►
From YouTube: SIG Cloud Provider 2023-07-19
Description
Meeting agenda: https://docs.google.com/document/d/1OZE-ub-v6B8y-GuaWejL-vU_f9jsjBbrim4LtTfxssw/edit#bookmark=id.61bwf783uejc
A
Okay,
welcome
to
Sid
cloud
provider
for
July
19th
2023.
We
are
governed
by
the
kubernetes
code
of
conduct
so
be
kind
and
have
fun.
We
will
take
a
quick
look
at
it
looks
like
none
of
us
have
put
anything
into
the
sub
project
updates,
so
we
can
check
with
folks
and
see
if
any
sub
projects
have
something
to
add.
I
know:
I,
don't
have
something
for
Azure.
Today,
let's
see
going
down
the
list,
Alibaba
Badoo
I
think
for
nothing
from
Google
this
time.
Okay,
Huawei
IBM
cloud.
B
A
A
I've
already
lost
track
of
how
many
so
many
versions
of
kubernetes
are
code.
Free,
yeah,
that's
right,
1.27
came
out
in
April
code.
Freeze
for
1.28
is
in
effect
now
so
I
think
everybody
is
pretty
busy
with
that.
Let
me
go
ahead
and
share
the
agenda.
Hang
on
just
a
second.
A
Okay,
so
the
first
item
on
our
list:
let's
just
pause
on
that
one
in
case
Walter-
does
end
up
joining
us.
Let's
talk
about
the
2022
Sig
report
and
do
you
wanna?
Do
you
want
to
add
a
link
to
the
doc?
If
you
have
a
moment
to
do
that,
yeah.
C
A
A
What
we're,
what
we're
figuring
out
that
I'll
click
on
that
as
soon
as
you
get
that
added
just
so
I'm,
not
trying
to
bring
something
over
from
another
machine,
but
what
we're
figuring
out
and
I
I
was
actually
right
before
this
meeting
I'm
just
talking
to
Mr
Bobby
tables,
to
figure
out
exactly
what
should
the
time
frame
that
we
cover
be,
and
he
did
say
that
we
do
want
to
actually
just
scope
it
to
2022.
A
So
this
part
is
going
to
be
kind
of
exciting
because
I
was
kind
of
thinking
like
oh
okay,
there's
stuff
that
we
nope
nope.
That
didn't
happen
yet
that
didn't
happen
yet,
so
we
basically
have
to
have
like
a
point
in
time
of
what
we're
talking
about
of
things
that
happened
up
to
the
end
of
December.
So
like
we
can't
talk
about
kubecon
this
year,
we
can't
talk
about.
You
know
leadership,
changes
in
the
Sig
there's
a
bunch
of
stuff.
We
can't
talk
about
so
like
I.
A
Would
love
to
have
especially
Andrew
and
anyone
who
was
more
active
in
leading
the
Sig
back
in
2022
to
maybe
put
some
comments
in
here
so
that
we
can
make
sure
that
we
cover
the
things
and
I
would
love
to
also
talk
about
it
live
here
just
for
a
minute
or
two
make
sure
we
cover
the
things
that
were
the
most
relevant
for
2022,
without
stepping
into
the
things
that
we
started
doing
in
2023.
A
That's
where
it
gets
a
little
tricky
I
didn't
realize
that
we
had
neglected
to
get
this
filled
out
back
in
I,
think
February
or
March
when
they
wanted
it,
which
is
like
oops
I,
guess
we
were
thinking
about
kubecon,
so
yeah.
So
what
do
we
want
to
make
sure
that
we
talk
about
and
highlight
from
2022
I
think?
Did
we
get
the
entry
providers
like
AWS?
Did
they
did
they
actually
make
progress
for
moving
in
2022?
Or
did
that
not
happen
until
the
turn
of
the
year?
A
B
A
B
I
think
the
I
I
think
the
cube
the
cubelet
credential
provider
that
removes
the
the
entry
cloud
provider.
Integration
with
container
Registries
I
think
that
I
think
that
went
J
in
2022.
A
Okay,
that's
one
for
us
to
look
at
that's!
Well,
that's
fine!
We
can
look
back
and
see
what
else
merged
question.
Another
question
for
ongoing
initiatives
from
2022
that
weren't
caps.
Anything
Michael
was
saying
that
the
ete
refactor
might
be
one
that
we
want
to
list.
