►
From YouTube: Kubernetes WG LTS 20190108
A
B
B
B
Wow
to
past
the
hour,
I
want
to
go
ahead
and
get
started
to
be
mindful
of
time.
This
is
the
working
group
LTS
meeting
the
first
of
2019
welcome
everybody
happy
new
year.
The
meeting
is
being
recorded,
asked
of
course,
that
everybody
be
mindful
of
our
community
kind
of
conduct
and
the
video
will
be
going
up
on
to
YouTube.
Afterwards
it
is
being
recorded.
So
if
you
could
drop
your
names
in
the
the
doc
just
for
the
record
of
who
all
was
attending,
that
would
be
appreciated.
B
Ok,
so
then,
the
the
main
things
that
I
wanted
to
talk
about
were
what
happened
at
cube
con,
but
does
anybody
else
have
anything
that
they
would
like
to
discuss
today.
D
You
go
Quinton
now
I
was
just
gonna
I.
Think
I
mentioned
it
to
some
of
you.
I
had
an
interesting
chat
with
JPMorgan
Chase
Seattle
about
this
and
found
out
that,
in
fact,
they
as
representative
a
large
international
bank,
had
no
interest
whatsoever
in
long
term.
Support
which
came
as
a
surprise
to
me.
So
I
was
just
wondering
whether
you're
worth
having
that
discussion
about
broadly
and
sorry.
I
missed
the
opportunity
to
have
the
discussion
in
face
in
Seattle.
E
B
So
I
tried
to
I
was
I
was
trying
to
be
engaged
in
the
conversation,
but
also
scribbling
down,
notes,
frantically
and
trying
to
capture
what
was
happening,
and
then
I
tried
to
kind
of
collate
those
into
some
some
topical
areas.
So
it
wasn't
just
cuz
the.
If
you
watch
the
video
back
it's
we
did
ramble
a
bit,
but
there
were
some
kind
of
themes
across
that
I
just
want
to
make
sure
that
the
notes
accurately
reflect
what
others
heard
there
as
well.
So
it's
a
Google
Doc.
Please
make
suggestions
on
changes.
B
B
I
think
the
the
general
couple
things
that
were
were
talked
about
are
things
that
we've
kind
of
already
been
talking
about
earlier
in
December,
or
maybe
even
slightly
in
in
November,
as
well,
around
better
defining
API
boundaries
and
then
doing
more
survey
to
collect
data,
because
the
the
voices
in
the
room
are
just
that.
It's
a
limited
subset
of
our
broader
community.
B
C
I
feel
like
the
I
feel
like
the
the
surveys
is
the
obvious
sort
of
first
thing,
lifting
it
up.
We've
got
a
lot
of
people
commented
on
the
survey
already
on
a
survey
draft
off.
So
if
we
can
collect
those
comments,
get
them
turned
into
an
actual
survey
and
get
the
survey
out
then
the
longer
we
can
run
that
the
better
you
know
and
given
that
we
want
to
keep
this
relatively
time
limited.
You
know
it
feels
like.
There
is
a
bit
of
time
pressure
to
get
a
rolling.
D
Yeah
I
agree
with
that.
The
other
offer
I
have
so
Michael
Payne,
who
believes
his
team
runs.
Most
of
the
kubernetes
clusters
for
JPMorgan
Chase
has
actually
offered
to
present
to
this
group
and
explain
why
they
would
love
ults
and
how
there's
a
bank
see
this
stuff,
which
I
thought
was
one.
It
was
inconsistent
with
my
understanding
of
that
industry
and
and
what
they
would
want.
D
C
E
D
I
absolutely
agree:
yeah
yeah.
The
main
reason
why
I
found
it
so
interesting
was
because
it
was
so
contrary
to
what
I
understood
the
requirements
of
that
industry
to
be.
This
is
a
real
person
running
real
stuff
for
real
banks,
so
presumably
their
their
opinion
accounts
way
more
than
mine
does,
at
least
for
a
subset
of
that
industry
and
and
I.
Think
Michael
was
pretty
candid
about
the
fact
that
they
probably
are
not
the
mainstream
banking
industry.
