►
From YouTube: Kubernetes 1.12 Release Team Meeting 20180905
A
C
A
So
if
you've
got
the
meeting,
minutes
open
you'll
see
that
today's
agenda
is
a
little
bit
different
format
and
it'll.
Look
familiar
for
folks,
who'd
been
around
on
the
prior
cycle,
but
probably
new
or
a
prior
cycle,
probably
new
for
folks
who
hadn't.
So
it
deals
like
when
we
go
into
a
little
bit
more
rigorous
burndown
format
and
the
notes
where
we're
tracking
a
set
of
specific
things.
Much
more
closely.
A
A
And
so
yesterday,
code
freeze
started,
and
hopefully
we
see
things
dialed
down,
I've
already
been
getting.
I've
already
been
getting
pings
on
slack
like
hey.
Could
I
get
this
in?
Is
it
okay?
So
people
are
people
are
at
least
acknowledging.
It
was
interesting
to
see
on
Monday
that
we
had
somebody
say.
Oh
hey,
I
have
a
feature
that
I
need
to
add.
What's
the
exception
process,
it's
so
it's
sort
of
a
mixed
bag.
A
A
Actually,
there's
discussion
going
on
in
cig
release
arounds
what
we
might
do
to
move
forward,
but
this
is
this
is
part
of
breaking
the
build
and
release
cycle
out
from
underneath
Google
and
and
figuring
out
how
we,
as
a
community,
run
it
so
I
would
expect
that
there
will
be
some
hiccups
there
over
time
and
what
we
will
work
through
them
and
hopefully
do
that
in
a
way
that's
is
minimally
destructive
or
as
possible.
Do
some
fail
fast
fix
back
fix
fast.
A
A
Okay,
so
see
I
signal
report
we
again
do
not
have
Mohamed
here
and
I
want
to
put
a
call
out
so
we're
struggling
here.
We
need
help
on
the
CI
signal
side,
so
I'll
be
looking
to
see
if
I
can
get
a
volunteer
and
be
peeing
some
people
individually
to
pick
up
the
slack
here,
we're
not
getting
issues
open
promptly,
we're
not
getting
them
assigned
to
people,
we're
not
getting
follow-up.
So,
basically,
the
the
fundamentals
of
that
role
in
the
risk
management
that
happens
there
have
not
been
happening.
A
A
We
we
have
a
bunch
of
issues,
we
have
a
couple,
people
looking
and
are
waiting
for
more
status,
and
then
we
have
other
issues
that
need
open.
So,
if
any
of
you
happen
to
know
somebody
who
might
be
suited
to
the
CI
signal
triage
role,
let
me
know
otherwise,
like
I
said
I'm
going
to
be
reaching
out
to
people,
so
then
I
guess
we'll
shift
over
to
issues
and
Gwen.
So
non
CI
signal.
A
D
Notes
here
we
are
okay,
so
we
still
don't
see
a
whole
lot
of
issues.
There
were
a
couple
ones
that
were
open
and
closed
pretty
furiously
again,
and
everyone
was
really
responsive
about
me.
Pinging
them
for
making
issues
critical,
urgent
and
most
of
them
haven't
looked
over.
There
is
one
issue
that
hasn't
been
removed
from
the
milestone:
I
talked
to
them
and
they
said.
Actually
we
are
not
getting
us
in
for
112
and
it's
fine.
D
Otherwise
it's
going
to
get
stuck
in
CI,
but
you
know
that's
kind
of
a
self-correcting
problem.
Eventually,
they'll
see
that
but
yeah
so
I
think
I'm
relatively
on
top
of
that
and
if
you
need
help
with
creating
issues
or
I'm
getting
really
good
at
bothering
people
and
I
think
people
are
getting
to
see
my
face
a
lot
so
so
I
can
probably
bother
people
about
CI
stuff.
Maybe
I
should
connect
with
eyes
on
that
yeah.
E
I
get
a
meeting
out
for
both
of
us
and
then
Tim.
If
you
could
give
us
a
list
of
issues
that
you
want
us
to
look
at,
we
can
split
between
us
and
take
some
action
items.
I
anyway
wanted
to
meet
with
gwen
to
understand
a
little
bit
more
about,
but
triage
now
that
113
will
soon
be
there.
So
I
put
something
on
a
calendar
square
awesome.
