►
From YouTube: Kubernetes SIG CLI 20210908
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
Good
morning,
good
evening,
good
afternoon,
depending
on
where
you
are
today
is
september,
eighth-
and
this
is
another
of
our
bi-weekly
six
cli
calls.
My
name
is
marcie
and
I'll
be
your
host.
Today,
our
agenda
is
fully
packed
with
topics.
So,
let's
quickly
jump
into
the
announcements,
the
most
important
and
pressing
one
is
that
the
cap
freeze
is
tomorrow.
A
I
linked
their
the
dates
for
123
release,
so
if
you're
working
on
a
cap
make
sure
to
have
it
ready
and
tag
by
tomorrow,
late
afternoon,
pacific
time,
if
I
remember
correctly,
that
was
what
we've
been
using
in
the
past,
I'm
pretty
sure
that
the
the
release
team
will
be
looking
carefully.
A
Although
looking
currently
at
the
dates,
it
only
is
the
date,
but
it's
usually
oh,
there
is
enhancement,
freeze
it's
midnight
pacific
time,
so
that
buys
me
time
all
the
way
until
friday
morning.
Actually,
but
that's
just
me,
okay,
eddie
shawn,
any
other
announcements
that
we
should
call
out.
B
A
Okey-Doke,
let's
move
on
to
the
introductions:
do
we
have
anyone
new
on
this
call,
if
so
feel
free
to
speak
up
right
now.
This
is
your
time
say
who
you
are
what
you're?
Looking
for?
Why
are
you
joining
our
call?
C
I'll
say:
hi,
I'm
carl
from
google
solution
architect
at
google,
and
I
was
looking
at
contributing
to
cli
utos,
so
just
wanted
to
hang
out
and
see
what
was
going
on.
A
A
Do
we
have
anyone
else
or
the
rest
looks
like
there
are
old
folks
that
would
have
been
with
us
for
quite
a
while
scrolling
through
the
participant
list.
Last
call.
A
Okay,
hearing
none,
let's
move
on
to
the
main
section,
so
the
topics
eddie
want
to
take
it
away
from
here.
B
Yeah,
so
we
had
this
on
the
agenda
and
we
cancelled
last
meeting.
So
this
isn't
quite
as
relevant
because
kept
freeze
is
tomorrow,
but
I
believe
the
extension
period
is
longer
than
the
opt-in
period.
So
please
get
your
caps
in
tomorrow,
but
if
there
are
any
that
won't
make
tomorrow,
we
have
a
generous
extension
period.
I
believe
so.
A
From
what
I
remember
for
123
personally
putting
on
the
enhancement
list,
there
are
two
topics.
The
cube
cuddle
events
is
something
that
I've
been
pushing
for
quite
a
while,
and
I'm
hoping
that
123
will
be
the
time
when
we
get
that
one
pushed
over.
I
reach
out
to
nikita
and
yara,
if
I
remember
correctly
his
name
about
the
sub
resource,
because
the
only
thing
that
I
wanted
from
them
was
to
have
an
update
to
the
latest
milestone
in
their
cap.
Other
than
that
that
that
is
ready
to
go.
I
haven't
heard
back.
A
I
do
hope
I'll
try
to
pinging
I'll,
try
pinging
nikita
tomorrow
morning,
when
it
will
be
more
more
time
wise
for
her
and
I'll
see.
If
we
can
get
that
one
going,
there's
still
cube
cuddle
debug
profiles
kept
from
lee
that
I'll
be
looking
at
later
today.
We're
hoping
that
that
one
will
still
be
will
be
able
to
to
be
included
in
123..
A
B
I
was
hoping
that
we
could
get
the
delete
kept
in,
but
it
seems
like
it's
sold
again.
I
had
a
good
call
with
tim
hawkin
and
we
brainstormed
a
bunch
of
different
ideas.
So
if
you
haven't
seen
that
email
yet
go
check
it
out
on
the
k,
dev
mailing
list.
A
Definitely
the
delete
is
one
one
of
the
topics
that
is
on
my
list
since
monday
and
it's
wednesday
already
and
the
time
flies
by
and
I'm
still
I'm
still
missing
a
lot
of
stuff
okey-doke.
Anyone
else
have
any
topics
that
I
want
to
call
out
for
123.
That
requires
our
as
in
sean,
eddy
or
myself.
