►
Description
A
It's
so
lucky:
yes,
okay,
hi
everybody!
Welcome
to
this
week's
alright,
let
me
try
it
again:
hi
everybody
welcome
to
Monday
March
eighteenth.
You
are
at
the
kubernetes
114
release
team
meeting.
You
are
all
being
publicly
recorded,
so
you
can
go
and
watch
yourselves
on
YouTube
later
adhering
to
the
kubernetes
code
of
conduct,
where
you
are
basically
not
being
jerks
I
in
in
honor
of
yesterday's
holiday
in
the
u.s..
A
C
E
E
Hi
everyone,
so
things
are
looking
better
than
last
week.
I
believe
that
master
logging
currently
has
no
failing
tests.
A
couple
of
dashboards
underneath
it
are
flaking
one
1:14
blocking
still
experiences
the
same
failures
as
master
blocking
did
its
specifically
the
HTTP
to
ingress
test
that
used
to
fail
on
master
a
bit
of
a
bit
of
context
here
and
things
that
happened
on
Friday,
so
sick
Network
has
decided
to
take
out
the
test.
E
First
of
all,
there
isn't
why
it
was
failing
is
because
the
go
client
used
to
use
within
the
test
to
target
HTTP
2
and
points
did
not
support
those
personality.
Tests
needed
signet
work
decided
to
take
the
test
out
of
master
blocking
Suites
on
the
grounds
that
it
does
not
cover
coronaries
functionality
and
and
therefore
they
decided
monitor
these
things
outside
sig
release,
but
live
in
a
conformance
test,
which
I
think
sounds
about
right
to
me.
A
So
I
can
answer
the
second
question.
First,
they
have
made
it
to
the
dashboard
test
grid.
Has
this
fun
feature
where
it
considers
a
job
as
failing?
If
there's
a
test
that
failed
and
then
didn't
have
any
green
runs
following
that
failure
until
some
amount
of
hysteresis
has
happened?
So
that's
how
like,
if
you
go
to
the
release
master
blocking
version
of
the
ingress
test,
there's
a
little
link
that
says
like
show
snail
tests
in
the
column
where
it
lists
all
the
test.
A
Names
right
above
that
there's
a
little
link
that
says,
show
snail
tests
and
you
can
see
that's
how
the
test
got
removed.
The
same
thing:
it
has
happened
for
release
114
blocking,
but
it
hasn't
done.
The
last
looks
like
it
takes.
Ten
runs
before
it
consider
something
stale
and
ages
it
out
of
the
display
in
computations.
So
a
few
more
runs
to
go
on
114,
interesting.
A
A
Could
be
that
is
very
fair
from
a
conformance
perspective,
I
think
I'm,
more
ok
with
that
than
blocking
on
correct
me
if
I'm
wrong,
I
think
what
you
are
telling
me
is.
We
had
a
specific
cloud
providers,
implementation
of
the
ingress
resource,
failing
and
so
we've
removed
that
specific
cloud
providers,
implementation
of
ingress
there.
A
E
These
are
so
the
this
used
to
pass
and
are
still
passing.
Nothing
has
really
changed
there.
The
only
one
that
was
failing
was
the
HTTP.
Two
specific
specifically
was
a
test
that
that
asserted
that
the
controller
was
able
to
switch
between
HTTP
and
82
if
I
remember
correctly,
so
a
CTS
and
HTTP
2,
and
that
was
the
one
that
was
removed
by
the
from
their
release.
Blocking
dashboards
I,
don't
think
that
ever
passed,
if
I
read
correctly.
G
H
A
That
in
the
abstract,
like
the
there,
there
is
a
bit
of
an
anti-pattern
that
happens
sometimes
where
people
are
like.
Oh
I'll
just
make
everything
go
green
by
removing
the
failing
test.
It's
like
yeah,
but
you
didn't
fix
the
problem.
The
problem
is
still
there
so
just
reading.
