►
From YouTube: 20200408 - Cluster API Office Hours
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
B
A
A
So
we
will
add
new
features
and,
like
I
tried
to
get
the
milestone
that
we
currently
have
for
do,
3x
in
the
issues
and
Link
kind
of
like
the
proposal
or
the
bigger
issues.
One
thing
that
I
wanted
to
point
out:
I
also
added
this
end-to-end
test
plan,
which
is
a
spreadsheet,
and
this
is
a
public
spreadsheet
and
we
started
like
kind
of
like
tapping
down
like
which
seniors
we
want
to
test.
A
What
are
the
sub
tests
for
the
scenarios
and
what
are
the
expected
results?
So
if
your
infrastructure
provider,
we
should
provider
or
a
user,
please
take
a
look
at
this
and
we
wanted
to
get
most
of
these.
That's
done
in
the
zero
3x
time
frame
so
that
we
can
like
become
a
stable
project
and
a
robust
project.
A
Okay,
other
things
to
note
like
we
have
some
more
testing
focus
so
enable
Web
books
in
tests.
This
is
probably
like
a
larger
issue
that
we'll
have
to
tackle
in
multiple
pieces.
We'd
have
a
few
odd,
kcp
adoption.
I
know
you
relic.
Folks
have
already
appear
open.
It
just
needs
another
review.
So
if
you're
interested
in
reviewing
it,
please
do
robustness,
there's
a
whole
plan,
and
this
issue
please
take
a
look
and
provide
feedback.
A
A
A
What
else
do
we
have
support
for
control,
plane
own
machine
health
check?
We
have
been
having
like
quite
a
bit
of
a
discussion
on
this
one.
How
should
be
implemented
but
yeah?
This
is
definitely
something
that
like
we
would
like
to
see
it
done
for
alpha
3
timeframe,
remote
workload,
cluster
watcher.
This
is
a
great
feature
that,
like
I,
think
Joe
and
NV
are
still
working
on.
There's
a
PR
open
for
this,
so
this
is
like
imagine,
you
have
a
management
cluster.
Do
you
want
to
watch
like
a
work
workload?
Cluster
nodes?
A
A
We
need
a
lot
of
work
to
be
done
on
this.
It
can
be
done
in
a
1:43
time
frames
if
you
have
experiences
or
we
want
to
provide
feedback
on
what
you
would
like
to
see
in
a
condition.
Please
add
your
comments
to
as
issues
spices,
support,
I,
think
the
disappear
of
the
proposal
married
jewel
or
still
open,
yeah,
okay,.
A
Is
a
PR,
opener,
I
I
think
it
was
almost
ready,
like
I,
think
we
just
need
some
approvals
and
things
like
that
machine
shop,
failure,
detection-
this
is
kind
of
gonna-
be
probably
a
tricky
one.
We
need
like
some
support
from
infrastructure
providers,
machine
health
check
and
it
needs
a
proposal
pretty
much
load
balancers.
A
This
is
probably
gonna
be
a
little
bit
of
a
stretch
unless
we
can
find
more
people
to
work
on
it,
but
this
is
pretty
much
to
have
like
a
way
to
have
pluggable
load,
balancers
and
a
primitive
time
for
post
repair
to
understand
what
balancers
and
then
extensible
machinery
delete.
Hooks
I
think
so
this
issue
that
it's
linked
here
it
it's
kind
of
more
specific.
What
we
would
like
to
have
like
more
like
generic
hooks
at
some
point.
This
also
needs
a
full
proposal.
Probably
questions
comments
concerns.
D
E
Wake
ya:
hey
thanks,
Vince
I,
guess:
I
had
kind
of
pinned
you
on
slack,
but
I,
guess
I'm,
trying
to
understand
like
what
what
type
of
topics
would
fit
best
into
the
roadmap,
because
we
have
a
few
issues:
we're
working
on
in
the
autoscaler
that
directly
relate
to
cluster
API
and,
as
I
said
in
the
issue
like
I'd
love,
to
get
more
visibility
on.
What's
going
on
there
so
I'm
curious,
like
does
it
make
sense
for
us
to
kind
of
cross
post
that
stuff
into
this
road
map?
I'm
just
curious,
yeah.
A
Like
I
mean
I
I,
don't
know
if
you
need
like
any
code
changes,
maybe
like
some
minor
ones,
but
well
definitely
some
documentation
right,
because
this
feature
will
be
supported
and
I
got.
A
lot
of
people
have
asked
for
it.
So
there's
very
interest
so
yeah
I
would
add
it
I
asked
for
like
an
issue
that
I
can
link
to.
So
let
me
just
pop
that
one
in
and
yeah
we
can
track
it
in
the
issue
very.
A
F
Yeah,
just
a
quick
update,
the
one
off
of
three
blog
post
is
that
yeah?