A
I
think
we
should
I
mean
you
worked
on
that
in
2022
anything
else
from
2022
standing
out
in
terms
of
things
that
weren't
a
cut
but
was
a
focus
of
the
Sig.
C
Just
to
go
back
a
little
bit
for
the
2022
highlights
as
well.
I
did
notice
that
it
looks
like
openstack
provider
was
removed
in
2022.
A
That
is
something
all
right:
okay,
so
that
is
current
as
of
December
2022..
They
generated
this
for
us,
so
we
can
go
check
these
and
see
if
we
want
to
write
about
these.
Does
anything
stand
out
for
people?
It's
like
okay,
web
hook,
hosting
added
to
CCM
service
type
load
balancer.
You
know,
class
field,
controller
manager.
B
B
B
Yeah
we
we
know
that
cert,
like
various
managed
offerings
for
kubernetes
or
using
this
as
part
of
their
like
migrating,
without
any
availability
loss
to
CCM.
B
A
B
A
Okay,
that's
a
good
call
out,
okay,
all
right
anyway,
so
if
anything
else
comes
to
mind
for
people,
they
can
add
it.
There
then-
and
these
are
really
hard
questions
to
answer
because
they're
questions
like
as
of
2022,
what
did
you
need
and
it's
like
well
as
of
2022,
we
needed
more
people
to
help
her
on
the
city
and
we
got
that
already
so,
like
I,
don't
know,
I,
don't
know
what
else
we
want
to
point
to
like.
B
There
areas
or
set
projects
I
have
kind
of
a
meta
question,
but,
like
okay
is
that
is
the
2022
annual
report
like
mandatory
at
this
point,
or
could
we
instead
use
the
same
amount
of
energy
and
just
write
the
2023
report
I.
A
Know
right
so
that's
kind
of
where
I'm
that's.
Why
I
was
asking
the
TOC
and
got
told
like
just
right
what
you
can
the
best
for
2022
and
I.
Guess
we'll
keep
some
notes
and
we'll
have
a
head
start
on
our
2023
one
anyway,
I,
don't
know
if
there's
anything
that
we
want
to
call
it
that
we
think
we
needed
help
for
at
the
time
that
we
still
do.
E
So
I
will
say
and
I'm
not
sure
how
to
go
into
this,
but
given
the
nature
of
this
sick
I
think
you
know,
we
always
need
better
Stronger
connections
to
the
cloud
providers
right.
So
you
know
having
sufficient
representation
from
every
cloud
provider
in
this
thing
has
always
been
an
issue.
C
Welter
is
much
more
diplomatic
about
this
than
me
because
I
was
about
to
start
naming
names,
but
I
think
it's
probably
better.
Not
to
do
that.
You
know
like,
but
but
I
mean
in
general.
We
could
use
help
with
the
specific
provider
implementations
right
like
that,
that's
where
we
could
use
the
most
of
them.
A
And
I
think
maybe
a
Meta
Meta
Meta
question
that
we
should
bring
up
that-
maybe
not
for
a
discussion
at
this
moment,
but
we
Mike
and
I
were
talking
about
maybe
staggering
the
meetings
at
some
in
some
way
to
make
them
more
accessible
to
people
in
other
GEOS
so
like,
for
example,
the
club
provider
Azure
people
are
in
Shanghai.
So
it's
like
they're
never
going
to
come
to
this
time
zone
or
this
time
slot.
So.
E
I
I
love
the
idea
I,
the
one
question
I
would
have-
and
this
is
not
you
know
meant
to,
but
we
stag.
This
is
a
staggered
time
that
was
meant
to
make
it
easier
for
some
folks,
so
the
real
question
I
would
have
with
that,
then
is
what
is
the
time
that
we
I
we've
got
a
24-hour
24
hours
worth
of
time
zones.
What's
the
time
zone
that
works
or
you
know
what
is
the
set
of
time
zones
that
would
work.
A
And
I
think
our
friends
like
Joel
in
the
UK
would
say
that
some
of
their
overlap
with
China,
while
not
at
all
ideal
for
the
U.S,
would
be
a
perfectly
fine
meeting.
And
so.
D
A
This
is
kind
of
a
some
of
us
and
I
put
myself
in
this
category
of
U.S
people
who
may
need
to
help
facilitate
meetings
in
non-us
Friendly
time
zones
and
or
make
sure
that
people
can
have
meetings
without
us
present,
and
so
that's
kind
of
a
me
and
Mike
this.