B
The
issue
is
that
we
require
image
as
a
mandatory
flag,
but
the
way
that
the
json
7396
rfc
merge
stipulates
is
that
if
the
patch
is
anything
other
than
an
object,
the
result
will
always
be
to
replace
the
entire
target
with
the
entire
patch.
B
Also,
it
is
not
possible
to
patch
part
of
a
target
that
is
not
an
object,
such
as
to
replace
some
of
the
values
in
an
array,
and
so
the
issue
here
is
this
user
is
trying
to
merge
in
their
limits
for
resources
into
the
containers,
which
is
an
array,
so
it
actually
blows
away
the
image
flag
that
they
specify.
So
the
only
way
for
this
to
work
is
for
them
to
actually
specify
the
image
in
their
json
patch
and
anything
else
that
is
like
set
by
default
or
otherwise.
D
Do
we
know
the
history
of
why
this
particular
patch
type
is
being
used
instead
of
the
strategic
merge
patch?
That
would
work
as
expected.
A
Overrides
introduced,
I
would
probably
check
when
the
overrides
were
introduced,
but
my
my
first
reaction
was
pretty
much
similar
to
katrina's,
as
in.
Let's
try
to
figure
out
to
a
patch
that
will
be
much
more
flexible
than
the
current
one.
A
It
seems
somehow
more
reasonable,
especially
that
we're
pushing
a
lot
of
folks
away
by
not
adding
additional
flags
but
showing
them
that
they
should
be
using
overrides
instead,
I'm
guessing
that
the
reason
that
what
we
picked
originally
the
the
current
patch
is
was.
It
was
because
that
one
was
the
simplest
one
and
the
first
implemented
and
nobody
changed
it.
Since
then,
that
would
be
my
first
gut
feeling
and
probably
we
would
have
to
double
check
with
the
code.
B
B
A
Seems
the
most
reasonable,
because
that
would
basically
mean
we
would
come
forward
towards
users.
I
wonder
my
only
worry
would
be
backwards.
Compatibility,
whether
the
strategic,
merge
patch
would
or
could
work.
B
A
So
if
that's
doable,
then
that's
probably
what
I
would
suggest
us
doing
in
the
worst
case,
I'll
be
even
inclined
to
specify
the
name
of
the
merge
type
as
the
first
as
the
prefix
from
the
actual
contents
alternately
introducing
a
new
flag
which
would
say
which
merge
type
and
by
default
we
would
pick
the
current
one
more
flags
yeah,
but
only
one
and
then
rather
a
general
usage
one.
No,
I
I
get
you.
B
I
I
would
like
to
see
these
commands
go
away
at
some
point,
but
I
know
that's
never
gonna
happen,
which
one
that
keep
cuddle
run
or
run
yeah
create
it's
just.
It's
like
jordan,
said
a
while
ago,
like
representing
structured
information
via
flags
is
just
no
bueno
like
it's.
It's
really
hard
to
maintain.
A
Well
that
nicely
reminds
me
the
question
of
how
is
your
work
towards
having
a
generic
create
command
off
of
open
api.
B
A
Yeah
I
mean
I
I'm
already
getting
complaints
about
removing
the
majority
of
the
run
functionality.
A
But
I'm
guessing
it'll
take
a
significantly
longer
time
to
switch
people
from
run
and
create
towards
something
more
like
generation
type
of
thing.
Yeah.
B
And
this
person
they
I
mean
they
do
have
a
valid
use
case.
I
understand
why
they're
doing
it,
they
said
we
leveraged
this
feature
to
give
our
learners
proof
of
implementation
style
tests.
Examples
show
that
your
network
policies
are
in
place,
spin
up
a
one-time
shell
in
the
namespace
and
then
execute
a
command
to
see
the
results
so.
D
Here's
the
second
problem
highlighted
in
this
issue,
much
more
minor,
obviously,
but
it
the
flag,
is
claiming
that
it
doesn't
work,
but
it
actually
does
work,
and
there
are
a
couple
that
are
that
are
like
that.
According
to
the
issue,
yes,.
A
Thank
you.
I
forgot
about
that.
Yeah,
that's,
definitely
something
that
we
we
should
make
it
not
work
if
they're,
if
we
mark
them
as
deprecated,
must
have
missed
that
one.
