►
From YouTube: 20181009 sig cluster lifecycle
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
B
I
think
most
of
you
would
have
heard
me
now.
There
calls
by
now
so
113
is
already
underway
and
we
are
starting
feature.
We
started
feature
collection
yesterday
and
the
dates
are
we
have
a
code
freeze
coming
up
on
the
15th
of
November,
so
this
is
a
pretty
short
cycle,
so
Kendrick,
who
is
the
enhancement,
lead
and
me?
B
That's
pretty
much
the
main
reason
why
I
wanted
to
be
here,
and
we
were
also
looking
at
the
features
from
cluster
lifecycle
for
113
and
so
far
we
just
see
one
on
the
and
I've
I've
noted
that
in
the
meeting
agenda
there
this
is
feature
number
11.
It
says
dramatically:
simplify
Kuban,
it
is
cluster
creation.
So
a
question
I
had
around
this
was:
is
it
more
of
a
cube
admin
tracking
issue,
or
is
it
specifically
trust
a
lifecycle?
Let's.
C
D
A
D
A
B
I
think
the
Danube
admin.
Let
me
quickly
make
sure
that
the
other
one
is
also
yeah.
A
B
A
B
A
B
I
was
hoping
I
do
if
it
Robert
or
anybody
else
might
be
on
the
call
just
to
get
yeah.
This
is
not
cube.
Admin
at
all.
This
is
the
Google
maintained
ones.
So
I
just
wanted
to
know
if
there's
a
point
of
contact
from
cluster
lifecycles
who
could
help
to
see
if
it's
a
test
in
from
failure
versus
sake,
GCP
we've
we
painted
suggest
beyond
it,
but
in
the
past,
I've
had
more
success
from
from
folks
in
faster
life
cycle,
then
from
CCP
in
debugging.
These
things
typically.
D
I
can
have
a
look
at
this
issue
and
I
was
just
scanning
through
the
logs
and
there's
nothing
like
it
looks
like
it
actually
runs
for
fifteen
hours.
It
doesn't
like
just
pause
this,
so
I
will
I
will
have
a
look
at
it
in
route
because
yeah
this
is.
This
is
definitely
said.
Cluster
dice,
I
call
responsibility,
first
triage
and
then
yeah.
B
D
D
B
Kind
of
chronically
times
been
timing
on
since
August,
so
that's
yeah,
that's
concerning
and
thanks
a
lot
Justin.
If
you
could
that's
all
I
had
unless
Kendrick
you
have
anything
else.
Nope.
G
B
B
A
I
think
the
problem
we
have
is
that
we
have
parallel
tracks
by
which
we
manage
things
and
they
can
all
get
routed
through
the
sig.
Even
the
sig.
Only
certain
parties
maintain
a
versus
B
I
would
love
it.
If
Google
would
switch
to
a
cuvette
DM
wrap
deployer,
because
then
they
would
then
there
could
be
only
one
and
we
would
be
betting
one
and
leveraging
that
yeah.
D
D
Things
will
be
at
the
next
level
up
cuz
QB.
Damn
is
perfect
right.
No,
it's
not!
You
know
what
I
mean
like
it,
I
I,
don't
think
it's
probable
I
think
that's
I,
think
that's
a
good
thing
to
do.
I,
don't
think
it
will
solve
the
problem
and
I,
don't
think
I'm,
not
sure
it
is
even
a
problem.
I
think
that
it
is
okay
for
like
different
people
to
route
different,
so
project
sure.
A
G
B
D
Like
just
because
GCP
is
the
last,
it's
very
fair
if
GCP
is
like
the
last
one
still
testing,
it
gets
like
all
the
test
failures.
You
know
what
I
mean
like.
So
if
we
turned
off
cops,
then
what
GCP
be
responsible
for
like
old
test
failures,
are
gonna
screw
bananas
project
right?
That
just
seems
like
not
the
right
thing
to
do
like
it's
sort
of
like
it's,
not
that
it's
the
GCP
specific
tests,
it's
this
is
only
tested
on
GCP,
so
I'm,
a
routing
thing
to
me
feels
no,
that's
not
true
in
it.
It's.