D
F
Do
you
mind
if
we
go
back
to
see
I
signal
so
because
I
wanted
to
maybe
call
out
that
perhaps
instead
of
the
status
being
yellow,
it
should
be
red
if
you're,
at
the
point
where
you're
asking
for
volunteers
I
also
wanted
to
point
out,
at
least
from,
like
my
perspective,
is
an
outside
observer.
It
looks
like
every
single
one
of
your
dashboards
are
red
or
whether
or
not
a
release
should
go
out
so
release
master,
blocking,
release,
master
upgrade
and
release.
One
is
called
blocking
all
red
in
some
way,
shape
or
form.
F
F
A
There
is
a
gke
angle
there
over
the
last
couple
days,
I've
seen
better
response,
so
I
want
to
give
it
just
a
little
while
yet
but
yeah
I
think
exactly
what
you're
saying
that
these
either
are
gonna
fall
out
or
will
get
the
attention,
but
I
think
that
that
is
the
stick
or
the
carrot.
Whichever
way
to
look
at
it
that
these
are
these
are
going
to
be
taken
out.
I
also
I
I
was
roughly
there
last
week
and
they
started
to
look
better
for
a
moment
and
then
they
got
really
worse
over
the
weekend.
A
F
Right
I
will
follow
up
you've
offline
about
documenting
criteria
in
some
way
shape
or
form,
because
you're
not
really
gonna
have
a
document
to
be
able
to
point
to.
But
we
can
talk
about
tribally.
What
we've
done
in
the
past
and
I
think
you
should
shout
this
loudly
during
the
community
meeting
and
you
should
give
a
heads
up
within
Kade
of,
and
you
should
give
folks
an
appropriate
amount
of
time
to
react
to
see
if
that
causes
husband.
G
E
Yeah
looking
at
that
looks
like
there
are
some
mostly
there.
Three
two
three
two
two
three
buckets
one
is
like
all
the
costs,
jobs
that
are
mostly
Jiki,
but
then
there
is
some
one,
GC
cost.
That's
also
like
failing
in
the
112
blocking
and
then
there's
a
cube
admin
stuff
which
I
don't
know.
If
there's
already
some
active
work
going
on
there
or
we
should
say
little.
G
A
F
Go
ahead,
I
know
just
to
bring
my
historical
context
for
some
perspective,
a
job
failing
on
release
master
blocking
that
concerns
me
is
the
GC
igce
serial
job.
If,
if
I
put
on
my
gke
blinders
and
assumed,
why
are
we
blocking
on
a
hosted
solution
and
then
I
assume
well
cube,
ATM
and
scale
have
always
historically
been
failing.
F
E
F
E
A
A
E
D
In
issues
in
general,
this
is
the
time
we
should
be
seeing
like
daily
work
on
it
right.
Yes,.
A
D
A
So
I
think
so,
there's
first,
if
you've
got
an
individual
penguin
on
slack,
try
to
get
a
response
raise
them.
Their
github
usually
is
more
lossy,
but
you
can
also
put
out
a
little
bit
broader
net,
both
in
the
slack,
the
appropriate
sig
slack
channel,
say:
hey
sig
leads
XY
and
Z
by
name
I've,
not
gotten
progress
on
issue
X
with
person
Y
like
a
lot
of
at
direct
app
mentions.
A
D
A
But
we
are
into
that
position
and
yeah
I
feel
like
whether
it
was
summer
or
Burning,
Man
or
conferences
or
whatever
the
heck
it
was
like
this
month.
Things
have
been
very
high,
latency
and
people
that
people
said
yes
on
it
and
then
silence
for
too
long
yeah.
A
And
that's
where
I
mean
as
a
team?
If
we
have
a
shortlist-
and
we
all
know-
and
that's
part
of
what
this
meetings
about
like
how
Quinn
is
listed,
specific
issues
we
can
notate
next
to
them
like
I
need
this
person's
attention
in
each
of
us,
then
any
of
us
could
say:
okay,
I
will
take
that
one
for
the
day
and
I
will
work
on
hunting
down
this
particular
person.
It
really
takes
that
one-on-one
connection
on
a
specific
issue.
To
do
it.
B
So
we
are
currently
at
63
features.
There
are
60
features
that
seems
to
be
on
track.
That
seems
wild
to
me.
It
seems
crazy.
I
need
to
mark
it
and
have
it
recorded
that
this
seems
crazy
at
this
point
in
the
game,
there
are
three
that
require
exceptions.