When
so,.
B
A
I
would
remove
the
pass
effect
piece
so
that
the
statement
is
correct
with
the
wording,
especially
that
the
wording
is
out
there,
since,
if
I
remember
correctly
past
release,
I
think
that's
why
I
actually
remove
a
feature
backing
it.
E
Down
so
one
other
quick
thing
I
wanted
to
note,
is
we
we
have
our
facts.
The
frequently
asked
questions
that
are,
you
know
typical
problems
and
I'd
like
to
see
a
fact
if
we
don't
already
have
one
that
basically
points
to
apply
as
being
the
the
command
that
we
suggest
for
creating
objects,
and
so
if
this
doesn't,
this
won't
solve
this
particular
problem.
E
But
I
think
that
if
part
of
the
problem
is
people
don't
understand
that
we
are
really,
we
would
much
rather
see
apply
to
actuate
resources
in
the
api
server
than
run
or
create.
Then
at
least
we
could
point
to
that
as
a
fact
and
say
also,
did
you
realize
apply
as
the
best
practice
for
creating
objects
at
the
api
server.
A
Right,
although
the
the
most
important
distinction
between
apply
and
create
or
run,
is
that
the
latter
are
basing
all
their
knowledge
on
the
flags
apply
works
only
on
yeah.
You
need
to
have
a
fully
fully
created
file
first,
to
be
able
to
create,
whereas
run
and
create
commands,
and
I'm
talking
create
sub
commands,
maybe
to
be
more
specific
because
create
the
row
create
where
you
pass.
The
file
name
is,
would
be
and
should
be
replaced
by
the
apply
usage
in
99
of
cases,
but
for
the
sub
commands
and
run.
A
E
Yeah,
I
think
our
a
lot
of
our
documentation
uses
run
because
it's
just
so
simple
to
get
something
in
the
api
server,
and
I
think
that
there
may
be
some
confusion
that
that's
actually
what's
recommended
when,
in
fact,
it's
basically
a
toy.
B
Let
me
let's
say
the
action
item
so
action
item.
I
will
pr
removing
that
confusing
wording
about
has
no
effect,
and
then
we
will
investigate
doing
a
6902
patch.
There.
A
Hold
on
the
action
item
would
be
to
not
remove
the
wording
to
remove
the
actual
changes,
because
the
flag
is
deprecated,
so
I
would
prefer
the
wording
be
accurate
with
what's
going
on,
because
we
will
be
removing
the
functionality
right
specifically.
B
A
Was
I
was
more
leaning
towards
actually
remove
the
the
functionality
so
that
the
wording
is
correct
with
with
the
deprecation,
because
the
deprecation
warning
is
there
since
121
so
for
at
least
two
releases?
A
A
B
F
What
was
the
decision
there
to
change
it
to
use
strategic,
merge
patch
or
was
it
to
add
a
add,
a
prefix
that
would
to
the
content
that
would
say
which
type.
A
Whichever
is
easier,
whether
a
separate
flag
allowing
you
to
specify
which
patch
type
or
or
a
prefix,
somehow
I'm
leaning
towards
separate
flag,
because
that's
just
simpler
and
then
make
it
explicit
in
the
dock
that
you
can
specify
a
different
patch
type.
But
you
need
to
set
this
flag
explicitly.
D
A
A
You
know
in
the
in
the
zoom
settings
in
the
audio
pane,
there's
a
check
box
below
your
mic,
setting
which
says
automatically
adjust
the
microphone
volume.
D
D
Okay
right,
so
this
issue
wants
to
add
the
option
for
cuphead
to
set
the
ca
field
in
a
tls
secret,
and
my
main
question
about
that
is
that
tls
secret
is,
is
documented
as
having
certain
specific
fields.
There's
built-in
validations
on
the
two
fields
that
it
has
documented
today,
because
both
are
required.
D
This
ca
field
appears
to
be
fairly
widely
used,
mostly
not
in
first
party
things,
but
fairly
widely
used
in
ingress.
For
example,
you
can
have
arbitrary
fields
in
secrets,
so
it's
not
like
this
is
inherently
canonical.
D
So
the
question
for
me
here
was
like,
if
we
add
a
flag
to
cubecuttle,
to
make
it
support
this
field,
then
we're
de
facto
making
that
a
canonical
field
of
the
tls
secret
type.
