►
From YouTube: Kubernetes SIG CLI 20161207
Description
Kubernetes SIG-CLI bi-weekly meeting with discussions about our roadmap for 2017 and specifically for the 1.6 release.
A
B
A
B
B
Should
we
so
we
could
continue
to
talk
about
this
more
or
we
can
go
through
the
agenda.
First
thing
come
back
to
this:
you
have
a
preference
yeah.
C
No
I,
don't
really
have
a
strong
preference,
but
I.
Think
of
giving
a
individual
status
quick
status
update
would
be
nice
if
that's
ok,
for
you
guys.
So
for
me,
like
this
best,
it's
it's
been
a
little
more
than
a
month.
Alright
already!
Actually,
since
our
last
meeting
in
this
past
few
weeks
for
me
specifically,
has
been
pretty
much
about
going
over
issues
and
doing
code
reviews
things
like
that,
I
did
work
a
little
bit
on
some
way
related
to
new
features,
specifically
about
cube
CTL
extensions.
C
So
you
guys
remember
there
is
a
couple
requests
from
laten.
We
it's
a
it's
a
doctor
request
about
cube,
CTL
extensions,
so
something
I
tried
to
do
was
start
some
some
drafts,
a
kind
of
proof
of
concepts
about
extensions
like
plug-ins
for
cube
CTL,
so
I
have
a
per
quest
for
for
the
apt.
Really,
you
know
an
early
draft,
not
even
a
proposal
yet
but
kind
of
a
proof-of-concept.
C
What
I
wanted
to
achieve
was
to
try
to
you
know,
set
up
some
framework
for
cube,
CTL
extensions
that
would
allow
extinction
writers
write
something
in
another
language,
meh
language
other
than
go
so
I
wrote
a
kind
of
a
little
bit
framework
and
an
extension
written
in
rubbish
script,
but
that's
really
not
even
a
profess
open
for
review
it.
It
was
really
a
proof
of
concept,
but
other
than
that.
C
C
Yeah,
I
agree.
I
personally
like
that
proposal
a
lot,
I'm
not
sure
if
we
will
be
able
to
achieve
a
final
version
or
you
know
a
ga
for
plugins
for
1.6,
but
in
my
opinion
we
should
at
least
have
a
proposal
right.
So
if
we
don't
have
an
implementation,
yet
we
shoot
for
116
at
least
understand
exactly
what
we
want
and
have
a
solution
proposal
in
a
document.
Something
like
that
if
we
are
able
to
achieve
a
min
accommodation,
that
would
be
really
great
I'm
personally
interested
in
that
in
the
matter.
C
But
not
sure
if
we
should,
you
know,
agree
on
really
have
that
as
a
as
a
goal
for
about
1.6,
because
as
far
as
I'm
seeing
from
discussions
in
that
/
quest,
there
are
a
lot
of
things
to
write,
discuss
and
to
be
come
to
an
agreement.
So
naturally,
if
you
will
be
able
to
get
in
time
for
116,
but
at
least
having
a
proposal
will
would
be
nice
yeah.
B
Who
I
can
have
a
quick
status?
I've
been
traveling
for
Leah,
the
last
five
weeks
and
then
was
also
Thanksgiving
was
out
so
I
haven't
had
a
chance
to
work
on
much
besides
the
events
I've
been
tending
one
thing:
I
was
working
on
last
week
was
the
the
2017
goals
for
my
team,
and
one
thing:
I'm
trying
to
do
is
thematically
group-specific
work
items
in
the
context
of
barter
broader
strategic
goals,
and
so
one
of
those
is
like
apply,
and
how
does
it
relate
to
arm
the
other?
B
B
How
do
we
allow
third-party
resources
and
other
extensions
to
define
how
they're
displayed
natively
using
annotations
or
something
like
that
up
I'll,
put
a
design
doc
together?
More
for
that,
and
then
according
like
third
party
resources
with
KU
control,
apply
which
doesn't
work
very
well
now
and
that's
written
thing
and
then
also
getting,
I
think,
the
technical
debt
or
engineering
excellence
or
whatever
you
want
to
call
it
there's
another
thing
that
we
probably
want
to
spend
some
time
on.
B
C
Yeah,
I
definitely
agree
and
I
I
personally
wasn't
able
to
attend
cube
con
this
year.
Unfortunately,
but
did
you
guys
had
any
discussion
about
client
tools
or
anything
like
like
that
that
would
you
know
required
a
follow-up
from
was
or
anything
like
that
I.
