►
From YouTube: SIG Cloud Provider 2023-08-02
Description
Meeting agenda: https://docs.google.com/document/d/1OZE-ub-v6B8y-GuaWejL-vU_f9jsjBbrim4LtTfxssw/edit#bookmark=id.klwajed3s0o5
A
All
right
welcome.
Everyone
today
is
Wednesday
the
2nd
of
August
2023,
and
this
is
the
Sig
cloud
provider.
Meeting
Sig
cloud
provider
is
a
sub-project
of
kubernetes
sigs
and,
as
such,
we
follow
their
meeting
guidance,
which
is
basically
means.
You
know,
please
treat
everyone
kindly
and
raise
your
hand
if
you'd
like
to
speak
and
I'll
call
on
you,
although
there's
only
four
of
us
today,
so
it
probably
won't
be
that
big
a
deal
so
I
guess
like
to
begin
with.
A
Let's
go
through
the
sub-project
updates
and
just
you
know
just
go
around
from
the
folks
who
are
here
and
yeah
Bridget.
Why
don't
you
go
first.
B
The
only
one
that
I
really
had
that
was
we're
talking
about
is
that
there
is
a
bug
fix
that
we're
trying
to
bring
in
after
code
freeze
to
the
Legacy
cloud
provider,
and
we
can
look
at
it
and
see
if
anyone
else
has
any
concerns
about
it.
It
looks
like
it
only
affects
Azure,
but
that's
the
only
provider
or
a
specific
item
that
I
had
to
talk
about.
A
A
Okay,
thanks
John
anything
from
you.
A
All
right
cool!
Thank
you!
That's
easy
all
right!
So
let's
look
at
the
agenda
first
and
then
we'll
go
back
to
triage
time
permitting.
A
A
I
have
a
feeling
we
might
need
Walter
or
Andrew
to
help
out
with
this
apparently
there's
some
way
we
can
log
into
the
zoom
console
that
we
have
associated
with
our
Sig
and
create
like
specific
URLs
for
sub-project
meetings
and
I'm,
bringing
this
up
because
the
IBM
sub
you
know
cloud
provider
had
been
using
a
zoom
Channel
that
I
guess
was
associated
with
an
old
Sig
account
and
those
all
got
shut
down.
A
So
I
was
looking
to
see
if
we
could
perhaps
get
this
arranged
so
that
we
could
create
a
new
room
for
IBM
and
I
think.
At
the
same
time,
we
also
learned
that
AWS
is
kind
of
reusing
the
zoom
link
for
this
meeting,
so
I
think
if
we
could
get
access
back
to
that
console,
we
probably
should
create
specific
Zoom
rooms
for
each
of
the
the
sub
projects,
but
I,
don't
I,
don't
have
access
to
any
of
that.
So
I'm
not
sure
what
the
next
step
here
is.
I.
B
Think
I
know
what
the
next
step
is
and
I
think
it
is
opening
a
cncf
service,
desk
ticket
I.
Think
and
yes
go
ahead.
Sorry.
D
Yeah
so
I
do
have
the
Zoom,
email
and
password
my
understand
so
I
guess
when
Walter
did
the
handoff
he
he
just
gave
you
the
host
key
and
didn't
give
you
the
zoom
password.
Is
that
correct.
D
B
A
Yeah
I
misunderstood
what
you
were
saying
about
the
service
desk.
There
yeah
I
had
reached
out
in
like
the
contrabax
channel,
and
this
is
like
one
of
what
Bobby
was
telling
me
was
like
Hey
like
get
the
whatever
okay
cool,
Andrew
I.
Guess
we
will
hook
up
out
of
band
then
and
or
we'll
sync
up
out
of
band,
and
then
we
can.
You.
A
Can
get
that
set
up,
so
thank
you,
okay,
so
the
next
topic,
for
also
for
me
I,
got
an
email
from
the
and
I'm
sure
Bridget
did
as
well
from
the
the
cncf
folks
or
the
kubecon
committee.
Folks
saying
that
we
have
an
opportunity
to
create
a
one
minute,
video
as
a
Sig
update
that
would
be
included
as
part
of
the
sigs
update
for
the
Keynotes.
A
The
downside
here
is
that
we
have
to
have
it
created
by
August
10th.
It
can
be
no
longer
than
a
minute
and
we
have
to
make
some
slides
for
it.
My
reading
of
the
email
and
the
related
links
was
that
this
is
not
mandatory,
but
we
could
we
could
take
it
up
if
we
wanted
to
so
I
just
wanted
to
bring
the
question
here
and
see
do
do
we
want
to
do
this
or,
like
you
know,
with
a
little
over
a
week
left.
