►
From YouTube: Kubernetes Community Meeting 20170601
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
Demo: Exec into pod; Release 1.7, 1.6 updates; SIG Testing, SIG UI;
C
All
right
welcome
everybody
to
the
weekly
kubernetes
community
meeting.
We
have
these
meetings
every
Thursday
and
you
should
follow
along
on
YouTube
I
just
subscribe
to
it,
so
you
can
follow
along.
What's
going
on,
we
have
a
fact
agenda.
Today
we
have
a
demo
from
Joe,
Nolan,
art,
I,
hope,
I
got
that
right
and
then
some
release
updates
and
then
sig
updates
from
sig
testing
and
sig
you
I
with
Dan
and
then
some
announcements.
D
And
take
it
away:
okay,
so
in
cui
a
good
couple
of
months
are
going
like
November
Ian
Lewis
started
working
on
on
products,
tech
feature
which
we
should
bring
some
capability
for
us
to
to
get
a
shell
on
the
dashboard
interface
and
I
picked
up
his
PR
and
a
little
bit
a
bit
further.
So
now
now
it
actually
works,
and
hopefully
it
will
be
over
soon.
So
I
will
demo
that
if
everything
works,
which
of
course
is
always
a
question,
so
I
will
share
my
screen
in
a.
D
So
just
a
quick
confirmation:
can
you
see
my
micro?
Yes
often
so
just
go
to
your
post
list?
Select
your
favorite
pause
and
then
you
have
this
option
execute
shell,
and
if
everything
works,
you
should
get
a
nice
terminal.
Now
this
is
H
term.
We
can,
you
know,
type
everything
whatever
you
want.
You
have
copy/paste
what
you
would
expect
from
his
term.
It's
colorful
it
can
resize
correctly.
D
D
Let's
see
what
else
now
previously,
the
problem
with
this
with
this
PR
I
think
the
main
problem
was
that
that
it
was
only
working
through
WebSockets,
which
meant
that
it
didn't
work
through
cube,
CT,
a
proxy
which
is
a
major
use
case
for
for
the
new
users,
and
now
this
demo
is
also
not
using
few
sepia
proxy,
but
I
have
one
running
and
it
works
to
that
as
well.
So
this
one
is
through
the
proxy.
D
Now
what
you
are
going
to
experience,
that
it
is
a
little
bit
dirtier
as-
and
you
tripod
of
course,
is
difficult
for
you
to
perceive,
but
but
I
can
feel
that
there
is
maybe
an
extra
hundred
millisecond
a
for
every
T
press,
not
intolerable,
but
a
bit
annoying.
That's
because,
instead
of
WebSockets,
if
you
are
using
cube
CT
a
proxy
we
are,
we
are
going
through
xhr
and
the
stream
and
that's
just
not
as
naughty
as
as
cute
as
'm
as
the
WebSockets
directly.
Now.
D
D
So
now,
if
from
this
shall
I
say
curl
the
llamo
Phi
for
this
particular
deployment,
then
we
should
see
in
one
of
them
yep
life
that
there
was
a
get
so
same
as
you
would
expect
from
cube
CTL
attach
now
we
can't
yet
send
signals
and
and
stuff
that
I
would
also
like
to
do.
But
what
that's
also
coming
another
interesting
thing
that
I
played
this.
D
So
from
inside
from
inside
this
container
I
can
and
can
of
course
expand
access
to
the
cluster
itself,
either
with
you
CTL
and
configure
that
one,
and
if
anyone
wants
to
try
this
at
the
moment,
this
is
not
yet
in
the
in
the
any
of
the
dashboard
releases.
But
I
built
a
dashboard
image.
That's
here-
and
this
is
this-
is
the
URL
for
the
for
the
llamo
file.
E
D
D
D
So
one
of
the
features
that
is
is
missing
at
the
point
is
that
we
are
simply
trying
to
start
a
batch.
If
there
is
no
batch,
we
thrive
it
with
SH
and
if
there
is
no
accident
at
the
end
of
it,
you
are
not
going
to
get
a
shell.
