►
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
So
this
is
a
meeting
for
this
meeting,
abide
to
the
cncf
cotton
controls,
so
please
be
kind
with
each
other
and
we
have
a
meeting
agenda
document.
I'm
posting
the
link
in
the
chat.
If
you
cannot
access
it,
you
have
to
subscribe
to
the
castle
cycle.
Many
many
lists
so
for
getting
started.
A
Let's
start
by
welcome
new
attendees.
If
it's
there,
I
will
pause
for
a
moment.
If
there
are
some
new
attendees,
basically
feel
free
to
speak
up,
introduce
yourself.
Why?
Why
are
you?
Why
are
you
interested
in
class
api.
B
Oh
hello,
hi,
can
you
hear
me?
My
name
is
bharti
and
I've
joined
this
meeting
for
the
first
time.
I
have
an
agenda
to
discuss
but
yeah.
That's
it.
I
work
for
red
hat
as
a
software
engineer
and
hello.
Everyone.
A
A
C
Yes,
jack
has
been
moved
over
to
a
different
project
for
the
moment,
so
I'm
still
working
on
the
add-on
proposal.
I'm
gonna
make
some
more
edits
this
week
and
add
some
sections
about
dependencies
and
upgrade
strategies,
and
hopefully
we
can
take
another
look
next
week.
C
Oh
hero,
yes,
so
we
created
a
pr
and
we
didn't
get
any
reviews,
but
since
our
previous
google
doc
was
open
for
like
three
months,
I
think
we
reserved
we
addressed
almost
our
comments
and
there
were
not
very
controversial
comments.
C
So
please
take
a
look
at
it
again,
but
I
know
that
the
oci
team
is
actually
trying
to
use
the
proposer
to
implement
their
managed
kubernetes
and
also
in
kappa
we
are
actually
we
started
the
discussion
about
eks
api
evolution,
so
yeah
it'll
be
great
to
get
it
merged.
A
Okay,
so
if
I
look
at
what
we
did
for
trying
to
fast
forward
the
some
proposal
recently,
I
suggest
you
to
go
and
basically
ping
all
the
people
that
you
set
up
as
a
reviewer
and
so
ask
explicitly
for
feedback
from
those
viewers.
Everyone
asked
I
in,
but
your
first
goal
is
to
get
a
plus
one
from
the
regulars,
so
yeah
go
and
and
back
then.
A
B
B
So
this
has
been
about
adding
the
feature
to
fetch
user
cube
config.
That
kappa
provides
with
the
help
of
cluster
cutting,
get
cube
config,
and
there
has
been
a
lot
of
discussion
going
on.
So
I
was
hoping
to
get
a
move
on
this
one.
A
I
I
can
give
my
opinion,
so
steph
stephanie
is
not
here,
but
basically
so
just
to
give
a
little
bit
of
the
ground.
Oh,
usually
all
the
provider
generates
one
one,
cuba
country.
A
For
some
reason:
kappa,
when
using
manager
kubernetes,
they
decided
to
have
two
cuba
confit
one
is
the
one
that
is
used
by
copy
and
it
has,
and
it
basically
uses
the
yam
authentic.
The
the
authentication
mechanism
of
kappa
and
the
other
one
is,
is
a
plain
config
that
is
meant
for
user
to
be
okay.
A
A
The
pro
the
problem
is
that
so
we
we
initially
tried
to
define
a
kind
of
behavior
that,
where
we
look
from
the
user
config
and
then
we
pull
back
if
it
if
it
exists,
we
give
this
one
back.
Otherwise,
we
give
the.
D
A
Config
this
this
was
a
behavior
design,
basically
to
make
this
special
case
transparent
to
the
user.
The
problem
is
that
stefan
raised
a
very
good
point:
if
you
basically
transparently
move
from
one
kuber
config
to
another,
there
is
the
risk
that
you
have
airbag
escalation
in
both
direction.
You
can
have
more
privilege
or
less
privilege
than
you
expect
due
to
race
condition.
A
I
don't
know
the
controller
is
taking
he's
creating
one
of
the
two
of
the
config
before
the
other
will
start
like
that,
so
this
kind
of
defeats,
the
idea
of
having
this
implicit
escalation
that
we
were
discussing-
and
it
makes
sense,
is
a
reasonable
concern
for
me
and
in
this
comment
I'm
trying
to
discuss
some
possible
alternatives.
One
is
user,
must
explicitly
ask
for
the
user
config,
which
is
kind
of
weird,
because
it
is
it
is.
We
are
telling
the
the
kappa
user
that
for
managed
kubernetes
they
have
to
use
a
different
command.
A
A
C
D
A
Also
discuss
another,
we
can
think
about
a
strict
contract
for
the
provider,
so
they
ensure
that
cuba
configure
provisioning
in
a
certain
order,
but
this
is
weak.
Probably
the
solution
will
be:
let's
stable
the
discussion
and
try
to
define
what
what
is
the
more
than
one
kuber
conflict
story,
and
so
all
the
provider
align
instead
of
kappa
being
a
use
case.