So
I
was
wondering
if
there's
anything
we
needed
to
do
in
advance
of
taking
that
step.
Eddie
brought
up
the
fact
that
we
would
probably
make
a
constant
to
reflect
the
that's
just
in
the
code,
but
that
seems
like
a
very
small
detail.
We
probably
adjust
the
documentation
as
well.
I
imagine
to
reflect
that.
D
D
A
B
A
And
katrina
and
jeremy
are
like
oh
yeah,
there's
jeremy,
two
related
customized
caps.
You
want
to
talk.
D
About
this,
this
is
something
that
I'm
really
excited
about.
We
have
two
caps
that
are
both
loosely
around
customizes
plug-in
story.
Right
now,
customize
has
a
whole
bunch
of
different
alphas
for
using
plug-ins
and
the
other
one,
not
the
one.
That's
up
on
the
screen
right
now
has
a
bunch
of
history
about
what
these
different
mechanisms
are
and
and
how
they
work
or
don't
work
today.
D
Some
of
them
are
available
in
cube
kettle,
and
some
of
them
are
not
so
the
the
overarching
proposal
is
sort
of
giving
that
background
and
explaining
a
path
forward.
That
would
allow
us
to
have
one
unified,
plug-in
story
that
we'd
feel
comfortable
promoting
for
folks
and
that
customize
itself
could
even
potentially
use
and
significantly
and
the
reason
for
including
it
in
the
enhancements
repo,
even
though
this
doesn't
follow
the
release
cycle
is
that
we
would
like
to
propose
something
that
folks
feel
comfortable,
including
in
cube,
cuddle
someday
once
it
graduates.
D
So
we
want
to
make
sure
that
we
have
buy-in
on
the
long-term
plan,
even
though
certainly
it
won't,
it
will
go
through
alpha
that
will
be
in
a
customized
stand-alone.
So
it's
not
going
to
show
up
in
cubepedal
tomorrow
or
anything,
but
we
want
to
make
sure
that
the
plan
that
we
have
will
will
be
acceptable
in
cube
cuddle
sunday.
D
So
there's
the
one
cap,
that's
the
over
the
overarching
proposal
and
then
the
catalog
cap,
which
I
believe
is
yeah
20.
The
one
on
the
left
have
that
one.
It
goes
into
a
lot
more
detail
about
a
key
piece
of
the
proposal
that
is,
on
its
own,
a
very
powerful
feature
for
for
customized
plugins,
so
that
that
one's
jeremy
so
I'll,
let
jeremy
say
a
few
words
about
that.
G
Sure
this
is
kind
of
based
off
of
some
work.
G
We've
been
doing
like
in
an
internal
prototyping
tool,
but
it's
really
a
way
of
facilitating
discovery
and
distribution
of
of
these
of
plugins
both
for
containerized
things,
but
also
looking
towards
how
we
might
also
support
things
like
exec
plugins,
for
places
where
you
might
not
be
able
to
run
containers,
and
I
think
that
it's
interesting,
because
it
allows
us
to
provide
a
mechanism
where
we
can
build
kind
of
an
ecosystem
of
plug-ins
for
folks
and
those
can
be
extensible,
enterprises
can
build
their
own.
G
G
Hopefully
today
I'll
push
an
update
with
a
lot
of
the
addressing
some
of
the
comments
that
have
been
pushed
through
so
far,
but
like
at
a
really
high
level.
It's
defining
a
new
resource
that
essentially
is
an
enumeration
of
plugins
that
can
be
consumed
with
identifying
information
like
where
they're
at
additional
metadata
and
then
some
additional
things
like
for
executable
plugins.
Maybe
what?
What
platform,
what
architecture,
verification
shots
that
can
go
along
with
that?
And
then
it
also
takes
a
look
at.
D
And
as
jeremy
was
saying,
this
would
solve
problems
not
only
for
people
who
want
functionality
from
customized
that
it
doesn't
offer
today
and
that
we
can't
really
build
in.
But
customize
has
some
issues
with
functionality
that
it
already
builds
in
or
has
built
in
past
and
has
had
to
remove
because
of
the
type
restriction
on
dependencies
that
are
introduced
to
cue
cuddle.
D
For
example,
the
helm
feature.
