►
Description
A
A
Don't
know
if
it
was
like
this
for
anybody
else,
but
when
I
try
joining
its
there's
another
meeting
already
in
progress.
It's
like
that's
weird,
because
this
is
the
meeting
that's
supposed
to
be
in
progress
and
I.
Think
somebody
clicked
the
European
burned
down
meeting
Lee
because
that's
scheduled
for
a
separate
day
at
a
separate
time
and
zoom
doesn't
let
the
have
same
look.
The
same
account
have
multiple
scheduled
meetings
going
on
concurrently
and
because
I
have
to
do
this
wonderful
thing
where
I
allow
everybody
to
join
the
four
the
host
joints.
A
A
Okay,
let
me
do
the
usual
spiel
and,
while
I'm
doing
a
general
spiel
I
will
increase
the
link
to
the
meeting
notes
in
the
chat.
So
you
all
can
fill
your
name
in
the
attendees
list.
Hi
everybody
I
am
Erin
of
sig
beard
today,
his
Monday
March,
the
4th.
This
is
the
kubernetes
114
release.
Team
weekly
meeting
this
week
is
week
9
for
all
being
publicly
recorded,
witnessing
and
talking
about
week.
9
as
we
adhere
to
the
CNC
f
code,
our
society
kubernetes
code
of
conduct,
which
boils
down
to
please
don't
be
a
jerk.
A
So
as
usual,
I
would
like
to
try
and
have
everybody
else
go
before
me,
so
that
I
don't
end
up.
Monologuing
I
will
say
up
at
the
top
that
I
do
appreciate
everybody
helping
out
with
memes
to
remind
everybody
that
it's
code
freeze,
I'm,
slowly,
building
up
my
tweet
thread
of
of
memes
to
remind
everybody
that
code
freeze
is
coming
up.
B
So
I
said
for
the
current
status
of
114
enhancements.
It's
feeling
it
yellowish-green
to
me
right
now,
leaning
on
the
yellow
side,
because
we
still
have
11
issues
that
are
marked
at
risk,
mostly
because
we're
waiting
for
Tusk
panniers
to
get
merged
and
like
there
are
just
a
lot
of
open,
PRS
and
realizing
that
code
freezes
on
Friday
but
green,
because
people
are
responsive
on
github
and
slack.
It
seems
like
at
least
when
we
do
ping
them
for
updates.
B
B
A
C
B
A
Yeah
I
pinged
in
slack
and
I,
got
a
lot
of
feedback,
and
so
I'll
just
throw
my
opinion
out
there,
but
I'm
curious,
like
I,
really
prefer
other
people's
opinions
ever
mine
again
a
lot
of
feedback
that
this
is
like.
Oh
we're
just
shuffling
code
around,
it's
just
a
refactor,
it's
not
user
facing
nobody
should
care
about
like
it's
not
worth
blogging
about.
It's
not
worth
bragging
about.
It's
not
worth
highlighting
as
a
theme.
A
Okay,
cool
I,
get
that
it
seems
like
it's
maybe
too
early
to
try
and
cluster
it
in
the
extensibility
and
stability.
But
there's
a
lot
of
code
related
to
it.
That
does
involve
moving
things
from
place,
a
to
place
B
and
sometimes
that
can
break
stuff.
Should
we
as
a
release
team,
be
tracking
these
sorts
of
things
or
should
we
as
release
team,
say?
Oh,
don't
worry.
A
C
D
E
E
F
E
Those
people
who
are
more
or
less
I
guess
it
would
be
I
guess
the
next
way.
You
can
say
it's
pre,
factoring
work,
I
think
there
is
still
the
same
level
of
scrutiny.
You
would
expect
for
anyone
shipping
any
changes
to
be
released,
elevation
branch
because
it
impacts
our
ability
to
stabilize
the
branch.
So
I
think
we
do
need
to
have
some
scrutiny
over
the
changes
going
and
writ.
F
F
E
A
Okay,
the
one
I
had
liked
about
this
cap
is
that
it
talked
about
a
very
phased
approach
to
implementing
things,
and
then
it
even
called
out
what
the
objectives
were
for
the
1:14
timeframe.
But
then
it
appears
as
though
the
state
of
today
is
that
that
objective
is
only
going
to
be
partially
complete,
and
so
in
that
partially
complete
state.
