►
Description
Kubernetes Public Steering Committee Meeting for 20230911
A
Hello,
everyone
welcome
to
today's
steering
committee
meeting.
We
are
under
the
cncf
code
of
conduct
and
short
of
the
excellent
to
each
other.
This
meeting
will
be
recorded
and
uploaded
to
YouTube
later
and
Bob
has
just
shared
the
agenda
Doc
in
the
chat.
A
B
C
Yeah
five
out
of
seven
with
Tim
and
I
pending
I'm,
it
looks
it's
looking
pretty
good,
but
I
wanted
to
make
sure
that
we
had
a
chance
to
do
any
discussion
for
this
meeting
that
we
didn't
have
Quorum
for
the
last
one.
Yeah
overall
looks
pretty
good
to
me.
Tim
any
comments.
D
B
Yep
they're
going
to
stay
in
the
SCD
org
and
we
will
just
adopt
the
org.
A
D
A
A
A
Yeah,
it
also
sends
a
better
signal
that,
like
you,
know,
we're
not
like
taking
over
at
CD
and
turning
it
into
just
a
kubernetes
thing
or
something
like
they're
becoming
part
of
the
organization
and
getting
support
from
us,
but
it's
still
kind
of
their
own
group
within
that
I.
I.
Think
all
the
technical
problems
aside
suddenly
announcing
that
we're
like
moving
all
the
repos,
would
be
a
lot
of
fuss
and
just
give
that
signal.
It
doesn't
really
help
us.
We
already
have
to
manage
a
handful
of
kid.
Hooker
works.
E
B
It'll
also
become
easier.
The
the
GitHub
management
team
has
been
granted
a
Enterprise
copy
of
GitHub,
so
we
can
now
sort
of
centrally
manage
multiple
organizations.
We
have
not
rolled
that
out
yet,
though,.
B
D
A
book
staying
where
it
is,
would
we
still
under
k.
Community
have
a
subdirectory
for
the
Sig
and
repo
and
Charter
at
least
kind
of
I
guess
tombstone's
the
wrong
word,
but
like
a
link
off
to
the
the
things.
So,
if
somebody
comes
in
from
our
our
normal
processes
and
is
expecting
the
normal
artifacts,
they
see
I'll
jump
jump
off
to
other
place,
yeah.
That
should
still
be.
C
On
the
VR
former
lesson,
okay,.
B
About
that
well,
I'll
actually
pass
this
into
two
numbers.
It's
he's
the
author
of
the
pr.
C
B
So
this
has
been
long
discussed
in
chairs
and
TLS
going
well
over
a
year
now
and
a
way
we
of
trying
to
create
an
additional
step
in
the
latter
for
people
to
sort
of
get
recognition
for
the
work
that
they're
already
doing
and
with
it
being
an
official
artifact
in
there.
It's
a
lot
easier
for
them
to
basically
point
in
and
justify
it.
For
you
know
internally
for
work
or
anything
of
the
like
in
general,
too.
It's
just
like,
even
though
you
might
have
multiple
people
in
an
owner's
file.
B
F
In
addition,
this
also
gives
a
good
way
for
the
six
to
identify
people
who
can
transition
into
technical
leads
in
the
future.
So
when
someone
is
a
project
lead,
you
identify
them
as
essentially
subject
matter
experts
for
that's
a
project
and
eventually,
if
they
creep
into
other
areas,
they
brought
in
their
expertise
into
other
sub
projects.
Eventually,
you
are
actually
identifying
a
technical
lead
for
your
Sig,
so
that's
how
it
also
makes
it
easier
for
the
six
to
identify
Future
Leaders.
C
A
A
Yeah
I
think
that
that's
kind
of
the
point
is
it's
not
a
huge
change.
It's
just
trying
to
more
explicitly
call
out
that,
like
we
have
people
that
are
like
running
these
sub-projects,
instead
of
just
interacting
out
to
our
like
owner's
file,
technical
implementation
of
who
has
approval.
A
C
Pr
saying
that
it's
in
a
net
new
role,
that's
that's!
Why
I'm
asking
that's!
That's
what
I'm
trying
to
dig
in
on
it
a
little
bit
because
it
seems
like
we
could
just
name
existing
sub-project
owners
like
if
you're,
if
you're,
not
covering
the
I
kind
of
feel
like
if
you're
not
covering
the
area,
then
you're
not
actually
a
project
owner.
