►
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
All
right,
hi,
everyone
today
is
Wednesday
May
3rd,
and
this
is
the
cluster
API
office
hours
thanks
for
joining.
If
you
are
new
and
joining
for
the
first
time,
just
know
that
you
can
get
access
to
the
agenda
by
joining
with
the
cluster
lifecycle
mailing
list,
and
also
please
remember
that
we
are
following
the
cncf
code
of
conduct.
A
Our
meeting
etiquette
is
to
use
the
raise
hand
feature
on
the
zoom
if
you
want
to
speak
and
I'll,
make
sure
to
watch
the
hands
and
call
on
you,
and
if
you
have
a
topic
that
you
want
to
bring
up
or
any
questions
or
any
general
feedback
feel
free
to
add
it
to
the
discussion
topics
under
the
agenda
and
as
well.
If
you
could,
please
add
your
name
to
the
attendee
list
that
helps
us
keep
track
of
who's
attending
these
meetings.
A
With
that
we
have
a
first
item
in
the
agenda.
We
usually
give
a
second
for
folks
who
are
new
to
introduce
themselves
if
they
want.
So
this
could
be
just
telling
us
a
bit
about
why
you're
here,
what's
interesting
to
you
and
yeah
just
a
bit
about
yourself.
So
if
anyone
is
new
and
wants
to
say
hello,
I
will
mute
and
now
is
your
chance.
B
Or
maybe
I
will
say
something:
Michael
Hill,
hello,
I'm
new
here
for
the
first
time,
I
was
told
about
those
meetings
on
keep
calling
Amsterdam
on
one
of
Cappy
talks
and
I'm
very
glad
that
I'm
joining
in
this
particular
meeting,
because
I
just
recently
stumbled
Upon,
A,
Helm
provider
and
I
have
a
few
questions
regarding
it,
and
also
since
the
company
I'm
working
for
is
at
the
moment
heavily
investing
in
development
of
the
platform
using
cluster
API
I
would
like
to
begin
somehow
contributed
to
the
project.
So
thank
you.
A
All
right,
let's
keep
going
then
so.
First
item
is
the
open,
Puzzle,
readout
I,
don't
know
if
we
have
anything,
there's
one
I.
Think
Jack
are
you
here?
Do
you
want
to
give
an
update
on
that?
One.
C
Yeah
I,
just
added
that
I
was
racing
and
I
looked
like
I
made
it.
So
this
is
coming
out
of
the
managed
kubernetes
feature
group
I've,
the
The
Proposal
is
still
a
work
in
progress,
but
there's
a
lot
of
discussion
so
I
think
it's
worth
moving
it
up
to
this
list
going
forward.
C
So
this
is
going
to
propose
a
change
to
the
info
cluster
requirement
that
providers
Implement
when
implementing
cluster
API
there's
a
lot
of
details
there,
and
there
is
some
active
discussion
on
on
maybe
some
some
of
the
scope
and
we
might
actually
change
the
title.
There
have
been
some
good
feedback
on
that,
but
yeah.
Please
link
to
the
click
on
the
pr
link
and
join
the
discussion
if
you're
interested
in
manage
kubernetes
and
how
we
can
make
that
a
little
bit
more
consistent
across
providers.
A
Thanks
I
added
during
their
feel
free
to
add
other
co-authors.
If
you
want
anyone
got
any
questions
for
Jack
about
this
proposal,.
A
All
right,
let's
see
discussion
topics.
All
right,
I
am
going
to
pass
it
over
to
the
release.
Comms
team.
First
to
tell
us
about
the
new
patch
roses.
D
Yes,
hi
everyone
Johannes
here
from
the
cons
team,
yeah
I
just
wanted
to
quickly
announce
that
copy
release,
one
for
two
and
one
three
seven
have
been
released
yesterday.
D
I
should
surely
most
of
you
have
already
read
it
in
in
slack.
But
if
you
want
to
yeah
read
about
the
details,
you
can
click
on
the
links
you
can
see
which
patches
and
Bug
fixes
have
come
into
the
release
yeah
and
the
next
one
143
and
one
three
eight
are
planned
for
in
about
four
weeks
on
30s
of
May.
So
these
are
the
next
ones
for
the
future
yeah,
that's
it.
Thank
you.
A
Great,
thank
you
so
much
and
thank
you.
The
really
Steam
and
congratulations
on
your
first
patch
releases
in
the
one
fire
cycle.
Great
work,
all
right.
Any
questions
about
the
release.