Through
your
conversation
on
that,
though,
it
seems
like
this
is
a
cloud
provider
saying
they
acknowledge
that
ingress
doesn't
work
quite
properly
on
their
cloud
at
the
moment
until
a
specific
bug
fix
has
gone
through
yeah.
E
E
E
So
I
don't
know
if
there's
much,
except
following
up
on
things
that
are
there
on
the
see
a
signal
boards
like
as
I,
said:
I,
don't
believe
that
there's
anything
new
I
haven't
seen
much
new.
There
is
one
issue
that
I've
added
to
the
report:
sorry
to
them
released
a
meeting
which
has
not
had
a
lot
of
recent
attention,
so
I'll
just
follow
about
just
following
up
on
things.
It's
really
it's
got
open
at
the
moment.
E
A
Can't
necessarily
think
of
anything
off
the
top
of
my
head
I
feel
like
we
are
preparing
ourselves
for
cutting
a
release
where
a
number
of
things
are
read
and
failing,
and
we're
going
to
have
to
interpret
that
as
business
as
usual
as
humans,
so
being
really
clear
what
the
flakes
are
that
we
have
documentation
for
to
point
to
say
yeah.
This
is
all
business
as
usual,
we'll
help
us
and
also
what
may
help
Hannes
assuming
we
can
stick
to
the
current
release.
Schedule.
A
Part
of
the
reason
I
want
us
to
stick
to
a
schedule
of
like
not
accepting
any
more
changes
by
Friday
is
to
give
us
plenty
of
CI
runs
against
the
same
commits
over
a
weekend,
which
is
something
we
just
went
through
and
remember
back
to
our
point
about.
We
want
to
make
sure
we
can
have
a
commit
past
the
same
test
three
times
in
a
row.
I
would
be
curious
if
the
boards,
as
they
exists,
work
today
after
this
weekend,
were
in
a
state
to
allow
us
to
do
that
and
then
like.
A
The
thought
it's
not
a
hard
to
ask,
but,
like
that's
those
are
the
things
that
come
to
mind
to
me,
because
that
I
viewed
like
this
weekend
as
a
dress
rehearsal
for
what
next
weekend
will
look
like
the
only
difference
being
next
weekend,
we're
not
going
to
be
in
code
freeze,
so
the
master
upgrade
board
becomes
meaningless
to
us,
and
the
master
informing
board
also
becomes
meaningless
to
us,
because
we
only
care
about
code.
That's
in
the
114
branch.
As
soon
as
we
left
code
freeze.
H
A
Okay,
the
main
outstanding
PR
there
is
the
killing
one
trying
to
get
I,
don't
have
the
historical
context
to
be
able
to
answer
the
question
that
is
asked
in
that
poll
request.
Does
anybody
else
on?
This
call
have
historical
context
to
answer
the
question
about
well:
Liggett
you're.
The
person
who
asked
a
question
I
think
we
need
to
bump
the
version
file
and
Jim's
is
uncertain.
I
have
reached
out
to
Jeff
Grafton
to
understand.
J
A
It's
kind
of
a
nice
to
have,
but
I
also
feel
very
strongly.
We
should
be
building
our
releases
with
actual
images
so
idea.
This
is
the
main
thing
that
would
be
an
impediment
to
lifting
code
freeze
and
cutting
and
release
candidate.
I
believe
I
have
whipped,
approves
an
LD
TMS
on
the
remaining
PRS
on
it.
How
is
our
test
and
for
looking.
G
Hey
the
testing
phase,
looking
pretty
good,
we
merged
42
years
last
week,
which
is
good
since
we
were
in
code
freeze
and
then
they
go
and
well
one
changes
are
looking
good.
We
will
move
to
using
official
images
now
that
they
know
that
we're
seeing
that
it's
working
and
then
the
only
other
thing
was
that
upcoming
milestone
is
code
thrush
tomorrow,
unless
we
decide
something
unless
we
decide
otherwise,
today
I
think
that's
you.