A
Okay,
so
my
suggestion,
unless
anybody
like
disagrees,
is
I'm
gonna
kick
this
enhancement
out
of
the
milestone
and
we
can
start
using
like
potentially
as
we
come
up
to
code
freeze,
we
could
start
being
like
hey.
Is
this
PR
actually
related
to
an
enhancement?
Could
you
like
us
to
be
enhancement?
Could
you
like
us
to
the
Catholic?
This
PR
is
definitely
like
helping
fix,
because
if
we
can't
see
those
that
doesn't
sound
like
a
justification
for
trying
to
land
after
code,
freeze,
okay,
so
the
next
one
I
think
has
to
do
with
huge
pages.
B
B
This
one
I
also
think
is
probably
likely
going
to
get
bumped,
because
the
tests
or
the
the
document
that
they'd
been
referring
to
as
they're
sort
of
community
proposal
isn't
really
in
line
with
a
cup,
and
we
can't
seem
to
be
getting
a
lot
of
updates
on
them
on
where,
like,
if
they're,
converting
anything
to
a
cup
or
if
they
have
test
plans
elsewhere,
I
can
also
just
king
sing,
while
I'm
picking
cig
note
on
huge
pages.
I
can
sing
them
on
this
one
as
well.
Okay,.
A
So
I
think
I'd
like
for
Claire
to
or
I
can
do
it,
but
since
I'm
thought,
I
want
to
empower
Claire
here,
Claire
I
think
you
should
make
the
call
that
127
is
getting
kicked
out
of
the
milestone
and
then
I
think
the
issue
triage
team
should
make
sure
that
any
PRS
related
to
this
milestone
are
held
with
a
link
back
to
the
comment
that
says:
we're
not
accepting
this
because
of
these
reasons,
I
think.
There's
sufficient
comment.
A
B
A
I
would
rather
have
that
list
soon,
so
we
can
start
kicking
things
out
soon.
I
feel
like
trying
to
kick
things
out
a
day
right
before
code.
Freeze
might
be
a
little
uncomfortable
for
people
and
I
about
the
opinion
that
at
the
start
of
the
work
week
is
a
much
better
time
to
give
folks
a
heads
up
that
we
are
moving
on
some
things.
G
Indeed,
hello,
okay,
so
a
couple
of
things
happening,
so
I've
put
I've
got
a
yellow
there.
My
main
my
thought
process
is
that
there
are
a
couple
of
consistently
failing
tests
in
blogging
suites,
which
makes
moving
away
from
green,
but
the
vast
majority
of
them
are
not
long-running
so
they
have
occurred
since
the
weekend
and
the
one
that
has
been
open
for
days
has
appear
open
to
to
solve
it.
So
in
terms
of
consistently
famous
sorry
consistently
failing
stuff
we're
in
relatively
good
shape.
G
The
thing
that
is
a
bit
more
worrying
is
that
there
is
a
significant
percentage
of
rounds
that
are
flaky
so,
while
technically
the
fast
at
least
once
for
commits
that
we
care
about
that
have
been
ported
on
to
114.
They
are
flaking,
some
of
them
the
majority
of
the
time
so
in
terms
of
just
numbers
in
in
failures
in
blocking
Suites,
we've
got
zero
at
the
moment
in
master
blocking
three
in
master
upgrade
one
has
an
open
PR.
G
One
of
the
issues
has
an
open
PR
to
fix
it,
and
the
other
couple
of
issues
started
during
the
weekend
and
we've
got
one
job
in
114
blocking,
but
it
seems
like
a
flake
which
is
not
much
better
but
seems
at
least
like
an
indication
of
that.
It's
not
a
bug
you
mentioned
before
that
you
are
freaking
out.
Quite
hungry,
I
am
a
little
bit
not
heavily
just
yet.
G
It's
mostly
the
the
amount
like
the
the
hinder
instability
that
comes
with
high
percentage
of
flakes
and
that's
going
to
be
I,
don't
want
to
say
the
top
priority,
but
the
joint
top
priority
over
this
week,
together
with
any
new,
obviously
like
failing
tests.