It's
not
yet
possible
for
you
to
provide
your
own
command
so,
like
I
think
it
would
be
nice
if
there
was
a
way
to
to
provide
that
you
want
team
access,
but
it's
not
possible
at
the
moment.
D
C
All
right,
thank
you
very
much.
Thank
you
all
right.
Moving
on
to
the
release
updates,
we've
got
one
four
one:
seven
one
six,
five
and
maybe
1.5
who's
here
for
1.7.
A
I
forget
here
so
goes
away,
is
the
are
called
50,
so
we,
the
pending
p.m.
it's,
going
to
merge,
I
fix
ACOG
and
Texas
Tech
typical
I'm
60
M
has
a
VidCon,
and
then
we
are
going
to
do
status
of
make
you
all
those
kind
of
things
and
the
there's
no
change
and
I
planned
so
far
and
and
I
saw
some
other
things
like
some
special
extra
edge
magazine
happen,
so
that
can
comic-con
and
so
yesterday,
advance
accommodate.
A
A
So
though,
so
now,
suddenly
QE
is
working
as
expected
by
professional
due
to
some
really
freaky
build
the
top
two
and
list
here
it
is
the
the
cops
a
doctor
is
and
effects
justin
and
the
bacchanals
and
things,
and
they
she
keeps
excuse
working
on
those
things
and
another
one
is
the
Federation
into
each
is
baby
and
into
now
look
the
figure
by
dilution
group
and
he's
never
seen.
Those
kind
of
things-
and
it
is
so
nice
because
I,
don't
have
a
face.
I,
don't
want
to
answer.
A
Why
actually
surprised,
but
to
me
it
is
the
note
I
believe
and
the
specs
is
just
the
nobody
to
impact
against
the
five
different
image.
So
there's
a
to
image
and
sort
of
looked
out.
The
problem
when
it
is
very
old,
see
all
the
to
be
deprecated,
come
in
and
wear
an
image
and
I'll
escort
you
so
so
by
the
rest.
Let's
us
to
arrest
the
street
in
which
actually
will
find
us,
Slovenia
and
so
well.
A
A
The
convective
losses
types
of
the
main
problems
so
because
the
peoc
will
emerge
because
it's
because,
because
the
flip
he
has
in
a
sudden
acute
flow,
another
one
is
real
relief
block
issue
and
caused
by
recent
merge,
so
merge
the
PR
for
the
debugging
and
the
pedestrian
release
broker
for
the
network.
So
so
so
I
just
merge
the
the
PR
to
revoke
that
circle
of
logging,
PR
and
metal.
The
pin
to
be
able
to
do
those
kind
of
things
and
s.
A
You
have
the
region
across
the
network,
genetic
issues,
so
they
have
been
located,
but
we
are
going
to
asset
me
able
to
kick
out
to
the
fact
s
0
and
the
wave
some
information,
only
issues
with
those
things
and
so
last
things.
What
I
want
I
think
I
send
email
to
the
community
to
the
lettuce
time
about
the
apps
actually
policy,
so
so
we
teach
which
it
is
effect
our
upgrade
the
past
automated
Africa.
So
we
merged
up,
ask
that
one.
A
So
we
have
the
in
the
documents
and
we're
going
to
clean
up
our
document
and
remove
those
and
figures
statement
about
outbreak,
but
I
think
I'm
going
to
be
the
cross
to
ocean
upgrade
path
like
the
previous
many
needs,
because
we
don't
want
to
just
do
the
dramatic
change
because
we
may
have
the
complemental
note
as
well.
It's
all
support
documents
problem
before
so
we
are,
we
need
to
understand,
what's
going
on
and
then
in
this
unit
at
least,
we
have
the
most
clear
things
why
you
need
a
refworks
and
get
the
customer
one.
A
G
I
want
to
point
out
about
the
submit
queue.
Is
that
there's
currently
no
due
date
associated
with
the
1.7
milestone
in
github,
which
means
that
Co
requests
that
are
in
the
submit
queue
for
1.7
do
not
take
priority
over
other
pull
requests
for
no
milestone
or
other
milestones.