F
K
K
A
I
Yes,
not
a
problem
and
the
reason
it's
yellow
is
a
I
feel
like
I'm,
at
like
kind
of
that
pause
area
where
we're
in
a
really
good
place,
but
I
need
some
additional,
like
approvals
or
green
lights,
from
the
release
team
before
I
feel
like
we
can
move
to
a
green.
So
the
next
steps
for
me
really
is
the
external
media
pitching,
and
so
this
is
where
I
get
really
sensitive.
Is
everybody
comfortable
with
the
messaging
in
here?
Is
this
accurate?
I
Can
we
share
it
under
embargo,
so
media
can
use
it
to
write
their
stories
off
up.
So
that's
why
it's
really
yellow
so
the
release
blog
is
in
a
really
good
place.
I
do
ask
that
the
release
can
go
through
the
blog
today
and
anything
in
there.
That
needs
to
be
changed.
Please
make
a
change
to
if
there's
things
in
there
that
are
going
to
be
changing
all
the
way
up
until
the
release
is
cut
on
Monday,
please
just
kind
of
make
a
note
in
there
highlight
it.
I
You
know
make
a
note
so
that
way,
I
know
like
we
want
to
include
this
in
the
final
release
version.
If
it
makes
it
through
the
cut
but
I'm
not
comfortable,
sharing
it
with
media
right
now,
so
that
so
those
are
kind
of
the
lens
that
I
need
you
to
go
into
the
blog
and
look
at.
Is
this
okay
to
share
with
media
to
start
writing
their
stories?
Is
it
not?
Okay,
are
you
comfortable
sharing
this?
Are
you
not
comfortable
sharing
this?
I
If
my
entire
blog
turns
yellow,
then
obviously
we're
not
ready
to
share
at
all?
But
if
it's
just
a
few
pieces,
then
I
can
make
a
version
where
it's
just
a
share
under
embargo,
and-
and
so
that's
really
important,
because
for
me
for
my
process,
it
I
would
reach
out
to
media
tomorrow
morning
and
say
you
know:
hey
even
covered
covering
super
Nidhi's
releases.
A
lot
are
you
interested
in
covering
the
1.14
release?
The
media
comes
back
asks
for
the
the
announcement
under
embargo,
so
that
way
they
can
read
a
little
bit
more
about
it.
I
That's
when
I
send
the
version
that
everybody
is
okay
with
me
sharing
and
then
from
there.
They
either
write
of
that
blog
or
they
request
interviews,
and
then
those
are
the
interviews
that
I
scheduled
Wednesday,
Thursday
Friday,
so
having
a
version
of
the
blog,
is
kind
of
like
my
poker
chip
to
play
in
so
that
way,
they'll
accept
it
and
want
to
write.
I
So
please
please,
let's
have
lots
of
conversations
the
release
blog
today,
I
can't
even
push
pitching
off
if
we
need
tomorrow,
but
let's
get
it
in
a
spot
where
they
were
comfortable
with
sharing
these
aspects,
the
media
messages
dock,
it
looks
like
there's
still
quite
a
few
responses
and
conversations
on
responses
so
again
that
this
is
a
document
that
is
100%
shell
for
media
to
write
their
stories
they
suffer.
So
if
there's
something
in
that
document
that
you
do
not
want
to
put
in
there,
please
remove
it.
I
If
there's
things
in
that
documents
that
are
missing,
please
try
to
fill
that
out
today.
This
document
has
detailed
about
the
release
that
are
not
relevant
in
the
release
blog
but
are
very
relevant
for
media
and
their
stories.
They
help
kind
of
tell
that
continuation
release
story
that
we
want.
So
again,
please
please
go
in
there
today.
Let's
have
lots
of
conversations
in
the
dock
today,
let's
get
that
in
a
solid
place,
I
will
not
and
Erin.
This
is
Hugh.