D
They
kind
of
consider
themselves
to
be
sort
of
Gorillaz
of
me
bleeding
in
edge
in
that
industry,
but
at
least
you
can
explain
why
they
do
things
the
way
they
do
and
whether
you
know
we
can
kind
of
speculate
about
whether
that
represents
the
trajectory
of
those
kind
of
industries
or
whether
it
is
an
anomaly,
either
of
which
will
be
useful.
I
mean
point
of
view:
yeah,
yeah,
III,.
E
E
B
One
of
the
things
that
I
think
I've
heard
some
folks
arguing
is
that
by
there
there's,
even
if
there's
multiple
modes
in
the
distribution
of
how
people
are
deploying
things
there
may
be
within
the
group
or
within
the
the
base
of
developers
and
communities,
a
preference
for
one
of
those
moving
forward.
So
so
cataloging
some
of
these.
Even
if
there
are
outliers
they
may
be
early
leaders
or
something
that,
if
well
documented,
becomes
a
guidepost
for
others.
That
that
makes
the
situation
better
for
all.
A
A
To
upgrade
on
a
two-week
basis
or
a
four
week
basis
or
a
three
month
basis
or
a
six
month
basis
kind
of
what
you
had
to
go
through
in
order
to
upgrade
to
a
supported
version
at
that
cadence
and
also
you
know
things
that
the
community
claims
are
supported
but
are
not
tested
by
anyone.
You
know
and
kind
of
trying
to
capture
the
current
state
of
the
world
and
you
know,
get
a
get
a
at
least
a
common
understanding,
so
that
you
know
anyone
can
go
to
this
thing
and
understand
it
at
a
glance.
A
A
Yeah,
just
something
to
kind
of
organize
our
thoughts
around.
You
know
what
what
the
community
claims
is.
You
know
supported,
for
example,
like
supporting
older
Kubla
versions
with
a
current
master.
You
know,
I,
think
that
is,
you
know
supposed
to
work,
but
you
know
I,
don't
think
it's
actually
tested
very
far
back
and
so
kind
of
getting
a
catalog
of
all
those
things
that
are
claimed
to
be
supported,
but
not
tested
or
things
that
are.
You
know,
differences
between
multi
master
cluster
versus
a
single
master
cluster.
D
Yeah
I,
agree
and
I
think
more
generally,
something
along
the
lines
of
the
white
paper
and
I
know
those
are
somewhat
maligned
in
some
industry
or
some
communities,
but
but
I
think
actually
just
making
distinction
between
you
know
reasonable
aspirational
goals
of
what
you
know
should
work
or
could
work
if
you
know
fixed
up
properly
versus
the
reality
today
and
also
versus
what
we
might
consider
to
be
unreasonable,
aspirational
goals.
You
know
when
one
could
make
an
argument
that
you
know
being
able
to
release
a
stable
kubernetes.
Every
two
weeks
is
an
unreasonable
goal.
D
I'm
just
you
know
using
it
as
a
totally
made-up
example,
but
I
think
breaking
all
of
this
duffel.
Those
sort
of
broadly
three
categories
might
be
useful
and
putting
down
the
facts.
These.
These
are
actually
the
things
that
do
work
today.
These
are
the
things
that
we
think
should
work
and
could
work
and
everybody
agrees
are
feasible.
B
Zoom
Chad
Jordans
mentioned
that
a
lot
of
these
topics
are
planned
follow-ups.
If
you
look
at
communities
website,
pull
1
106
0,
so
Jordan
when
you
say
planned,
is
that
a
documented
plan
or
liquid
I
know
we,
like
this
kind
of
been
talked
about
and
I,
don't
know
it's
definite,
I,
don't
know
it's
definitely
in
your
mind,
yeah.
F
F
F
F
A
Matt,
my
hope
is
that
if
we
add
some
sort
of
chart
like
this
than
any
sort
of
ideas
that
this
group
came
up
with
regarding
LTS
or
any
changes,
you
could
evaluate
against
that
kind
of
matrix
and
say
this
idea
is
going
to
improve
these.
You
know
areas
of
the
matrix
and
you
know-
or
it's
not
going
to
improve
anything
or
there
are
other
non
LTS
solutions
that
you
know
improve
the
matrix
that
you
know
accomplish
the
same
goals
without
you
know
introducing
another
release
type.
