
trouble,
but
it'll
be
good,
and
then
the
next
piece
is
that
I've
been
talking
to
contribex
about
this.
B
Obviously
the
catacota
thing
died,
and
so
one
direction
is
to
kind
of
revamp
and
revive
the
at
least
the
goals
of
what
pedocodo
was
about
using
code
spaces
to
create
kind
of
playgrounds
where
people
can
try
things
out
and
learn,
learn
the
ropes
of
kubernetes
and
then
the
other
thing
is
that
contribex
is
working
on
a
new
contributor
experience,
education
program,
training
program
and
in
partnership
with
the
consulting
firm
that
built
that
we've
got
at
least
preliminary
agreement.
That
codespace
is
code.
B
Spaces
is
a
natural
platform
for
enabling
that
kind,
of
course,
material,
and
so
the
practical
way
of
doing
that
is
actually
to
create
these
custom
configurations
in
the
kk
repo,
so
that
anybody
who's
using
code
spaces
and
anybody
who
comes
to
the
kk
repo,
would
then
actually
be
able
to
take
advantage
of
exactly
those
same
configurations.
B
You
know
and
then
essentially
code
spaces
and
the
way
it
is
in
github
would
naturally
become
a
part
of
the
way.
One
at
least
has
an
option
to
work
with
the
codebase,
that's
kind
of
where
we
are
right.
Now
we're
just
kicking
off
that
curriculum
development
process
and
actually
have
a
meeting
next
week,
wednesday,
with
dee
vibrata
to
kind
of
kick
off
the
kadakota
work.
C
I'd
have
I
have
one
question
craig:
if
that's,
okay,
and-
and
so
so
one
of
the
one
of
the
use
cases
that
will
be
awesome
to
to
tease
out
would
be
people
collaborating
on
fix
working
on
end-to-end
tests.
So
do
you
think
it
would
be
possible
to
run
end-to-end
tests
for
for
two
people
to
to
collab
on
that
and
work
on
that?
Absolutely
cool?
Okay?
Yes,.
B
I've
actually
just
played
around
with
running
ede
tests
in
the
codespace,
just
to
prove
it
out
to
make
sure
that
I'm
not
barking
up
a
bad
tree
yeah
and
it
works
great.
The
the
the
way
it
works
with
code
spaces
is
a
couple
of
fold.
B
One
is
you
can
obviously,
as
an
individual,
I
can
run
the
ed
tests
in
my
code
space
and
then
there's
a
technology
called
live
share,
which
you
can
kind
of
co
code
using
the
same
same
web,
browser
interface
into
it
and
see,
and
do
that
together,
which
certainly
helps
the
the
code
space
is
actually
owned
by
an
individual
user.
So
it's
not
like
co-owned
per
se.
The
resources
are
actually
assigned
to
someone
with
their
permissions,
and
so
you
have
to
use
the
web
browser
capability
to
do
the
sharing.
A
C
A
And
craig
to
be,
to
be
quite
honest,
we've
been
talking
about
godzilla's
for
communities.
For
three
years
now
we
basically
had
the
idea
when,
when
it
came,
when
code
failed
went
out,
we
tried
to
do
that,
for
I
remember
there
was
a
possibility
to
use
that
for
contributor
work.
Subject
contributing
summit,
but.
A
Yeah
and
we,
I
think
we
add
performance
issues
and
boundaries
and
stuff
like
that,
so
we
never
were
able
to
to
use
that.
I
I
think
your
the
big
concern
is
how
we
basically
gonna
land,
the
dev
container
fire
in
the
kk
repo.
That's
gonna
be
the
biggest
concern,
because
if
we
basically
provide
the
possibility
to
people
to
bootstrap
code
space
with
the
kk
ripple,
the
rest
is
just
spilling.
B
That's
very
sensible.
I
think
the
the
product
has
changed
and
improved
quite
a
bit
in
the
last
couple
of
years.
I
I
heard
the
stories
of
the
failed
efforts
a
couple
of
years
ago,
and
I
was
very
pleased
when
you
know
before
I
started
this
effort
of
reaching
out
to
the
kubernetes
community.
When
I
joined
github,
I
I
tried
to
replicate
what
you
guys
were
trying
to
do
using
you
know
our
current
configurations
and
product
and
had
a
very
different
and
good
experience
using
a
four
core
machine
now
with
the
product.
B
You
can
provision
anything
from
a
two
core
machine
to
a
32
core
machine
with
plenty
of
ram
and
this
space,
and
so
the
experience
can
be
dramatically
different
than
what
you
could
do
on
your
local
dev
box.
You
know,
since
you
have
a
machine
in
the
data
center
with
great
network
bandwidth
available
to
it.
B
A
A
D
It's
so
good
to
see
you
again
greg
I
I
we
should
catch
up
more
often
one
of
the
things
that
iii
does
a
lot
of
is
pairing
and
that's
the
whole.
Even
the
representation
of
an
eye.