There's
a
there
was
a
hash
core
url
go
getter
feature
that
enabled
a
bunch
of
different
remote
urls
and
the
starlark
runtime
for
functions.
D
All
three
of
those
things
are
problematic
and
could
be
provided
more
easily
as
third-party,
especially
the
former
two
plugins
or
as
built-ins,
that
aren't
compiled
into
customized
that
are
provided
through
some
sort
of
official
catalog.
That's
out
of
doing
actually
doing.
That
is
out
of
scope
with
the
cap,
but
I
think
it's
a
powerful
possibility
that
it
also
inherently
introduces
another
thing.
That's
of
note.
Jeremy
mentioned
that
we
are
getting
some
good
feedback
from
the
kept
folks
like
kpt.
D
It
sounds
like
kepler,
when
I
say
like
ke
pisa,
just
to
clarify
and
anyone
who
is
working
with
the
configuration
function,
specification
which
is
detailed
in
the
customized
repo
and
followed
by
both
kml
and
customized
transformers
and
kept.
It
could
be
interested
in
this
proposal
because
it
is
based
around
it
like
the
or
both
of
these
proposals.
D
In
fact,
the
proposal
for
plug-in
graduation
story
is
proposing
consolidating
on
on
the
functions
specifications
on
the
plugins
that
follow
the
function
specification
because
I
think
we're
going
to
get
the
most
leveraged
there
and
and
they
have
a
lot
of
benefits
as
well
in
terms
of
how
they
work
and
their
ability
to
work
with
this.
D
D
It's
the
size
of
them.
If
I
had
to
identify
one
overarching
region,
a
bunch
of
those
yeah,
it's
staging
that
there's
there's.
There
are
different
problems
with
different
dependencies.
The
ones
that
I
mentioned
in
particular,
like
the
hash,
go,
give
her
get
her.
If
I
remember
correctly
and
also
helm,
they
they
have
they,
they
import
a
lot
like
to
bring
in
that
piece
of
functionality.
D
You
you
just
you
have
to
get
a
lot
more
than
you
would
expect
in
terms
of
the
code
that
you're
bringing
in
and
just
the
volume
and-
and
some
of
some
of
the
details
in
particular
are
are
problematic
for
inclusion
in
in
a
tool
like
cuddle.
D
We
just
don't
want
to
bring
all
that
in
so
right
now
it
they've
been
extracted
and,
like
the
helm,
relies
on
the
user
to
install
the
helm
binder
themselves
and
to
provide
extra
flags
to
allow
us
to
shell
out
to
it,
and
the
go-getter
was
wholesale
removed
when
we
upgraded
to
customized
4.0,
so
yeah
they're
problematic
it
because
of
the
extent
of
them.
B
A
Scope
of
the
problem
and
the
other
one
was
the
licenses
of
the
transitive
dependencies
that
were
being
brought
in
which
were
not
compatible
with
the
licenses
allowed
by
the
linux
foundation
for
the
kubernetes
project.
So
we
had
issues
that
some
of
the
dependencies
that
were
that
were
that
were
being
brought
in
were
just
not
compatible.
So
both
the
scope
and
the
license
was
the
main
issue.
E
So
customize
is
imported
into
coupe
control
and
so
any
of
its
dependencies
end
up
being
dependencies
of
coup
control,
and
we
have
strict
limits,
having
spent
significant
amount
of
time,
refactoring
coupe
cuddle
to
put
it
into
staging
and
that
that
entire
process
was
about
cleaning
up
dependencies
so
that
coupe
control
can
be
pulled
out
of
kubernetes.
A
Yeah
eddie
linked
in
the
chat
for
those
interested
in
the
pr
I'm
guessing.
That's
the
bump,
cue
cuddle,
customizing
pedal.
It
has
all
the
details
with
all
the
information.
A
Okay,
I'm
hearing
none
with
that.
I
think
we
can
move
over
to
the
standups
customize
katrina.
I
guess.
D
A
Okay,
sean
clio,
hotels,.
E
Yeah,
so
I
just
wanted
to
spend
two
minutes
to
bring
the
cigs
and
the
community's
attention
to
some
of
the
applied
library
development
that's
been
going
on
in
the
cli
utils,
and
so
this.
This
link
that
I
put
there
is
the
closest
that
exists
to
documentation.