A
B
B
B
B
One
of
the
questions
that's
been
in
my
mind
is
what
what
is
sufficient
here,
what
is
sufficiency
and
how?
How
far
we
go
towards
the
perfect
survey
that
captures
everything,
because
at
some
point
we
need.
We
want
to
actually
put
a
survey
out
there
and
get
some
feedback.
So
what
do
people
think
about
the
the
state
of?
What's
there
currently
and
the
commentary
around
it
like?
Are
we
starting
to
bike
shed?
Are
there
major
refinements
needed
still
I.
C
Feel,
like
the
the
main
thing
that
it
needs
is
just
probably
a
couple
of
people
to
sit
down
and
go
through
all
to
sit
down
together
or
in
person
or
on
a
call
coming
through.
All
the
comments
and
sort
of
you
know
make
a
call
about
them,
like
I
feel,
like
it's
kind
of
stuck
at
a
moment
in
the
you
know,
hey
we're
seeking
feedback
mode
and
not
know
one
sort
of
gone
in
like
okay.
These
feedback
is
good.
This
feedback
is
incentive.
This
feedback
is,
you
know
now
we're
not
gonna.
Do
that.
C
You
know,
like
I
mean
that
that
needs
a
relatively
small
team
of
people
to
just
make
some
calls
and
be
like
okay,
you
need
to
getting
something
out
we're
doing
this
this.
This
isn't
this
do
resolve
those
comments,
fold
them
into
the
dock
and
then
actually
go
and
do
that
and
then
get
and
then
so.
These
are
the
final
one.
You
know
speak
now
or
forever
hold
your
papers.
D
C
B
We
should
we
go
ahead
and
do
that
kind
of
editorial
voice
scrub
through
because
for
the
the
simpler
comments
that
went
by
I
think
we
were
sort
of
accepting
and
nudging
things
around,
but
some
of
the
ones
that
remainder
ones
we're
a
little
bit
bigger
choice
needs
made.
If
we
go
ahead
and
do
that,
then
we
could
come
back
at
the
meeting
two
weeks
from
now.
B
That
would
be
that,
like
the
20
22nd
23rd,
something
like
that
and
and
share
what
we
believe
to
be
a
final
draft
with
folks
and
its
needs,
a
thumbs-up
and
and
then
as
Quentin
says,
be
able
to
put
it
out
there
by
the
end
of
the
month.
You.
D
No
I
was
just
gonna,
make
another
statement,
which
is
that
this
doesn't
have
to
be
the
last
survey.
Obviously,
so
we're
not
sending
out
a
new
survey
every
month,
you
know
once
or
twice
a
year
is
fine.
So
if
this
thing
keeps
us,
it
gives
us
useful
information
for
the
next
six
months.
I
think
it's
good
enough
agreed.
B
B
So
the
next
thing,
the
the
charter-
PR
I,
don't
believe
so.
Similarly,
a
cube
con
threw
that
out
there
just
is
a
link
for
for
broader
request
for
review
and
I
know.
I
saw
people
pulling
it
up
at
the
time.
I
haven't
seen
any
more
comments
on
it.
Anybody
here
have
anything
that
they
think
is
particularly
blocking.
There
should
be
revised
and
proved
as
a
sufficient
should
we
actually
ping
staring
and
say.
Yes,
please
move
forward.
B
The
idea
of
talking
about
some
experience
reports,
folks
that
people
talk
to
a
cube
consequent
you've
kind
of
talked
a
bit
about
Michael
pain
and
in
the
notes
there
did
I
spell
his
name
correctly:
I'll
check
its
BA
y
and
E,
the
okay
cool,
so
I
also
I
briefly
talked
with
a
couple
people
there,
but
basically
just
enough
to
say
like
hey,
this
working
group
exists
and
they
were
intrigued
and
I
told
them
in
the
new
year.
I'd
kind
of
cycle
back
and
engage
with
them,
so
I
owe
them
a
reaching
out.
C
Yeah
I
didn't
have
really
very
very
many
conversations,
but
I
was
kind
of
like
the
co-chair
who's.
Also,
a
user
I,
probably
shouldn't
put
my
name
like
my
sort
of
he's,
what
we
do
for
upgrades
and
you
know
why
we
care
and
what
we
fund
all
that
stuff
at
some
point,
I,
don't
know
if
that's
the
right
place,
how.