A
It's
that
this
specific
way
that
the
tests
are
failing
is
wholly
specific
to
GCP,
because
cou
bad
and
it
has
its
own
upgrade
test.
We
have,
we
have
a
set.
We
have
a
suite
across
multiple
versions
for
doing
a
career
test.
If
our
our
signals
green
and
that
signals
red,
that's
telling
you
something
yeah.
A
G
A
It
seems
like
I
started
to
see
this
now.
I
saw
the
email
go
through
I,
don't
know
if
there
are
I
think
it's
probably
just
a
PSA
if
you
have
non
code,
specific
contributions
or
documentation
or
other
things
that
you'd
have
like
this
separate
location
for
Ana
guide
to
follow,
I,
don't
know,
I,
think
PSA.
This
exists
then
go
read.
The
link
I
have
not
read
it
before
this
meeting
and
they're
not
here
to
talk
about
it
right
now.
So.
D
So
Justin
cool,
I
guess
I
have
shared
the
dock
with
the
with
this
group,
but
I
will
share
my
screen
just
so
that
people
not
to
like
click
through,
but
this
is
sort
of
a
draft
of
something
we've
talked
about
in
the
past,
which
is
that
I
would
like
to
propose
a
new
sake.
Sub
sink
repository
for
xcd
management,
I
think
last
week,
Daniel
from
platform
9,
who
I
think
is
just
doing
on
errands,
but
might
be
here
soon.
D
I
demonstrated
at
CD
a
DM
which
is
sort
of
like
coop,
a
DM
for
Etsy
D,
and
we're
really
talking
about
Etsy
d
management.
Our
management
of
an
STD
cluster
that
is
used
for
the
kubernetes
control
plane,
specifically
and
then
I've
also
been
working
to
like
have
a
sort
of
self-driving.
What
a
mission
type
thing
and
the
proposal
is
that
we
create
a
new
secret,
posit
Ori.
That
is
essentially
the
platform
9
work
and
then
I
rebase
my
work.
D
On
top
of
that,
the
idea
being
that
we
have
sort
of
the
same
approaches
QbD
and
we
have
the
toolkit
which
works
great
and
is
very
focused
and
more
unup,
enya
nated.
And
then
we
can
have
things
that
build
on
top,
that
sort
of
add,
more
automation
or
like
cross
machine
collaboration
and
our
more
opinionated.
D
A
So
how
does
this
sound
as
a
mode
of
pushing
this
forward?
Because
typically
I
don't
remember,
call
if,
if
Robert
did
a
kept,
PR
or
not
but
I,
think
we
did
the
first
pass
feedback
on
an
initial
dock
that
was
well
vetted
and
then
there
was
a
cap
and
then
there
was
a
just
a
single
up
down
vote
and
that
seems
like
I,
told
I'd
like
to
follow
the
same
process
we
did
with
cluster
API,
because
that
was
wildly
successful
way,
I
think
of
engaging
all
parties.
A
D
A
D
Yeah
I
mean
I,
will
I
will
yeah
I
think
that
that
works
I
mean
I?
Think
there
shouldn't
be
anything
outrageous
in
the
dock
itself,
so
I
will
work
on
mccafé
YES
on
the
understanding
Billy
how
people
can
sort
of
the
feedback,
but
yeah
I
also
need
to
run
this
by
signifi
machinery
to
make
sure
that
they
don't
want
to
take
on
this
responsibility,
which
I
am
pretty
sure
they
will
not
want
to,
and
then
yeah.
D
If
anyone
has
any
comments
now,
I
very
much
welcome
them,
but
very
much
positioning
this
as,
as
for
the
kubernetes,
the
difference
between
like.
Why
does
an
SUV,
for
example?
Do
this
is
it's
for
the
kubernetes
control
plane
right?
It
is
not
a
generic
@cd
tool
and
sure
it'll
probably
work
if
for
like
other
things,
but
that's
not
the
intended
use.
Kids.
Oh.
D
So,
like
the
HCDA,
DM
command
is
sort
of
like
cube
a
DM
where
it
is
a
something
that
will
be
scriptable.
Is
my
understanding,
we're
gonna
it's
unclear
yet
whether
the
thing
which
was
previously
called
Etsy
D
manager,
the
automation
layer,
will
do
CLI
commands
or
will
interact
with
it.