You
know
implementation
detail,
but
I
think
we
need
to
do
something
about
that.
E
Oh
I'm,
all
for
doing
it
I'm
just
saying,
but
let's
not
I
mean
we.
We
picked
the
time
zone
for
this
by
putting
a
question
out
to
the
the
Sig
and
we're
like
hey,
which
time
zone,
and
this
is
what
we
settled
on
and
absolutely
if
this
is
the
answer
is
this
is
not
the
correct
answer.
Then
we
probably
need
a
better
way
to
pick
the
time.
A
Yeah
and
I
also
think
that
we
might
need
a
a
staggered
meeting
which
would
make
it
so
that
people
could
come
to
at
least
half
of
them
as
opposed
to
none
of
them
laughs.
C
A
Yeah
so
anyway,
so,
let's,
let's
time
box
that,
but
yes,
that
is
hey
look
at
that
we
got
some
stuff.
We
can
actually
write.
That's
awesome!
Okay,
we
kind
of
talked
about
Community
Health.
You
know,
feel
free
to
put
comments
in,
or
you
know
to
reword.
It
I
always
think
this
one
I
think
we
should
probably
make
some
edits,
and
this
is
something
my
client
can
do
to
the
contributing
dad
MD.
A
But
I
also
think
you
know,
as
Walter
was
pointing
out
like
yeah
Club
predators
are
the
ones
who
are
writing
cloud
provider.
Implementations
we're
not
necessarily
going
to
have
drive-by
people
who
rewrite
the
cloud
provider
so
yeah,
it's
kind
of
a
generic
question,
but
I
like
I
we
had-
and
this
is
is
interesting
to
me
recently-
because
Michael
and
I-
both
you
probably
saw
in
the
chat
in
slack,
were
answering
some
drive-by.
Hey
I
would
like
to
make
babies
first,
PR
I
think
a
club
provider
is
ideal
and
it's
like.
A
Oh,
oh
honey,
no
go
over
some
place
where
they
have
good
first
issues.
The
reason
we
don't
have
good
first
issues
is
because
literally
there
are
no
good.
First,
issues
for
like
I'm
brand
new
to
kubernetes,
I
think
I'm
going
to
work
on
the
cloud
provider.
I,
don't
know
like
I'm,
not
saying
I
want
to
be
unfriendly,
but
I
think
maybe
our
contributing
to
MD
should
like
explicitly
point
to
KK
new
contributor
stuff
and
just
be
like.
If
you
really
want
to
work
on
Club
provider
pick
the
club,
whoever
you
want
to
work
on.
A
E
I
I
think
the
way
you're
thinking
of
it
you're
right
but
I
think
there's
some
edge
cases
that
we
may
be
skipping.
So
if
you
want
to
work
on
a
cloud
provider
yeah,
you
probably
need
to
be
very
involved
in
that
cloud
provider,
and
you
should
probably
pick
one
but
I
think
there's
some
interesting
things
around
the
periphery
that
we
frequently
skip.
E
So
as
an
example,
if
you
are
a
docs
person,
not
like
a
tech
writer
I
would
love
your
help
to
talk
about
some
of
the
more
General,
how
CCMS
work,
how
the
and
I'm
not
sure
that
you
actually
need.
You
know
I
think
there
may
be
some
things
where
Docs.
E
Release
engineering
QA:
we
could
really
use
some
help
in
those
areas
that
we,
you
know
for
things
like
generalized,
like
a
lot
of
the
stuff.
Almika
is
doing
around
trying
to
build
cloud
provider,
testing,
I'm
sure
El
Miko
can
find
a
very
good
set
of
first
issues
for
someone
with
a
testing.
You
know
mindset
that
are
not
actually
specific
to
a
cloud
provider.
C
C
That's
where
that's,
where
we're
putting
the
common
tests,
they
all
live
in
KK,
so
like
I,
think
Walter
is
spot
on
about
about
the
docs
and
the
testing
like
we
could.
We
could
very
I
think,
like
I'm,
getting
to
the
point
where
my
I'd
like
to
show
off
this
proof
of
concept:
I've
been
building
and
once
that
gets
in,
and
if
people
are
okay
with
it,
then
yeah
we'll
have
a
runway
for
people
who
want
to
do
testing
work.
You
know
that
there'll
be
lots
of
work
to
be
done
there.