A
Okay,
if
not
moving
on
for
britsu,
you
have
the
next
one.
Oh
sorry,
Stefan,
please
go
ahead.
E
Yeah,
just
one
comment
just
want
to
highlight
at
1.4.2
supports
kubernetes
127
now
I
think
you
didn't
mention
it
right,
Johannes,
just
for
people
who
are
looking
to
use
it.
F
F
So
first
I
saw
someone
trying
to
assess
the.
G
F
Meeting
notes,
if
you
want
to
assess
you,
have
to
subscribe
to
the
sequester
excitement
list,
I
posted
the
the
link
to
the
mailing
listing
to
the
into
the
chat.
Okay.
Moving
to
my
my
topic
in
agenda.
So
there
is
an
issue
in
copy
which
is
linked.
It
is
about
migrating.
The
CSI
driver
out
of
three.
This
issue
was
opened
by
folks
working
on
the
migration
and
it
is
pretty
generic.
It
was
like
a
call
for
Action
to
Cluster
API.
Folks,
hey,
please
get
ready
for
this
change
that
is
happening.
F
Similar
change
are
planned
for
CPI
or
already
happen
for
CPI,
and
so
there
is
some
discussion
around
back
and
forth,
but
this
is
mostly
not
a
topic
for
carcass
API,
but
is
a
topic
for
advanced
solution
like
the
end
provider
or
eventually
for
people
using
the
other
tools
to
install
a
tone
like
guitars
or
for
a
people
using
clusters
set.
But
it's
not
a
change
or
requirement
because
I
really
said
it
is
only
that
Forks
has
to
be
ready
for
the
for
the
for
the
immigration.
F
Which
is
provider
specific
in
in
those
migrations?
F
Also
the
fact
that
a
different
provider,
different
infrastructure
Cloud
are
doing
this
migration
at
different
timing,
but
at
the
same
time
it
is
a
common
problem
across
all
the
cluster
API
provider.
So
this
is
a
call
for
action
for
people's
in
providers
working
different
clouds
working
on
it.
F
If
you
face
some
problems
or
you
find
some
smart
Solution,
please
report
back
to
the
team
because
we
can
work
on
Saturday
night
in
survey
program
or
we
can
share
knowledge
and
make
the
life
easier
for
the
people
that
are
going
to
migrate
after
you
want
to
add
something.
On
top
of
this.
A
Yeah
no
I
think
he
summarized
it
well.
I
think
it's
something
that
we
can
still
I
think
do
some
thinking
on,
because
the
current
user
experience
of
having
to
know
to
install
cnis,
CSI
drivers,
copywriter,
is
not
necessarily
the
best,
especially
since
usually
Cloud
providers
or
infrastructure
providers
will
have
an
opinion
on
what
users
should
install
for
what
version
Etc.
A
A
But
then
we
also
want
to
allow
modularity
around
add-ons,
so
I
think
there's
yeah
some
definitely
some
thinking
we
can
do,
but
for
now
I
think
treating
the
CSI
drivers
the
same
way.
We
treat
cni
and
cloud
provider
makes
the
most
sense.
A
All
right
was
this:
mostly
a
just
a
call
for
awareness
for
reserved.
Did
you
want?
Is
there
any
action
that
we
need
to
take
on
this?
No.
F
It's
for
awareness,
because
there
is
decision
in
Castro
piaga.
Basically,
we
cannot
it's
not
actionable
for
for
us,
so
it
is
a
moral.
A
folks,
I
know
that
you
are
working
on
it.
Let's,
let's
work
at
this
on
this
as
a
community
as
much
as
possible.
A
Makes
sense
anyone
else
has
thoughts
or
questions
or
comments
on
this.
A
Moving
on
next,
one
is
also
ubizu
go
ahead.
Yeah.
F
So
there
is
a
cap
in
kubernetes
which
is
getting
near
to
approval,
so
it
already
has
a
summer
program.
This
cap
is
from
Journal
legit
and
it's
basically
even
if
the
title
is
not
super
clear,
but
what
is
is
basically
defined
in
advocating
foreign.
F
The
skill
test
for
kublet,
so
the
SQL
support
for
kubernetes,
so
kubernetes
support
Ministry,
like
API
server.
Does
the
interesting
part
is
that
by
extending
this
support
and
the
Happy
by
itself,
it
is
interesting
read,
but
the
interesting
part
of
it
is
that,
given
that
we,
as
kubernetes
release
three
version
of
per
year
with
this
cap,
it
will
be
possible
to
do
what
Jordan
is
calling.