So
we
may
want
to
assign
a
due
date
if
we
want
the
one
seven
PRS
to
get
higher
priority,
we.
G
A
E
Thanks,
thank
you
and
I
hold
up
the
last
week.
Last
week's
notes
as
well
just
to
mention
again
at
the
exception
processes,
all
deadlines
for
exceptions
are
June
fish
pick
alright,
and
the
decisions
for
exceptions
are
going
to
be
made
at
burndown
meeting
on
June
7
I
think
it
wouldn't
have
any
other
questions.
I
Yeah
I
just
wanted
to
offer
to
help
Donna
regarding
the
Federation
tests
causing
problems
I,
just
a
look
at
the
dashboard
I
see
they
were
all
stable
until
yesterday
and
then
something
happened.
Last
night
we
have
an
on-call
engineer
who
is
responsible
for
keeping
that
clean
and
unfortunately
they're
in
India
at
the
moment
and
that's
closing
in
on
midnight.
So
if
you
want
to,
if
you
mean
your
system,
Stream
courses
today,
either
myself
or
Nikhil
or
mother,
you
should
be
able
to
help
you.
G
G
G
B
K
L
G
I'm
just
trying
to
clarify
not
just
for
me
right
but
for
people
who
are
trying
to
get
in
like
bug
fixes,
as
we
start
burning
down
we're
going
to
be
changing
this
to
make
you
so
it
only
accepts
PRS
that
have
to
be
1/7
milestone
attacks.
So
you
should
go
we're
just
because
I
guess
we're
burning
down.
We
should
know
who
to
go
mag
to
get
about
Stan.
That's
right!
Yes,.
E
L
M
C
G
A
M
H
Me
Anthony,
oh
so,
for
one
stage
we
do
have
a
pretty
big
backlog
of
cherry
picks
that
didn't
make
it
into
one
to
explore,
because
I
was
a
hot
pick.
So
I
want
to
I
do
wandering
through
these
165
they're
a
spot
on
thinking
for
next
week,
but
I
still
need
to
track
down.
There
are
blocking
tests
that
are
broken.
So
if
you
submitted
a
traffic
on
one
six
branch,
maybe
take
a
look
at
the
test
grid,
because
if
your
PR
look
suspicious
I'm
coming
for
you.
B
C
G
G
Hi,
okay,
I'm
Aaron
curtain,
burger
I'm,
signed
in
as
state
testing,
because
I
now
have
a
cig
testing
zoom
account.
Thank
you
so
much
Cameron
and
Sarah
for
putting
us
together.
I
have
this
zoom
account
so
that
our
meetings
can
potentially
run
longer
than
30
minutes.
We
hold
them
weekly
Tuesdays
at
1:00
p.m.
Pacific
time,
4:00
p.m.
G
G
G
That's
used
to
drive
the
variety
it
has
dashboards
that
we
have
I'm
not
super
familiar
with
this,
but
there's
a
directory
called
kettle
that
basically
is
in
charge
of
scraping
and
extracting
all
the
test
results
out
of
GCS
and
turning
those
into
their
local
Superlite
database
for
JSON
files
that
you
can
use
for
bigquery
and
then,
as
you
may
remember,
from
my
update
on
flakiness
a
couple
weeks
ago,
Eric's
put
together
a
couple
of
bigquery
things
that
are
run
periodically
to
come
up
with.
Excuse
me
a
variety
of
metrics
about
jobs.
G
So,
for
example,
if
I
want
to
see
what
were
the
latest
jobs
this
week,
I
can
click
on
this
JSON
file.
I
can
see
that
yeah
that
sed
three
full
request
job
was
definitely
super
flaky.
It
was
the
second
least
consistent
job
or
whatever
reason,
our
unit
tests
or
even
less
consistent
than
that.
This
past
week,.
G
One
of
the
coolest
things
since
the
last
update
I
gave
is
that
the
triage
board,
which
again
just
sort
of
tries
to
cluster
together
groups
of
failures
now
in
FERS
different
six
for
some
of
these
clusters
of
failures,
I
wish
I,
could
tell
you
what
code
is
actually
used
to
infer
these
six,
but
I'm
not
actually
super
familiar
with
that.