As
a
library
I
don't
know,
would
you
use
it
as
a
library
or
yes,
I.
A
D
D
A
A
H
A
Folks
to
folks
provide
the
capabilities
to
currently
with,
instead
of
coops
free
to
use
any
arbitrary
versions
container,
because
part
of
the
CI
automation
will
be
that'll,
be
a
forcing
function
right
because
you
want
to
do
the
tip
of
Master,
which
means
that
you
build
all
these
containers
with
the
SHA
and
then
use
that
shod
version
to
do
the
deployment.
Yes,.
H
A
H
A
Done
I
was
gonna
move
on
to
the
next
topic,
so
is
there
any
other
updates
that
folks
want
to
make
from
any
other
sub
projects,
goop
spray
or
mini
cube
or
etc?.
D
D
A
D
G
Ruggiero
we
can
do
it
from
cubed
M
and
also,
if
you
expose
it
to
to
the
users
in
general,
they
can
use
it
as
well,
because
if
you
only
have
it
as
a
CA
Y,
you
know
if
you
limit
the
the
calls
that
the
users
can
use
in
basically
we're
talking
about
a
package
structure.
You
have
it
exported
cause.
It
makes
it
much
easier
for
users
to
like
use
it
as
a
library.
So
there
are
a
library
part
of
in
the
go.
Language
is
a
bit
ambiguous
in
general.
A
So
one
thing
that
be
really
interesting
to
talk
about,
maybe
not
today,
because
I
don't
think
we
have
up
the
right
people,
all
the
people
that
wouldn't
need
to
be
here,
talk
about
it,
but
for
any
given
version
of
kubernetes.
Now
that
we
have
all
these
different
sub
projects
releasing
at
all
these
different
times
to
have
a
single
jump
point
location
for
us
to
be
able
to
discover
all
of
these
tools
for
a
specific
version
of
kubernetes,
because
from
a
end-user
consumer
perspective
it
is
confusing,
as
all
heck
can
use
other
words.
A
But
this
thing
is
recorded
so
III
know
other
folks
have
come
to
us
with
this
topic
before,
but
I
think
we
need
to,
because
we
are
probably
at
the
forefront
of
having
so
many
sub
projects
that
we
might
have
opinions
before
some
of
the
other
folks.
So
are
there
any
topics
with
regards
to
that.
D
A
But
the
problem
is
useful
from
a
manifest
perspective,
but
the
problem
is
like:
where
do
what
I
did?
Where
do?
I
get
all
the
bits
for
kubernetes
110
latest
right?
That
includes
cops.
That
includes
creep
spray.
That
includes
everybody
cube
ATM,
though,
all
the
tools
like
where
do
where
does
a
user
or
consumer
of
corinna
DS
go
to
to
find
all
this
stuff
as.
D
I
understand
Josh
is
bundle
proposal.
The
idea
is
that
the
the
bundle
would
be
a
specification
of
everything,
I
guess
underneath
the
tooling
so
not
knock
up
some
cubes
rate
but,
like
you
know
which
version
of
Ruby
Nettie's
in
which
version
of
cube
DMS
and
all
these
other
things
and
your
your
tooling
would
would.
D
A
D
A
D
I
mean
so
cops
is
a
good
example,
in
that
we
it
exposes
like
two
of
the
really
big
problems.
So
we
have.
We
have
that
issues
we
have
where,
when
we
currently
have
the
idea
of
like
a
stable
channel
on
alpha
channeling
cups,
so
like
110
committees,
110,
something
is
the
one
in
our
stable
channel
for
cups
110,
which
is
the
current
released
version,
and
so
you
that's,
we
sort
of
navigate
the
other
problem
which
the
bundle
sells
is
currently
we
have.
A
H
A
D
Is
a
pointer
to
a
bundle,
type
of
concept
which
I
think
is
the
Coptic
logical
channel
last
people
call
it
a
channel,
but
you
know
that's.
That
is
a
solution.
I'm,
not
saying
it's
your
decision,
but
I.
It
seems
as
a
meta
point
which
is
I.
You
think
we
should
someone
someone.