The
streamlined
upgrade
does
mean
that,
once
a
year,
people
can
do
a
simple
upgrade.
F
F
Nodes
and
five
and
fifty
workers,
basically
the
50
workers
roll
out
only
one
time,
I
think
this
is
interesting.
I
don't
know
so.
The
question
that
I
have
for
the
team
here
is:
is
this
an
operational
practice
that
people
follow
and
if
the
and
I'm
trying
to
make
up
of
my
mind
is
if,
if
it
is
something
that
we
have
to
start
thinking
to
automate
in
Caster
API
as
well,.
H
Yeah,
so
I
don't
do
this
anymore,
but
earlier
versions,
when
we're
doing
more
active
first
migrating
kubernetes,
we
used
to
do
this.
We
would
fall
behind
and
we
would
upgrade
our
control
plane,
verify
everything
worked,
upgrader
control
plan
again
and
then
replace
the
nodes
to
catch
up
to
the
control
plane.
Just
because
you
know
we
had
the
N
minus
two
and
cubelet
would
work
fine
on
those
compute
nodes,
but
we
we
didn't
get
more
than
more
than
that.
A
C
I
think
it
would
be
terrific
if
Kathy
could
have
first
class
test
that
ran
through
the
scenario
after
this
kept
lands.
C
I
think
we
have
The
Primitives
in
place
in
the
end-to-end
suite
and
cluster
API
that
we
could
compose
a
test
like
this.
Is
that
the
seem
feasible.
E
I
E
Where,
basically,
we
upgrade
from
I
don't
know:
120
21,
121
2022,
those
are
separate
shops
that
we
have
in
core
copy,
and
then
we
have
one
which
you're
just
running
whenever
we
run
all
end-to-end
tests
and
that's
basically
from
the
previous
version
to
the
to
our
current
newest,
supported
version,
so
I
think
we
just
have
to
pick
basically
what
we
want
to
do.
But
I
guess
we
just
pick
the
newest
one
that
we
have
and
then
we
go
back
minus
three
and
then
we
can
do
this.
E
I
think
we
have
to
either
write
a
new
test
or
just
extend
the
current
upgrade
test
that
we
have,
but
I
think
it
shouldn't
be
too
complicated.
I
think
we
can
do
it
with
and
without
Dusty
class.
As
far
as
I'm
aware
of
what
cluster
class
can
do
and
yeah
we
might
need
like
one
or
two
utils,
but
just
guessing
you're
a
bit,
but
it
shouldn't
be
too
hard.
C
Yeah
that
that
sounds
about
right,
I
mean
I,
assume
it's
not
going
to
be
completely
easy,
but
any
additional
things
that
we
would
add
would
probably
be
minor
and
could
be
a
part
of
that
Spike
do
do.
We
have
M
minus
three
right
now
from
a
single
version
of
Cappy,
or
is
that
gonna
be
something
that
we
actually
change
when
this
kept
lands.
E
F
F
So
if
I'm
I'm
doing
n
plus
one
I
do
one
way,
if
I'm
doing
n,
plus
two
or
n
plus
3
I,
do
the
control
plane
first
and
then
work
out
notes
and
I
go
through
intermediate
version,
so
it,
let
me
say:
automating
end-to-end
test
is
the
first
step,
but
then
out
automating
the
operation
making
the
operation
simple
for
users.
My
penis
is
the
the
next
one,
but
yeah
please
take
a
look.
Eventually,
we
can
open
I
can
open
an
issuing
copy
as
well.
A
F
I
I
answered
quickly,
so
we
are
not
enforcing
skill
in
general
and
we
have
a
couple
of
issues
around
these.
What
we
have
is
in
cluster
class,
we
don't
we
do
not
upgrade.
We
don't
want
to.
We
do
not
allow
to
upgrade
out
of
this
queue
so.
G
A
I
got
it
thanks,
let's
see
I,
don't
know
who
is
first
but
Vince
since
you
haven't
spoken
yet
go
ahead.
G
J
Don't
know
if
you
saw
like
there
is
LDS
support,
like
a
working
group
like
trying
to
be
formed,
so
I
came
in
later.
It
was
already
discussed,
but-
and
we
talked
to
you
yesterday
in
the
closure-
life
cycle
and
I
am
really
interested
in
the
feedback
that
this
community
can
give
into
LTS.
For
a
number
of
reasons
like
there's,
nothing
decided
to
get
necessarily
like
this
first
cap
that
that's
happening.
