►
From YouTube: Kubernetes 1.14 - Week 9 - Burndown 2019-03-06
Description
A
A
B
A
C
So
I'm
still
cheating
and
calling
it
yellow
green.
Instead
of
actually
picking
a
color,
we've
got
33,
I'm
inclined
to
say
really.
We
should
be
at
31,
but
I'll
get
to
that
only
get
to
the
notable
ones
at
risk
of
those
17
are
at
risk
due
to
either
having
missing
test
plans
or
having
open
pr's
inflate.
Most
of
those
17,
though,
don't
have
more
than
two
open
pr's
inflate
and
the
notable
enhancements
at
risk
that
we've
called
out
issue
1
to
7
this
one's
been
at
risk
for
a
very
long
time.
C
It's
one
of
the
ones
I'm
inclined
to
punch
at
this
point
because
there's
been
no
activity
on
the
issue,
we
have
no
kept
that
we
can
find
peers
are
still
open
and
when
I
poked
in
sig
node,
there
was
no
one
commenting
about
any
activity
for
1
to
7.
Does
anyone
think
punching
was
a
bad
idea
at
this
point?
Oh
cool
sounds
like
a
decision
for
that.
One,
no.
C
All
right
and
then
the
next
issue
at
risk
116.
This
is
the
windows
2ga.
This
one
is
at
risk
right
now,
just
because
there
are
a
large
amount
of
PRS
that
are
in
flight
and
when
I
asked
in
the
saved
window
slack
channel,
they
indicated
they
even
anticipate.
Some
of
these
PRS
might
get
merged
post
code
freeze.
So
that's
one
just
to
watch
out
for
if
those
PRS
don't
get
added
I
see,
Michael
Michael
is
on
I
know,
he's
one
of
the
main
people
responding
out.
If
you
have
anything
you
want
to
add,
we.
D
A
D
C
Move
on
issue:
seven
five,
one:
this
one's
been
talked
about
before
being
at
risk,
because
test
plans
are
missing,
but
all
their
PRS
are
merged
and
to
us
tests
are
linked
in
the
github
issue,
but
not
in
the
cup
and
talking
to
Derek
Wayne
at
Derek,
Wayne
Carter
in
hash
tag,
sick
note
he
was
responsive
and
they
do
want
this
issue
in
I
guess.
My
question
is:
is
having
the
end-to-end
test
that
he
also
say
to
the
end-to-end
tests
haven't
changed
since
this
issue
is
being
promoted
to
it's
going
beta
to
GA.
C
A
E
A
A
C
Then
the
last
one
is
seven,
eight
three,
this
one
is
at
risk
again
because
of
missing
cup
plans,
but
when
talking
with
at
Luke's
us
and
at
stts
in
both
github
and
in
slack
they
said
it's
a
the
only
changes
that
were
merged
and
for
114
we're
really
small
in
the
bulk
of
the
work
is
actually
targeting
115.
So
they
agree
that
possibly
punching
this
makes
sense.
A
C
F
F
The
short
description
there
is
that
the
cluster
itself,
the
cluster,
upgrade
rather
stops
Midway
and
seems
to
be
leaking
resources
and
that's
persistent
volumes,
firewall
rules
etc,
etcetera,
causing
a
bunch
of
other
test
cases
to
fail.
Tim
Sinclair
helped
with
triage
last
because
last
night
for
me
last
afternoon
for
West
Coast.
He
seems
to
believe
that
the
problem
is
with
the
tests
that
are
owned
by
Google,
and
this
is
where
I
would
like
a
bit
more
of
help
as
to
who
to
reach
out
to.
F
A
Google
thing
just
makes
me
really
sad,
just
real,
quick
context.
It
used
to
be
Robbie.
Robbie
has
moved
away
from
working
on
this
particular
aspect
of
the
project,
and
I
have
yet
to
actually
find
anybody
who
can
pick
up
and
carry
the
torch.
So
these
set
of
scripts
that
stand
up
the
majority
of
clusters
that
all
of
our
tests
depend
upon.
Tim
st.
Claire
is
adamant.
A
F
Sounds
good
and
then
there
is
another
there's.
A
couple
of
issues
on
failing
jobs
that
seem
to
beyond
by
signaled
is
failing
consistently
and
seems
to
just
fail
to
stand
up
all
the
pods
that
it's
expecting
to
stand
up.
I
reached
out,
okay,
they
surely,
or
today,
and
reached
out
to
signal
I
haven't
heard
back
yet
it
hasn't
been
ages,
but
I've
put
it
as
read,
because
it's
a
consistently
failing
test
in
a
blocking
suite.