Someone
has
to
create
a
repo
which
has
these
bundles
or
something
something
similar
in
it
like
when
we
add
it
to
cops,
we're
gonna
start
it
on
the
cops
repo,
but
I.
Don't
want
it.
I
don't
want
it
to
live
there.
It
would
be.
A
A
I'm
glad
actually
glad
we're
talking
about
this,
so
I
can
point
in
Senate
later
on
and
then
to
have
them
manage
that
those
bundles
over
time
are
pointers
to
bundles
would
be
the
ideal
scenario.
They'd
also
have
to
have
managed
actual
artifacts
themselves,
which
hopefully
would
be
90%
plus
containers,
which
is
a
lot
easier
to
manage
than
the
the
Deb's
in
the
RPM
is
themselves.
H
If
we
have
a
central
page,
sorry
to
intrude,
if
I
wasn't
repaid
for
the
see
and
say
like
a
small
mattress
with
care
of
this
version
is
going
to
be
the
opposite.
Like
amenities
112,
you
need
to
use
care
of.
This
version
kill
spray.
This
version,
qbm
discussion
and
for
111,
would
be
this
way
and
we
have
like
a
small
page
and
people
know
where
to
navigate
depending
on
who
they
want
to
pick
up
and
and
schooling,
to
abandon
everyone.
A
Think
we
could
probably
have
just
a
simple
I,
don't
know
if
it'd
be
a
page
in
the
main
documentation.
Maybe
that
might
be
a
little
bit,
maybe
a
link
on
the
main
page
that
points
to
a
repo
that
we
maintain
that
we
can
update
over
time
because
the
problem
with
the
documentation,
the
main
page,
is
that
PRS,
the
main
box
is
painful.
So
I,
don't
it's
kind
of
we're
kind
of
running
chasing
your
tail.
C
D
Machine
readable
would
be
I
mean
it
was
to
be
super
helpful
for,
like
you
know,
cops
and
I.
Imagine
other
tooling
to
be
able
to
have
a
specification
of
what
it
should
be
installing
and
if
people,
if
humans
want
Docs,
that's
great
I
would
I
would
hope
we
could
get
to
the
point
where
it
there
is
a
tool
for
everyone
that
mamas
are
the
same
tool.
The
tool
of
everyone
that
like
meets
their
needs
right.
A
A
I
A
Think
because
sequester
lifecycle
is
so
tightly
coupled
with
the
release
like
our
artifacts,
are
basically
how
you
stand
up.
Everything
right
that
we're
gonna
be
carrying
a
lot
about
this
I
think
one
of
the
we
might
even
feed
requirements
into
that
group,
such
as
like
we'd,
ideally
like
a
testing
stable
and
nightly
I.
Think
I
speak
for
everyone
in
saying
that
we
want
this,
because
we
don't
really
have
that
with
our
current
situation,
and
that
is
a
as
a
perpetual
pain.
F
F
And
we,
we
did
have
a
discussion
yesterday
regarding
net
Debian
problem
on
the
weekend,
so
our
main
discussions
have
been
around
OPM's,
Deb's
and
less
about
the
containers.
At
the
moment,
it's
basically
wherever,
where.
G
F
I
I
So
like
at
the
beginning
of
the
dock,
you
can
see
kind
of
roughly
prioritized
by
like
what
order
we
think
is
most
feasible
to
work
on
moving
things.
So
like
right
now,
there's
been
a
bunch
of
worker
I'm,
trying
to
get
the
DNS
managed
by
config
in
a
repo
and
by
this
team,
instead
of
like
tamaak
and
himself.
G
I
So
at
least
for
now
there
was
discussion
of
that
and
B
we'd
opted
to
go
with
I,
think
just
the
kids
I
already
before
this
stuff,
because
we
weren't
yet
fully
cleared
as
like
a
working
group
or
a
cig
or
anything.
Yet
we
thought
it
like
I
think
only
as
of
maybe
the
past
week
or
two
as
we
fit
into
the
governance
at
all
I
believe
it's
a
working
group
now.
A
All
right
so
I
think
the
TLDR
of
this
is:
we
need
to
think
about
it
more
and
plug
in
with
the
Cades
and
for
team
with
regards
to
how
we
would
version
artifacts
and
be
able
to
track
them
over
time
and
be
able
to
modify
that
list
potentially
using
bundles
where
possible.