B
B
B
Oh,
how
can
we
do
it
in
a
way?
That's
consumed
can
be
consumed
by
as
many
clients
as
possible.
So,
for
instance,
the
dashboard
needs
to
have
a
display
for
third
plate
resources
and
federated
API
resources
and
to
control
will
probably
also
want
to
have
customized
display
around
them,
and
so,
whatever
solutions
we
design
should
be
supported
by
both
those
two
extent.
That's
possible.
C
D
You
had
me
come
in.
The
only
thing
I
was
going
to
mention
was
I'm
working
on
exposing
config
maps
as
environment
variables,
which
is
going
too
well,
so
there's
a
tie
where
they
want
to
extend
how
freakin
big
map
works
so
that
you
can
pull
in
the
Dockers
environment
piles,
but
right
now
we're
getting
wrapped
up
into
how
much
it
will
look
like
or
not.
Look
like.
D
C
Okay,
I
guess
that
sits
in
terms
of
status
updates,
real
quick
about
the
1.5
release.
We
don't
have
any
any
issues
listed
as
blockers
as
up
now.
So
that's
pretty
good.
Any
other
comments
about
the
1.5
early
is
anything
we
still
have
to
address
before
the
release,
but
from
up
from
my
end
I
don't
think
there's
anything
we
we
have
to
do
so.
C
Okay,
I
guess
that's
it
then
then,
moving
to
to
the
1.6
release,
we
were
already
talking
about
it.
As
I
said,
I
try
to
start
real
quick
road
map
document
for
the
1.6
release.
Link
is
on
the
chat,
really
not
a
proposal
yet,
as
I
said,
but
I
try
to
pick
some
items
mentioned
by
Brian
in
our
last
meeting
and
some
other
things
that
are
important
from
our
issues
backlog.
C
C
So
that
would
be
one
idea.
We
have
a
lot
of
stuff
in
a
our
backlog,
both
in
terms
of
issues
bugs
and
also
type
that.
But
then,
anyway,
if
we
decide
to
work
on
new
features,
we
can
add
that
to
the
document
and
we
could
come
up
with
a
proposal.
You
know
roadmap
for
1.6,
I,
guess
I
guess
at
some
point
it
would
be
nice
for
us
to
there
that
you
know
the
six
CLI
proposal
roadmap
for
1.6
and
things
we
are
looking
forward
to
work
on
for
that
release.
B
C
Don't
know
you
know
the
way
it
works
better
for
you
guys,
maybe
maybe
we
don't
need
to
you
know,
do
that
right
now,
in
the
meeting
casa
I
know,
there
are
a
lot
of
items
that
requires
to
take
a
look
at
the
list
of
issues
on
ye
tubs
and
pink
github,
and
things
like
that,
so
you
know
wicked.
We
could
either
you
know
each
people
do
it
individually
and
we
try
to
sync
up
by
mail
or
in
the
the
other
meeting.
C
B
Don't
why
don't
we
finish
the
press,
the
meeting
and
then
with
the
time
that
we
have
at
the
end,
then
we
can
see
how
far
we
get
on
that.
Maybe
yeah.
C
Sure
yeah
works
for
me,
so
the
other
right
item
I
had
is
the
self-assessment
for
cigs.
So
that
was
also
something
that
came
up
from
sara
is
a
proposal
for
each
seek
doing
these
assessment
and
Tony
further
farrar
dead
and
the
mail
order
today,
so
I
believe
it's
nice
for
us
to
go
over
each
bullet
point
and
make
sure
everything
is
addressed.
So
initially,
that's
in
the
the
meeting
agenda
notes
documents
so
initially
does
your
sig
on
code
in
Cuban,
Eddie's
or
elsewhere.
C
If
we
do
q
Burnett
is,
is
it
reflected
in
out
the
honors
files
so
specifically
about
the
honors
files?
I
think
we
did
that
a
few
weeks
ago
right.
So
there
was
a
review
of
our
owners
files
for
the
cube
CTL
package,
specifically
so
I
believe
in
that
topic.
Specifically,
we
are
looking
good,
so
anyone
has
anything
to
add
to
that.
C
Okay,
so
the
next
one
does
your
sig
spend
time
responding
to
user
issues
in
github
on
stack
overflow
or
the
Sigma
only
list
so
for
me,
specifically,
I
really
don't
spend
too
much
time
on
stack.