It's
like
basically
like
just
a
way
to
simplify
today
and
allow
basically
like
worker,
note
to
live
at
least
a
year.
J
J
J
All
of
these
things
are
still
open,
but
if
you
just
call
out,
if
you
do
have
any
feedback
like
or
like
experiences
in
like
managing
upgrades
like
how
to
make
them
more
simple
again,
this
is
a
very
good
first
effort
or
also
like
what
problems
like
you
all
have
like
gone
through
updates,
like
I,
can
think
a
few
myself
like
and
they're
most
related
to
feature
dates,
but
but
yeah
so
like
that.
J
That
would
be
great
in
terms
of
what
you
said
before,
like
about
what
we
want
to
do
like
I
say
like:
let's
strive
to
do
like
the
simplest
things
for
users
as
much
as
possible.
Like
you
know,
the
gold,
the
project
ultimately
like,
is
to
make
cluster
life
cycle
easy,
which
is
kind
of
a
big
challenge.
Anyways
safe,
maybe
like
in
the
short
term.
We
could
support
like
hey
I
wanna,
like
upgrade
the
control
plane.
First,
like
three
versions
ahead.
J
Sure
you
could
do
it
like
tremendous
apologies
like
we
don't
block
you
anymore,
but
then,
if
I
want
to
like
automatically
like
skip
upgrades
through
worker
nodes,
like
do
we
require
them
to
go
one
by
one.
You
know
like
things
like
that,
so
we'll
have
to
take
those
edge
cases.
E
Yeah,
just
at
this
point
short
spoiler,
I
guess
so
we're
kind
of
thinking
about
the
skew
problem
of
it
and
how
we
can
make
this
better
than
this
today.
So
there
will
be
probably
something
coming
up,
though
I
guess
the
next
one
or
two
weeks
or
something.
But
we
are
still
thank
you
for
some
details.
I
would
say.
A
Thanks
cool
I
think
just
bouncing
on
what
Vince
said.
It
would
actually
be
really
interesting
to
hear
more
from
Costa
Rica
users.
How
they're
doing
upgrades
today
like
in
production
like
what
are
people
actually
doing?
Are
you
building
new
clusters?
Are
you
replacing
your
note
pools?
Are
you
upgrading
in
place?
Are
you
using
cluster
class
I'd
love
to
hear
more
like
either
here
or
maybe
we
can
serve
an
issue
just
together?
Some
feedback
and
stories
Mike
go
ahead.
H
I
can
comment
in
the
issue
if
you
don't
want
to
hear
it
now,
but
I
can
do
it.
We
do
version
upgrades
one
at
one
version
at
a
time
and
then,
when
possible,
we
are
not
swapping
load.
Node
definitions,
we're
just
upgrading
in
place,
but
we
do
try
to
when
possible
utilize
another
tool
that
we
have
called
cage
Shredder.
H
It
is
in
the
Adobe
GitHub
if
you
wanted
to
take
a
look
at
it,
but
it
kind
of
helps
with
some
of
the
draining
with
a
little
bit
of
business
logic
around
you
know,
crash
loop
back
offs
and
pdbs
that
cluster
API
can't
really
handle,
but
yeah
one
one
version
at
a
time
just
for
trying
to
keep
up
to
date
on
a
more
regular
basis,
but
we
do
swap
node
pools
when
necessary.
A
Got
it
thanks
for
sharing
I,
think
it'd
be
great
to
yeah,
get
more
answers.
So
if
we
can
start
like
a
discussion
or
survey
or
something
that'd
be
good.
A
All
right
cool
Vince:
do
you
want
to
take
sending
out
some
questionnaire
for
the
discussion
whatever
it
is,.
J
Yeah
somebody
wants
to
collaborate
that
would
be
great.
I
could
prepare
like
a
list
of
questions,
then
see
I
think
you
could
probably
do
both
like
actually
like
a
questionnaire.
We
can
send
out
to
the
mailing
list,
so
people
can
do
it
I
think.
Currently,
it's
more
guided
I
would
think
like
for
folks
that
are
like
you
know
that
don't
use
GitHub,
that's
easier
to
just
like
push
it
out
and
then
also
link
to
the
database
discussion
and
also
we
can
post
the
results
publicly.
G
A
Know
it
can
also
be
a
sharing
thing:
knowledge
sharing,
Stefan
I'm.
A
Okay,
let's
maybe
follow
up
on
the
details
of
the
how
but
I
think
just
yeah.
Let's
get
some
feedback
from
people
cool
anything
else
on
this
topic,
the
cap
or
LTS
or
anything.