B
C
I
think
something
with
some
sort
of
fields
for
essentially
metadata
about
the
use
case
to
enable
us
to
classify
them.
A
bit
would
be
useful,
but,
aside
from
that,
it
will
be
mainly
pretext
but
you're.
Having
like
a
you
know.
Who
are
you
where
you
come
from
had
how
many
clusters
do
you
have?
How
often
like
that
somewhat
some
a
lot
of
the
stuff
that
we
have
into
survey,
as
you
say,
would
be
useful
to
sort
of
bucket
the
bucket
the
bucket
the
use
cases
at
some
point.
B
D
Problem
is
to
publish
white
paper,
you
know,
do
you
do
all
of
this
background
work
to
get
all
the
information
have
meetings,
whatever
the
end
of
the
day
entered
into
a
single
thing
that
people
can
go
and
consume
and
say
this
is
everything
that
was
found
out.
You
know
structured
properly,
I,
don't
know,
I
assume
that's
one
of
our
planned
out
words.
B
So
and
that's
where
I
say,
if
sort
of
like
a
pre
Kepesh,
the
the
informing
background
information
in
a
way,
that's
feels
kept
ish
without
the
proposal.
Part
person,
perhaps
or
the
I
feel
like
the
the
other
working
groups,
white
paper
type
documents
that
I've
seen
have
kind
of
laid
out
a
lot
of
background
and
then
a
set
of
possible
proposals.
So
that
being
ahead
of
the
kept
where
a
proposal
is
made,
it
kind
of
lays
the
the
groundwork
in
potential
directions,
but
still
has
that
plurality
to
it.
C
C
But
yet
I
think
in
the
I
think
practically
gathering
the
data
for
for
a
white
paper
will
involve
a
you
know,
a
form
of
some
sort
that
will
be
that
will
end
up
in
a
spreadsheet.
So
you
know
it
might
be
easier
to
just
cut
out
the
little
creases
and
have
people
put
it
straight
into
a
spreadsheet.
But
a
million
tation
detail
not
important
yeah.
D
C
C
Mr.
I
I
agree
the
that,
having
a
you
know
effectively
a
point
in
time
document
of
his
he's
where
we
are
at
you
know
to
to
enable
people
to
start
from
that
point
rather
than
having
to
roll
back
through.
All
the
meeting
minutes
is
useful.
Speaking
of
someone
who
often
has
to
pick
up
where
I
see
girl
working
group
is
up
to
by
reading
them
many
minutes,
because
I
can't
attend
the
meetings.
It
is
very
useful,
I
think.
B
One
of
the
things
that
the
weather
spreadsheet
specifically
or
something
else
that
the
notion
of
some
basic
schema
so
that,
when
we're
talking
to
somebody,
we
say
like.
Oh,
we
haven't
asked
about
their
upgrade
path,
or
we
haven't
asked
about
that
that
there
are
a
certain
key
set
of
buckets
that
we've
identified
as
the
ones
where
we
want
to
pull
data
so
that
it's
not
that
things
are
comprehensive
enough.
That
we
don't
then
retroactively
Li
have
to
discount
somebody's
information
because
we
well
so
it
seems
irrelevant,
because
we
don't
even
know
how
they
do
X.
D
B
Yeah
that
makes
sense
and
I'm
thinking.
That
also
is,
is
Nick
and
I
scrub
through
the
survey,
and
maybe,
if
we
think
about
some
of
the
the
buckets
of
discussion
that
I
kind
of
collated,
the
notes
from
the
cube
Cardon
off
into
that
that
maybe
those
inform
are
our
key
areas.
But
we
could
also
kind
of
try
to
transition
that
into
something
like
a
an
initial
table
of
contents
and
see
if
that
makes
sense
to
be.
C
B
B
All
right,
I'll
move,
Nick
I
will
give
you
and
perhaps
Honda's
also
the
floor
to
discuss
having
a
additional
alternate
time
meeting.
C
So
I
think
we
discussed
this
briefly
in
a
couple
of
the
previous
meetings.
I
have
run
a
yeah.
The
usual
seems
like
the
only
way
to
try
and
do
this
across
time
zones
is.