Overflow,
I
really
don't
know
if
there's
anyone
that
tracks
that
in
our
sig
about
user
issues
on
github,
so
as
I
said
in
my
status
update,
that's
something
I've
been
doing
a
lot
this
past
this
past
release,
so
I
believe.
Overall,
we
are
looking
pretty
good.
What
I
feel
like
is.
We
are
still
liking
some
tooling
around
it.
C
So
so,
for
example,
there
are
some
discussions
trying
to
address
that
already,
but
things
like
we
being
able
to
to
put
the
poor
requests
to
tread,
looks
good
to
me
to
progressed
for
it,
for
example.
So
so
that's
one
issue
we
had
today
if
issues
so,
for
example,
for
me
and
Tony
specifically,
if
issues
are,
if
we
requests
are
assign
it
to
us,
we
can
add
the
looks
good
to
me
flag,
but
we
can't
set
the
release.
C
Note
label,
so
that's
an
issue
I
believe,
that's
being
you
know,
adjusted
but
anyway,
there
still
a
few
things
in
terms
of
tooling
to
address
that,
like
I,
don't
know
a
receiving
patience
when
there
are
issues
related
to
our
see
something
like
that.
So
there
are
some
things
in
terms
of
the
community
that
we
still
have
to
address,
but
overall,
every
time
I
am
able
to
to
check
their
new
issues.
New
progress,
things
related
through
our
sig
on
github
I
personally
have
been
working
on
them.
So
with
you
guys
comment
anything
about
that
topic.
A
A
B
Have
a
little
more
structured
process
for
triaging
the
issues
I
know
I
skin,
both
of
you
up
triage
and
naming
commenting
on
issues
and
I,
try
and
triage
all
the
ones
labeled
with
Q
control,
I,
don't
know,
I
guess
I'd
be
interesting,
getting
a
feeling
for
how
you
think
the
process
is
working
for
us
in
it.
If
it's
too
much
to
have
all
of
us
reviewing
them
all
the
time
or
that's
a
that's
working
for
us,
yeah.
C
Something
something
we
do
eventually
here
in
an
open
shift
is
a
is
a
bug,
triage
meaty,
so
you
know,
meetings
are
sometimes
are
not
always
the
solution,
but
eventually,
once
in
a
week
or
once
in
every
couple
weeks,
we
do
some
kind
of
triage
meeting.
Sometimes
there
are
just
a
couple
of
people.
C
It
would
be,
may
be
nice
that
could
be
even
done
individually,
but
to
take
some
time
to
take
the
entire
list
of
issues
take
a
look
at
their
at
them
and
at
least
close
what's
already
addressed,
and
you
know,
because
I'm
pretty
sure
that
in
in
depth
amount
of
stuff
500
there
are
probably
stuff
fix
already
are
things
that
are
no
longer
initial
or
anything
like
that.
C
But
yeah,
maybe
coming
up
with
a
better
process
and
tooling
around
doing
that,
so
maybe
receiving
some
notifications
or
being
able
to
take
some
time
to
triage,
along
with
having
a
the
required
access
to
do
that
like
close
issues,
something
like
that
could
help
us
to
to
address
that
anyway.
Otherwise,
it's
just
it
will
just
you
know
getting
worse,
probably.
B
Yeah
I,
don't
even
know
what
our
reasonable
goal
is
there
like
we
could
try
and
keep
the
bug
countdown
to
a
specific
number
could
just
grow
linearly
over
time,
or
maybe
we
want
to
say
that
you
know
have
some
labeling
system,
so
we
keep
bugs
around
just
for
tracking
purposes,
but
label
them
as
such
with
no
intention
of
closing
them
are
yeah.
I,
don't
know.
C
Yeah,
naturally,
there
anyway,
I
think
that's
that's
really
a
problem
not
only
to
us
right.
It's
a
problem.
You
know
all
over
the
deeper
project
today
and
I
guess
there
are
some
discussions
happening
in
the
community
repo
and
from
brian
about
how
we
can
address
that.
So
anyway,
we
could
check.
Maybe
you
know
check
what
they
come
up
with
for
after
the
e
1.5
release.
C
C
Okay,
anyway,
moving
to
the
next
bullet
point
is
the
balance
between
new
features,
work
and
stabilization
good
over
the
last
two
releases.
Do
you
feel
the
need
to
focus
on
stabilization,
then?
Actually
something
else.
So
we
really
don't.
Don't
have
history
in
this
sig
to
compare
right
so
I'm,
not
sure
if
you
Phillip
have
any
information
about
that,
but
anyway,
definitely
some
work
on
stabilization
would
be
good.