B
B
Just
like
we've
got
right
here,
it's
like
okay,
whatever,
but
the
the
ones
where
they
have
a
call
to
action
are
pretty
valuable,
so
I'm
kind
of
wondering
if,
if
we
could
kind
of
take
this
as
an
opportunity
to
almost
do
a
trailer
for
the
The
Talk,
this
is
just
basically,
you
know
whether
it
be
El,
Miko
or
a
couple
of
the
tech
leads.
Just
you
know
on
screen,
saying
we're
going
to
talk
about
this
and
we're
going
to
talk
about
this
and
we're
going
to
talk
about
this.
B
This
is
all
you
know
like
getting
people.
This
is
why
you
would
want
and
get
get
engaged
with
cloud
provider,
because,
like
think
about
it
from
the
point
of
view
of
everyone,
who's
watching,
the
Keynotes
will
see
this
and
they
may
or
may
not
care
about
a
list
of
bullet
pointed
updates,
but
they
may
be
very
interested
in.
Oh,
that
looks
like
something
that
is
interesting
to
me,
for
whatever
my
reasons
are
anyway,
that
that's
kind
of
I
think
that's.
B
A
B
E
B
E
I
I
love,
Bridget's
idea,
I
will
say,
I
mean
it
kind
of
depends
on
which
call
to
action.
We
want
one
suggestion
I
would
have
is
engagement
from
anyone
who
has
a
who
is
a
cloud
provider
or
works
for
a
cloud
provider
might
be
one
of
the
actions.
E
We've
done
a
lot
to
throw
the
cloud
providers
out
of
tree
and
we're
adding
a
lot
of
capability
in
terms
of
el
Miko,
your
testing,
some
of
the
CCM
stuff,
some
of
the
credential
provider,
stuff
I,
wonder
if
maybe
the
call
about
call
to
action
is
hey.
Do
you
really
want
to
be
a
cloud
provider
who
is
taking
full
advantage
of?
You
know
everything
we've
built
if
so
show
up
to
the
meetings,
help
us
build
out.
E
You
know
a
good
Cloud,
Cloud
controller
manager,
a
good
credential
provider,
make
use
of
the
network
proxy
make
use
of
the
web
hook
system
we've
built
for
you
make
sure
that
you
are.
You
know
your
code
is
being
tested
every
time
a
build
is
done
so
that
you
know
that
future
releases
won't
break
you
I
mean.
Maybe
that
is
the
call
to
action.
We
want.
A
B
A
The
only
thing
we'd
have
to
be
worried
about
then,
is
like.
If
we
wanted
to
make
a
slide
deck
to
go
with
it,
I
don't
know
if
the
slide
deck
is
a
hard
requirement
or
not,
it
didn't
look
like
it
was,
but
I
couldn't
tell.
B
A
Okay,
so
yeah
think
about
it
and
maybe
like
maybe
next
week
or
like
Monday
or
something
we
could.
We
could
at
least
in
the
in
the
slack
Channel.
If
anyone
has
ideas
or
thoughts,
we
could
put
see
if
we
could
throw
up
some
scripts
or
something.
A
A
A
Chicago
too
right:
okay,
cool
thanks,
I
guess
like
yeah,
maybe
I'll
try
and
start
an
email
and
send
it
out
like
this
week.
Just
so
we
can
get
some
action
going
cool,
okay,
Bridget
you've
got
the
next
couple
topics,
so
please
take
it
away.
B
Just
real
quick
updates
we
did
put
in
that
say:
cloud
provider
update
at
kubecon,
China
and
pangfei
is
going
to
give
this
update
and
it
is
basically
a
remix
of
the
update
that
we
did
in
Amsterdam,
but
his
version
of
it
and
be,
and
we
can't
have
any
of
the
other
Azure
people
join,
because
it's
the
same
problem
where
we
couldn't
have
just
Joel
and
El
Nico
speaking,
which
is
two
from
the
same
employer
and
no
additional
employer.
B
But
if
anyone
has
people
from
their
employer
who
are
going
to
be
in
the
Shanghai
event
and
would
like
to
participate
in
this
session,
we
can
add
speakers
to
it.
We
just
you
know
yeah
like
please
reach
out
to
me
like
ASAP.
If
you
have
people
who
are
going
to
be
there
and
would
like
to
participate,
because
it's
always
good
to
get
voices
for
more
than
one
provider
if
at
all
possible.
B
Okay,
so
there's
that
then
there's
a
cherry
pick
that
I
mentioned
right
at
the
up
at
the
top.