A
Well,
if
you're
posting
links
in
the
chat,
those
don't
stay
their
ephemeral.
So
if
you
want
any
of
the
links
that
are
shared
to
be,
you
know
discoverable
after
this
meeting
ends.
Please
share
them
in
the
meeting
notes.
A
All
right,
Mike,
you
have.
The
next
topic
go
ahead.
H
Yeah,
it's
just
a
quick
one.
I've
been
writing
up
my
notes
on
the
managed
control
plane
importing
I
know.
Quite
a
few
people
ask
about
that
on
a
kind
of
a
regular
basis.
H
I
got
the
Kappa
stuck
together
and
I'm
working
on
the
cap,
z
ones
today,
but
I'm
just
not
sure
what
the
the
best
way
to
kind
of
share
these
more
with
the
team
to
allow
commenting
and
feedback
generally
PR's
by
the
way
that
that
kind
of
goes
but
I
wouldn't
know
where
the
pr
this
to
or
if
a
Google
doc
somewhere
is
better.
A
I
guess
what
kind
of
feedback
are
you
looking
for?
Is
this
meant
to
be
like
a
guide
at
the
end
or
is.
H
This
yeah
I
think
a
guide
is
kind
of
where
I'm
looking
to
go
just
because
I
I
do
notice.
In
slack,
you
know
at
least
at
least
once
a
week,
if
not
more
often,
people
are
kind
of
asking.
How
can
I
get
my
eks
or
my
AKs
cluster
into
cluster
API
and
so
I'm
just
trying
to
figure
out
if
there's
a
a
way
to
share
this?
Somehow
just
you
know
at
some
point
I
like
to
share
it
with
the
Greater
Community
I
just
don't
know
what
that
would
be
exactly.
C
Yeah,
this
certainly
reminds
me
of
a
I
think
was
earlier
this
week,
a
conversation
in
cap
Z
about
how
to
better
integrate,
manage
kubernetes
docs
into
the
cap
c
book,
so
I
think
Mike
is
totally
right.
There
is
a
lack
of
like
really
detailed
reference
materials
for
this
and
Cappy
overall
and
putting
it
I.
Think
Cecile
made
an
observation
in
a
slack
thread
that
putting
it
into
the
cluster
API
book
is
is
hard
because
there's
a
lot
of
sort
of
like
in
the
there's.
C
A
lot
of
this
stuff
is
already
sort
of
covered
in
the
existing
materials,
and
so
there's
a
lot
of.
If
then,
if
you're,
using
managed
kubernetes
consider
this
else,
do
it
the
default
way?
So
integrating
all
this
into
the
existing
cluster
API
book
flow
or
the
Kappa
or
the
capsu
book
flow
might
be,
might
make
the
the
overall
documentation
just
sort
of
harder
to
read,
especially
for
a
naive
viewers,
so
yeah
really
glad
you're
working
on
this.
It's
an
unsolved
problem,
I
think.
E
I
think
regarding
integrating
it
into
the
a
normal
book.
I
think
it's.
The
question
is
what
we're
aiming
for.
I
mean
if
there's
like,
that's
a
that's
a
actual
guide
on
how
it
works,
and
someone
will
make
sure
over
time
that
it
keeps
on
working.
That's
another
level,
then
oh,
here's,
a
blog
post
about
how
you
can
do
it
and
it
works
as
of
now
and
linking
blocks
like
that
in
the
book
should
be
an
issue
at
all.
We
just
have
to
find
a
nice
place.
It's.
G
E
A
Yeah
I
think
that's
a
great
Point,
Stefan
I
was
actually
going
to
say
something
similar
I
think
this
looks
more
like
a
blog
post
to
me,
where
it's
like
try
this
at
your
own
risk.
It
worked
for
me,
but
it
might
not
work
for
you
more
than
like
a
supported
feature
of
infra
providers
right
and
in
order
to
make
it
a
support,
feature
and
document
it
in
the
book.
It
would
need
end-to-end
test
to
make
sure
it
doesn't
regress.
A
It
would
need
commitment
from
the
providers
to
support
it
long
term
and
maintain
it,
which
we
don't
have
right
now.
This
is
kind
of
like
a
hack
like
you
figured
out
how
to
get
it
to
work,
and
a
lot
of
people
would
be
interested
in
using
it,
but
we
don't
have
a
guarantee
that
it
will
always
work
right.
A
H
Yeah
I
was
going
to
put
up
a
personal
blog
post
about
it.