I
have
run
a
doodle
with
some
alternate
times,
notably
I
did
get
one
I
have
I
have
had
one
actual
other
person
who
lives
in
Australia,
who
also
is
interested
in
coming
tick.
C
Some
boxes,
that's
John,
Slee,
and
so
you
know
it
looks
likely
that
what
we're
talking
about
is
middle
of
the
day,
Thursday
Australia
time,
with
the
intent
being
that
for
now
for
now,
I
was
kind
of
suggesting
doing
this.
In
a
way,
that's
like
you
know,
I
will
go,
I
will
come
to
this
meeting
and
then
I
will
go
to
that
meeting
in
anyone
who
wants
to
dial
in
you
know:
hey
you
that
way.
I
can
I
can
be
the
information
conduit,
Hannes
I,
think
it
doesn't,
it
would
be
cool
to
be
able
to
swap.
C
You
know
to
be
able
to
alternate
times
and
stuff
when
I'm
sort
of
talked
to
that
about
poking
people
about
that
in
the
past.
A
lot
of
the
time
it
seems
to
be
hard
to
put
yeah
because
there's
more
people
in
u.s.
timezone,
it's
much
harder
to
also
know
times
and
stuff.
Then
then
to
just
have
another
meeting.
C
So
I
was
having
suggesting
having
in
on
the
meeting
just
to
get
something
happening
quicker
because
there
are
I
know
it's
like
I'm
I've,
been
to
a
few
like
Enterprise,
kubernetes
meetups,
and
things
like
that
in
Australia,
where
you
know
I
can
sort
of
read.
I
know
know
a
few
people
who
work
in
which
and
say
hey
is
call
to
come
and
talk
about
long-term
support
and
I
can
do
all
do
the
same
sort
of
experience
gathering
stuff
once
we
have
a
framework
for
it.
C
C
C
Given
that
I've
got
one
other
person
so
far,
I'll
try
I'll
be
helping
a
couple
of
people
that
I
know
you
are
in
Australia
who
use
cube
and
see
if
I
can
get
them
to
commit
to
a
time
as
well.
I,
don't
think
I'm,
possibly
so.
I
can
try
and
get
a
couple
of
other
Australia
opinions,
and
then
I
will
just
pick
one
given
that
tomorrow,
I
will
yeah
the
best
sort
of
the
proposed
first
date.
C
C
C
B
And
for
that,
for
the
proposed
time,
I
could
probably
also
potentially
join
and
and
cover
and
make
sure
that
we're
bouncing
stuff
between
meetings
as
well
I
think
it's
really
important
when
we
have
when
we
start
splitting
meetings
that
we
have
good
conduits,
that
it's
it's
not
a
primary
and
alternate
meeting,
but
both
become
solid
forums.
So
yeah.
B
C
Think
you
that
I
I'm,
a
strong
supporter
of
that
I,
probably
won't
be
able
to
go,
but
I
used
up
all
my
travel
credits,
I
suspect,
but
yeah.
Absolutely
my
experience
with
you
big
thing.
Has
it's
much
easy?
It's
really
good
to
get
the
you,
people
like
people
who
live
in
the
EU
and
find
it
hard,
because
timezones
I
also
really
appreciate
you
bringing
this
sort
of
discussion
to
them
and.
B
Presumably,
at
that
point,
we'll
have
formulated
a
lot
of
the
the
initial
documentation
and
information
and
be
also
perhaps
starting
to
to
drift
and
much
more
in
the
direction
of
proposals
as
well
and
be
able
to
kind
of
vet
those
against
people
to
see
like
no
we'd,
never
use
that
or
yeah.
We
could.
We
could
like
that
or
kind
of
start
getting
a
pulse
of
that
as
well.
B
D
D
So
I
mean
there's
a
good
kind
of
forcing
function
to
say
we
publish
the
white
paper,
so
we
can
present
it
both
in
EU
and
Shanghai,
and
then
consider
that
sufficient
distribution.
You
know,
in
addition
to
whatever
other
email
mechanisms
and
presenting
here
and
and
we
consider
that
draft
now
final
and
then
the
proposal
is
based
on
the
finalized
thing
that
has
been
you
know,
presented.
A2Q
comes.