That
I
was
trying
to
I
chatted
with
the
release.
Team
Grace
seems
fine
with
getting
this
into
1.28,
but
I
did
say.
I
would
bring
it
to
the
meeting
and
mention
it
to
people,
and
this
is
a
bug
fix
to
Legacy
Club
provider
that
they
are
to
code
for
you.
So
that's
why
it's.
B
If
you
want
to
pop
that
one
out,
then
it's
a
bug
fix
to
the
Legacy
cloud
provider.
It
only
affects
Azure,
you
can
go
ahead
and
look
at
files
changed
and
you'll
see
what
I
mean
but
yeah.
B
If,
if
anyone
who
has
more
familiarity
with
these
code
bases
wants
to
look
at
this
and
see
if
it
looks
like
this
is
a
safe
cherry
pick
or
not,
that
would
be
super
valuable
because
somebody
can
either
you
know,
because
if
we
can
get
this
into
the
1.28
release,
that
would
be
great,
otherwise
we'll
have
to
get
into
a
patch
later.
B
Could
this
be
a
different
one,
then,
because
this
is
not
merged?
Currently.
D
B
D
D
If
it's
a
bug,
fix,
I
think
yeah
defer
to
Azure
maintainers
on
whether
it's
on
the
risk
and
whether
we
should
merge
it.
But
from
my
initial
view
of
the
code
change
it,
it
does
sound,
it
doesn't
sound
low
risk.
It
sounds
more
like
a
medium
risk,
but
that's
really
like
any
sort
of
like
from
past
bugs
that
we've
seen
like
any
change
in
the
load.
Balancer
logic
where
we're
changing
the
the
Readiness
check
for
nodes,
like
that's
all,
just
like
in
general,
pretty
high
risk
change.
D
Okay,
Goods,
like
any
incorrect
evaluation
of
nodes,
is
obviously
gonna
like
lead
to
downtime
and
service
disruption.
So
by
Nature
it
just
seems
like
high
risk,
but
but
yeah
again,
I
think
up
to
Azure
folks
to
decide
that
it.
E
F
F
So
I
would
mark
this
as
pretty
high
priority
for
fixing,
but
yeah
you're
right.
There
is
some
risk.
The
issue
is
the
way
it's
interacting,
so
Alex
came
on
the
cool
few
few
weeks
ago.
F
The
issue
is
how
it's
interacting
with
the
changes
that
he's
been
making
through
some
of
the
Caps
to
improve
the
the
node
health
checks
across
the
board,
our
providers
and
it
just
the
way
it
was
working
with
the
jaw.
Just
just
stopped
the
members
being
added
I.
D
B
Okay,
good,
okay,
so
what
I
think
I'm
hearing
is.
This
is
probably
a
good
idea
to
get
into
1.28
because
it
is
broken
and
it
wasn't
broken,
but
I
became
broken
and
this
would
Unbreak
it.
In
theory,.
A
This
is
a
pretty.
This
is
a
pretty
short
change
here,
but
I
mean
do
we
want
to
put
the
backboard
label
on
in
that
or
not
the
backboard
label?
Do
we
want
to
put
the
approval
label
on
it
now
or
I
guess
Walter
did
so
yeah
I
guess
we're
good
to
go
right.
B
Yeah
I
think
we
just
have
to
get
it
yeah.
This
is
if
anyone
else
on
the
call
has
other
thoughts
or
like
hey.
This
is
good
for
this
reason,
or
this
is
risky.
For
this
reason,
please
go
ahead
and
comment
on
the
issue
which
is
linked
in
the
in
the
meeting
minutes.
This
I
need
to
just
go
back
over
on
the
channel
and
say
we
discussed
it
on
the
call.
We
believe
this
is
worth
going
ahead
and
putting
in
the
Milestone
go
ahead.
B
F
B
D
B
Okay,
cool
I'll
probably
be
the
one
putting
in
the
back
part
request,
but
it's
good
to
put
a
comment
on
there.
So
I,
don't
forget
okay,
what
else?
Oh
yes,
I
seek.
Yes
and
then
I
see
Chris
has
joined
in
the
next.
When
we
get
back
to
the
agenda.
The
next
item
that
I
had
was
hey.
Welcome,
equinix
metal
I
see
that
they
are
working
on
finishing
your
onboarding.
It
looks
like
Mr,
Bobby
tables
and
Company
were
working
on
some
slightly
manual
processes,
but
do
we
have
anything
more
that
needs
to
happen?
B
B
Do
you
need
anything
pr'd
anywhere
that
says
hello?
