►
From YouTube: k8s 1.15 Burndown 20190607
Description
See https://github.com/kubernetes/sig-release/tree/master/releases/release-1.15 for more information
A
As
a
reminder
to
everyone,
this
meeting
is
being
recorded
and
will
be
posted
to
a
public
YouTube.
So
please
follow
the
kubernetes
code
of
conduct,
which
basically
boils
down
to
don't
be
a
jerk.
I
will
pass
it
over
to
Kendrick
for
and
enhancements
update.
B
C
Having
a
fantastic
Bailey
I'm
today
see
a
signal:
it's
green,
a
master
block
in
115
blocking
in
no
per
system
failures,
master
blocking,
I,
I
thought
I
told
you
all
that
yesterday
the
way
it
was
around
a
fail,
and
it's
instead
between
yesterday
and
today
they
come
away
some
of
it.
The
pills
were
removed
from
account
the
other
ones
they
they've
been
clearly
clearing
out.
So
it
will
keep
on
say,
will
kept
up
their
windows.
C
It
was
couple
tests
to
see
that
everything
keeps
going
keeps
going
well
and
now
by
but
chemistry
informing
so
must
a
master
informing.
There
is
an
issue
from
six
scalability
in
doing
a
conversation
during
a
conversation
with
with
them
a
was
mentioned,
and
the
failure
rate
that
the
errors
were
related
to
prowl
restarting
pots
too
often,
although
we
had
non-canon
and
take
from
them
a
won't
reach
out
today
see
that's
still
the
case
or
something
that
we
should
worry
about.
C
Then
the
favor,
a
the
favorite
issue
of
everyone
for
the
last
couple
weeks,
Guardian
s,
1,
3
or
1/5,
so
Bucky
a
lot
of
important.
A
lot
of
a
support
on
this
one
seems
to
be
the
simple
life
cycle
in
general:
is
leaning
or
reverting
the
cuba
puts
a
method
or
both
stop
in
a
cluster.
To
use.
Codeine
is
one
which
is
a
right
now,
the
PR
that
is
open
to
do
the
Robert.
He
needs
they're
waiting
for
unknown
or
you
want
to
go
board
to
approve
or
denied,
and
yes,
so,
just
wait.
C
Just
waiting
on
a
problem
from
mine,
actually
from
an
actual
owner
puts
across
from
my
cycling
swimming
towards
their
rebirth,
then
flaking
a
flake
him
a
the
must
is
some
of
the
reboot
test
from
secretly
cycle.
They
failed
again
I
think
earlier
today,
a
we
still
need
to
to
create
a
natural
issue,
and
in
fact
we
tend
to
see
if
there's
anything
that
we
can
do
to
make
that
to
make
the
less
flaky,
but
nothing
nothing
to
worry
about
right
now,
besides
that,
thank
you.
C
Thanks
to
the
feedback
from
Jordan
widget
yesterday,
we
are
working
on
moving
some
of
the
cubic
to
a
test,
a
back
to
back
to
the
sacral
back
to
a
stick
release
dashboard
a
currently.
They
we
are
waiting
to
merge
the
PR,
that's
gonna!
That's
gonna,
bring
you
back
to
the
secretly
stash
for
a
due
to
the
fact
that
we
are
waiting
for
the
actual
fix
to
be
merged
and
the
fix
is
a
cherub.
It's
a
cherry
pick
and
it's
a
I
listed
in
the
update
it
78-77
to,
and
that
is
forward
today's
CI
signal.
A
C
D
A
E
E
I
I
noticed
it's:
it's
been
going
on
so,
okay
welcome
again,
once
again,
everybody
to
the
wonderful
world
of
there's
a
script
called
Cuba
SH
that
doesn't
actually
have
an
owner
and
we
all
keep
calling
it
separated.
But
it's
the
script
that
runs
the
maturity
of
our
CI
and
I,
don't
know
who
actually
uses
it
live
in
the
wild.