So.
A
Awesome,
okay,
so
I
think
good.
We
have
some
solid
content
for
that,
then,
obviously
we
can
rewrite
it
some,
but
I
think
this.
This
has
some
very
pointed
questions
that
kind
of
just
hint
towards
we
need
to
go
rewrite
the
contributing.md
to
point
to
the
things
we've
been
talking
about,
which
is
fine.
A
We
can
do
that.
Hey,
hey
one
place
where
we
shine.
Look
at
that.
We've
got
a
whole
class
cross
vendor
collaboration
here,
which
is
awesome.
Then
we
kind
of
this
kind
of
asked
that
same
question,
but
from
a
different
angle
that
we
can
talk
about.
This
I
did
mention
to
Mr
Bobby
table,
so
we
can't
go
back
in
time
and
count
this
stuff
at
the
time
that
it
would
have
been
better.
So
we.
A
C
A
I,
like
that,
all
right,
and
then
you
gotta
double
check
that
and
yeah
anyway.
So
if
anyone
wants
to
put
more
comments
on
this,
I
just
wanted
to
make
sure
that
that
was
surface,
that
yes,
we're
aware
this
need
to
get
done
and
we
have
made
some
progress
and
I
appreciate
everyone's
time
on
it
today,
any
more
thoughts
or
questions
on
that
before
we
move
on
okay
Walter,
we
had
started
there
because
we
were
we
were
holding
for
the
house
and
by
the
house.
I
mean
Walter.
E
Yeah
so
I
I,
a
couple
of
the
members
of
the
the
steering
committee
related
to
the
testing
changes
and
things
that
are
going
on
on
that
front.
E
The
cncf
is
trying
to
rationalize
a
lot
of
their
testing
and
how
they
do
billing
for
testing,
and
they
essentially
want
to
get
to
a
place
where
all
of
the
testing
accounts
are
owned
by
the
cncf
and
as
weird
as
this
may
seem
to
some
of
us
today
that
isn't
actually
true.
E
Some
of
the
larger
contributors
like
Google
and
Microsoft
actually
have
their
own
test
accounts
and
then
put
large
amounts
of
money
into
those
test
accounts
as
donations
to
the
cncf
and
then
Delta
cncf
go
do
your
testing
with
these
accounts
and
here's
the
testing
we
want
you
to
do
and
the
cncf
is
moving
towards
a
model
where
they're
saying
no,
no,
no,
no!
No,
we
own.
E
We
we
own
our
resources,
we
own
our
the
usage
of
our
resources,
and
for
that
to
happen,
we
have
to
own
our
testing
accounts,
and
so
all
of
this
is
the
testing
is
moving
away
from
the
cloud
provider
owns
tests
to
being
something
that's
actually
cncf
and
and
they're
just
going
to
turn
off
tests
that
are
still
running
at
some
point
on
those
old
accounts,
and
so
they
would
they.
E
A
Yeah
I
see
that
there
is
this,
this
old
issue
and
the
person
that
is
being
played
in
it
is
actually
my
manager
and
we've
had
a
conversation
about
it
recently
and
the
tldr
on
on
that
particular
one
is
there's
a
a
way
that
accounts
can
get
set
up
where
they
can
be
used
for
by
internal
people
and
then
there's
a
different
way
to
segment
the
accounts
correctly,
and
it
would
probably
have
to
be
a
different
kind
of
account
in
a
different
account
and
an
account
with
a
different
scope
in
order
to
just
like
it's
not
like,
we
can
just
change
the
password
and
give
it
to
someone
like.
A
A
A
You
have
to
figure
out
exactly
how
to
implement
it
within
all
of
the
constraints
that
you
can
imagine
exist
in
a
corporation
I
do
I
I'm
curious,
just
because
I
was
going
in
trying
to
find
exactly
which
tests
I
know,
there's
some
stuff
that
was
being
triggered
in
one
way
and
then-
and
this
is
specifically
around,
as
you
know,
of
course,
there
are
some
there's
some
tests-
infra
that,
for
example,
in
cluster
API,
we
were
able
to
switch
over
to
start
using
the
eks
stuff
instead
of
the
Google
stuff,
and
then
it
was
like.
A
Oh
okay
for
the
new
account.
Certain
things
are
not
turned
on.