This
is
super
meta,
but
those
who
listen,
the
recording
might
be
able
to
parse
that
with
some
level
of
sanity,
I.
C
Had
a
vague
idea:
I,
don't
know
if
it's
useful
or
not
maybe-
and
maybe
we
could
somehow
annotate
a
github
releases
so
that
this
could
be
this
information
to
be
obtained
programmatically.
So
whenever
cops
version
comes
up,
that
much
is
certain
the
Bernese
release
that
could
be
picked
up.
Just
I
thought
I'd.
A
Prefer
to
have
a
version
bundle
in
a
separate
repository
than
to
have
like,
because
you'd
have
to
retrofit
tag
annotations
on
the
main
repo,
and
that
would
be
super
weird
right
so
like.
If
you
already
had
an
annotated
tag,
usually
you
don't
have
any
tended
tags
be
fungible
usually,
but
they
are
fungible
if
you
really
wanted
to
so
that
gets
funky
I
think
from
a
release
perspective.
G
A
G
I
have
a
another
major
topic.
Sorry,
it's
about
the
materials
report
and
this.
This
is
a
good
forum
to
discuss
this,
because
we
have
a
lot
of
sub
projects
and
the
original
idea
by
Lucas
was
to
like
every
team
hosted
under
copper86
or
kubernetes
or
kubernetes
incubator,
to
eventually
support
multimode
petitions
like
those
architectures
that
we
support
for
the
images
that
are
currently
in
the
GCR
bucket.
Like
you
know,
the
IBM
mainframe
arm,
I'm,
64
and.
G
I
I
will
find
out
again,
though,
that
whatever
we
want
like
like
if
this
sig
wants
something
from
the
Cates
m14
we're
going
to
need
manpower,
almost
everyone
who's
like
joined
that
especially
outside
of
Google,
are
people
like
DIMMs,
who
are
doing
way
too
many
things
and
are
not
really
gonna
have
time
to
do
a
lot,
but
that's
definitely
going
to
be
a
big
problem
going
forward.
If
we
want
any
kind
of
velocity
in
that
group,
I
want.
A
One
of
the
things
I
would
ideally
like
to
hope
for-
and
this
is
a
governance
question
which
I
might
take
it
back
to
the
high-level
folks,
is
a
lot
of
people
throw
money
at
the
CN
CF
and
that's
great,
but
the
problem
is
we
need
bodies
to
do
the
work,
so
the
CN
CF
has
a
ton
of
money,
but
they
don't
have
enough
people
who
are
long
term.
Maintainer
z'
of
some
of
these
things,
yeah.
I
A
I
were
to
use
an
analogous
reference.
Apache
has
an
Apache
infra
team
for
better
for
worse,
if
you've
never
seen
it
so
I
know
is
recorded,
so
my
apologies
but
I've
seen
it
so
the,
but
they
are
sourced
by
different
companies
right
like
they
actually
dedicate
people
to
work
on
the
infra-
and
you
know,
I
think
I
think
has
folks
contributed
this
project.
It's
not
just
enough
to
extract
value
that
they
need
to.
Actually
you
know,
chop
the
wood
and
carry
the
water
to
get
the
bits
out.
G
Yes,
so
about
this
issue
that
Lucas
created
about
the
boatyard
support.
Do
you
think
that
we
should
like
simply
transfer
it
to
seek
release
at
this
point
yeah
because
yeah,
because
Lucas
added,
like
a
pointed
out
earlier,
Lucas
added
this
big
goal
of
having
everything
multi
arching?
But
it
is
including
cold
sub
projects
and
everything-
and
that's
I
mean
who
is
going
to
monitor
this
demand
for
the
sub
projects?
A
Think
it's
in
release
like
there.
They
are
in
the
release
artifacts
because,
like
the
I,
don't
think
it's
this
the
architecture
related
to
other
than
the
requirements
you're
like
of
what
architectures
we
should
support,
but
I
I
do
think
this.