E
E
Okay,
so
I
would
just
like
Oh
with
with
the
fact
that
I
do
have
approver
rights
inside
of
the
Koster
directory
I,
guess
that
gives
me
some
responsibility,
even
without
that
I'm
always
going
to
err
on
the
side
of
what
makes
it
easier
for
users.
This
seems
like
upgrade
path,
isn't
quite
clean
enough.
It
seems
like
we
have
a
very,
very
well
exercised
path
with
the
older
version
of
Corti
ns.
Can
somebody
explain
to
me
why
we're
so
pushy
to
move
forward
with
the
latest
version
of
Gordian
s
so.
C
My
understanding
is
that
there
were
some
scalability
tests
that
were
failing
with
way
accordionist
1/3.
Whenever
your
warning
clusters
I
have
5,000
or
more
notes,
they
memory
the
memory
resources
you
need
to
be
increased,
otherwise
Corvinus
is
gonna,
keep
a
crashing
and
that's
gonna.
That's
just
gonna
put
a
heavy
weight
on
the
cluster.
E
Okay,
so
I
feel
like
we
need
input
from
scalability
here
to
make
the
most
effective
decision.
But
it's
also
unclear
to
me
how
you've
been
using
core
DNS
131
for
a
while
I,
don't
think
the
scalability
tests
were
failing
for
a
while.
So
it
could
be
that
there
was
a
regression
introduced
that
caused
poor
DNS
to
be
a
problem,
but
like
I'm,
always
going
to
err
on
the
side
of
let's
roll
back,
but
I
feel
like
what
is
needed
to
make.
The
most
effective
decision
here
is
to
include
scalability
in
the
loop.
C
E
E
E
E
A
A
D
D
A
D
D
Sounds
good
all
right,
my
main
thing
so,
once
those
three
issues
are
resolved,
we
have
a
few.
We
have
about
five
open
issues.
The
other
two
are
minor
and
could
probably
be
just
cherry
picked,
Garin
code
thaw
and
then
the
remaining
open
issues
it
doesn't
seem
like
any
of
them
are
released,
blocking
either
and
settle.
G
A
Nice
was
somebody
trying
to
say
something:
I
wasn't
sure.
I
thought
I
heard
two
voices
at
one
point:
I
was.
E
G
F
H
We
have
the
Google
Doc
up
and
going.
We
have
shared
it
with
everyone.
We
have
alerted
all
the
sig
leads
most
of
the
work.
The
editorial
work
is
going
to
happen
over
the
weekend
like
updating
dependencies
and
actually
going
through.
It
sounds
like
the
themes
meeting
went
well
and
we
will
be
able
to
take
those
themes
and
put
them
into
the
release
notes
as
well
over
the
weekend.
So
pretty
much
autopilot
checking
on
Monday.
A
A
Cool
I
can't
for
that.
Like
third
open
slot
you
mentioned
I,
know
in
the
past,
that's
been
brought
up
that
a
lot
of
the
groups
who
are
doing
things
like
outside
of
KK
might
be
interested
in
like
taking
some
blog
space
to
talk
about
the
cool
things
they're
doing
so,
we
could
possibly
reach
out
to
one
of
the
six
I
think
it's
cloud
provider,
if
I'm
not
mistaken,.
A
Cool
yeah
for
the
interviews,
I,
remember
last
cycle:
it
was
mostly
myself
Aaron
and
then
Natasha
and
then
one
person
from
the
cig.
If
we
had
any
well,
we
had
Windows
people
joining
for
a
lot
of
them,
because
that
was
a
big
one.
I
don't
know
if
we
have
enough
big,
if
we
have
any
big
enhancement,
we'd
like
to
pull
a
rep
from
a
cig
into
specifically
and
then
I
think
lucky
also
mentioned
interest
in
joining
for
a
couple
of
them
and
I.