Next
to
an
eye
is
people
working
together
and
so
we've
ever
since
we
logged
on
to
some
unix
boxes,
solaris
boxes
back
in
university
and
shared
a
t-mux
or
a
gnu
screen
back
then
a
session
together
like
how
do
we
get
people
to
do
it
together?
D
Well,
and
so
we
one
of
the
evolutions
of
that
has
been
pear.sharing.io
and
that
is,
has
been
available
to
the
cncf
community
to
to
click
on
and
create
a
sharing
environment
where
we
bring
up
an
actual
kubernetes
cluster
on
top
of
the
infrastructure
provided
by
equinix,
and
so
they
actually
get
a
dedicated
physical
hardware
machine
with
that
kubernetes
cluster
and
they're
actually
working
within
a
pod
deployed
to
that
cluster,
and
we
most
of
the
time
people
connect
via
ssh
into
the
host
and
getting
a
shared
tmc
emux
teammate
session
and
inside
that
we
drive
it
with
usually
with
emacs
on
the
ii
team,
primarily
because
there's
a
thing
called
org
mode
and
organ
mode
is
this
type
of
executable
literature.
D
It's
storytelling
in
a
in
a
very
many
language
form,
including
they
actually
have
a
library
called
the
org
babel
and
it's
from
from
babel.
It's
how
they
get
all
these
different
people
speaking
languages
together
to
work
together.
Well
and
it's
primarily
been
used
internal
to
our
team
and
when
we
pair
with
someone
else
for
a
period
of
time,
one
of
the
things
we've
been
looking
at
is
trying
to
make
that
available
to
others.
D
And
so
we've
been
looking
at
ways
to
use
the
cncf
ability
to
spin
up
infrastructure
underneath
projects
to
support
something
like
your
the
piece
of
software
working
on
dot,
github
user
dot,
the
name
of
the
project,
dot
cncf.
That
is
very
similar
to
code
spaces,
but
it
hasn't
involved
any
work
with
vs
code.
Yet,
but
I
just
submitted
to
bs
co
or
through
the
lfx
mentorship
stuff
to
get
people
to
I've.
D
Had
a
few
people
apply
to
go
through
that
to
be
kind
of
this
initial
phase
of
instead
of
of
of
bringing
up
just
our
terminal
environment
to
go
ahead
and
using
something
like
coder.com
to
bring
up.
They
currently
don't
target
cluster
api,
but
he
thought
about
this.
This
initiatives
to
add
cluster
api
support
to
the
provisioner
for
coder,
so
that
anyone
anywhere
can
target
a
cluster,
and
if
we
go
for
v,
cluster
and
and
and
kuba
admin,
I'm
sorry
cooper.
D
D
How
old
oregon
emacs
are
yeah,
but
there's
really
well
defined
engineering
things
that
could
be
brought
forward
if
we
can,
and
so
that's
why,
with
the
code
spaces,
that's
what
it's
called
right
in
the
github
trying
to
find
some
of
these
ways
to
overlap
some
of
the
things
we're
doing
and
pull
them
out
so
they're
available
to
a
wider
community
and
also
things
from
what
we're
doing
to
maybe
be
brought
into
workspaces.
So
they're
stuff
we're
doing
this
cross-pollination
of
ideas.
I
think
it's
particularly
well
for
this
onboarding
conversation.
B
B
I
guess
are
you
thinking
that,
like
what
would
be
inappropriate
for
a
new
contributor
using
this
sharing,
I
o
capabilities.
B
Yeah,
like
I
guess,
I'm
wondering
like
why,
wouldn't
you
invest
then
in
doing
using
that
approach,
as
opposed
to
you
know,
invest
like
why?
Why
do
the
effort
of
trying
to
find
where
the
synergies
are
between
code
spaces
and
that,
as
opposed
to
investing
mostly
in
the
pair
sharing,
I
guess
what's
what
what's
complementary,
that's
conflicting?
What's
absolutely.
D
D
D
We
do
have
a
little
bit
of
that
in
vs
code,
where
there's
a
terminal
down
there,
but
we
don't
have
that
code
coda
code
thing
where
here
is
the
document
that
drives
the
the
the
terminal,
and
we
also
don't
provide
people
the
ability
to
edit
the
code
inside
while
they're
working
on
it,
which
is
what
we
do
like
our
template
for
getting
started
for,
identifying
a
new
endpoint
in
kubernetes
that
doesn't
have
a
test,
yet
is
all
org
mode.
You
bring
it
up
and
there's
the
sql
block.
D
You
execute
the
sql
block,
says
here's
the
things
defined
by
all
of
our
ci
engines.
This
ain't
done
yet
choose
one,
and
then
they
choose
ones
where's
the
documentation.
That's
a
little
more
manual,
take
the
docs
and
then
write
a
little
small
snippet
of
golem
code
and
it's
not
in
a
file.
It's
just
there
and
it
goes
through
and
hits
it
and
then
another
thing
to
check
the
db.