So,
if
anybody's
really
curious
about
that,
we
can
come
have
a
conversation
about
that
in
the
state
testing
sock
channel
in
a
little
bit.
The
idea
is
these
are
not
supposed
to
be
perfect.
G
G
So
sorry,
jumping
around
a
bunch
in
the
jobs
directory
has
tested
for
their
whole
bunch
of
files,
I'm
interested
in
looking
fake
JSON,
because
config
dot
JSON
now
has
a
field
called
sig
owners
which
we
can
start
parsing
out
and
using
this
actually
more
effectively
triage
like
which
SIG's
are
associated
with
which
jobs
we
could
start
using
this
to
build
six
specific
dashboards.
So
they
can
look
at
all
of
the
jobs
that
they
care
about.
G
In
the
meantime,
I
believe
six
CLI
has
a
demonstration
of
how
you
can
create
a
dashboard
that
only
your
signatures
about,
but
this
is
a
more
like
specific
thing
to
creating
a
yellow
file,
I
think
inside
of
the
test,
Grid
directory
and
seeing
all
the
jobs
that
this
specific
state
cares
about.
But
if
we
do
a
better
job
of
and
we'll
probably
discuss
this
at
the
leadership
summits
of
actually
getting
States
to
associate
themselves
with
all
of
these
different
jobs,
we
can
auto
create
a
lot
of
these
things.
G
Velodrome
I
just
wanted
to
call
out
as
an
alternative
view
of
showing
well
maybe
I
can't
call
that
out
that
doesn't
seem
to
be
working
right
now.
Maybe
I
can
do
submit
queue.
I
wanted
to
show
it
as
an
alternative
view
of
the
submit
queue
you
can
tell
we're
coming
up
to
code
freeze,
because
the
lines
that
were
agreeing
a
couple
days
ago
has
started
turning
red,
so
we're
all
in
for
a
fun
ride,
or
maybe
things
are
going
to
be
super
smooth,
I,
don't
know,
but
I
hesitate.
G
G
G
We
have
munchkin
hub,
which
sort
of
pulls
github
for
things,
and
then
we
have
prowl,
which
has
the
best
push
to
it
from
github,
so
Munsch
github
recently
just
had
a
feature
added
to
it,
where
it
swept
through
all
of
the
issues
that
did
not
have
any
sig
labels
associated
with
them,
and
how
did
they
need
cig
label
somewhere
here?
It
should
have
commented
why
it
had
a
bit
needs
sake
label
in
this
case
I.
Guess
it
didn't
here.
G
But
if
you
are
watcher
on
a
number
of
issues,
you
probably
notice
yesterday
around
3:30
Pacific
that
the
Box
started
asking
people
to
triage
issues
appropriately.
The
idea
here
is:
we
anticipate
that
adding
sig
labels
to
issues
will
encourage
six
to
actually
do
something
about
the
issues
associated.
So
this
probably
since
its
tagged
with
area,
API
and
API
server
might
be
associated
with
the
API
missionary
sig
or
something
to
that
effect.
So
what
are
all
the
things
that
much
github
can
do?
It
uses
things
called
lenders,
so
I
apologize?
This
isn't
like.
G
We
don't
have
a
tremendous
amount
of
documentation
on
all
this,
but
for
developers
who
are
capable
of
reading
code,
each
of
these
files
corresponds
to
a
single
munzur
that
can
do
something
for
every
issue
or
every
pull
request
in
a
given
repo.
What
does
the
lender
actually
do?
Well,
let's
take
a
look
at
an
example:
config
map,
that's
used
for
the
submit
queue
that
is
deployed
today
for
kubernetes,
and
you
can
see
that
there
are
a
variety
of
configuration
options
that
are
passed
through
to
each
of
the
munchers
again.
G
I
would
have
to
encourage
you
to
reach,
read
the
code
to
understand
what
all
these
do,
but
some
of
them
are
a
little
self-explanatory
like
these
are
all
of
the
jobs
that
must
pass.
You
know
when
you
go
to
your
pull
requests
and
see
the
list
of
required
jobs.