F
All
in
all,
I
wanted
to
go
for
like
a
reddish
yellow,
but
as
we
don't
have
these
I'm
going
to
be
decisive
and
say
right
at
the
moment.
Another
thing
that
might
be
interesting
is
that
I've
started
opening
issues
for
the
top
flaky
jobs,
so
that's
rates
over
both
70%
of
flakiness,
a
tiny
like
over
the
last
couple
of
weeks.
F
A
F
F
But
it
is
quite
interesting
so
when
I
was
doing
some
data
digging
yesterday,
it's
quite
interesting
that
tests
themselves
flake
about
I
think
five
to
six
percent
of
the
time,
which
is
still
significant,
but
is
nowhere
near
the
80
percent
of
flake
rate
that
we
see
for
the
job.
So
it's
it's
quite
sad
that
many
kind
of
like
unrelated.
If
you
want
tests,
because
they
don't
all
flake
at
once,
because
a
lot
more
mess,
then
than
they
do
individually.
F
F
So
we
dissing
some
weekly
reports.
It
isn't
so
much
a
list
of
flaky
tests
and
jobs.
We
could
start
sending
that
a
bit
more
explicitly.
There
are
so
in
the
weekly
report
that
we
send.
There
is
now
a
link
to
the
project
board
that
we
look
at
and
the
issues
are
linked
there,
but
I
think
we
could
improve
that
and
make
it
even
more
actionable
in
a
way.
That's
good
idea.
Thank
you.
G
Quick
update
regarding
bug
triage,
so
we
have
started
actively
working
on
this
and
going
to
all
issues
and
PRS,
and
so
far
we
have
figured
most
of
them.
There
are
some
workup
England
to
see
was
the
state
was,
but
they
were
all
too
reactive
and
we
are
going
to
do
that
today
or
tomorrow
and
then
we
will.
Then
we
went
to
Ali's
recipe.
G
Everything
looks
fine
so
far
and
we
can
say
that
we
are
between
yellow
red
green,
considering
the
status.
The
most
of
your
choice
and
PRS
are
related
to
sing
windows,
and
that
is
maybe
the
biggest
features,
but
we
will
continue
monitoring
that
and
see.
Is
it
going
well
and
is
going
to
make
into
1.14
so
far
about
the
matrix?
We
have
some
open
and
closed
issues
compared
to
the
last
meeting
we
had
on
Monday
there,
eighteen
issues,
West
open
and
for
TBR
sliced,
open,
open
to
Monday
and
the
total.
G
A
A
H
Yeah,
so
it's
pretty
good
and
I'm
going
to
follow
every
lead
in
saying
it's:
yellow,
green,
yellow,
because
tide
was
pretty
unhealthy.
Today,
unhealthy
in
the
terms
of
only
user
facing
the
project,
wasn't
able
to
set
the
statuses
correctly,
and
that
was
because
of
the
tub
rate
limit
abuses.
The
testing
of
jobs
was
running
fine,
and
that
was
the
fact
by
that
we
still
were
able
to
merge
71
peers
in
the
last
two
days.
So
that's
pretty
good
right
now.
Tide
is
looking
pretty
healthy,
I
would
say
even
healthier
than
before.
H
A
A
A
H
Yeah
I
would
say
considering
that
code
freezes
tomorrow,
that
type
who's
getting
larger
is
expected.
Historically,
what
we
will
mostly
be
worried
about
is
the
tile
processing.
Time
which
seems
to
have
gone
is
seem
to
have
improved
a
lot,
so
it's
still
picking
up
batches,
but
it
will
take
the
same
amount
of
time
for
testing,
because
the
batches
are
longer.
D
By
the
way,
we're
getting
a
lot
of
reports
about
code
freeze
being
on
Friday,
so
other
folks
are
saying
Thursday
another
Thursday,
it's
Thursday,
that's
what
it
wasn't
the
dogs
are
reaching
on
me.
So,
what's
gonna
happen,
I
remove
experience
this.
We
had
like
six
or
seven
PRS
yesterday
that
were
all
back
that
because
of
the
system.
You
know
some
of
them
failing
and
due
to
the
flakiness
and
having
to
resend
and
through
you
know,
obviously
the
fragility
of
the
system
is
impacting
our
ability
to
to
get
some
PR.
D
A
A
What
you
can
do
to
really
really
help
yourself
out
is
add
the
beat
114
milestone
to
PRS
that
you
want
to
land
in
a
v1
14
release,
because
that's
what
we're
gonna
be
using
as
countries
right
like
if
your
PR
doesn't
have
that
milestone.