But
I
understand
this
is
much
wider
discussion.
E
Yeah,
I
think
I
agree
with
fabrizio
on
this
one
I
think
like
so
solution
wise.
I
think
this
is
definitely.
E
This
definitely
can
lead
to
some
security
issues,
especially
if
someone
understand
on
which
specific
conditions
you
can
fall
back
to
one
cube,
config
or
the
other
so
yeah,
I
think
I'd
rather
like
probably
the
next
steps
would
be
something
like
I
don't
know
getting
getting
at
least
some
a
proposal
or
some
sort
of
document
to
discuss
what
would
be
the
contract
with
respect
of
like
provisioning
or
creating
configs,
so
that
at
least
like
we
can
have
like
discussions
about
the
trade-offs
of
enabling
multiple
cube,
config
one
cube
config
and
like
any
mechanism
to
select
one
or
over
the
other.
B
A
I
I
have
a
background
question,
however,
so
if
I
look
at
kappa,
kappa
today
support
basically
two
two
modes:
kappa
with
vuem
and
kappa
with
benz
and
kubernetes
the
question
my
question
and
sorry.
If
it
is
a
silly
question,
what
are
you
trying
to
understand
is
why
it
is
okay
to
use
a
plain
secret
for
kappa
with
vm,
while
instead
we
we
choose
to
use
a
secret
with
aws
authentication
for
magic
burnishing.
It
cannot
be
the
same.
B
Okay,
but
I
mean
not
with
the
cup
of
maintainers,
this
is
my
first
pr
with
cappy,
but
I
guess
I
can
take
that
back
with.
Maybe
richard
and
or
maybe
you
know,
copper
slack
general
then
ask
them
or
if
you
could
articulate
it
there.
That
would
help.
F
Yeah,
I
don't
know
if
richard
is
on
this
column,
but
I
think
the
differences
were
that
cap
eks
specifically
has
constraints
around
authentication
mechanisms.
You
can't
use
anything
other
than
I
am
authentication
unless
you've
configured
an
oidc
authentication
provider.
F
A
A
A
It
is
kind
of
hard
to
make
an
exception
on
our
own
inside
capital.
We
have
to
work
on
a
proposal
problem.
Probably
I
like
I
just
seen
I
was
saying
otherwise,
I'm
open
to
any
suggestion
how
to
make
a
user
experience
nice
if
we
can
prevent
any
security
issue.
B
C
C
A
Thank
you
thank
you
and
yeah
welcome
again
to
this
meeting.
So
okay,
moving
on
next
point
is
mine.
So
last
week
we
discussed
that
we
are
changing
a
little
bit,
how
we
are
doing
issue
triage,
so
we
we
started
with
the
new
approach.
So
now,
basically
we
have.
A
If
we
look
at
our
issue
backlog.
If
you
look
at
milestone,
we
have
only
one
milestone,
which
is
the
1.3
and
we
are
going
through
the
backlog
and
do
two
things.
A
There
are
one
person
assigned
and
there
is
a
a
kind
of
commitment
to
get
the
issue
completed
before
the
next
release,
which
is
not
planned
yet,
but
we
can
keep
some
time
and
end
of
the
year
as
a
reference.
So
what
we
decided
last
week
in
last
week
we
had
on
friday
the
first
triage
session.
A
We
decided
to
have
issue
triage
session
every
friday.
A
10
am
pacific
time
using
the
same
link
that
that
we
are
using
now
and
we
decide
to
go
through
the
entire
backlog
to
apply
transcend
that
our
clause,
whenever
whatever
we
can
close,
we
are
at
the
20
now
more
or
less
at
20
percent
of
the
effort.
So
there
is
still
a
way
to
go.
This
will
keep,
will
take
some
fridays
and,
as.
A
D
Yeah
real
quick,
I
mean
just
call
out
like
that.
Claustrophobia
is
having
another
momentum
already
like
at
the
u.s
north
america
kubecon.
There
is
one
two
three
four
five,
six
seven
eight
talks
like
about
cluster
api.
I've
strolled
down
a
little
bit
and
you'll
see
them.
B
D
D
I
don't
know
what
happened
there
and
I'll
reach
out
to
cncf
to
see
if,
like
they
didn't,
receive
the
submission
or
something
something
similar,
but
but
yeah
at
least
we
have
the
sickle
cycle
intro,
which
we'll
do,
and
the
last
thing
to
mention
is
that
at
the
at
this
kubecon
cappy
will
be
introduced
and
I
think
we'll
actually
be
introduced
on
stage.
So
that's
pretty
exciting.
A
Thank
you,
and
just
to
follow
up
to
another
topic
that
we
talked
last
week.
So
it
was,
we
decided
the
the
name
for
the
vpn
and
family
character
that
we
represent.
Cluster
bi
cncf
is
trying
to
to
make
this
happen
for
the
next
kubercon,
and
if
this
will
happen,
usually
the
new
characters
are
being
announced
on
stage
and
so
yeah.