Let's
figure
out
exactly
what's
going
on
with
that,
and
so
anyway,
there's
there's
things
that
it
would
be
good
if
we
can
figure
out
exactly
which
tests
are
currently
the
broken
ones
or
are
the
soon
to
be
broken
ones
just
yeah.
Basically,
it
would
be
I
feel
like
that
would
be
valuable
if
we
could
even
make
a
subset
of
a
dashboard.
A
E
I'm,
actually
not
sure
it
would
be
I
can
reach
out
again.
The
my
contact
is
actually
on
vacation
right
now,
so
I
won't
be
able
to
answer
this.
That
question
this
week.
Sure
of.
D
Course,
but
this
might
be
Complicated
by
some
issues
like
I,
know
the
cloud
provider
equinix
metal.
We
use
equinix,
metal
right
and
obviously
we're
not
in
the
the
kasigs
repo
yet,
but
there
was
no
chance
to
plans
to
change
that
once
we
get
there
and
we
don't
use
prow,
we
use
GitHub
actions
right
now,
so
I
this
seems
really
complicated
if
they
want
to
ensure
everything's
on
Pro
and
and
cncf
owned.
E
So
I
mean
I
think
it
is
worth
being
very
clear
here.
I
think
what
we
are
talking
about
is,
if
you
want
your
tests
paid
for
by
the
cncf,
and
you
want
them
as
part
of
KK.
Okay,
so
that
you
know
you
get
some
verification
of
incoming
changes
and
whether
or
not
they
work
with
your
cloud
provider
that
as
they're
as
they're
occurring,
then
you
need
to
get.
You
need
to
basically
adhere
to
the
cncf's
rules.
E
If
what
you
want
to
do,
which,
frankly,
I
think
all
of
the
cloud
providers
do
no
matter
whether
they're
integrated
in
the
way
I
set
or
not-
and
you
want
to
consume,
released
cake,
even
candidate
KK
images
in
your
own
repo
and
use
them
in
your
own
testing
environment.
You
are
obviously
free
to
do
that
and
know
what
is
going
to
object.
This
is
strictly
about
trying
to
get
your
tests
integrated
into
KK
yeah.
E
I
will
give
it
another
go
and
then
I
may
try
borrowing
someone
else.
I
I
have
tried
to
ask
you.
I
I,
believe
I
have
bypass
students
for
exactly
what
you
you
asked
for,
because
it
seems
rational
and
reasonable
to
me,
but
sometimes
I
get
a
a
different
answer.
An
answer
to
a
question.
I,
don't
feel
like
I
asked,
but
I
will
give
it
a
minute
ago.
A
C
Yeah
totally
so
this
might
be
holy
Red
Hat,
slash
openshift
problem,
but
it's
something
that
we're
noticing
across
a
lot
of
CCM,
so
I
wanted
to
kind
of
bring
it
up
here
and
see
if
others
had
thoughts.
C
So,
basically,
you
know
openshift
runs
on
a
lot
of
different
Cloud
platforms,
and
so
we
are
working
to
integrate
more
CCMS.
You
know
to
make
it
easier
to
run
openshift
wherever
people
want
to
and
on
openshift
you
know
we
tend
to
have
higher
than
I.
Think
vanilla,
like
security
standards
right.
C
So
we
don't
populate
Kube
system
namespace
and
we
have
some
special
rules
around
like
the
security
context
there
and
everything,
and
so
what
we're
noticing
is
that
a
lot
of
you
know
just
a
lot
of
cloud
providers
in
general
and
I'm
talking,
maybe
like
a
dozen
or
so
of
these
that
we've
looked
at
you
know
recently.
C
C
You
know
offering,
so
this
wouldn't
really
be
a
big
deal
to
them,
but
as
we're
looking
at
helping
people
to
install
a
CCM
from
like
a
user
perspective,
what
we're
noticing
is
the
Assumption
seems
to
be
putting
it
in
Cube's
system
for
everything
and
so
I'm
wondering
if
it
might
be
beneficial
for
us
to
kind
of
take
more
of
a
position
that,
like
you
know,
should
we
advise
these
things
to
be
put
into
Cube
system,
or
should
we
make
it
a
little
bit
more
open
and
just
say
maybe
advise
people
hey
like
you
might
want
to
put
the
CCM
in
its
own
namespace
or
you
might
want
to
consider
the
namespace.