They
are
real
here.
They
are
on
the
readme
like
I,
don't
know
if
we
I
haven't
checked
to
see
if
we're
missing
you
from
any
of
the
lists
of
things
yeah
so.
C
B
C
D
Do
you
have
any
plans
on
integrating
your
platform,
what's
like
with
prowl
and
kubernetes,
like
Ci,
tooling
and
whatnot,.
C
D
D
A
Okay,
anything
else
on
that
topic.
Are
we
good
roll
on
here.
A
A
All
right,
sorry,
if
nothing
else
on
that
one
then
Joel
you've
got
the
next
topic
here.
F
Yeah,
so
more
bugs
I
got
certain
private
recently
and
it's
come
from
a
customer
of
ours.
Who's
looking
at
AWS
and
they've
noticed
that
if
they
change
a
service
from
a
classic
low
balance
to
a
network
load
balancer,
the
service
controller
isn't
removing
the
classic
load
balancer,
it's
just
kind
of
orphening
it,
which
obviously
is
bad.
So
I
need
to
go
look
into
this,
but
before
I
do
I
was
just
going
to
ask.
F
D
It's
possible
that
there
is
an
annotation
for
this
I
know.
I
know
for
for
AWS,
like
there's
a
very,
very
long
list
of
annotations.
You
could
put
on
the
service
for
various
things,
so
I
would
just
check
those
and
see
if
it's
like
a
known,
known
issue,
that's
kind
of
been
worked
around
with
the
label
for
annotation.
F
D
D
I
think
it's
yeah
like
for
cases
where
you're
switching
the
type
of
the
load
bouncer
for
this,
like
for
assuming
you're
using
the
same
service,
resource
and
you're,
just
changing
the
underlying
infrastructure,
it's
kind
of
up
to
the
provider
when
you,
when
you
do
the
update
that
you
clean
up
the
old
one
but
but
yeah,
but
because
we're
not
actually
trading
a
delete
of
the
service.
D
There's
no
like
Generico
path,
where
you
can
call
the
delete
path.
So
it's
kind
of
one
of
those
weird
Corner
cases
where,
if
you
are
supporting
switching
the
type
you
can
you'd
have
to
kind
of
add
that
yourself.
But
we
did
add
a.
D
This
is
specifically
why
we
also
added
the
load
balancer
class
field,
to
support
specific
cases
like
this,
where
single
load,
balancer
controller,
supports
like
multiple
types
of
load,
balancers
and
you
use
the
class
to
differentiate,
and
for
this
exact
reason
we
also
made
the
class
field
immutable
because
we
didn't
like
you're
not
supposed
to
be
switching
like.
If
you
want
a
different
load
balancer,
you
should
just
create
a
new
service,
but
but
yeah
prior
to
the
class
field.
A
All
right
cool
thanks
for
looking
into
that
Joel,
okay,
so
that's
kind
of
the
end
of
our
regular
agenda.
I
guess
we
should
look
to
see
if
there's
anything
that
needs
to
be
triaged
at
this
point.
Unless
people
have
other
topics
for
the
agenda,
they
wanted
to
bring
out.
B
B
D
D
B
E
D
Maybe
just
reach
out
to
Kishore
I
think
maybe
there's
a
backlog
on
the
uks
side
on
it.
A
Thank
you,
okay.
Well,
that
brings
us
to
the
end
of
our
triaging
and
the
agenda.
Is
there
anything
else
people
would
like
to
talk
about,
or
should
we
take
back
like
you
know,
25
minutes
of
our
day
here.
D
I'm,
sorry,
just
one
one,
quick
question:
what
would
the
maintainer
track
session
in
Chicago
is?
Is
that
everything
finalized
there
or
anything
else
you
need
on
that.
A
Well,
I
mean
we
just
have
to
write
it
at
this
point,
but
no
like
as
far
as
I
know,
it's
like
the
submission
is
in
if
anyone
would
like
to
join
or
if
we
want
to
have
a
different
set
of
presenters
on
there.
I
think
I'm
going
to
recommend
you
know,
Joseph
be
added
as
a
presenter
and
Joel
and
I
were
talking,
perhaps,
like
you
know,
Joe
might
want
to
join
as
another
presenter
I'm
perfectly
fine
to
take
myself
off
the
presenter
list.
A
A
A
Okay,
anything
else.
We
should
talk
about.
A
All
right:
well,
thanks
for
coming
out
everybody
and
I!
Guess
we'll
see
you
at
the
extraction
meeting
next
week,
if
you're
around
or
the
network
server
proxy
meeting
so
later,.