One
is
around
CRI
I'll,
put
links
in
the
chat
as
well.
What
is
around
CRI
that
was
close.
That
I,
don't
think,
should
have
been
closed,
but
it
looks
like
it
was
around
logging
in
set,
so
maybe
incremental
improvements.
B
The
second
one
was
a
round
of
release
architecture,
so
this
is
something
that
we
need
to
bring
to
SiC
release
and
we
haven't
had
a
sick
release
meeting
in
maybe
three
weeks
now,
so
the
one
next
week
is
the
one
that
we
need
to
bring
what's
dearest.
So
essentially,
this
is
a
task
that
was
owned
by
cluster
lifecycle,
but
it
shouldnt
be
owned
by
cluster
lifecycle.
They
believe
secretly
should
hold
it
and
I
I.
B
B
B
So
it's
been
filed
and
there's
activity
happening
on
the
other
feature,
but
I'm
trying
to
wrangle
people
to
actually
put
it
where
it
needs
to
be
outside
of
that
they
don't
actually
seem
at
risk
at
risk,
just
that
exception
should
be
filed
for,
for
these
special
cases,
cases
I'm
doing
a
sweep
of
the
features
right
now
just
to
see
if
anything
has
kind
of
bubbled
up
to
the
surface,
I
have
a
view
Zack
heavy
I,
don't
know
if
Zack
is
on
this
call.
Actually,
okay.
A
B
H
A
H
I'm
happy
to
say
mean
stuff
to
people
in
a
loving
way.
Just
I
guess,
I
didn't
really
know.
I
didn't
really
know
when
I
said,
like
officially
cuz
yesterday
was
Doc's
deadline.
I,
don't
you
know
I,
don't
want
to
be
like
immediately.
Everything
that's
not
in
is
gone
with.
Just
like
an
iron
fist.
I
just
I
was
gonna.
Ask
like
generally.
When
should
we
start
saying
no.
A
H
A
Release
really
I
mean
we,
we
have.
We
have
three
weeks
until
the
release
date,
but
two
weeks
until
master
reopens
I
mean
if
they're,
if
their
stuff
is
still
questionable
or
under
way,
it's
gonna
be
landing
in
the
1/13
branch
and,
if
they're
seriously,
proposing
at
this
point
that
it
should
be
cherry
picked
back
to
the
release
branch.
A
No
you
you've,
you've
missed
the
strain
and
especially
like
I
would
be
a
little
more
willing
to
consider
things
if
we
had
a
clean
CI
signal
and
they
had
their
code
on
a
branch
and
had
testing
happening
on
it
and
could
say
look.
This
is
not
going
to
destabilize
things,
but
we
already
have
enough
instability
that
we
do
not
have
the
luxury
of
taking
on
extra
risk.
At
this
point,
if
you
don't
have
your
code
and
you
haven't
filed
your
exception,
describe
it,
you
haven't
filed
your
Doc's,
PR,
you've
missed
and
you
are
my
hero.
A
This
is
just
I
mean
that
this
isn't
like
being
mean
or
anything.
This
is
just
pragmatically.
We
are
out
of
time
and
we
have
a
scary
amount
to
do
to
figure
out
if
and
how
we
can
stabilize
CI
signals
so
that
because
because
really,
what
we're
realistically
looking
at
this
point
increasingly
I
would
say,
as
of
this
weekend,
that
the
increase
in
CI
failures
this
weekend
that
this
releases
late
and
if
this
is
releases
late
1:13
has
issues
because
it's
already
a
short
cycle
with
all
of
the
holidays
and
things.
A
C
A
Right
now
so
I
is
when,
when
Gwen
mentioned,
that
she
didn't
have
milestone,
maintainer
access
I
was
surprised
and
I
started.
Since
people
were
talking
and
typing
notes,
I
started
going
off
to
grant
access,
but
we
will.
We
will
start
exercising
that
in
a
simple.
We
remove
the
milestone
label,
with
a
notation
of
why
it
time
has
gone
by
and
if
they're
really
serious
and
have
proof
and
evidence
they
can
come
back
and
bring
it,
but
practically
speaking
how
this
happens
from
every
path
cycle.
I've
been
involved
on
people
say:
yes,
ok,
we
get
it.
A
C
A
Yeah
we
can
collate
that
list
first
and
give
them
one
more
layer
of
announcement.