I
A
If
you
can
get
that
content
in
before
your
end
of
day,
I
know
some
of
you
aren't
in
the
Pacific
time
zone,
because
the
folks
love
to
write
stories
about
the
inside
workings
of
the
release,
and
you
know
what
went
what
went
well
so
if
you
can
help
them
tell
some
interesting
stories,
that's
great
and
I'm
sure
you
have
stories
to
tell
that
I
wasn't
necessarily
around
for
so
that
could
be
really
helpful,
but
Natasha
I
do
appreciate.
I
will
act
as
final
green
light
for
that
and
I'm
hearing.
I
Yeah,
if
possible,
if
we're
getting
towards
the
end
of
day
today,
Pacific
time
and
you
guys
are
like
hey
I-
need
all
day
tomorrow
to
work
on
this
as
well.
Just
slack
me-
and
let
me
know
so
that
way.
I
know
I'm,
not
waiting
for
that
green
light
tomorrow
and
I
can
definitely
push
till
Wednesday.
I
would
prefer
not
to
push
farther
than
Wednesday,
because
that
only
leaves
a
couple
days
for
media
to
write
their
stories
and
a
couple
days
to
do,
interviews
and
I
something
that
happens.
I
A
lot
is
companies
who
have
people
who
have
worked
in
the
sig
or
are
working
on
the
release.
Sometimes
there
appear
people
really
want
to
be
the
folks.
Have
their
company
be
the
spokesperson
for
a
release,
and
so
they'll
start
pitching
media,
and
then
there
are
the
ones
that
do
the
interviews
and
sometimes
they're
not
doing
the
interviews
based
off
of
the
source
of
truth.
This
happened
a
lot
in
1.7
and
1.8,
so
I
try
to
get
ahead
of
that,
and
so
that's
why?
A
I
A
I
And
then
the
other
thing
that
I
need
is
my
spokespeople,
so
the
enhancements
leads
Claire.
This
be
you
Aaron,
if
you're
comfortable,
being
a
spokesperson
for
the
overall
release
and
then
usually
the
sig
covers
leading
office.
Thing
is
a
spokesperson.
The
ones
that
we
mainly
highlight
in
the
blog,
so
like
windows
is
going
to
be
a
big
one.
I
Cube
cube,
CTL
is
going
to
be
a
big
one.
It
looks
like
storage
storage
is
always
going.
I
Spokesperson
from
you
can
just
black
message
me
who
this
person
is,
and
it's
helpful
to
know
what
company
are
at,
because
then
I
can
kind
of
go
reach
out
to
the
peer
person
on
your
behalf
and
just
say,
like
you
know,
for
example,
for
Microsoft
hi
Molly,
like
so-and-so
from
Windows
big,
is
going
to
be
a
spokesperson
for
us
on
on
1.14
they're,
not
going
to
have
their
Microsoft
hat
on
they're
gonna
have
their
thick
hat
on,
but
just
wanted
to
give
you
a
heads
up
that
you
know
they
may
have
a
couple
of
media
interviews.
I
We
do
that.
You
know
not
only
you
know.
Out
of
respect
for
the
different
companies,
but
that
way
we
don't
want
any
of
the
PR
people
coming
back
and
saying
wait
a
minute
they
weren't
approved
to
be
a
spokesperson
period
so
and
I
know
most
of
the
PR
people
for
all
the
companies
that
are
working
on
this
release.
So
I
could
just
send
quick
emails
out.
That's
fine.
I
just
need
to
know
who
my
people
are.
I
Thank
you
so
much
yeah
and
I
I,
don't
I,
don't
anticipate
a
ton
of
interviews,
but
we
usually
get
about
four
or
five,
and
sometimes
the
reporters
are
like
I
just
want
to
talk
to
them
about
everything,
that's
going
on
with
the
release,
and
so
that
may
be.
You
know
three
or
four
of
us
on
the
phone
or
sometimes
they'll,
hone
in
and
say
you
know,
I
really
just
want
to
talk
about
the
windows,
enhancement
and
so
then
I
would
just
set
up
the
meeting
you
know
with
with
the
window.