Think
that's
a
good
idea
just
so
he
gets
exposure.
F
Muted
again
so
I
was
saying
released
candidate
one
would
be
released
on
June
11,
which
is
the
Tuesday.
As
for
the
thing
I
reported
about
yesterday,
it
was
mostly
cuz.
The
bucket
the
kubernetes
gcd
bucket
was
not
showing
on
the
consoles,
but
if
you
went
to
the
URL
by
just
changing
you
know
the
Nantucket
name,
you
would
go
to
it
and
the
artifacts
are
over
there,
so
we're
wearing
the
green
okay,
okay
cool
other
than
that
yeah
Michael.
Very
welcome
to
do
a
good
fight.
That's
it
for
me!.
I
I
I
A
Yes,
strong
plus
one
we're
about
t
minus
I,
think
six
working
days
at
this
point
until
our
targeted
release
date,
so
we're
getting
really
close
to
the
end.
So,
hopefully,
everyone
has
some
sort
of
plan
in
mind
for
their
secession
sense.
We
want
to
make
sure
that
Waukee
gets
to
go
into
his
release
with
a
fully
staffed
team.
A
So
and
I
know
that
in
a
few
moments,
we're
going
to
talk
about
changes
to
test
in
for
us,
so
I
understand.
I
do
well
why
you
might
have
been
holding
off
for
why.
Anyone
who
you'd
want
to
coordinate
with
around
that
might
be
holding
off
since
I
know
we're
merging,
there's
or
there's
an
open
PR
to
merge
that
role
with
other
roles,
but
yeah
try
to
get
your
secession
plans
finalized,
very,
very
soon
cool
and
then
a
Josh
for
the
shadows.
Survey
for
116.
I
A
A
A
A
Okay,
I
feel
like
I'd,
like
someone
who
wasn't
part
of
the
release.
Just
so
they
have
like
the
third
party
thing,
I'm,
assuming
possibly
just
asking
and
said
contra
bucks
I
should
be
able
to
find
someone
who's
able
to
help.
Yes,.
E
A
So
that
is
my
action
item.
To
get
us
a
facilitator,
then
secession
plans
as
Josh
should
just
mention
but
like
want
to
make
an
additional
plug,
and
then
our
next
upcoming
milestones
on
June
10th.
We
have
Doc's
last
call
on
June.
11Th
is
what,
when
we're
planning
to
do
code
saw
June
11th
is
then
the
115
are
see
as
well.
June
13th
is
the
cherry-pick
deadline,
and
then
our
target
date
is
June.
17Th.
E
A
J
This
is
Vivek,
I
was
shadow
for
branch
management
with
Cheryl,
and
if
those
two
roles
are
getting
merged,
we
might
need
an
overview
on
tests
infra
as
well.
If
both
the
rules,
if
we
will
have
to
shadow
both
the
roles
together
and
yeah
I
mean
this
was
the
major
thing.
Otherwise,
I
lied
something
too
directly
as
well.
E
E
Nobody
who
has
staffed
the
tested
for
role
has
dry
run
the
changes
either.
So
we're
understanding
that
mistakes
and
we're
understanding
of
doing
it
live,
and
so
I
mean
that's.
That's
my
understanding
of
it
and
I
don't
mean
to
be
the
dude
who's.
Speaking
with
the
voice
down
from
on
high
I,
just
feel
like
looking
at
the
docs,
it's
kind
of
unclear
to
me
exactly
why
this
role
exists.
E
F
Could
I
get
some
more
context
and
what
you
do
that
is
probably
asking
so
I
brought
up
saying
about
you,
know,
shadows
not
having
enough
commissions
and,
and
only
when
they
become
a
leave
they're
actually
going
through.
You
know
the
excellent
practical
work
of
it.
So
I
think
what
these
I
was
trying
to
say
was
that
we
want
commissions
for
shadows
and
length
cycles
and
that
will
be
brought
up
in
the
retrospective.