C
C
E
I'm
actually
somewhat
neutral
on
this,
but
I
I
that
the
interesting
thing
is
I
know
exactly
what
that
particular
namespace
field
is
used
for
so
I
I
have
a
couple
of
questions
if
you're
good
with
it,
so
the
first
would
be.
Where
do
you
run
your
KCM.
C
I'm
trying
to
remember
I
think
we
have
a.
We
have
a
specific
names
like
we
have
a.
We
have
like
a
a
raft
of
namespaces
that
we
run
everything
in
and
I'm
trying
to
remember,
like
I
think
we
have
KCM
in
its
own
controller,
manager's
namespace,
perhaps
I
don't
know
Marco.
Do
you
remember
yeah
hi,
it's
openshift
group
controller
managers
are
dedicating
namespace.
E
Okay,
then
I
I,
then
I'm
entirely
agreeing
with
what
you're
saying
El
Miko
the
two
things
I
will
point
out
related
to
this.
One
has
to
do
with
one
of
the
Caps
that
Andrew
and
Bridget
were
talking
about
when
I
joined.
E
So
there
is,
there
is
a
and
you
may
be
passed
the
need
for
this,
but
when
you're
dealing
with
clusters,
the
there
is
a
cap
that
talks
about
how
you
transition
controllers
from
one
kind
of
controller
manager
to
another,
because
when
you're
doing
the
cloud
provider
extraction,
the
various
node
Service
and
networking
controllers
that
we're
running
in
the
KCM
need
to
now
run
in
the
CCM.
But
you
want
to
make
sure
you're
never
running
two
copies
of
the
controllers
and
so
a
chunk
of
that
leader
election
configuration
you
were
just
looking
at.
E
There
are
three
leader
election
locks.
There
is
a
KCM,
only
lock,
there's
a
CCM
only
lock,
and
then
there
is
a
shared
lock
for
doing
the
handoff,
and
so
this
actually
is
the
CCM
only
lock.
But
there
is
another
lock
which
is
the
and
off
lock.
So
with
that
handoff
lock,
you
want
to
think
about
what
namespace
it
belongs
in,
because
it
needs
to
be
somewhere
that
if
you're,
using
that
handoff
lock,
both
controller
managers
have
access.
C
Great
context,
thank
you,
Walter,
like
I,
I
and
I.
Think
we
have
this.
You
know
I,
think
we
solve
it
in
a
different
way
in
openshift,
but
I'm.
Just
trying
to
think
about
like
is
there.
Can
we
be
more
open
or
more
Dynamic,
I
guess
how
we
advertise
people
to
do
these
things,
but
I
totally
get
what
you're
saying
about
like
you
have
to
be
careful
when
doing
the
handoff.
If
you're
coming
from
the
other
way,
it
would
have
been
assuming
that
Kube
system
was
where
this
stuff
was
being
you
know
placed
and
so
yeah.
E
That
makes
sense
to
me,
and
then
the
other
thing
I
will
say
is
that
you
know,
because
both
the
KCM
and
the
CCM
are
dealing
with
things
like
setting
node,
IP
addresses
and
various
you
know,
and
the
life's
life
cycle
of
nodes
as
well
as
services
and
a
few
other
key
things
you
do
have
to
be
somewhat
privileged.
Now.
I
will
also
say,
I
think
you
know,
and
this
is
a
larger
problem
that
is
not
cloud
provider
specific.
E
But
if
we
can
avoid
having
the
permissions
of
the
KCM,
we
should
the
KCM
for
one
of
one
or
two
of
the
controllers
in
the
KCM
that
are
actually
generic
they're,
not
cloud
provider
specific,
but
they're,
actually
setting
the
permissions
for
controllers,
which
means
that
they
are
incredible.
The
KCM
is
incredibly
privileged
when
it
comes
to.
You
know
how
much
permissions
it
needs.
E
C
Okay,
I
think
maybe
like
in
you
know,
unless
people
have
objections
or
whatever
what
I'll
probably
do
is
do
some
more
research
and
if
I
can
maybe
come
up
with
a
PR
against
the
cloud
provider,
repo
or
something
you
know
if
anything
comes
out
of
the
research
I
do
because
it
may
be
that
this
is
not
a
big
issue,
but
yeah
I
appreciate
I,
appreciate
the
the
context
and
whatnot
thanks.
E
Yeah
and
in
fact
I
I
think
you
know
I
mean
the
least
privileged.
You
know.
Philosophy
says
that
we
should
do
exactly
what
you're
saying
it's
just
never
been
the
higher
priority.