The
problem
with
cig
release
is
there's
no
continuous
group
of
people
that
actually
work
on
cig
release,
cars,
every
single
cycle,
there's
a
couple
of
people
that
do
part-time
work
across
every
release,
but
there's
no
one
that
really
owns
it
for
the
long
haul
to
make
sure
that
these
things
are
refined,
yeah.
G
G
D
I
think
you
know
there
are
some
like
there
are
these
Cube
ABM
test
now
and
I
think
there's
kind
which
I
think
Ben
wrote
as
well.
If
we,
if
we
ran
some
of
their
tests,
some
of
the
system,
like
you
emulate,
allegedly
as
390
I,
didn't
even
know.
That
was
really
a
thing,
but
there
we
are
like
we
get
some
signal.
I
Would
get
signal
you
know
we
get
some
I
I
do
think,
though,
like
one
model
I've
been
thinking
about
more
is
especially
as
we
have
all
these
different
cloud
providers,
and
things
is
that,
like
we
should
so
I
I
know,
there's
a
there's,
an
ongoing
effort,
open-source
test
grid
itself,
and
now
that
we
have
the
CNC
F
work,
we
should
have
an
easier
time
setting
up
like
buckets
with
different
people,
have
access
to
them
directly.
I
think
we
might
want
to
look
at
pushing
some
of
these
groups
to
say,
have
arm
run
tests
more
regularly.
I
They're
kind
of
doing
it
now
arm
is
at
least
if
we
can
get
someone
with
s
390x
hardware
to
do
the
same.
We
can
kind
of
offload
a
little
bit
of
the
like.
How
do
we
run
all
of
this
because,
like
for
example,
even
if
you
handed
me
an
s3
90
bucks,
I,
don't
know
anything
about
how
to
do
anything
with
one
so
yeah.
A
I
think
there's
a
problem
with
like
the
people
who
are
the
consumers
of
these
artifacts
should
be
the
ones
who
are
publishing.
The
CIO
signal
right
so
like
packet,
is
a
good
example
of
arm
64
work.
They
actually
have
a
product
there's
no
reason
why
they
couldn't
contribute
resources
to
push
that
signal
same
with
IBM
for
mainframes
and
PowerPC,
for
that
matter.
I
G
There
is
also
something
else
that
our
m64
is
a
consumer
level,
like
general
public
level,
architecture
for
the
raspberry,
PI's
and
stuff.
So
maybe
we
should
have
this
at
least
I.
Don't
really
have
any
demand
for
it
personally
or
my
employer,
like
the
question
is:
should
we,
how
should
we
populate
this?
I
G
I
Yeah
but
I
like
then
we're
still,
then
we're
still
having
to
maintain
all
of
it,
and
it's
probably
going
to
be
slower,
and
it
won't
be
as
authentic
of
signal
as
if
we
like
run
a
day
because
I
think
so
with
something
like
kind.
I'm
thinking,
mostly
targeting
things
that
don't
need
kubernetes
to
be
as
real
or
things
that
are
like,
say,
pre
submit
if
we're
running
in
C
I,
just
like
periodically
against
the
repo,
and
we
want
signal
for
the
release
team
I
think
we
want
to
cover
as
many
like
real
scenarios
as
possible.
G
G
A
A
Well
before
we
remove
it,
which,
let's
point
the
issue
that
made
a
topic
back
to
sig
release
and
say,
like
you
guys
need
to
maintainer,
we
should
probably
PR
to
get
that
rid
of
a
sequestered
lifecycle
out
of
our
name
cuz.
We
don't
want
to
own
that
stuff
and
just
call
it
multi-platform
or
something
or
just
remove
it
outright
for
it
now
and
let
sig
release
illness
yeah.
G
I
Like
one
contribution
now
from
army
and
I
need
to
move
where
it
is,
it's
only
in
the
six
no
dash
board
currently
and
it
looks
like
it's
actually
publishing
enough
results.
Now
previously
it
was
uploading
some
of
the
wrong
data
and,
interestingly,
at
least
with
how
they've
got
it
set
up.
A
few
failed.
The
one
room
we
have,
but
we
do
have
looks
like
maybe
twice
a
day
or
something
we'll
see.
It's
very
new.
I
G
A
G
J
Great
I'll
do
a
three
count:
the
city
mayor,
sessoms,
oh
I,
can
propecia
topic,
so
I
started
to
create
slides
for
kid.