F
1.15
so
I
think
that's
a
portion
of
what
the
event
was
trying
to
ask
for
bring
up
bring
up
about.
Is
that
right?
We
like
is
that
what
you
made?
Okay,
yes,
so
that
will
be
brought
up
in
the
retrospective
and
right
now
we're
trying
to
take
apart
those
responsibilities,
really
test
infra
point
and
put
it
on
to
branch
management.
E
G
Think
we
used
to
have
a
casks
of
tests
and
Ron
called
in
the
test
and
first
job
that
is
no
longer
part
of
the
handbook,
so
that
removed
from
the
role
of
testing
for
veto
not
require
anything
else
that
the
test
and
for
folks
should
do
specifically
so
I.
Don't
see
that
as
a
requirement
anymore.
It's
just
I
think
what
what
as
I
have
also
been
a
shadow
under
testing
from
before
and
what
did
what?
What?
G
The
experience
is
that,
as
a
shadow,
you
kind
of
are
trying
to
understand
what
is
happening
in
the
space
station
from
Earth,
because
you
don't
get
to
fit
in
the
space
station.
You
don't
have
a
mock
of
space
station,
so
you
don't
understand
what
the
job
is.
So
you
don't
have
enough
confidence
to
graduate,
but
that
discussion
is
not
relevant
actually
with
respect
to
whether
we
should
split
and
merge
the
jobs
of
tested
fry
into
other
roles.
That
is
with
respect
to
that
concern.
E
Maybe
this
is
a
special
case,
but
I
think
a
lot
of
the
work.
That's
been
done
to
sort
of
automate
this,
the
the
toil
away
has
been
done
by
Kathryn,
who
is
your
shadow,
and
so
what
you've
been
doing
as
a
lead
is
sort
of
letting
a
shadow
do
the
work
and
then
providing
the
oversights
make
sure
that,
yes,
the
work
is
being
done
in
an
appropriate
manner
and
maybe
that's
a
good
model
for
others
to
think
about
in
terms
of
delegating
responsibilities.
E
E
E
I
I
wanted
to
mention
one
detail
that
I
already
commented
on
the
proposal,
but
nobody
responded
to
which
is.
There
was
a
weird
proposal
that
reporting
on
PRS
would
become
responsibility
of
the
CIA
signal
and
I.
Don't
really
understand
that
I
don't
understand
how
it
became
part
of
testing
fur
to
begin
with.
Okay,
the
last
time
I
was
bug,
triage
or
CI.
Signal
reporting
in
PRS
was
the
responsibility
of
bug
triage,
wouldn't.
E
It
be
great
if
the
report
was
like
an
auto-generated
thing
that
we
could
all
look
at
instead
of
humans
having
to
read
out
numbers
at
the
meetings
randomly
I'm.
Sorry,
y'all
I
started
this
when
I,
let
them
release
and
I
asked
people
to
like.
Could
they
give
a
metric
that
represented
from
their
roles
perspective
what
they
thought?
The
health
of
the
release
was
I
feel
like
that's,
cargo
coltan
now
to
people
like
have
to
read
out
numbers
and
like
maybe
that's
not
working,
maybe
that's
not
the
best
use
of
everybody's
time.
E
I
E
G
One
after
I
mean
I
think
I
commented
Josh.
Do
you
think
that
we
we
started
reporting
this
number
as
a
way
of
explaining
tight
and
browse
health?
If
you
have
a
huge
backlog
in
much
food,
then
your
number
of
PR
Smerch
would
be
less
in
that
last
week.
So
I
think
that
was
one
of
the
reasons
I
got
also
from
coal.
When
we
started
this
so
yeah,
maybe
we
should
look
at
from
that
perspective.
It
gives
up
metric
yeah.
Yes,.
E
Holistically
I
would
just
I
think
it's
about
when
things
went
a
little
pear-shaped
on
Thursday
and
Friday
of
whatever
week,
it
was
now
sorry.