I
was
just
because
it's
asked
so
often
I
was
just
trying
to
figure
out
a
way
to
say:
hey,
you
can
try
this
approach,
you
know
just
something
that's
so
we
can
help
people
more
as
it
goes
along
and
I
did
want.
Other
people
to
you
know
hopefully
solicit
some
help
in
testing
the
southern
in
other
environments.
So
it's
not
just
saying
hey.
This
works
for
Adobe
and
good
luck.
You
know
sort
of
thing
someday.
H
You
know,
as
the
working
group
on
managed
cluster
gets,
you
know.
Finalized
yeah
it'd
be
great
to
have
this
be
more
official
but
I'm
not
looking
for
an
official
place
to
put
it
now,
I'm.
Just
looking
for
a
place
to
you
know
it's
like
more
unofficial
stuff
to
help
people
along
as
they
try
to
do
that
stuff.
So
that's
that's
kind
of
what
I
was
hoping
to
to
get.
A
Yeah,
okay
yeah,
that
sounds
great
I,
think
personal
blog
post
probably
was
best.
You
know
in
terms
of
hosting
it
would
probably
make
sense
and
then
either
linking
it
in
the
adoption
of
clusters
like
issue
or
you
know,
if
you
want
it
to
be
more
discoverable,
more
visible,
I
think
I
can't
speak
for
Kappa,
but
in
cabs
e
would
definitely
be
open,
I
think
to
adding
a
page
in
the
book.
For
you
know,
cluster
adoption
and
then
saying
something
like
this
is
not
currently
officially
supported.
A
However,
here
are
some
you
know
blog
posts
from
users
who
have
found.
You
know
attempted
this,
so
we
can
find
out
the
wording
but
yeah
essentially
just
saying
this
isn't
a
feature
but
like
here's
an
example
of
someone
doing
it.
A
Someone's
commenting
in
the
chat
that
sounds
like
Mercedes-Benz
was
planning
a
larger
block
with
on
this
topic
as
well,
so
it
might
be
good
too.
Maybe
we
can
add
their
blog
post
to
the
page
too.
Johannes
go
ahead.
Yeah.
D
Yeah,
you
said
the
right
words:
yeah
I
wanted
to
yeah
just
quickly,
teaser
that
we
have
planned
to
write
this
also
in
yeah,
like
a
blog
post
fashion.
So
we
can
align
this
with
your
post
for
copper
as
well,
so
yeah.
That
would
be
a
good
idea.
A
I
Yeah,
it's
just
something
that
came
up
at
kubecon
a
couple
of
weeks
ago
that
when
we
we
did
the
talk
and
asked
go
out
the
whole
room.
He
was
used,
Kappa,
oh
Cappy,
sorry,
there's
only
a
few
hands.
The
rest
of
them
were
completely
new,
too
happy,
and
so
we
generally
have
to
repeat
at
the
same
things
with
all
of
our
talks.
So
then
it
got
me
into
thinking
and
actually
and
actually
had
Stefan
made
a
comment
after
the
talk
as
well
about
do.
I
A
I
will
say
the
one
thing
that
makes
that
difficult
is
things
tend
to
change
quickly
around
here
and
every
year
we've
had
coupon
talks
of
introduction
to
Cappy
and
every
year
it's
very
different
like
two
years
ago.
Cluster
class
wasn't
a
thing
nowadays,
three
years
ago,
kcp
wasn't
a
thing
you
know.
So
if
we
were
to
create
standard
material
now,
I'm
pretty
sure
it
would
be
outdated
by
next
year,
but
I
don't
know
if
you
have
any
thoughts
on.
You
know
how
to
mitigate
that.
I
Yeah,
that's
best
interesting
point
yeah,
so
I
guess
there
is
that
we
would
have
to
keep
it
updated
so
whether
we
we
produce
it
into
smaller
chunks,
so
that
we
only
had
to
update
a
small
part
of
it
or
add
to
it,
and
it's
essentially
a
playlist
that
someone
could
go
through
in
small
little
parts.
I
So
we
could
build
up
over
time,
maybe
and
just
re-record
small
parts
of
it
just
seems
that
there
is
some
some
desire
there
for
an
intro
and
definitely
there's
a
there
seems
to
be
a
lot
of
interest
in
how
do
I
contribute.
I
You
know,
take
me
through
how
I
set
up
I
don't
know
Stefan
from
British.
Here
did
the
let's
talk
about
so
it's
just
following
on
from
that
I
think
as
well.
A
I
think
at
some
point
we
had
some
stuff
in
the
book
for
like
learning
material
right.