Kong
China,
where
me
and
Steven
were
going
to
be
hosting
we're
six
sessions.
So
if
there
are
some
projects
underway,
seek
who
wants
to
was
a
highlight
some
of
the
features
of
some
of
us
part
of
the
project.
Please
send
us
your
information,
so
we
can
include
it
in
the
slides.
J
I
will
actually
allotted
in.
Let
me
lock.
A
On
a
similar
note,
I'll
say
it
here
and
I'll
say
it
in
the
Canadian
office.
Hours,
Robert
and
I
are
between
two
separate
sessions
at
coupe,
con
North
America,
and
if
there
are
other
folks
who
would
like
to
cope,
resent
or
interesting
co-presenting
on
either
cluster
API
work
or
the
Covidien
work,
if
you
are
planning
to
go
to
that
venue,
then
we'd
be
happy
to
chat
to
try
and
get
other
folks
who
are
active
in
the
sig
to
get
involved
in
the
presentation,
or
you
know
in
the
content
itself
too
as
well.
A
What
are
you
sessions
call
means,
there's
there's
deep
dives
or
there's
there's
two
tracks
at
any.
Could
con
the
one
track
is
the
actual
talks
that
are
been
submitted.
The
other
track
is
basically
like
the
contributor
track,
so
every
sig
has
the
option
to
be
able
to
present,
and
we
have
two
tracks
that
we
are
presenting
at
the
next
coop
con.
One
is
akabane,
iam
talked
the
other
one.
Is
a
cluster
API
talk.
K
G
Documentation
of
like,
like
you
see,
documents
but
documentation
in
general,
is
a
big
problem
for
projects
like
coaster
API,
who,
which
pretty
much
doesn't
have
any
documentation
or
at
all
right
now,
and
also
cubed
iam
oftentimes
needs
help
with
documentation.
So
documentation
in
general
is
something
that
we
can
get
help
with
in
terms
of
non
code.
Contributions.
H
G
For
cube
videos,
specifically,
we
have
pages
on
the
kubernetes
website
and
we
do
pull
requests
against
the
kubernetes
websites
Rico
for
question
a
pi/2
minus
ending
right
now
they
mostly
maintain
Google,
Docs
and
markdown
pages,
a
whole
set
and
repositories,
and
that
also
needs
poor
request.
Basically,
but
it's
not
against
the
communities
website.
L
D
G
So
one
way
to
to
get
attention
or
dislike,
because
if
you
have
contributors
who
want
to
help
one
way
to
get
the
attention
of
the
people,
who
can
point
you
out,
a
specific
problems
is
to
posting
slack,
join
the
sequester
lifecycle
channel
or
the
Costa
Rica
channel,
or
cops
and
just
posts
that
you
can
help
with
dogs.
And
this
is
going
to
get
the
attention
of
the
maintainers.
K
G
Yes,
so
there's
something
else
that
it
that
they
can
think
of
at
the
moment
like
if
we
have
kubernetes
it
needs
this
in
general.
If
we
have
people
that
monitor
desk
grids
like
the
on
a
daily
basis
and
if
they
post
on
slack
like
a
comet,
a
this
grid
started
failing
for
this
particular
test,
and
even
if
we
have
notifications,
I
know
that
we
have.
I
Will
say,
Aaron
and
I
are
looking
into
having
mailing
lists
specifically
for
those
notifications
and
discussing
more
around
how
we
can
improve
ownership.
I
do
think
like
well.
That
could
be
helpful.
It
is
kind
of
a
little
sad
to
ask
a
human
to
go,
look
at
a
thing
and
then
tell
you
if
it's
changed
constantly
that's
something
that
computers
are
really
good
at
doing
for
us.
Yeah.
I
I
G
A
We
are
nearly
out
of
time,
so
I
think
we'll
probably
call
it
at
that
one.
We
can't
believe
bring
up
any
more
topics
for
you're,
not
gonna
have
time
to
talk
about
it.
So
thanks
for
filling
the
agenda
list,
I
thought
we
read
a
short
agenda,
but
we've
managed
to
take
up
little
hours.
So
thanks
everybody.