It's
not
gonna
merge.
So
if
you
add
it
now,
then
we
can
get
a
good
idea
of
what
you
are
trying
to
merge.
I'm.
A
I
Cool
yeah,
thank
you,
so
is
able
to
go
through
all
the
various
kind
of
stats
attack,
sync
up
with
Claire
and
yourself.
So
we
do
have
some
good
metrics
here.
I
would
say
we're
greenish-yellow,
so
the
way
I
ended
up
doing
it
was
recalculating
how
we're
actually
categories
in
the
docs.
We
can
get
a
clearer
picture
of
exactly
where
we're
at
so
I've,
actually
added
some
functionality
in
the
features
stats
tab
for
the
docs
stats.
I
If
you're
in
the
enhancement
spreadsheet,
you
can
look
at
it,
but
essentially
I,
grouped
together,
the
PRS
that
are
ready
to
merge
and
they're
ready
for
they're,
the
ones
that
are
are
merged
and
the
ones
that
are
ready
for
review
into
the
green
category.
I
consider
those
near
done
very
shortly
going
to
be
done,
Docs
and
then
I
categorized,
and
even
that
we're
in
draft
as
being
yellow
now
draft
could
be
just
an
open
blank
PR.
It
can
also
be
a
fully
completed
doc,
with
some
minor
knits
needed
to
be
adjusted.
I
So
it's
a
little
inaccurate
number
I
figured
to
be
a
play,
safe,
categorize
it
as
yellow
and
then
obviously,
if
it's
light,
it's
gonna
be
red.
So
we're
looking
at
right
now,
roughly
60%
of
the
release
is
green.
35%
is
yellow
and
that's
the
8%
is
late.
We're
going
to
Claire
resolving
some
of
those
in
the
two
pr's
that
are
late
and
the
next
deadline
is
Monday
March
11th
for
having
all
the
PRS
ready
to
review.
So
looking
at
having
roughly
10
come
up
to
that
state.
A
B
Kaitlin,
are
you
sorry
I'm
on
the
double
mute
got
me
again
so
I
found
out
yesterday
we
actually
hadn't
aligned
on
the
two
additional
features
that
we
want
to
highlight.
So
Claire
I
sent
you
an
email
on
that
now
that
I
think
things
are
kind
of
starting
to
shake
out
what
will
make
it
in
what
won't.
So
if
we
could
think
on
that,
maybe
today
I
think
we're
still
on
track
for
our
to
get
this
done
by
the
12th.
But
we
need
to
figure
that
out.
Yes,.
C
C
A
Okay
and
I
know
we
had
originally
talked
about
like
trying
to
walk
through
something
today
or
have
used
something
today,
I'm
gathering,
you
saying:
that's
not
really
ready.
Do
you
think
we
could
allocate
some
time
for
this
on
Friday
just
to
share
with
the
team
sort
of
the
direction
that
we're
headed
yep.
A
Hannes
said
he
wasn't
going
to
be
able
to
make
it
so
in
my
best
robot
voice,
apparently
he's
cutting
beta
to
on
what
is
the
next
Tuesday
and
he's
moving
around
all
of
the
caps
and
that's
it
so
then
it
falls
to
me,
oh
boy,
what
do
I
choose?
What
do
I
choose.
This
is
so
hard
I
feel
like
I'm
gonna,
go,
yellow
I'm,
still
kind
of
a
yellowish
red
in
my
heart,
but
I
did
say
my
criteria
for
not
going
red
was
to
see
that
there
was
actually
a
plan.
A
I
feel
like
all
of
the
hard
work
that
Maria
has
done
to
reach
out
to
the
individual
SIG's
and
make
them
aware
that
they're
flakes
nor
tests
are
blocking.
Release
has
helped
oh
and
Maria's
amazing
team.
Thank
you
all.
So
very
much
so
I'm
still
super
worried
about
CI
signal,
so
I'm
staying
on
top
of
that
and
trying
to
help
escalate
and
raise
visibility.
There
is
kind
of
my
top
priority
right
now.
My
next
priority
is
helping
with
that
risk
items.
My
next
priority
is
based
on
those
two
things.
A
A
Actually
I
did
have
one
other
thing:
I
really
fell
down
on
the
job
of
like
actually
posting
my
daily
code
freeze
meme
yesterday,
so
I'll
have
to
get
one
out
there
today.
If
you
all
have
any
code
freeze
memes,
you
want
me
to
share
with
the
community
during
tomorrow's
community
meeting,
update
I
will
gladly
accept
input
on
the
sig
release
channel.