The
the
action
plan
that
I
had
in
mind
there
is
to
try
and
address
a
little
bit
the
tragedy
of
the
comments
that
comes
with
failing
jobs
like
a
lot
of
the
time,
especially
with
flakes
things
with
flake,
but
because
of
two
different
tests
that
belong
to
different
SIG's
and
whatnot.
G
G
A
A
A
But
to
my
eyes
there
really
haven't
been
that
many
instances
of
that
in
the
past
two
weeks,
which
is
about
as
far
back
as
the
test
grade
history
goes
and
I
suspect
if
I
invested
the
time
in
digging
up
longer,
reaching
metrics
out
of
bigquery
and
tried
to
turn
that
into
a
graph.
It
maybe
wouldn't
look
a
whole
lot
prettier.
This
is
really
concerning
to
me
because,
as
of
right
now,
I
feel
like
we're
not
going
to
be
able
to
cut
on
release.
A
So
why
is
this
happening
so
much
if
I
look
at
both
of
those
boards
I
feel
like
I,
actually
can't
point
to
any
excuse
me.
Anyone
Ingle
pressed
to
use
Maria's
tragedy
of
the
Commons
thing,
but
if
I
look
at
the
prefixes
on
those
test
names
this
this
B's
board
sort
things
in
order
of
flakiness.
You
can
also
sort
in
order
failures.
They
all
seem
to
be
sake,
storage
or
sing-off.
A
So
my
suggestion
would
be
that
we
literally
get
somebody
from
each
of
those
SIG's
to
show
up
the
next
burndown
meeting
and
that
you
reach
out
to
those
things
directly
and
let
them
know
this
is
really
a
problem
and
it's
really
blocking
their
release,
because
I'm
not
sure
if
this
is
a
question
of
do,
we
need
to
kick
some
of
these
tests
out.
Are
they
just
poorly
written
tests?
Is
there
some
larger
systemic
issue
that
is
causing
like
over
30
different
test
cases
just
to
the
clients
here
to
all
flake
in
one
way
or
another?
A
A
E
Suggest
as
much
as
I
hate
ad-hoc
process,
if
there's
challenge,
if
there
are
challenges
in
SIG's
storage,
that
is
a
sig
that
is
coach
hit
well,
both
with
SIG's
origin
cigar.
There
are
opportunities
to
pull
more
people
more
eyes
on
this
lease
from
the
Google
side,
because
we
staff
those
places
pretty
heavily.
If
Maria,
you
have
some
like
a
procedure
that
say
an
engineering
manager
could
direct
himself
or
herself
or
their
team
to
use
to
help
triage
this
as
kind
of
a
fire
drill
exercise.
F
A
I
think
that's
why
I'm
advocating
for
escalating
to
the
state
directly,
because
I
need
I,
think
I'm
looking
for
some
acknowledgment
from
the
say
that
they're
aware
of
this
broader
issue,
Saad
and
Michelle
are
both
amazing
super
productive
people
that
sig
is
amazingly
productive.
They
helped
unblock
us
with
difference
between
Otis
that
you
know
their
their
tests
aren't
locked
leaking
volumes
everywhere.
It's
become
really
clear.
Just
how
much
the
rest
of
the
flakes
are
affecting
us.
I
also
see
there's
a
signet
workless.
A
Yeah
I
mean
just
if
I'm
just
trying
to
focus
on
one
thing
right
now:
I'm
focused
on
that
one
job.
There
probably
a
couple
others
but
like
that
one
job
is
super
red
and
the
way
the
board
looks
right
now:
six
storage,
his
name
pops
up
a
lot
and
so
does
sing-off.
So
those
are
the
two
that
I
would
start
with.
Yep
make
sense.
I
I
mean
most,
if
not
all,
of
the
issues
that
we've
identified
in
the
triage
tracking
spreadsheet
have
been
labeled
with
114
milestone
and
have
been
pinged
by
Nikko
and
yourself,
amusingly
and
so
I'm
going
to
spend
some
time
this
week
and
trying
rally
the
troops
to
reaping
these
issues.
They
need
another
update.
There
hasn't
been
much
of
an
update
since
the
pinging,
at
which
one
do
we
need
assignment,
and
not
these
issues
are
going
to
get
into
114
at
all.
I
A
J
Good
good
does
infer
from
the
distance
aside.