The
given
even
you
know,
given
the
amount
of
issues
we
have
in
the
backlog.
B
Yeah
I
mean
I
I
think
we
definitely
need
to
do
some
stabilization.
The
number
of
issues,
I'm
not
sure,
is
wow
yeah,
I'm
not
sure
how
to
interpret
that
metric
on
its
own,
but
definitely
some
of
the
issues
like
the
all
issues
we're
seeing
with
apply,
which
we're
telling
people
to
use
is
pretty
good
indicator.
Yeah.
C
Yeah
and
actually
I
liked,
I
liked
your
the
way
you
you
were
tracking
that
keeps
Adela
apply
umbrella
issue,
so
maybe
one
idea
and
I
could
volunteer
to
do
that.
We
could
take
I,
could
take
a
look
at
our
back
lot
of
issues
trying
to
identify.
You
know
the
most
important
ones
and
things
that
would
be
really
interesting
to
get
to
have
addressed
like,
for
example,
in
the
1.6
timeframe
and
try
to
create
a
few.
C
You
know
of
similar
umbrella
issues
listing
the
ones
we
would
like
to
address,
and
then
we
could
come
up
with
either
1.6
road.
My
plan
we
could
list
on
a
number
of
umbrella
issues
like
issues
in
keep
CTL
I,
don't
know
help,
for
example,
issues
and
keep
CTL.
You
know
create
subcommands
things
like
that.
Have
a
number
56,
something
like
that
umbrella
issues
listing
the
individual
issues
we
would
like
to
address.
So
I
could
volunteer
to
do
that
and
in
the
meantime
you
know
automatically.
C
I
will
kind
of
triage
some
some
things,
because
I
will
eventually
find
things
that
are
ready,
fixed
or
really.
We
don't
want
to
don't
want
to
address
so,
but
anyway,
I
think
that
that
would
be
a
nice
idea
have
a
number
more
of
those
umbrella
issues
and
use
those
in
our
planning
for
1.6.
If
that
sounds
good
to
you,
guys,
yeah.
B
C
Yeah,
absolutely
okay,
so
I'll
start
doing
some
work
about
that.
Let's
say
until
the
end
of
this
week
trying
to
come
up
with
at
least
you
know
some
better
organization
of
at
least
the
most
important
ones.
You
know
trying
to
do
it
in
for
a
real
bugs
and
take
that
things
like
that.
So
at
least
that's
you
know
a
list
of
500
things
will
be
a
little
more
manageable
for
us
to
to
come
up
with
a
plan.
C
Okay
after
that,
overall,
how
do
you
feel
good
about
the
level
of
test
coverage
for
code
you're
responsible
for
docs,
completeness
number
of
open
bug,
bug
reports
so
about
the
test
coverage?
That's
something
I
had
in
mind
this
past
week,
so
on
which
we
have
an
issue
about
making
cube
CTL
the
the
actual
features
separate
of
the
dependency
of
Cobra
command
and
having
those
more
reusable.
C
So,
along
with
that
probably
comes
with
a
better
testing,
because
today
we
really
don't
have
much
unit
tests
for
cube
CTL
command.
It's
right
in
that!
That's
something
we
do
here
in
open
chief
that,
but
it
kind
of
requires
you
to
have
the
feature
really
separated
from
the
command
the
commander
codebase
right.
So
that's
one
of
the
topics
in
terms
of
deck
tech
depth
that
I
personally
I
think
it
is
important
so
being
able
to
work
a
little
more
on
this
separation
of
concerns.
C
You
know
the
actual
command
code
base
and
what
it
does
and
by
having
it
separate,
you
will
probably
be
able
to
better
unit
test
them.
Show
up
a
few
comments
already
respect
one
of
those
proposals
we
have
with
their
complete
validating,
run
command
and
that
kind
of
gives
us
this
separation
and
makes
it
unique
testable.
So
something
I
think
that's
something
nice
for
1.6.
You
know
being
able
to
separate
these
things
a
little
better
and
with
that
come
up
with
more
unit
tests.
C
Think
that
I
think
I
even
listed
that
as
one
topic
in
the
road
map
document
yeah
reduce
the
amount
of
dependency
on
Cobra
and
make
comments
available
at
the
library,
even
because
another
issue
we
have
is
a
moving
stuff
to
the
server
side.
So
the
client
keeps
each
other's
kind
of
too
much
stuff
in
some
cases,
so
at
least
being
able
to
move
them
out
of
Cobra
and
have
that
as
a
separate,
reusable
piece
would
be
better
right.