C
Well,
right,
right
and
I'm
getting
some
pushback,
not
pushback,
but
we're
seeing
a
bug
internally
or
like
we're,
seeing
some
difficulties
as
we
try
to
bring
more.
You
know:
Cloud
providers
into
openshift
and
we're
just
trying
out
different
CCMS
and
we're
kind
of
seeing
like
oh,
like
these
people
expected
that
this
Coupes
this
CCM
was
always
going
to
be
deployed
into
Cube's
system.
So
what
I
was
thinking
was
like?
C
Is
it
possible
for
us
to
generate
maybe
a
little
more
documentation
around
this
and
I
think
the
context
that
you're
talking
about
Walter
would
be
like
would
be
perfect
to
put
into
a
discussion
about
like?
Where
do
you
want
to
run
your
CCM
and
like
here's,
the
things
to
think
about,
and
here's
maybe
some
guidance
that
you
could
follow?
Then
you
know
just
because
we're
seeing
a
lot
of
people
using
the
cloud
provider
repo
as
an
example
to
start
with,
and
then
they
just
kind
of
follow
the
advice
that
comes
out
of
there.
A
Okay,
and
while
we
were
talking
about
that
I
remembered
that
I
should
mention,
we
did
get
the
kubecon
maintainer
sessions
submitted.
A
We
got
one
submitted
for
China
and
it
will
be
very
much
an
overview
with
detail
about
the
entry
out
of
tree
stuff
somewhat
similar
to
the
one
we
gave
in
Amsterdam
pangfini
from
Azure
is
signed
up
to
give
it
right
now,
he'll
be
happy
to
have
a
co-presenter
in
Shanghai
from
a
different
Cloud.
If
anyone
is
available
I'm
just
going
to
put
in
the
notes
here
as
well,
we
have
an
Azure
speaker,
any
other
question
mark.
A
It
would
be
great
if
we
have
somebody
from
another
provider
who's
going
to
be
at
kubecon
in
Shanghai
that'd,
be
awesome.
I
submitted
that
session,
so
I
can
add
more
speakers
to
it.
If
we
can,
if
you
can
reach
out
to
me
on
that
one
and
then
El
Nico,
you
want
to
tell
us
about
North
America.
C
C
I
I
also
I
talked
to
Andrew
a
little
bit
about
this
and
he
got
me
in
touch
with
Joseph
Hall
who's,
one
of
the
contributors
to
the
API
server,
Network
proxy,
so
I
think
Joseph
is
gonna.
I'm
gonna
put
his
name
as
well,
and
it
I
I
think
it'll
probably
be
me
and
Joseph
presenting
it,
and
there
was
a
talk
about
connectivity
last
year
in
Detroit,
but
I
think
we
can
kind
of
differentiate
ourselves
enough
from
that
talk
that
this
will
still
be
like
good
value
for
the
community.
A
Okay,
cool
and,
if
and
again,
for
North
America,
if
anyone
else
is
going
to
be
there
and
wants
to,
you
know,
go
to
that
session
or
perhaps
assist
with
it
or
even
co-present,
for
you
may
recollect
that
for
maintainer
sessions,
unlike
talks
where
you
need
to
have
two
speakers,
or
it
starts
becoming
a
panel
for
maintainer
sessions,
we
are
allowed
to
have
several
speakers
in
order
to
represent
the
whole
maintain
artists
of
the
project.
So
if
which
which
organization
is
Joseph
Amigo.
C
You
know
wider
Community,
watching
the
recording
please
reach
out
if
you're
interested
in
being
part
of
this,
because
we're
always
looking
for
new
faces
to
kind
of
you
know
be
part
of
our
maintainer
tracks
and
whatnot
and
I'd
be
happy
to
help
someone
else.
You
know
co-present
with
Joseph.
If
they'd
like
to.
A
Fantastic
okay,
so
we
got
that
in
there
is
there
anything
else
we
should
talk
about.
I.
Think
we've
reached
the
end
of
our
pre-prepared
agenda.
Does
anyone
have
any
other
topics.
A
Okay,
going
once
going
twice,
I
think
that
we
are
at
the
end
of
this
call.
Thank
you
all
so
much.
We
will
continue
working
on
that
report
and
thanks
for
all
the
input
and
hopefully
it'll
make
it
and
I'll
be
able
to
get
that
in
before
the
next
meeting.