Wait
looks
pretty
good.
We
Mazda
158
piers
last
week,
which
is
looks
like
it's
under
standard
trajectory.
After
fixing
the
go-ahead
failures
on
Friday,
we
have,
we
have
successfully
placed
on,
go,
went
well
and
I.
Don't
you
sent
out
an
email
to
ke
dev
regarding
that
regarding
prowl
health?
A
So
I
am
personally
of
the
opinion
that
it
is
really
really
really
important
that
we
get
a
green
run
of
the
scale
performance
job
or,
if
somebody
can
point
me
to
a
green
run
of
that
job,
since
we
migrated
to
go
112,
because
that
was
sort
of
the
last
sentence
in
the
email
that
I
sent
out
to
kubernetes
des
about
migrating.
To
this.
A
As
we've
discussed
in
meetings
in
the
past,
sig
scalability
has
the
ability
to
reject
large
changes
that
impact
the
performance
of
kubernetes
and
prevent
it
from
meeting
the
two
SLS
they're
accountable
for
have
to
deal
with
like
pod,
latency
and
scheduling
very
good.
So
if
migrating
to
going
112
did
in
fact
break
that
we
would.
It
would
be
really
helpful
to
find
that
out
now,
as
opposed
to
during
code
freeze.
J
K
A
I
I
see
a
couple:
green
runs
of
the
master
scale
performance
job
over
the
weekend.
March
1st
March,
2nd
those
are
green.
That's
super
cool,
but
the
scale
correct
this
job
I
don't
see
any
green
since
February
27th,
which
means
I,
haven't
seen
a
successful
one
of
that
job
with
Caroline
112.
So
hopefully
we
can
get
that
happening
before
code.
Freeze.
K
A
A
So
if
anybody
has
any
opinions
or
is
interested
in
discussing
about
discussing
that
stochastic
meeting
tomorrow
at
1:30-
and
you
are
welcome
to
add
whatever
items
on
the
agenda,
you
would
like
as
well,
if
you
can't
make
the
meeting
and
we'll
make
sure
that
we
discussed
those
ok,
Jim,
I'm,
so
curious.
Let's,
let's
hear
about
the
effectiveness
of
the
bring
out
your
Doc's
mean
it.
I
The
yellowish
is
my
fault
for
just
not
keeping
up
and
keep
as
close
of
a
metric
on
our
documentation
that
we're
missing
so
we'll
have
a
deliverable
and
a
day
today
to
know
exactly
what
we're
missing
and
what
we
need
to
do.
As
far
as
any
problem,
once
I
clear
was
mentioning
earlier.
We
have
one
that
already
has
some
issues
in
addition
to
missing
the
docs
PR.
So
when
I
get
the
hammered
out
to
end
a
day
today,
I
don't
have
a
further
update
for
the
next
burndown
meeting.
B
Jim
I
don't
know
this
is
a
question
for
Jim
or
if
it's
just
kind
of
I,
don't
know
like
the
normal
process
around
dogs,
but
if
enhancements
miss
the
docs
freeze,
do
they
like?
Is
it
normal
for
us
to
bump
them
off
the
release
because
they
miss
that
or
do
we
kind
of
say
get
these
in
I
like
this
particular
extension?
Do
we
make
them
submit
exceptions
so.
I
B
F
Yep,
so
latest
PR
is
in
with
the
latest
release
notes.
What
we're
kind
of
working
on
this
week
is
really
updating
external
dependency,
known
issues,
doing
some
editing
on
the
release,
notes
and
also
kind
of
discussing
external
dependencies
going
forward.
I
know
Jeff
had
a
few
ideas
around
that,
but
that
be
something
kind
of
more
of
a
looking
forward
thing.
A
A
So
I
guess
help
me
understand.
You
know
we
talked
a
little
bit
last
week
about
kind
of
wanting
to
give
release,
notes
more
of
a
editorial
voice
and
the
authority
to
try
and
make
things
a
little
more
user
facing.
Do.
We
have
any
concrete
plans
for
what
we're
going
to
do
there
or
am
I
hearing
we're
just
gonna
like
the
drafts
is
what
the
release
notes
are
going
to
look
like
so.