Is
that
still
there
is?
That
being,
do
you
know?
Oh.
G
F
Yeah
I'm
not
supposedly
I'm,
just
a
little
bit
scared
about
the
amount
of
work
in
the
sense
that
if
I
look
at
the
at
the
book,
it
is
already
complex
to
maintain
these
amount
of
information.
F
Definitely
is
not
enough.
Even
they
ask
that
that
we
have
about
a
working
on
this.
This
is
already
a
job,
and
we
are
already
at
the
point
where
the
structure
of
the
book
is
exploding
by
the
size
of
the
content,
and
we
probably
need
a
better
tool
just
to
organize
the
content,
so
I
I
also
suggested
after
the
last
callback
on
the
matter.
The
material
that
that
we
are
producing
is
is
great.
F
Not
sure
how
we
can
combine
this
with
the
current
organization
of
the
book,
but
this
kind
of
problem
are
kind
of
clashing
one
on
on
the
other,
but
yeah.
Let's
start
off
our
first
step.
Let's
make
the
material
available
from
the
kubicon
things
more
visible
and
let's
start
also
pointing
users
to
it,
because
I
when
I
see
folks
in
the
chat
in
the
slack
channel
to
point
the
people
to
the
materials.
F
When,
basically,
we
create
this
wheel
of
people
reading
it
people
improving
it
to
people,
fixing
it
and
then,
as
soon
as
we
can
fix
up
the
booking
copy,
we
have
to
think
how
to
make
probably
this
book
extensible
so,
ideally
also
provider
books
can
be
a
session
of
it.
F
E
Okay,
yeah
I
think
we're.
Definitely
the
point:
I
mean
that
might
be
a
different
format.
I,
don't
know
if
you
want
to
make
like
short
videos
or
something
for
for
part
of
what
which
was
suggesting
but
I,
think
regarding
a
written
documentation.
You
definitely
hit
a
limit
where
it
just
the
structure
is
one
of
our
biggest
problems.
I
think
we
have
a
lot
of
information
there.
A
lot
of
it
is
even
like
on
multiple
pages,
but
it's
just
hard.
E
E
Basically,
rework
the
book
to
something
which
I
think
about
different
structure
and
probably
another
tool
that
she
generated
in
order
to
be
sorted
like
some
experimentation,
like
half
of
your
back
or
something.
G
E
I
think
we
just
never
got
to
it
to
make
more
progress.
A
All
right
that
brings
us
to
the
end
of
the
discussion
topic,
so
we
just
have
a
few
provider
updates
and
feature
group
updates
to
go
through.
So
let's
get
right
into
it.
Matt
borsma!
You
have
the
first
one
with
Captain.
L
A
Good
stuff,
thanks
all
right
cap,
M3,
Adil,
sorry.
K
Yes,
yes,
so
we
had
a
minor
release
last
Thursday
and
also
batch
release.
So
far
everything
is
working.
Fine
CI
is
going
good,
so
no
surprises.
Yet
so
it's
going
good.
Just
an
update.
A
Great,
we
like
no
surprises
release,
awesome
all
right,
any
other
badoc
Stefan
go
ahead.
E
Maybe
just
a
general
question
towards
providers,
any
kind
of
updates
on
adopting
the
latest
task
replay
104
release
any
feedback
that
I
don't
know.
We
broke
things
badly,
any
upcoming
plant
releases
which
are
tested
against
it.
I,
don't
know
if
just
spontaneously,
someone
has
something.
I
was
away
for
a
few
weeks,
so
I'm
not
even
sure
the
current
status.
L
We
had
a,
we
had
a
somewhat
difficult
time
releasing
one
four,
because
we
were
stuck
on
that
well
I'm,
going
to
forget
the
context
now
serbicide
apply
bug
that
eventually
led
to
141.
other
than
that
it's
been
relatively
painless
and
and
upgrading
from
141
to
142
was
kind
of
a
no-op
so
as
it
should
be.
G
Yeah,
sorry,
it's
pretty
much
what
what
Matt
said,
but
also
we
also
took
the
effort
to
remove
the
older
API
versions
because
they
were
deprecated
and
that
went
smoothly
for
1.9
release
so
yeah.
Besides
the
SSA
error,
which
is
fixed
now,
everything
was
good
for
us
on
Camp
Z.
M
Hey
yeah,
like
everybody
else,
we
didn't
run
into
any
issues,
but
we
did
end
up
going
from
140
to
141
again
without
any
issue
so
good
on
our
end
as
well.