Honestly
I
don't
expect
that
to
change
much
of
anything,
because
the
a
good
portion
of
the
people
aren't
reading
notices
like
that
anyway,
when,
when
they're
there,
when
their
particular
thing
does
not
merge,
they
ask
why.
C
A
C
A
Alright,
next
on
the
list
is
PRS
is
we're
transitioning
from
slush
freeze.
It
becomes
a
little
more
important
to
monitor
PRS
themselves
and
this
used
to
fall
under
issue
triage
and
is
sort
of
now
a
kind
of
a
team
collaborative
issue,
but
also
Gwen
has
volunteered
to
take
point
on
it,
a
bit
watching
what's
going
by
the
there's,
a
worry
in
that-
and
this
has
been
articulated,
particularly
in
sig
p.m.
but
since
we
don't
have
a
formal
like
kept
to
issue
to
PR
tight
coupling
on
feature
things.
A
All
of
these
people
that
we'd
be
talking
about
right
now,
potentially
dropping
their
things.
There
there's
nothing
to
stop
them
from
advocating
to
their
sig
leadership.
To
say,
yes,
add
the
milestone
back
on
to
this
thing
and
because
things
automatically
merge
and
things
just
flow
in
so
as
a
release
team,
we
need
to
watch
the
PRS
going
by
a
day
to
day
now
to
make
sure
that
we
don't
have
excess
and
unknown
things
being
merged
by
well-intended
people
that
continue
to
to
complicate
our
task
of
settling
down
CI
and
getting
a
quality
release
out.
So.
B
I
B
B
The
stand-up
so
a
note
for
for
everyone,
we
are
actively
Jase
and
I
are
kicking
around
some
ideas
on
on
sick
p.m.
and
I
know
that.
Maybe
some
of
you
have
seen
the
two
proposals
that
were
we're
doing
both
of
those
proposals
will
probably
not
happen.
We're
going
with
something
entirely
different,
which
is
the
fakes
the
kept
process
first
and
then
start
moving,
possibly
moving
the
design
proposals
and
caps
over
from
que
community
over
2k
features,
renaming
que
features
to
KK
EPS
and
starting
to
resolve
some
of
that
tech
debt.
B
A
So
that,
indirectly
reminds
me
we
do
have
a
retrospective
document
open
and
at
the
point
that
we
are
done
with
the
release
we'll
go
through
a
retrospective,
it's
useful
to
drop
things
in
as
we
go,
and
especially
during
the
kind
of
panic
of
the
current
two
weeks.
It's
easy
to
lose
track
of
things
so
day
to
day,
if
you,
if
you
notice
something
particularly
annoying,
please
drop
those
in
the
retro
dock.
A
Thank
you,
okay.
Thank
you.
Thank
you
for
noting
that
also
related
to
that
I
have
been
putting
out
a
kind
of
a
shout
out,
so
we,
if
nothing
else,
hopefully
we
we
will
not.
Hopefully
we
will
get
a
quality
release
out
here,
but
one
of
the
things
that
I
take
a
little
bit
of
pride
in
at
least
for
this
release
is
that
we've
done
a
whole
bunch
of
documentation
on
the
process.
A
There's
a
lot
more
to
do,
but,
as
you
think
about
what
you've
done
in
your
role
this
cycle,
please
give
another
read-through
to
the
role
handbooks
that
we've
gotten
updated
significantly
this
quarter
of
it
all
and
go
through
and
see
if
retrospectively.
If
there's
anything
else,
you'd
like
to
additionally
change
there
to
hand
off
to
the
next
folks
in
front
of
us,
yeah.
E
A
A
If
you
could
send
me
that
link,
then
all
are
on
that
list,
then
I'll
be
I'll
collate
the
set
of
those
things
and
similarly
Steven.
You
give
me
your
couple
and
go
in
if
there's
any
that
you
are
worried
about
that
you've
seen
as
well,
so
that
and
if
everybody
hands
me
that
sort
of
stuff
I
will
scrub
through
and
communicate
it
later
today
and
tomorrow
again
in
the
community
meeting,
sounds.
J
Are
right
on
schedule?
The
blog
post
is
now
in
review,
so
Tim
I
Steven.
If
you
could
go
into
that
dock
and
review
by
September
11th
as
well
as
input
your
comments
for
the
media
messaging
doc,
both
of
those
should
be
in
your
email
and
then
the
link
to
the
blog
post
is
in
the
notes
from
today.
If
anybody
else
wants
to
take
a
look
in
review.