That's
it
that's
the
preview,
there's
a
pull
request,
it's
being
reviewed
by
the
blood
of
the
Covenant
blog
team
right
now.
You
know.
If,
if
anybody
has
last
minute
comments,
please
please
leave
them
on
the
pull
request
and
yeah
I
want
to
say
you
know
thank
you
to
everybody
who
gave
feedback
wouldn't
be
possible.
You
know
to
write
this
with,
went
out
without
your
feedback.
Without
your
help
and
yeah
that
just
wound
up,
they
said
so
I
think
it
will
be.
A
A
The
next
item
on
the
agenda
is
general
questions.
This
is
just
like
a
space
for
everybody
like
if
you're
new
you're
implementing
your
provider
feel
free
to
ask
any
questions
like
small
or
big.
It
doesn't
matter
like
where
you
want
to
help
so
I'll
just
like
stop,
and
if
anybody
wants
to
raise
their
hand
or
like
a
mute
like
feel
free
to
do
so.
A
A
B
B
I
have
done
this
was
this
question
was
entered
by
Rudolfo,
who
is
the
ship
architect?
He
we
had
been
asking
questions
on
specifics
last
two
weeks
and
then
I
concurred
with
them.
The
advice
you
give
that
you
go
ahead
and
use
just
whatever
cluster
API
provides
0.3
now
probably
dot
for
whatever
comes
the
latest
will
pick
up
and
we
go
use
that
so
whatever
substitutions
and
whatever
layered
we
have
to
work
through
customized
will
do
that,
so
that
is
resolved
right
now.
B
So
this
question
is:
if
we
have
a
cluster
and
if
the
cluster
has
kubernetes,
which
is
some
version
X,
and
then
we
have
to
move
to
ocean.
Why?
How
does
cluster
API
handled?
That
is
it
underneath?
Because
cluster
API
is
obviously
committed
to
using
kubernetes
api
correct
if
I
am
right?
If
that's
the
case,
if
the
kubernetes
itself
is
changed,
the
updates
occur.
Does
it
handle
that
or
what
does
it
do?
That's
the
question:
is
it
the
responsibility
of
cluster
API
to
update
the
kubernetes
components
which
configures
and
runs
in
the
cluster?
A
Yeah,
yes,
like
some
clarification
like
taking
a
step
back
like
closure,
API
right
now,
like
works
with
the
assumption
that
you
are
infrastructure.
Is
it
I
understand
it's
like
a
big
statement,
but
we
working
towards
like
to
get
immutable
infrastructure
there
first
and
then
like
for
mutable
infrastructure?
That's
like
a
separate
topic
in
like
we
cannot
defer
that
for
now.
So.
B
A
So
we'll
regard
with
this
question,
I
said
rather
like
like.
Is
it
across
Turkey
a
responsibility
to
update
current
components
of
a
running
cluster?
Immediate
answer
is
no
it's
not
and
like
we
don't
that's
like.
Actually
something
like
a
cube
idiom
operator
could
do.
There
is
a
proposal
out.
If
somebody
can
link
into
here
for
bits,
you
could.
Could
you
get
in
here?
Maybe
look,
there's
a
proposal.
A
Thank
you.
So
it's
not
necessarily
the
role
of
cap.
You
could
do
that
for
mutable
infrastructure,
so
lay
that
means
like
what
I
mean
by
that
is.
Like
you
have
a
running
machine,
that's
running
one
16
and
you
want
to
take
the
same
running
machine
164,
closure.
P
I
can't
do
that
today.
What
we
can
do,
though,
is
to
replace
it,
so
we
create
a
new
machine
with
a
new
version
and
remove
the
old
machine,
and
this
is
how
like
machine
deployments,
works
and
kcp
works
as
well.
B
B
I
think
that
answers
the
question.
I,
don't
think
I
have
anything
more
than
that,
because
that
justifies
the
reason
is
because
we
will
have
a
ship
100
running
on
some
kubernetes
and
then
2.0
in
some
other.
So
the
migration
has
to
address
this
portion.
If
we
do
migrate,
then
we
have
to
include
the
replacement
as
part
of
this.
B
A
B
A
I
mean
compatible
with
deserve
the
the
right
to
do
breaking
changes
between
API
releases,
but
we
will
document
them
as
we
did
with
the
with
the
alpha
2
to
alpha
3.
So
the
book
has
this
whole
document
for
infrastructure
providers
or
like
well
actually
any
providers
that,
like
a
documents,
all
the
change
that
we
need
from
one
version
to
another
and
we'll
do
the
same
when
we
go
from
at
4012
before.
But
I
cannot
speak
to
that
right
now
because
we
don't
know
what
that
entails.
Yet,
okay,.