In
that
context.
These
are
all
the
jobs
that
have
to
pass.
If
you
want
to
see
which
particular
Munger's
this
this
deployment
is
using
against
the
curators
repo
you
can
see.
The
list
here
is
comma
separated
order
matters.
That's
why
blunderbuss
comes
first.
G
Okay
so
proud
same
thing:
what
can
it
do?
Prow?
Basically,
instead
of
having
munchers,
has
plugins
in
the
plug-in
directory.
Each
of
these
directories
corresponds
to
a
plug-in
that
can
accomplish
something
so,
for
example,
the
work
that
Garrett's
been
doing,
it's
probably
in
the
label
plugin,
which
can
respond
to
certain
commands
and
then
add
either
area,
priority
kind
or
sig
labels
right.
G
How
can
I
determine
the
plug-in
that
everybody
probably
most
cares
about?
Aside
from
the
labeling
plug-in,
is
the
trigger
plugin,
because
this
is
the
one
actually
given
a
github
event.
Those
often
creates
a
job
which
ultimately
gets
turned
into
a
pod
which
gets
run
on
a
kubernetes
cluster
or
the
job
ends
up
getting
ends
up
corresponding
to
something
that
goes
and
triggers
of
Jenkins
job.
G
So
what
does
prowl
actually
do?
I
can
take
a
look
at
the
plugins
file
to
see
which
plugins
are
activated
for
which
repos
or
which
organizations.
So
you
can
see,
for
example,
the
the
thing
that
lets
me
do.
/Lg
TM
is
active
for
any
single
repository
in
the
kubernetes
organization,
as
well
as
the
plugin
that's
responsible
for
looking
for
whether
or
not
somebody
is
C&C.
Fc
only
approved
the
last
people
to
assign
issues
you
can
also
you'll
notice.
G
The
trigger
plug-in
is
not
turned
on
for
all
repos
in
the
kubernetes
org,
but
it's
enabled
specifically
for
certain
treatments,
the
other
plug
the
other
configuration
file
of
interest.
Is
this
config
dot
memo
file?
This
is
used
principally
by
the
trigger
plugin,
and
this
is
the
thing
that
defines
for
a
given
repo.
What
are
the
jobs
that
I
care
about
for
that
repo,
as
well
as
a
variety
of
options,
to
configure
that
job
things
that
folks
might
be
interested
in
are
what
are
the
commands?
G
I
can
use
to
comment
in
a
given
github
issue
to
ask
the
bot
to
rerun
just
the
specific
job
miner
point
of
interest,
the
things
here
that
look
pretty
small
enough,
corresponding
to
a
Jenkins
job
somewhere
things
that
are
a
little
bit
larger.
This
looks
a
little
bit
more
like
a
pod
stack
right.
This
actually
ends
of
corresponding
to
a
pod
that
gets
run
on
a
kubernetes
cluster.
This
is
a
symbol
of
us
really
trying
to
move
away
from
Jenkins
and
more
towards
kubernetes
native
test.
G
Finally,
both
prow
and
lunge
github
responds
to
commands.
I
was
alluding
to
like
the
label
command
earlier.
This
is
the
webpage
you
can
go
to
that.
We
keep
up-to-date
with
all
the
latest
and
greatest
commands
that
both
of
the
bots
support,
so
you
can
see
the
assigned
command
is
handled
by
prowl,
while
the
approved
command
is
handled
by
munch
github
for
historical
reasons,
that
I
can't
quite
explain
just
yet
another
way
of
getting
to
these
is
on
any
arbitrary
pull
request
where
a
bot
tells
you
to
do
something
that
gives
you
some
information.
G
Yeah
you'll
generally
notice
that
there
is
a
detailed
section
here.
You
can
click
that
and
you
can
see
the
bottles
you
excuse
me,
but
the
commands
are
available.
Sorry
that
the
command
I
understand
the
commands
that
are
listed
here.
You
can
click
that
that
will
take
you
to
the
same
page.
It'll
also
tell
you
of
generally
speaking
what
the
flow
for
a
pull
request
is.