A
Great
Adil.
K
A
Awesome
right,
yeah,
there
are
two
SSA
issues:
I
think
the
first
one
was
attaching
the
node
status
dot
addresses
when
there
are
multiple
IP
addresses
and
the
second
one
was
the
machine
pool
patch
or
no
or
for
annotations,
wasn't
working
properly.
F
Yeah
I'm
not
sure
that
this
will
help
to
prevent
the
regression
around
this,
but
Killian
is
working
to
dual
stack
and
to
end
test,
and
maybe
this
can
help
and
prevent
this.
So,
first
of
all
kudos
to
Killian
for
doing
this
work.
And
second,
maybe
this
can
help
also
around
the
the
specific
case.
A
Yep
I
think
this
will
definitely
help
for
the
specific
case
scenario
and
then
in
terms
of
more
General
like
prevention.
We
have
also
an
issue
to
discuss,
like
maybe
integrating
continuous
testing
between
some
providers
and
Cappy
and
then
testing
the
better
earlier.
C
So
yeah
I
wanted
to
kind
of
pose
the
question
about
retirement.
Do
we
feel
like
a
feature
group?
That's
scoped
to
a
particular
proposal
is
an
appropriate
thing
and
when
that
proposal
emerges,
is
it
an
appropriate
time
to
retire
that
feature
group
I
do
think
that
just
for
you
know
resources,
we
probably
can't
accommodate
feature
groups
forever.
C
A
From
the
discussions
earlier,
it
sounds
like
we
want
to
retire.
This
managed
kubernetes,
API,
feature
group
and
start
an
adoption
feature
group.
A
Any
I
guess
any
thoughts
or
feedback
on
how
you
know.
Having
that
feature
group
helped,
since
it
was
the
first
time
we
were
doing
this
in
Cappy,
do
you
have
any
advice
for
future
feature
groups,
or
was
this
a
smooth
process
overall.
C
C
Right
yeah
I
I
mean
it
was
hugely
helpful.
It
produced
a
proposal
which
is
the
kind
of
best
case
scenario,
and
it
was
a
nice
mix
of
very
focused
discussion
aiding
the
velocity
of
that
proposal,
with
sort
of
like
more
traditional
office
hours,
where
folks
hopped
in
and
we
just
had
a
ad
hoc
chat.
C
I
think
you
know
we've
we
talked
about
the
intro
to
Cluster
apis
thing,
and
maybe
one
thing
that
I
would
take
away
from
the
feature.
Group
experience
is
that
for
folks
who
are
kind
of
kicking
the
tires
on
something
just
having
time
where
maintainers
are
in
a
zoom
room
and
folks
can
hop
on
and
ask
any
question
and
there's
not
necessarily
a
set
agenda.
That
can
be
helpful,
but
it
obviously
it
takes
a
lot
of
time.
So.
F
So,
first
of
all,
retirement
I
think
that
it
is
really
up
to
the
team
if
they
think
that
they
achieved
the
goal
is,
is
fine
to
retire
it
with
regard
to
the
overall
experience,
I
I
think
that
it
was
positive,
let
me
say,
adding
a
little
bit
of
formality,
which
is
for
defining
it
as
a
feature
group
eating
crazy
visibility
and
also
increased
in
in
encourage
the
team
in
the
future
group
to
report
back
to
the
brother
nonsense.
So
I
think
that
this
is
good.
F
The
thing
is
visible
to
the
community,
and
the
people
in
the
team
are
somehow
committed
to
airport
back.
This
makes
the
work
of
the
feature
group,
not
something
happening
in
the
dark,
and
it
is
a,
let
me
say
the
main
risk
when
you
have
a
specialized
group
working
on
something,
and
so
in
my
opinion,
it
is
a
good
experience.
I
think
that
we
should
keep
going
at
the
end.
Opening
a
feature
group
is
a
is
getting.
C
Yeah
I'll
say
one
more
thing
that
I
didn't
include
in
my
assessment.
One
thing
that
definitely
I
think
achieved
is
getting
a
lot
of
those
conversations
out
of
this
meeting,
which
I
think
helps
this
meeting,
and
that
was
one
of
the
original
motivations
for
me
is
to
not
have
to
take
up
10
minutes
week
after
week
after
week
in
the
general
meeting
with
these
very
detailed
topics,.
A
Great
all
right
that
brings
us
to
the
end
of
the
agenda.
Thanks
all
for
coming
and
see
you
all
in
flag
and
online
have
a
good
Wednesday
all
right.