B
G
B
Not
be
as
it's
coded,
but
that's
what
people
think,
and
so
this
pulling
it
out
and
having
this
sort
of
separate
designation
is
a
way
to
recognize
the
you
know
the
people
from
those
sub
particular
owners
that
are
the
ones
that
are
organizing
the
work,
because
not
everyone
that's
an
owner
is
usually
the
people
that
are
truly
leading
that
sub-project.
B
C
Okay,
like
I,
just
got
a
funk
on
it,
I
guess
because
it
it's
like
I'm
hearing,
yeah
I'm
hearing
that
there's
not
much
of
a
diff
I'm
hearing
that
they're
functionally
similar
roles,
but
it
it
seems
like
it's
the
same
thing.
So
I'll
I'll
look
over
the
pr
in
in
a
bit
more
detail
and
try
to
understand,
but
it
doesn't
seem
like
it's
a
it's
a
it's.
A
new
change.
F
It
would
also
help
if
you
read
commitment,
because
there
are
some
logical
changes
in
there.
A
I
definitely
recall
this
as
sort
of
subsuming
sub-project
owner
I
think
that
there
aren't
that
maybe
there
are
some
sub-projects
where
it
makes
sense,
but
I
imagine
there'll,
be
a
bunch
of
places
where
it'll
be
kind
of
similar
to
the
chair.
Tl
split,
where
you'll
just
wind
up
copying
the
list
of
people
from
the
owner's
file
into
subproject
lead.
A
C
A
A
A
B
Not
just
from
there,
it's
mostly
just
a
way
of
officially
recognizing
generally
the
fewer
people
out
of
the
subset
of
people
that
are
in
an
owner's
file
that
are
truly
the
ones
driving
the
work
like
you.
You
might
have
someone
as
an
approver,
that
is,
you
know,
good
from
a
technical
standpoint
and
and
that,
but
they
aren't
the
ones
necessarily.
You
know
doing
issue
triage
and
just
again
sort
of
doing
everything
to
keep
that
sub
project
going,
and
that
would
be
the
like.
A
Well,
and
if,
if
if
subproject
owner
is
just
documenting
owner's
files,
that
will
be
places
like
KK,
where
it's
really
hard
to
make
own
like
an
individual
owner's
file,
a
line
to
everything
in
scope
for
a
project
or
like
there's
overlap
from
people
that
aren't
there's
like
shared
spaces
and
whatnot,
and
there
may
not
be
like
a
clear
like
here-
is
the
owner's
file
for
this
project.
Well,.
B
B
There's
a
couple
others
that
span
like
repos
and
locations
and
KK
to
also
like
other
places
in
kasigs,
and
it's
again
we're
just
listing
the
various
different
owners,
the
owner's
files.
And,
if
you
have
this
list
of
scattered
owners,
which
one
of
those
people
are
you
going
to
contact.
If
you
have
a
thing
for
that
sub
project.
A
Okay,
well
to
me
at
least
that
sounds
reasonable,
but
I
I
definitely
need
to
like
read
this
and
comment
up.
You
know
I'll!
Do
that
I'll
commit
to
doing
that
today.
I,
don't
know
about
everyone
else.
A
A
Going
once
going
twice:
okay,
let's
move
to
the
funding
topic;
I
believe
that
I
saw
that
was
Arno.
Adding
this.
G
G
Yeah
so
basically
some
some
of
the
maintenance
reach
out
to
me
about
getting
in
environment
where
they
can
set
up.
You
know
some
kind
of
basically
walk
on
setup,
CI,
automation,
to
run
tests
or
on
Windows
instances,
so
I
think
the
ID
firms
will
get
an
environment
where
they
can
work
on
that
automation,
stuff
and
move
that
to
testing
from
so
the
idea
to
get
funding
specifically
from
the
AWS
donation,
give
them
creating
account
and
provide
them
access
to
that
and
having
virtual
dollar
to
run
this
yes,
Crystal.
E
So
two
things
first
a
question,
so
there
it's
links
to
a
previous
issue
where
we
had
a
an
LFX,
an
LFX
intern
that
was
working
on
a
project
for
a
duration.
Do
we
know
if
we've
like
tracked
and
reaped
that
account,
if
that
account
is
no
longer
needed
like
do
we
do?