If
you
have
a
question
about
why
your
PR
is
stuck,
there
was
recently
some
work
done
on
unifying
a
number
of
pull
request.
Docks
into
this
single
dock
and
I.
G
G
Let's
see,
I
think
that's.
Basically,
all
I
had
so
I
apologize
for
the
super
disorganized
whirlwind
nature
of
all
of
that
Eric
beta
co-organizer,
cig
testing,
we'll
be
giving
a
presentation
at
stick
testing
next
week
to
governance,
Tuesday
at
1
p.m.
Pacific,
where
they'll
be
walking
through
what
is
prowl
with
actual
diagrams
and
boxes
and
arrows
and
then
further
describing
once
prowl
actually
kicks
off
a
pod.
G
What
happens
inside
of
that
pod
to
spin
up
a
kubernetes
cluster
run
tests
against
it,
so
on
and
so
forth,
like
I
was
alluding
earlier,
we
really
care
deeply
about
getting
6
donors
turning
saved
into
members
of
jobs
so
that
we
can
start
getting
some
more
accountability
when
Brandon
Jobs
failed
and
we
hope
to
start
tying
back
to
an
automated
SLA
for
flakiness.
You
may
have
noticed
that
Eric
did
not
send
out
his
carefully
slow-roasted
and
artistically
crafted
reports
on
the
greatest
hits
for
flakes
and
test
failures.
G
Last
week,
and
you
know
to
me,
it
was
a
something
I
really
missed.
I'd,
really
like
to
see
us
automate
that,
in
the
future,
after
we've
had
a
chance
to
look
back
and
see
whether
that's
having
effects
on
steering
failures
down.
That's
all
I've
got
I,
see
a
couple
things
in
the
chat
and
stop
sharing
any
questions.
N
Right
so
updates
in
cig
UI
we've
got
a
2017
roadmap
established,
finally,
which
is
exciting
to
have
dan
pike
thanks
for
helping
out
with
that
process
feel
free
to
check
that
out.
That's
mainly
what
we're
going
to
be
working
on
the
next
few
quarters,
one
six
one
is
our
most
recent
release
and
in
that
we
incorporated
global
search,
so
I
can
look
for
a
variety
of
different
resources.
Up
and
I
might
have
actually
shut
down.
N
My
mini
cube
instance
all
right
so
well,
my
global
search
is
running
I'm,
going
to
grab
a
screenshot
here
now
we
have
filtering
per
card
per
resource
as
well
as
precisely
sorting
as
well
as
filtering
per
resource
and
we've
aligned
our
navigation
to
the
API
Docs.
So
a
couple
slight
changes
in
ordering
now
things
are
alphabetical
and
going
forward
and
try
to
be
consistent,
API
Docs,
which
I
think
will
help
clarify
things
for
people.
N
Looking
ahead
to
one
seven
as
jános,
very
generously
demoed,
we
should
have,
except
back
in
the
pod
I
didn't
realize
code
freeze,
post
today,
I'm,
not
sure
that's
going
to
affect
that
or
not,
but
that's
the
hope
is
to
get
the
exact
into
pod
functionality
into
one
seven.
We
have
some
bug
fixes
around
issues
that
arose
in
one
six,
one
with
access
control
and
that
kind
of
breaking
pages.
If
you
didn't
have
full
access
control
then
and
tried
to
view
one
resource,
then
the
whole
page
would
crash
and
then
any
breaking
changes.
N
The
comp
in
the
main
kubernetes
project
we're
also
working
in
role-based
access
control,
that's
in
progress
but
isn't
functional
yet
and
we're
putting
together
integrations
page.
So
we
have
a
place
in
dashboard
to
put
things
like:
we've
scope
or
third-party
resources
kind
of
in
their
own
section
on
dashboard,
and
that
is
most
of
my
updates.
Does
anyone
have
any
questions.
E
L
N
O
Working
then
pitches,
so
I
said
it
again.
I
would
believe
scope
stopped
working,
that's
just
through
an
iframe
okay.