L
Y'all
looking
good
the
next
bigger
thing
is
happening
tomorrow.
At
least
it's
planned
for
that
I
would
cut
the
release
candidate
and
off
my
day,
if
you
have
any
comments
on
the
timing,
please
let
me
know
and
I'm
happy
to
push
around
a
bit
and
in
general,
I
will
check
on
a
go/no-go
signal
tomorrow
anyway,
either
in
the
burndown
meeting
or
after
that
and
I
guess
it's.
It's
gonna
be
a
point
of
discussion
now
anyway,
but
other
than
that
all
good.
All
green
yeah.
A
Awesome
so
to
that
point
of
discussion
I,
we
had
a
bunch
of
discussion
on
Friday
about
the
thought
that
the
only
real
major
hurdle
for
us
was
going
112,
potentially
introducing
some
kind
of
regression
that
we
were
experiencing
solely
in
a
few
of
our
upgrade
tests.
So
we
moved
forward
to
going
112
one
and
then
I
was
supposed
to
give
people
a
heads
up
that
we
may
be
pushing
the
code
thaw
date
back
a
little
bit
to
give
us
room.
Then
I
went
in
promptly
for
got
to
tell
anybody
that,
but
that
seems
okay.
A
This
is
why
I'm
so
lucky
right
now,
because
it
seems
as
though
going
112
one
from
my
evaluation
has
solved
the
problems
that
we
thought
it
would.
I
went
through
and
tried
to
chase
down
a
number
of
other
flakes
that
people
were
trying
to
attribute
to
the
move,
to
go
lying
112
and
either
it
seems
like
they
weren't
related
or
that
112
one
has
fixed
them,
but
I
kind
of
want
to
just
take
a
quick
poll
of
all
of
my
different
leads
and
see.
A
H
A
E
So
I
mean
from
the
things
that
are
not
stable
at
the
moment.
I
there
isn't
something
there
that
I'm
afraid
of
in
terms
of
I.
Don't
think
that
there's
something
that
we
look
at,
that
hints
at
a
other,
consistent
failure.
Famous
last
words,
I
am
a
bit
uncomfortable
with
the
amount
of
instability
still
I
would
definitely
appreciate
some
time
to
stabilize
things
more,
but
I
think
we
can
always
make
the
case
for
that.
So
I
think
I'm
I'm
definitely
way
more
comfortable
sticking
up.
E
A
Like
code
freeze
has
been
all
about
making
it
really
easy
to
land
changes
in
the
114
branch
if
you're
change
lanes
and
master
using
the
usual
process
so
that
you've
always
used
to
contribute
to
kubernetes
it's
going
to
get
into
114
as
soon
as
Hannes
presses
the
fast-forward
button.
It's
great
code
thaw
is,
when
things
become
a
little
more
complicated
or
if
you
want
your
change
to
land
on
114,
it
has
to
be
cherry-pick
into
the
114
branch
and
especially
during
code
thaw
period.
A
When
there's
this,
like
horde
of
PRS,
that
have
been
waiting
to
land
in
master,
we
may
find
ourselves
competing
with
all
of
those
PRS
when
we
want
to
try
and
land
something
in
master.
So
we
can
then
cherry-pick
it
into
the
114
branch.
The
way
we
have
dealt
with
this
in
the
past
is
to
suggest
that
somebody
put
their
change
directly
in
the
114
branch.
A
A
A
E
A
B
So
I
have
something
if
we
have
some
time
can
I
get
more
eyes
on
this,
and
this
mainly
take
a
look
at
the
first
thing.
That
is
an
out-of-date
copy
of
the
114
release
notes
by
the
way,
and
then
that
is
the
issue
that
I'm
using
to
track
just
progress
for
this,
because
I
would
kind
of
like
to
try
and
get
this
done
by
1:15.