So
you
know
apologies
for
not
having
much,
but
I
thought
it
was
much
more
important
to
break
to
make
sure
that
we
surface
this
sooner
rather
than
later.
E
It's
basically
a
wrapper
around
coupe
control
apply,
and
so
I
just
mentioned
earlier
how
we
spent
a
significant
amount
of
effort
into
refactoring
coop
control
to
put
it
into
staging
and
that
allows
us
to
reuse,
coop
control
code,
and
so
what
this
is
is
kind
of
an
experiment
to
to
see
how
we
can
reuse
the
control
apply.
E
And
so
this
library
wraps
food
control,
apply
and
there's
some
a
couple
of
experiments
to
to
add
some
functionality
to
coop
control,
apply
like
like
so
coupe
control,
applied
prune,
is
problematic
and
has
many
issues
it's
been
alpha
for
years
and
so
there's
an
experiment
to
to
try
to
implement
a
different
type
of
pruning.
E
But
anyway,
the
yeah,
the
tldr,
is
just
that
there's
some
apply
library,
development
and,
if
you're
interested.
If
that's
your
jam,
please
contact
me
and
the
next
time
we
actually
do
have
some
some
decent
documentation,
then
I'll
bring
it
up
again
and
then
you'll
be
able
to
actually
see
some
documentation
instead
of
just
some
some
code.
E
If
there's
enough
questions
I'll,
let
nick
talk
about
kui.
C
E
Yeah,
it
is
actually
so
the
the
kept
live
apply,
imports
parts
of
this
library.
H
Hello
yeah,
two
quick
things.
I
realized
just
right
now.
The
link
I
put
there
for
this,
for
the
survey
is
not
correct.
Don't
fix
that!
I'm
sorry
about
that
piece
of
the
wrong
and
the
link
is
valid.
That's
actually.
If
you
want
more
detail
on
the
study,
you
can
read
that,
but
that's
probably
boring
so.
The
first
thing
is
that,
as
far
as
remember
cora
coleman,
we
had
a
summer
intern
who
was
doing
some
ux
research
into
all
the
coup
cuddle
tooling.
H
Basically,
an
open,
open
shift,
console
coupe,
cuddle,
coulee,
trying
to
figure
out.
We
can
craft
a
story
around
that,
and
so,
as
part
of
that,
we
thought
we
would
we.
We
thought
this
would
already
been
prior
art,
but
a
study
of
what
tools
do
people
actually
prefer
to
use
for
various
kinds
of
tasks?
You
know
crud
tasks,
debugging
tasks,
monitoring
tasks.
H
We
thought
that
would
be
a
published
survey
on
this
already.
What
people
actually
for
use
group
cuddle
and
we
couldn't
find
anything-
the
cncf
red
hat.
We
couldn't
find
any
existing
studies.
If
you
guys
know
of
any,
let
me
know,
but
what
we
have
started
is
in
place
in
the
absence
of
that
a
study
of
asking
this
question.
So
it's
a
quick
two
minute
survey
we've
contacted
sid
usability
yesterday
to
see
whether
we
could
sort
of
publicize
this
under
their
auspices.
H
They
seem
interested
I'll
fix
the
link,
if
you
guys
have
time,
maybe
submit
your
preferences.
The
hope
is
that
we
can
start
driving
some.
Some
actual,
you
know
some
actual
improvements
based
on
people's
hardcore
preferences.
Second,
quick
thing
is:
we'll
represent
cuya
devx.
I
will
be
there
virtually
not
in
person,
but
I
have
to
finish
up
the
video
by
saturday.
So
just
a
quick
thing,
that's
all.
A
E
E
E
But
to
be
honest,
I'm
not
aware
of
a
whole
lot
of
work
done
elsewhere
to
surface
those
headers.
H
One
of
the
things
that's
driving
this
is
that
you
guys
have
may
have
different
experience
in
your
companies,
but
experience
I've
had
with
the
companies
I've
been
with.
Is
that
from
a
design
perspective,
there's
a
huge
amount
of
effort
put
behind
the
consoles
for
for
good
reasons.
Those
are
definitely
valid
use
cases,
but
what
we
tend
to
find
is
that
the
the
more
seasoned
you
get
with
kubernetes,
the
more
you
gravitate
towards
the
cli
and
so
that
that's
an
anecdotal
observation,
so
we
wanted
to
see
whether
that
was
substantiated
in
the
large.