We
have
a
method
for
doing
this
and
tracking,
like
temporary,
assigning
of
developer
accounts.
G
Not
we
track
extends
in
distance
like
we
can
set
order
that
if
a
specific
threshold
is
Rich,
we
we
can
do
that.
That's
possible,
but
tracking
assignment
to
an
individual
is
currently
not
really
possible.
G
So
basically,
we
kind
of
give
them
access.
We
provide
authorizational
at
the
project
level.
That's
what
we
do.
E
So
then,
the
second
part
of
this
is
a
comment.
We've
now
had
multiple
requests
for
this
kind
of
thing,
and
there
has
been
also
historical
asks
to
either
the
kubernetes
project
or
at
the
cncf
level,
for
access
to
Cloud
resources
in
order
to
support
development
and
in
those
historical
cases,
there's
also
been
cases
where
it's
been
abused,
where
we
have
granted
access
to
Cloud
credentials
and
Cloud
credits
and
accounts,
and
they
have
not
been
used
for
the
thing
that
they
were
supposed
to
be
used
for.
E
So
this
is
an
area
where
number
one
it's
my
personal
opinion
that
I
don't
think
steering
should
be
involved
in
this,
like
I
think
we,
we
should
make
sure
that
it's
clear
that
we're
delegating
Authority
out
to
a
Sig
I
think
in
this
case,
Kate's
info
makes
sense
to
me
that
Kate's
infra
can
make
these
decisions
and
not
bottleneck
on
steering,
but
also
I.
E
Think
that
the
Sig
that's
responsible
for
this
needs
needs
a
policy
for
when
and
where
and
how
these
requests
can
be
can
be
granted
and,
like
maybe
there's
a
threshold
of
like
hey.
If
the
request
is
for
this,
under
this
duration
or
under
this
amount
of
money,
the
Sig
can
just
grant
those
credits
at
their
at
their
their
discretion
in
accordance
with
their
policy
or
if
it's
above
a
certain
threshold,
then
it
needs
to
come
up
to
steering
as
far
as
like
assigning
resources
but
I.
E
Whether
this
is
a
thing
and
a
system
that
we
want
to
invest
credits
in,
because
I
can
understand
us
wanting
to
ensure
we
have
CI
jobs,
that
utilize
Cloud
credits,
but
assigning
developer
accounts
out
of
our
Cloud
credits
gets
messy
because
you
need
to
have
some
kind
of
policy
about
when
was
to
ensure
fairness
like
when
and
where
and
how
we
assign
those
those
credits
to
people
how
we
reap
them.
All
of
that
kind
of
stuff.
A
If
This
Were,
the
Kate's
in
for
meeting
I,
would
dive
into
how
we've
managed
that
and
that
have
thoughts,
but
I
think
we
probably
should
take
that
there
I
think
we
still
need
the
funding
ask
procedure,
but
just
for
things
that
we
don't
already
have
so
it
like.
A
If
we're
generally
saying
we're
in
favor
of
using
the
existing
credit
supplies
for
developer
accounts,
then
this
does,
you
know,
go
to
Kate's
infra
going
forward,
but
that
might
be
like
a
more
General
guidance
from
maybe
even
steering
unless
we're
just
saying
we're
totally
always
delegating
100
every
aspect
of
how
you
might
use
the
Cloud
credits,
which
I
think
would
also
be
fine,
but
we
should
probably
make
that
clearer.
A
I
also
agree
that
there's
some
thorny
bits
with
handing
out
developer,
accounts
and
I'm,
not
sure
who
should
deal
with
that
or
what
that
policy
should
look
like
or
if
we
need
to
overthink
it,
because
also,
if
it
isn't
a
big
cost
and
and
like
infra
or
steering
or
whoever
manages,
does
a
good
job,
not
letting
it
get
out
of
hand.
Then
you
know,
should
be
fine,
like
I.
Think
the
stuff
that's
happened
in
the
past
has
been
a
drop
in
the
bucket.
G
They're
also
a
I
think
one
of
the
concerns
is
basically
how
we
Define
that
selling
for
each
individual
trying
to
access
to
a
specific
project,
because
until
now
we
never
really
saw
abuse
of
those
credits,
but
assuming
it's
possible,
it's
difficult
to
basically
Define
ways.