So
there's
like
there's
specific
code
going
into
the
open
source
UI
for
weave
scope,
yes,
I,
believe
so
all
right,
yeah
I
mean
it
seems
like
we're
going
down
the
same
path
that
we
went
down
for
cloud
providers.
If
we
go
down
that
that,
so
something
to
think
about.
C
N
E
Didn't
see
either
of
them,
there
is
a
proposal
that
they're
looking
for
comments
on
and
it's
linked
off
of
the
dock.
It's
another
way
that
we
might
be
able
to
activate
contributors
within
the
community
and
I
think
I
saw
no
I.
Don't
I
saw
a
Ryan
on
here
earlier
and
I
wanted
to
give
a
plug
for
cake
support
and
many
people
to
take
a
peek
at
that,
as
well
as
a
fun
way
to
interact
with
committees
and
our
project.
G
I
was
going
to
bring
this
up
in
person
with
some
sense
of
Leadership
Summit,
but
I'll
just
throw
it
up
with
broader
community
as
well
I'm.
Finding
that
the
how
kubernetes
is
actually
built
he's
a
bit
of
a
gap
right
now,
as
far
as
take
ownership
and
I
was
going
to
propose
that
maybe
stake
release,
take
those
sorts
of
things
on
so
two
examples.
O
Actually
say
I
think
sig
release
is
probably
the
right
place
to
engage
in
this
stuff,
because
there
is,
you
know
it's
a
little
bit
sort
of
slow
rolling,
but
there
is
a
lot
of
talk
around.
How
do
we
build
releases?
How
do
we
do
packaging
across
different
types
of
packaging?
I
would
consider
the
container
images
being
part
of
that.
That
brings
in
some
of
the
build
stuff,
which
is
that
partisan
release
or
not
I
think
we've
had
discussions
around
that
I,
unfortunately
have
been
able
to
attend
those
that
say
guys.
O
G
K
This
has
been
near
the
top
of
my
list
of
things
that
need
to
fight
at
home.
For
a
long
time,
we
are
trying
to
make
sure
everything
on
the
project
has
a
home
in
some
sig.
The
obvious
candidates
are
similarly
sick,
testing
or
contributor
experience.
Sig
release
essentially
makes
sense.
We
should
discuss
with
them.
Some
other
products
do
have
like
CI
working
groups
which
puts
build
and
test
because
those
are
the
tests
are
unit
tests.
K
C
G
G
E
Of
the
box,
you
searched
several
months
in
15.
We
would
love
comment
on
them.
There
will
be.
You
mentioned
the
the
Leadership
Summit
tomorrow.
There
will
be
a
conversation
in
person
there
about
them
as
well,
and
then
we
will
be
synthesizing
a
lot
of
this
work
and
adjusting
tuning
up
what
what
our
Doc's
are.
E
Currently
as
far
as
the
process
getting
to
a
steering
committee
and
the
Charter
for
the
series
committee,
we've
gotten
a
lot
of
really
good
feedback
and
then
all
of
the
rest
of
the
docs
really
are
suggestions
or
thought
processes
are
on
what
we've
been
discussing
for
the
steering
committee
to
figure
out
what
to
after
that.
So
we
have
another
two
weeks
roughly
of
commenting
time,
so
please
feel
free
to
go.
E
Read
them
ask
questions,
give
comments,
hold
us
accountable
for
what
you
think
we
may
not
have
thought
of,
but
I
swear
with
with
the
number
of
days
that
we
spent
talking
about
this
emailing
about
this
over
time.
We
thought
of
a
lot
of
things.
We've
come
to
this
fairly
thoughtfully,
but
there
are
always
new
perspectives.
We
always
love
to
hear
them.
C
We
have
no
no
more
agenda
items
then
thanks
everyone
for
joining
us,
as
always,
we
we'd
like
it
when
other
people
also
MC.
So
your
name,
the
kubernetes
community
or
you've,
never
chaired
one
of
these
we'd
love
to
have
you
as
well
as
we'd
love
to
have
volunteers,
helping,
take
notes
and
that
sort
of
thing
help
help
us
wrangle
cats.
So
with
that,
thank
you,
everybody
and
we'll
see
you
all
next
week
and.