►
From YouTube: 20210819 Cluster API Office Hours
A
A
So
I
think
we
typically
start
this
meeting
with
introductions.
Is
there
anybody
new
on
the
call
that
would
like
to
introduce
themselves.
B
C
Cool,
I
think
I
heard
someone
else
hi
hi.
This
is
luchan.
I
am
from
spectro
club
yeah.
So
I'm
a
software
engineer
at
spectrocloud,
yeah,
it's
more
or
less
about
me
cool.
I
haven't
heard
of
spectrocloud
what
what
you
do
there.
C
So
we
are
a
startup.
We
also
provide
communities
or
for
data
operations
across
multiple
cloud
platforms.
A
Yeah
I
first
I
mentioned,
but
we
also
try
to
raise
our
hands
when
we,
when
someone
wants
to
speak
using
the
zoom
feature
there.
So
as
we
get
into
further
discussions
here,
well
I'll
call
on
you
as
you
raise
your
hands.
Thank
you
and
make
sure
you
add
yourself
to
the
attendee
list
up
above
all,
right
psas,
I
think,
welcome
matt
you're.
Now
our
reviewer
congratulations
and
ace
you're.
Now
a
maintainer,
fantastic
news.
Thank
you
for
all
your
your
awesome
contributions.
A
Correct
so
the
other
really
exciting
news
is
there's
some
hard
work
to
rename
the
default
branch
to
main,
and
if
so,
if
you
have
any
open
branches,
you're
gonna
need
to
update
your
remote
and
so
there's
some
instructions
here
on
how
to
do
this,
and
if
you
have
any
questions
or
concerns,
please
reach
out
on
slack
cecile
you
and
matt
you've
been
monitoring
ede
tests.
I
think
everything
looks
pretty
good
there.
I
know
for
the
windows
side,
the
switchover
was
was
smooth.
D
A
Do
awesome
all
right
anything
else,
anybody
wants
to
mention
there.
D
Yeah,
this
isn't
really
a
discussion
topic,
but
more
of
a
quick
marketing
from
my
proposal.
Just
if
you
haven't
taken
a
look
yet
I
put
a
lazy
consensus
set
to
friday,
which
is
tomorrow.
So
if
you
haven't
taken
a
look-
and
you
want
to
take
a
look-
and
you
don't
think
you're
going
to
have
time
by
tomorrow,
can
you
please
let
me
know,
I'm
happy
to
push
it
back
a
bit?
D
If
anyone
wants
time
to
review
it,
but
otherwise
yeah
it's
there
and
I
opened
the
poc
pr
for
a
few
of
the
first
like
a
couple
of
services
to
show
how
this
works.
So
you
can
also
take
a
look
at
that
if
you're
interested.
D
A
It
doesn't
look
like
we
have
any
other
agenda
items.
Does
anybody
want
to
bring
anything
up
ad
hoc
here
or
go
ahead
and
add
something
you
can
also
discuss.
D
Cecile
yeah
I'll,
add
a
quick
one
since
we're
here.
What
should
we
do
about
the
experimental
folder
owner's
file?
So
we
talked
about
adding
owners
for
managed
clusters
and
the
thing
is:
we
can't
really
do
that
right
now,
because
the
managed
cluster
files
are
all
mixed
in
with
like
the
machine
profiles
in
the
same
directory
and
when
you
add
an
owner's
file,
it
has
to
be
per
directory.
D
E
I'll
just
talk
yeah,
I'm
I
realize
that
separating
the
folders
is
kind
of
from
my
organization
perspective
a
little
bit
annoying,
but
I
mean
for
ownership.
It
seems
like
a
clean
way
to
do
it.
So
I'm
fine
with
that.
D
I
think
the
good
thing
about
separating
the
folder
is
that,
since
they
are
experiments,
it's
also
easier
to
get
rid
of
one
of
them.
If
we
decided
to
delete
one
instead
of
having
to
because
right
now,
some
of
the
stuff
is
even
sharing
files
like
some
of
the
unit
tests,
have
both
unit
tests
for
machine
tools
and
managed
clusters,
and
so
I
don't
know
it
would
make
it
a
little
more
isolated
and
easier
to
just
rip
off
one
of
the
directories.
If
we
decided
to
close
an
experiment.