Where
is
the
spelling?
Because
people
come
to
us
and
say
I
would
like
that
specific
amount,
but
never
wish
never,
basically
get
90
of
that
never
used.
90
percent
of
that,
so
we
think
is
fine,
but
I
think
it's
hard
to
basically
Define
the
threshold
between
hey.
A
About
what
like
possible
problems
with
this
particular
request,
I
think
that
by
policy
we've
previously
permitted
similar
requests
and
that
it's
short
term-
and
we
should
probably
just
let
them
move
forward
and
async
start
figuring
out
like
yeah.
A
Group
on,
like
oh
wait,
we
haven't
been
doing
this
with
rules
because
we've
let
other
people
before,
and
this
should
be
small
and
temporary,
but
we
should
have
rules
yeah
to
the
specific
one,
though,
when
we've
done
it
in
the
past.
A
That
has
mostly
been
on
gcp,
where
we
can,
where
we
have
like
named
projects-
and
it's
really
obvious
on
the
spend
that,
like
oh,
someone's
Dev
account
is
using
x
amount
and
we
can
keep
an
eye
for
that
and
close
it
out,
and
we
have
had
Dev
accounts
that
have
kind
of
lasted
a
little
bit
longer
than
we
thought
and
closed
them
out
with
the
way
that
we
view
AWS
visibility,
I,
don't
think
we
have
good.
A
We
have
much
better
tooling
for
like
this.
Is
the
services
that
we're
spending
on
and
and
not
the
accounts,
so
it'll
be
a
little
bit
harder
to
keep
an
eye
on,
but
I
think
that's
something
Kate's
Emperor
will
have
to
solve,
and
hopefully
not
again,
not
a
blocking
problem
for
this
request.
I,
don't
think,
but
something
to
consider
long
term
is
not
just
having
policy
but
actually
being
able
to
follow
through
on
it
and
like
like.
We
don't
have
like
spend
controls.
A
A
So
that's
also
worth
considering
when
you
when
you,
when
you
hand
people
accounts
that
happens,
but
the
same
thing
can
happen
through
CI
code.
If
you
give
a
subproject
access
to
an
account
indirectly
through
their
CI,
they
can
still
go
and
run
up
a
huge
bill.
Anyhow,
I.
A
C
Yeah
I
think
it
would
be
good
to
have
something
of
a
work
back,
at
least
for
something
like
this.
Where
scale
testing
is
in
the
mix,
and
it
has
the
potential
to
get
out
of
control
and
I'm
like
what
do
you
need
to
do?
How
long
do
you
need
to
do
it
for
and
what
like,
what
are
your
dates
for
Success
like
knowing
that
I
think
would
be
helpful,
but
but
I
I
agree,
I.
C
Think
in
the
like,
on
the
long
term,
we've
we've
talked
about
these
requests
in
the
past
a
few
times
it's
during
meetings
and
I
think
we've
landed
roughly
on
this
should
be
within,
say,
Gates
and
first
Charter
to
execute.
A
Yeah
I
think
that
we
should
be
available
if
they're
looking
for
advice
on
like
how
is
best
but
like
it
seems,
like
you
know,
in
theory,
the
this
is
not
necessarily
a
technical
committee
anyhow
and
Kate
semper
is
the
like
technical
group,
we've
delegated
the
accounts
to
so
they
should
just
manage
this,
and
our
know,
I'll
show
up
in
the
Kate's
infra
meeting
to
discuss
it
there
with
that
hat
on.
D
Should
we
make
a
point
of
formally
asking
that
the
Sig
work
up
something
for
a
plan
like
I
would
imagine
it
even
just
pragmatically
that
we
start
with
a
spreadsheet
of
like
these
people
are
assigned,
and
at
least
on
occasion,
go
back
and
say:
hey?
Are
you
still
using
that
or
has
access
been
turned
down
like
something
simple
doesn't
have
to
be
initially
massive
and
beautiful
automation,
but
to
to
not
lose
track
of
them?
I
think
would
be
important
somehow.
A
I
mean
as
it
stands,
we're
not
even
really
enforcing
like
spin
limits
on
anything
until
it
gets
out
of
hand
like
no
one's
no
one's
saying
you
know,
signode,
you
get
this
much
credit,
and
this
is
your
cap
like
we.