This.
B
B
I
see
exactly
what
you're
saying:
okay
yeah
we
actually
do
have
there
are
extensions
to
visual
studio
code
that
enable
essentially
exactly
what
I'm
understanding
from
your
description
of
that
experience.
What
I
want
to
go
do
now
is
do
that
and
then
yeah
do
the
compare
and
contrast
and
see
what
I'd.
D
Love
to
see
what
you
found
currently
I've
only
found
things
that
that,
in
the
document
itself,
support
the
the
highlighting
and
all
of
the
tree
collapsing
and
all
that
of
org
mode,
but
don't
support
because
it's
kind
of
essential
as
all
those
code
blocks
have
parameters
and
things
that
need
to
be
passed
on.
And
I'm
assuming
at
this
point
to
a
lisp
engine.
A
Sorry
for
us
we
have
four
minutes
left
in
the
meeting.
I
see
a
behind
the
call.
Do
you
want
to
talk
about
something
we
lost
ep
also.
B
And
we
okay,
but
that
was
a
really
helpful
kind
of
insight
for
me
yeah.
We.
C
C
Yeah
you
could
you
can
reach
out
either
to
hippie
or
myself.
I
I
shared
some
links
for
a
hippie
there
just
to
give
some
refs
on
what
he
was
talking
about
and
if
you
want
to,
I
think
the
synergies
will
be
to
see
what
can
be
done
right
now
in
pair
and
then
look
at
look
at.
What's
good
look
at
what's
might
be
limiting
and
then
just
just
learn,
learn
from
that
experience
and
see
what
we
can,
how
we
can
marry
the
two.
D
Robert
there's
one
last
thing:
robert:
did
you
cover
the
the
api
that
you're
working
on
and
the
pga
idea.
D
Yeah,
so
real
real
quick
if
you'll
drop
a
link
to
the
issue,
we're
trying
to
get
prow
hook,
actions
plugins
to
run
as
github,
and
we
have
an
issue
open,
and
my
this
seems
like
the
primary
scary
thing-
is
the
hard
limit
across
an
entire
org
for
github
actions.
Api
calls
being
a
thousand
per
hour,
and
I
want
to
check
and
see
if
there's
a
way
that
that
can
be
limited
higher.
D
I
think
for
most
projects
that
would
allow
most
proud
plug-ins
to
work
as
they
won't
hit
that
limit.
But
I
want
to
find
some
way
that,
when
we
do,
projects
are
more
successful
without
having
to
provide
prow,
but
just
continue
to
provide
this
github
action
that
we
don't.
We
have
a
way
forward
when
they
get
successful.
C
Yeah,
so
so
so
so
so
the
high
level
architectural
thing
here
is
to
have
to
to
leverage
proud
plugins
that
exist
to
run
on
a
on
on
a
pro
instance
and
have
them
kicked
off
by
a
github
action
and
we're
I'm
teetering
on
the
edge
of
I'm
teaching
on
the
edge
of
instantiating
a
proud
plug-in
from
within
this
this.
C
This
this
custom
action
called
pga,
which
is
proud,
github
action,
and
so
the
hippie
is
is
concerned
about
hitting
that
rate
limiting
at
that
rate
limit,
I'm
concerned
about
instantiating
the
proud
plugin
right
now,
because
I'm
I'm
to
my
eyes
and
trying
to
implement
that
so
I'll
worry
about
rate
limits
when
I
get
the
brow
plug
and
running
so
yeah.
But
that's
that's
that's
where
I'm
at
at
the
minute.
A
E
C
E
C
Yeah,
it's
27
150
on
it's,
I'm.
C
Yeah
on
kate's
testimo,
it's
issue
number
27150
I'll
share
a
link
in
the
chat.
A
A
A
Oh
okay,
I
think
it's
better
to
do
that
anything
because
we,
after
the
time
I
would
say
for
you,
priyanka,
opened
an
issue
about
getting
a
token
for
this.
We
agree
to
you
an
existing
one
yep.
We
can
follow
up
on
that
on
on
the
slack
channel
and
see
what's
happening
basically,
because
I
think
once
we
get
your
pro
request
merge
we
can
work
on
set
up
the
project.
E
E
This
would
save
a
lot
of
time
for
at
least
one
of
the
sub
group
and
segregates,
and
also
it
will
be
like
a
good
demo
for
or
a
poc
for
enhancement
and
the
other
groups
as
well,
so
yeah,
I'm
working
with
priyanka,
so
she
was
taking
over
this
because
I
was
in
vacation.
So
would
love
to
you
know
like
continue
helping
moving
forward
with
that.
A
Okay,
just
being
me
on
on
on
slack,
I
will
walk
I'll
work
with
you
about
this
perfect.
Okay,
we
out
of
time.
Thank
you,
everyone
for
coming!
Thank
you.
This
was
a
nice
conversation
enjoy
the
rest
of
your
day,
bye.