F
You
know
that's
kind
of
what
we'll
be
working
through
this
week
is
kind
of
finalizing.
What
that
will
look
like
I
think
we've
got
consensus
on
what
they
look
like.
So
some
of
the
things
we
discussed
and
can
I
decided
on
was
around
major
themes,
not
really
being
part
of
the
release
notes.
Just
for
the
fact.
If
we
look
at
major
themes
from
previous
release,
notes
and
I
think
you
mentioned
this
too-
they
kind
of
just
look
like
a
cig
update.
Rather
than
actual,
you
know,
changes
to
core
kubernetes.
F
The
other
thing
that
we
need
to
be
done
this
week
is
really
just
kind
of
what
that
changelog
looks
like
because,
right
now,
still,
if
you
kind
of
go
in
there
and
look
at
it,
it's
really
just
what's
ever
generated
and
needs
more
of
editorial
theme
to
it.
Rather
than
what's
generated
from
the
release
notes,
tool.
A
A
That
would
be
a
useful,
forcing
function
just
to
make
sure
that
Bach
that
dock
is
actually
getting
populated,
and
it
would
also
be
useful
information
for
this
group
to
know,
as
we
look
especially
hard
at
what's
trying
to
land
right
before
code
freeze
or
not,
if
things
do
seem
to
be
lining
up
with
that,
dock
would
probably
really
want
to
make
sure
they
land
if
they
don't
seem
to
be
lining
up
with
that
dock.
That
may
be
another
piece
of
signal
that
we
should
edge
them
away
from
the
code.
Freeze,
clip
yep.
L
A
M
Sure,
well
not
not
too
much
news
and
with
any
fast-forwards.
Not
the
next
cut
will
be
someone
next
week
for
the
beetle
other
than
that.
I
have
I'm
working
on
a
small
PR,
notifying
the
publishing
pot
people
when
we
cut
a
new
branch
so
that
we
don't
miss
that
in
future,
hopefully,
and
also
I'm
checking
a
bit
on
the
caps.
There
is
a
movement
right
now
on
the
caps
regarding
packaging.
M
F
A
Okay,
so
in
sum,
I
do
not
ascribe
to
the
philosophy
that
chartreuse
is
an
acceptable
stoplight
color,
so
my
color
is
yellow
and
it's
it's
probably
going
to
be
red.
If
I
don't
see
a
substantial
improvement
on,
or
at
least
a
plan
on,
what
we're
gonna
be
doing
with
our
CI
signal
I,
just
like
I'm
really
consider
just
so
many
flakes
and
I
can't
point
to
any
one
single
source
of
failure.
I'm,
really
hoping
we
can
figure
out
what
we're
gonna
do
about
that.
A
A
Also,
we
don't
actually
have
a
release
team
meeting
on
Thursday
get
we
don't
start
the
real
fun
of
daily
burn
down
meetings
until
next
week.
So
my
proposal
was
that
we
consider
whether
or
not
you
know
those
of
you
who've
been
shadows,
whether
or
not
you'd
like
to
try
some
some
other
role
either
the
role
you've
been
shadowing
for
or
something
else
that
we
all
kind
of.
Think
about
that,
and
then
we
do
that
on
this
time
next
week.
A
The
reason
I'm
suggesting
that
is
I'm
cognizant
of
the
fact
that
you
happened
were
of
European
timezone
folks
here,
so
I
would
really
like
to
make
sure
that
he
meet
at
a
time.
That
is
good
for
you.
The
least
team
selection
has,
in
the
past,
also
happened
at
the
cig
release
meeting,
which
happens
in
Tuesdays
at
1:00
p.m.
I.
Don't
feel
like
that's
a
really
acceptable
time
for
European
time
zones,
so
I
just
wanted
to
throw
that
out.
A
A
So
if
that's
the
case,
I
can
speak
to
who
I
would
want
to
nominate
and
have
that
happen
without
me,
but
I
just
wanted
to
throw
it
to
the
audience.
You
know
and
I
guess
asked
him
and
Steven
if
they
have
more
color.
They
would
like
to
add
just
because
I'm
approaching
this
from
this
perspective
of
I
am
about
to
go
heads
down
on
the
114
release
and
I
really
need
other
people
to
take
ownership
of
the
next
release.