Time
is
a
little
kingless
to
me
at
these
points.
At
this
point
like
when
you
know
we
had
fun
with
prowl
and
then
fun
with
github
API
like
it
was
useful
to
have
some
representation
of
that
some
somebody
speaking
to
that
in
the
release
team
meeting,
so
that
we
were
aware
it
was
happening
I.
E
Don't
necessarily
think
that
means
that
a
human
has
to
say
things
every
meeting,
that's
all,
and
so
to
me.
That
would
be
clear
and
effective
communication
from
the
testing
for
on-call,
letting
the
community
ISM
as
a
whole
know
that
something
is
up
that
ensures
that
the
release
team
is
aware
that
something
is
up
and
doesn't
require
a
specialized
role.
Just
for
the
release
team
to
know
that
something
is
a
that.
K
L
In
general,
we
do
email
v's
to
kubernetes,
dev
and
I
think
we
did
when
this
happened
last
week
four
week
before
I've
also
forgotten
I,
don't
know
if
anyone
actually
reads:
kubernetes
dev,
because
most
of
emails
I
send
to
it
seem
to
not
do
anything,
but
that
is
a
channel
we
try
to
use.
For
this
reason.
E
Try
to
be
pretty
clear
about
it,
so
we
we
always
have
room
to
improve
on
our
communication.
I
would
welcome
constructive
suggestions
in
the
form
of
tooling
that
allows
us
to
do
this
in
an
easier
manner,
because
it's
kind
of
difficult
for
us
to
both
keep
the
system
running,
prevent
it
from
exploding
and
also
build
tooling,
to
facilitate
our
communication.
E
I
aspire
to
grow
the
pool
of
people
who
can
help
us
out
with
one
side
of
that
equation
by
moving
the
infrastructure
over
to
someplace,
that
more
of
the
community
can
support,
but
like
tooling,
to
help
us.
You
know,
communicate
status,
report
status
and
whatnot
if
somebody's
got
one
of
those
they
can
pull
down
off
the
shelf
and
show
how
to
use
in
a
super
low
friction
grade
manner.
That
sounds
cool
in
the
meantime,
it
seems
like
our
stances.
E
To
me,
largely
the
the
other
like
Wiggly
thing,
that
testing
from
maybe
sort
of
used
to
know
about
was
like
all
of
the
jobs
that
get
cut
for
branches
and
why
those
jobs
exist.
I
feel
like
that's
now,
starting
to
be
kind
of
questionable,
and
that
CI
signal
probably
has
better
a
better
feeling
for
what
the
jobs
are
and
why
the
jobs
are
and
why
they
exist
and
I
continue
to
push
on
trying
to
like
better
document
that
as
well
for
other
humans
and
I.
Think
between
that
and
the
tooling
the
caffeine
is
created.
E
K
A
Give
him
the
final
final
say
of
what
to
do
for
1:16.
Does
that
sound
fair
to
everyone
on
this
call
I
guess
that
puts
like
a
another
delay
in
this
decision
until
Monday,
but
does
everyone
think
it's
fair
to
give
lucky
the
final
say
since
it'll
be
his
release
team
in
theory,
I
say
in
theory
like
it's
not
mostly
finalized
at
this
point,
but.
K
A
Yes,
and
if,
if
anyone
wants
feedback
on
communication
channels
around
announcements
of
this
info,
you
can
reach
out
to
me
offline
and
I'm,
happy
to
happy
to
give
what
was
helpful
for
me
in
that
state
as
a
release
lead
finding
out
about
the
proud
failures,
real
close
to
code
freeze,
but
we
don't
need
to
discuss
it
with
everyone
on
call
cool.
So
we'll
pass
this
info
over
to
locky,
we'll
hear
back
on
Monday
anything
else.
Anyone
wants
to
discuss
today.