B
No,
the
reason
why
is,
if
you
use
cin
well,
which
is
the
versioning
a
major
version
only
like
immutability
to
mutability.
Let's
say
if
that
is
going
to
be.
That
is
a
major
one.
I
mean
that
cannot
happen
minor
versions,
but
in
minor
versions.
If
you
keep
compatibility,
that
will
be
great,
but
if
not,
then
we
will
have
to
see
how
we
go
about
it
as
a
classic.
A
B
E
Yeah
I
just
wanted
to
speak
to
what
Prakash
was
talking
about
in
terms
of
the
strip's
mber
following
strict
cember
like
what
you
know.
What
I,
what
I
realized
from
looking
at
this
project
is,
even
you
know,
from
the
v
1
alpha
to
2
V
1
out
for
3
I.
Don't
think
I,
don't
think
we're
following
some
very
the
way
that
it's
specified
upstream
so
I,
just
I,
wouldn't
I,
wouldn't
expect
that
you
know
this
to
follow
cember
as
it's
specified
at
cember
org.
E
G
H
I'll
go
out
on
a
limb
and
ask
a
question
for
a
co-worker
mine.
We've
been
looking
at
the
the
copy
for
OpenStack
and
we're
trying
to
figure
out
how
to
fill
in
the
different
options.
For
that.
So
the
first
question
is:
is
that
an
appropriate
thing
to
bring
up
here,
but
the
second
is:
is
there
a
way
to
help
automate
getting
the
default
values
that
you
need
in
order
to
connect
to
an
open
stack.
A
H
H
A
So
for
the
environment
variables
like
we
see
if
you
look
at
the
cluster
template
without
I'll,
be
here.
If
you
don't
know
this,
you
will
find
you
in
here
that
there
will
be
environment
variables.
So
if
they're
not
documented
I
would
say
like
let's
open
an
issue
and
like
make
sure
they
like
it
their
document
here,
because
I'm
not
sure
that
yeah,
the
other
providers
have
already
thought
amended
like
yes,.
A
But
yes,
I
like
it,
that's
that's
like
where
it
gets.
The
environment
variables
is
in
this
cluster
template.
So
if
you
open
it
up
and
like
look
at
the
environment
variables,
you
probably
find
the
one
that
you
need,
but
my
first
suggestion
is
to
actually
like
open
an
issue
or
make
sure
that
we
update
the
big
start.
F
Yeah
III
forgot
to
ask
I
in
the
blog
post.
There
are
a
few
places.
Where
are
you
know
where
we
think
they're
like
major
contributors
to
you
know
some
of
those
some
of
the
big
features
I,
you
know,
I
I
wrote
those
sort
of
from
from
you
know
my
perspective,
my
memory.
If,
if
you
feel
that
you
know
you
should
be
included
in
you
know
in
the
in
those
clouds,
please
you
know
you
can
ping
me
on
slack.
F
A
A
A
I
A
Yeah
I
I
would
agree
sort
of
like
do
you
think
you
think
we
can
close
this.
I
I
think
like
I
mean
worst
case
I.
Can
you
can
just
set
it
to
an
empty
variable
or
like?
If
you
want
to
just
see
their
configuration,
you
know
turned
it
like
I
mean
we
could
have
it
like
I,
don't
know
what
dry
run
option
that
like
prints
it
out,
but
I
think
that's
just
like
what
you
would
download
right.
I
G
A
J
I
took
a
while
to
find
the
reset
button,
but
if
you
can
just
come
back
to
the
other
issue
I,
so
it's
it
I
do
get
the
point
of
the
issue
because,
what's
confusing
is
that
when
you
do
cluster
cuddle,
config
provider
and
the
same
thing
applies
with
Azure
as
well?
All
you
get
as
info
from
there.
Those
aren't
those
variables,
so
it's
kind
of
like
you
have
to
set
them
to
be
able
to
know
which
ones
you
needs.
So
it
doesn't
really
make
sense.
I
There
is
a
they
issue
about
templating,
where
basically,
we
are
discussing
how
to
improve.
They
are
the
user
experience
around
around
variables,
shown
valuable
validation
and
so
on
and
so
forth.
So
I
I
suggest
to
discuss
disappointed
there
for
what
is
experimenting
now,
which
is
a
really
simple,
is
basic.
I
J
A
E
I,
just
one
of
the
issues
you
cover
it
early
on,
there
was
like
I
think
it's
actually
the
docs
issue
and
I
think
look
that
the
person
who
added
that
I
think
was
asking
questions
in
chat
the
other
day
and
I'm.
Looking
at
our
Doc's
and
I.
Think
I
think
that's
an
odd
name.
I
think
we
can
actually
close.
I
was
just
trying
to
put
a
link
there.
His
cap
T
this
one
about
cap,
T
air
running
cluster
cuddle
in
it.