Because
that's
not
my
problem,
this
one
is
alright.
H
So
comments
there
I
would
say
the
first
thing
I'd
want
everyone
to
do
is
to
check
out
the
document
that
that
Aaron
wrote
I
mean
that
Aaron
linked
in
the
in
the
notes
around
release
team
selection
notes
that
that
is
a
living
document.
So,
if
you
feel
like
there
are
things
that
are
inaccurate
about
it,
we
should
change
it.
The
the
general
thought
about
the
selection,
starting
at
the
beginning
of
code
freeze,
is
more
of
a
guideline
than
it
is
a
rule.
H
This
is
a
good
time
to
consider
who
is
going
to
succeed
you
once
you
decide
who
is
going
to
succeed
you
they
can
go
and
run
with
really
seam
selection
itself
release
team
role
leads.
All
of
you
should
be
considering
who
you
think
should
be
succeeding.
You
who
have
been
performing
really
great
during
the
release
cycle
and
the
lead
should
feel
free
to
discuss
that
either
on
the
mailing
list
or
privately
I'm
just
to
get
the
ball
rolling
on
this.
This
is
not
a
it's,
not
a
hard
deadline.
H
This
is
just
let's
start
chatting
about
it.
The
the
second
part
of
it
is
the
survey
right.
So
once
we
select
leads,
and
we
have
to
start
looking
at
at
Adam
shadows
for
those
leads.
So
this
cycle,
we
decided
to
kick
out
a
survey.
The
survey
got
lots
of
responses
and
almost
overwhelming
amount
of
responses
for
everyone
who
was
involved
in
in
one
parsing,
the
the
raw
data
and
then
also
reaching
out
individually
to
each
of
the
the
prospective
shadows.
H
K
H
H
While
Tim
is
right
that
they're,
it's
nice
to
have
some
continuity,
it's
also
an
opportunity
to
step
outside
of
the
release,
team
and
move
into
release,
engineering
or
licensing,
or
how
do
I
handle
package
like
package
release
process
like
things
like
that
right
so
start
to
consider
that
as
well
and
and
if
you
feel
you
want
to
talk
to
any
of
us,
definitely
feel
free
to
reach
out.
On
slack
about
that.
A
Something
else
to
keep
in
mind
is
that
historically
there's
it
you
can.
It
can
take
a
little
bit
to
draft
like
the
schedule
for
the
next
release.
So
there's
generally
about
like
a
week-long
pollen
in
between
releases
I
have
personally
found
participating
in
the
release
team
to
be
an
incredibly
rewarding
experience.
That
has
allowed
me
to
see
so
much
of
the
project
and
yeah.
A
Cool,
so
my
proposal
is
I'm,
gonna,
try
and
time
box,
our
usual
Monday
meeting
to
a
burned-down,
so
just
30
minutes,
and
then
we
will
take
through
the
latter
30
minutes
to
run
through
release
team
selection
in
an
effort
to
be
as
friendly
as
possible
to
those
of
us
who
can't
regularly
attend
this
meeting.
You
are
totally
free
to
start
this
discussion
on
asynchronous
mediums,
such
as
email
and
slack,
and
make
sure
that
that
information
finds
its
way
finds
its
way
here.
H
A
Right
I
will
just
as
soon
as
they
stop
freaking
out
that
I
suggested
that
maybe
they
they
step
up
so
I,
don't
know
about
the
rest
of
you,
but
it
appears
every
any
time.
I've
asked
anybody
about
doing
something
with
the
release
it
kind
of
freaks
them
out,
and
they
really
need
to
to
have
a
good
long
think
about
it.
That's
sort
of
the
other
reason
that
I
think
should
it
be
good
for
us
to
spend
the
next
week.
Thinking
about
this.
A
A
Now,
maybe
I'm
speaking
to
the
wrong
group
of
people,
because
this
isn't
the
European
team
meeting.
But
I
was
thrown
out
there
like
when
we
kick
up
to
daily
burn
down
meetings,
I'm
happy
to
say
that
that
European
time
is
like
the
burned
out
meeting,
and
we
can
also
consider
having
the
Thursday
burned
down
meeting,
also
being
at
the
European
time
gives
us
three
like
Pacific
Centrex
times
to
European
centric
times.