C
B
C
C
Well,
there
are
still
things
we
would
like
to
address
like
extensions
and
a
few
other
things,
but
it
could
be
more
about
stabilization
for
us
not
sure
if
you
guys
agree,
but
I
wouldn't
make
it
a
release
entirely
for
stabilization.
I
think
it's
it's
nice
to
to
deliver
new
feature,
especially
because
you
know
we
are
what
end
users
face.
Keep
cgl
is
what
end
users
use.
So
it's
always
nice
to
leave,
deliver
new
stuff,
but
yeah.
We
could
give
more
focus
to
stabilization
in
1.6.
If
you
guys
agree.
C
B
Think
one
thing
that
might
be
nice
I
had
it's
like
kind
of
a
stretch
was
a
like
diff
for
coop
control.
Apply
I
think
would
be
really
helpful
when
you
to
have
a
mode
where
it
shows
you
that
actual
changes
is
going
to
send,
instead
of
just
doing
something
she
can
validate
their
correct,
but
automatically
goes
with
fixing
many
control
apply
issues.
B
And
then
the
other,
the
other
piece
I
have
on
my
list,
was
more
generally
around
figuring
out
what
the
big
management
solutions
are.
People
have
been
using,
so
I
canno
box
uses
Jason
net
factoring
and
so
I
don't
lose
anything
to
put
there
right
now.
D
C
Okay,
yeah
I,
remember
something
Brian
also
mentioned
in
our
previous
meeting
was
a
the
tool
distribution,
so
I'm
not
sure
it
would
be
a
goal
for
now,
but
anyway,
everyone.
That
was
something
he
mentioned.
C
Things
like
you
know
how
to
install
it
and
keep
keep
keep
CTL
up
to
date
when
you're,
not
using
you
know,
other
means
for
for
the
as
a
distribution
mechanism,
so,
for
example,
a
installation
and
upgrade
through
your
operating
system
so
think
about,
for
example,
Windows
users,
probably
also
Mac
users.
So
how
do
they
better
install
and
keep
keep
CTL
up
to
date?
So
not
sure
if
that
would
be
something
of
interest,
but
anyway,
Brian
mentioned
it
at
some
point.
D
B
B
Maybe
maybe
one
way
we
could
approach
the
problem
is
once
we
have
some
of
those
of
umbrella
issues
set
up.
We
can
stack
rank
within
them,
which
ones
are
the
most
important.
Maybe
that's
the
most
important
and
technical
debt
issue,
yeah.
C
C
C
Okay,
wow.
C
B
D
B
B
This
was
essentially
building
out
I'd
be
like
right
now.
What
we
have
is
like
tools,
we
get
people
these
like
tools
and
say,
look
at
like
all
the
awesome
features
and
commands,
but
we
don't
give
great
prescriptions
or
how
do
you?
How
does
your
team
actually
manage?
Your
application
can
face
right
like
how
do
you
get
your
initial
config?
B
Should
you
be
writing
them
from
scratch,
using
like
copying
them
up
a
blog
popping
them
off
docs
for
kami
an
example
and
a
git
repo
like
what
do
people
do
and
then,
after
that,
it's
like
how
do
you
back
to
them?
What's
your
directory
structure,
look
like
do
you
store
them
and
get
and
push
them
there?
Do
you
automatically
push
from
get
to
the
cluster,
or
is
that
done
by
users
that
doesn't
apply
or
create,
or
some
other
mechanism
and
I
think
I
have
a
document
I'm
working
on
that
kind
of
outlines?
B
I
think
there's
about
have
about
eight
different
pieces
to
it,
but
so
thematically,
I
think,
there's
that
viet
another
area
that
mushrooms
focusing
arms
we
could
either
have.
We
could
have
similar
ones
or
we
could
does
is
they
can
come
up
with
those
are
but
I
said
having
four
things
for
no
more
than
five
things.
I
think
the
probably
the
way
of
contract,
a
bomb.
C
C
B
B
B
Guess:
yeah?
What?
Let's
see,
why
don't
I
add
some
more
details
about
what
that
is
to
thee
to
the
roadmap
that
fabiano
stood
out
and
then
we
can
when
we
next
week,
when
we
look
at
all
the
items
on
the
list
and
try
and
decide
what
we
want
to
get
done
in
2017
in
what
we
want
to
get
done
for
1.6,
we
can
decide
where
that
best
fits
in
okay.