D
E
A
Any
other
thoughts
on
that.
What
else
do
we,
so
we?
What
else
do
we
still
have
in
there,
which
is
it
just
the
aks
work
or
is
there
other
things
in
in
the
experimental
folder
it's
machine
pulls
and
yes,.
F
Yeah,
sorry,
I
don't
have
a
ray
scan
feature
so
yeah.
I
agree.
It's
a
good
idea
to
separate
those,
so
it
would
make
a
managing
life
cycle
of
an
experimental
feature
easier.
Definitely,
you
know
like
even
graduating
it
to
next
level
or
something
like
that.
You
know
refactoring
with
our
reviews,
so
organization
wise.
It
could
be
nicer.
Definitely.
A
G
Yeah
thanks,
I
finally
found
the
raise
hand
button.
This
is
kind
of
trivial,
but
this
is
probably
the
only
group
would
care
we
had
a
and,
if
you
weren't
paying
attention
to
one
issue
that
came
up,
you
wouldn't
know
we
decided
kind
of
ad
hoc
a
month
or
so
ago
to
stop
publishing
ubuntu
1804
reference
images,
because
you
know
publishing
all
everything
is
a
hassle.
G
C
Sorry,
it's
fine
so
for
the
managed
aks
clusters,
I
think
most
of
the
things
said
within
the
security
case
baseline
with
capi
capsi
are
coming
to
a
closure.
You
just
wanted
to
know.
Are
these
the
only
features
what
we
are
having
in
plan
for
for
capsi
or
are
there
more
features
down
the
line
I'll
share
that
okay,
this
baseline,
maybe
in
the
chats.
A
E
I
missed
the
question:
what
was
the
question
specifically.
C
Yeah
I
was
just
asking
like
with
respect
to
managed
clusters
the
secure
aks
baseline,
whatever
features
what
were
discussed
over,
there
are
mostly
coming
to
closure,
most
of
them
so
down
the
lane
down
the
line.
Apart
from
whatever
issues
are
present
here,
do
we
have
any
other
features,
what
we
are
planning
or
is
there
a
roadmap
for
it?
I
just
wanted
to
know.
E
E
I
think
yeah
you're
right,
I
mean
they're
pretty
much
coming
to
a
close
from
my
perspective,
like
I
kind
of
want
to
play
around
with
the
unmanaged
machine
pools
and
like
integrating
that
with
the
managed
stuff,
because
that's
something
that
I'm
interested
in.
But
I
don't
know
how
what
other
people
want
to
work
on
or
what
other
people
need.
So.
A
Yeah,
if
you
have
anything
in
particular
that
you
want
to
see,
I
think
it's
okay
to
open
up
an
issue,
sicilian
wrist.
D
C
Work
on
I
mean
like
kind
of
mixed,
maybe
because,
like
I
think
there
are
more
features,
what
eks
kind
of
have
and
a
case
does
not
have
maybe
one
such
feature.
What
I
can
remember
right
now
is
like
the
auto
scaling
feature,
and
there
are
still
more
things:
what
the
aks
supports
and
caps
he
does
not
have
kalistan.
C
E
Yeah,
so
I
would
probably
want
to
see.
Like
I
mean
I
can
open
some
issues
for
things
that
I
see
as
gaps
so
like.
I
think
one
would
be
maybe
like
upgraded
test,
which
is
a
pretty
straightforward
one,
but
it's
probably
something
would
be
nice
to
see
for
like
day
two
operations
and
then
yeah.
I
think
the
in
my
model
it
was
like.
I
think
there
was
some
stabilization
work
that
we
needed
to
do
and
then,
like
adding
features,
I
don't
think
is
necessarily
like.
E
I
don't
think
a
lot
of
that
will
be
breaking.
I
think
it
will
just
be
like
add
a
field
like
populate
it.
You
know,
but
I
think
once
we
finish
kind
of
the
current
set
of
issues
that
we
talked
about
previously,
we'll
be
in
a
pretty
stable
place
to
do
that
to
just
be
adding,
like
you
know,
here's
a
new
set
of
fields
like
instant
validation,.
A
A
All
right,
I
think,
that's
probably
the
end
of
it,
then
it's
nice
seeing
everybody
and
we'll
catch
you
in
slack
or
see
you
next
week
or
two
weeks.