We
don't
have
any
of
that
today
for
anything
it's
just
when
we
notice
something's
going
up
a
lot,
then
there's
some
kind
of
pragmatic
discussion
within
the
Sig
on.
How
can
we
cut
it?
And
is
this
reasonable?
B
The
other
thing,
at
least
with
regards
to
like
Dev
account
versus
CI,
is
with
just
a
straight
Dev
account
it's
a
lot
easier
to
spin
up.
You
know
other
unrelated
things,
and
that
has
been
things
that
have
happened
in
the
past,
so
that
is
just
the
you
know.
The
the
broader
concern
around
just
a
straight
Dev
account
versus
you
know
again
using
RCI.
A
One
last
way
to
bring
it
back
in
the
focus
a
little
bit.
I
think
the
initial
ask
isn't
specifically
about
what
accounts
will
be
used
and
or
Earth
things,
and
it
even
mentions
that,
like
they'd
like
to
do
Azure
at
some
point.
So
for
those
sorts
of
things
we
still
have
funding
requests,
because
we
we
may
do
things
outside
of
the
accounts.
A
But
since
there's
a
suggestion
to
just
use
the
accounts
that
make
I
think
that's
what
makes
it
allocated
to
Sig
Kate's
infra,
but
there's
still
the
possibility
for
people
to
say
like
I
need
to
test
on
another
platform
or
there's
some
reason
that
it
shouldn't
be
on
those
accounts
and
for
steering
to
review
that
I
think
the
funding
template
is
useful
even
for
infrastructure.
It
just
should
be
defaulting
to
asking
suggestion
for
first
and
coming
to
us
if,
if
Kate's
infrared
doesn't
think
it's
a
fit
for
the
existing
funds,.
G
D
I
tried
to
capture
what
it
was
sounding
like
might
be
consensus
in
the
notes.
But
do
you
want
to
look
at
that
and
see
if
you
agree
I,
don't
want
to
write
down
something,
that's
perceived
as
our
steering
decision
when
it's
sort
of
vague
forming
of
notes
based
on
what's
been
discussed
in
a
very
back
and
forth
way.
A
Yeah
I
only
see
one
very
minor
point.
It
says
unsure
reject,
but
there's
also
the
case
where,
like
they
just
like,
like
this
thread
mentions
Azure
in
the
future
and
it's
a
kitten
surfera
can't
provide
them
that,
because
we
don't
have
like
we
don't
actually
have
community-owned
Azure
accounts
to
Grant
access
to.
G
A
Oh,
the
the
current
issue
doesn't
need
to
come
to
steering.
I
was
just
saying
that
that
they
like,
as
an
aside,
mentioned,
something
where
that,
like
would
be
a
case
that
might
come
to
us
where
they
went
to
Kate's
in
front.
It.
Isn't
the
Kate
SIM
for
rejected
it.
It's
just
kids
and
for
currently
can't
provide
those
resources,
which
is
a
little
bit
different
from
being
rejected,
as
as
opposed
to
what
was
written
in
the
the
consensus.
But
Tim
put
an
update
for
that.
C
No,
no
I
I
knew
that
you're
going
to
complete
her
sure
thought
so
so
I'm
I'm,
fine
I,
think
you
know,
given
we
don't
have
the
capacity
to
to
vote
here.
Let's
just
comment
on
the
issue:
if
we
need
to
close
this
out.
A
Yeah
I
was
just
gonna
leave
a
comment
to
the
effect
of
that.
The
current
ask
is
the
sort
of
thing
that
would
go
to
Kate's
him
for
first,
and
it
sounds
like
they
can
handle
that.
B
The
other
thing
we
can
do
as
a
follow-up
is
like
just
draft
some
language
to
update
the
like.
You
know
issue
template
like
if
you
have
an
infrastructure
request,
you
know.
Actually
it
might
be
good
to
create
an
issue
template
in
you
know
kids.io
or
someplace
for
the
sort
of
thing,
and
then
we
can.
You
know
point
to
that
from
there,
for
you
know,
start
here
for
infrastructure
requests.
D
A
Thanks
Tim
I
definitely
got
bind
on
the
notes.
There
I
think
that's
it
for
the
agenda
today,
I
see
a
few
other
people
in
attends.
That
didn't
say
anything.
Do
they
do?
Do
you
have
anything
the
rest
of
y'all.