H
So
that's
was
part
of
the
motivation
for
the
survey.
If
it
indeed
is
like,
if
anecdotal
evidence
is
and
plays
out,
I
think
hopefully
that
can
help
inform
some
of
our
companies
in
terms
of
war
to
start
directing
effort
we've.
I
think,
as
you
can
see
by
the
size
of
this
community
here
right,
it's
pretty
small,
but
but
it's
but
relative
to
the
size
of
the
community.
It
supports
it's
a
very
it's.
It's
an
interesting
ratio
relative
to
the
ratios.
H
B
H
Surprised
I
read
their
survey
asked
a
lot
of
good
questions,
but
it
asks
nothing
at
the
level
of
detail
of
what
you
actually
the
daily
driver
questions
right.
What
do
you
actually
use
from
a
tooling
perspective
and
nothing
even
really
close
and
there's
there's
2020
like
the
latest
one?
I
found
very
surprising.
A
I
mean
honestly,
if
you
think
about
it,
cube
cuddle
from
the
cube
cuddle
standpoint
or
literally
any
other
every
gui,
whether
that
be
a
web-based
or
a
standalone
application.
A
At
some
point
in
time,
you
will
want
to
regain
the
control
for
the
cost
of.
Yes,
it
will
take
me
a
little
bit
more
typing,
but
if
I'm
spending
90
of
my
time
with
my
with
my
keyboards
and
my
console
window,
I
somehow
lean
towards
using
the
cli
tools.
There's.
H
Definitely
that
that
that
control
part
of
it,
which
is
really
important,
I
think,
interestingly,
they
found
core-
also
ran
a
second
study
which
we'll
probably
we'll
get
this
all
up
there
very
soon.
But
the
second
study
was
trying
to
actually
track
trying
to
quantify
those
kinds
of
questions
about.
H
Is
the
preference
more
a
question
of
you
know,
sort
of
a
cerebral
question
of
control.
You
know
grabbing
by
control,
which
is
an
important
one,
or
is
there
actually
is
coop
kubernetes
kukato,
actually
somehow
more
optimal
from
from
from
an
actual
sort
of
effort
perspective?
H
So
I
don't
think
much
of
you
guys
actually
use
consoles,
because
probably
none
of
us
use
them
very
much
because
we're
all
on
the
season
end
of
the
spectrum.
But
if
you
actually
try
to
use
something
like
the
openshift
console
for
some
fairly
straightforward
tasks,
the
number
of
clicks,
the
number
of
actual
operations
you
have
to
perform
to
do
fairly
straightforward
tasks
is
pretty
high
and
actually
higher
than
what
you
have
to
do
for
couple.
H
So
it's
sort
of
an
interesting
twist
that
anyway,
so
I
just
want
to
inform
the
develop
the
design
community
because
I
think
they're
they're
operating
kind
of
in
a
bit
of
a
their
own
little
silo.
I
just
wanted
to
bring
them
into
this
discussion.
I
think
these
kinds
of
surveys
would
help
that
anyway,
I've
talked
enough.
A
Yeah
I'll
be
very
curious
about
learning
the
the
results
from
the
survey
both
from
personal
interest
and
and
from
the
sexy
life
perspective
as
well.
H
Great
yeah,
so
I
linked
the
survey
there.
I
think
I
updated
the
the
in
the
meeting
notes
as
well
so
and
I
think
tasha
drew
from
cpu
music
ability
said
we
can
blog.
We
can
get
some,
maybe
maybe
get
some
extra
publicity
and
not
it
on
that
end
from
the
top
down.
If.
A
A
That,
apparently,
she
is,
but
that
still
I
think
we
we
should
be
able
to
to
push
the
the
survey
even
further
outside
of
the
the
kubernetes
community.
H
I
That
is
like
I
came
from
a
you
know.
I
came
from
a
fairly
large
company,
you
know
as
a
consumer
of
kubernetes
and
openshift,
and
the
one
thing
that
I'll
say
having
been
part
of
establishing
some
of
the
training,
for
you
know
the
people
within
the
organization
to
use
it.
I
The
consoles
are
really
good
for
for
new
people
coming
in
and
like
trying
to
understand
the
concepts
and
sort
of
the
structure.
You
know
how
the
different
pieces
fit
together,
but
yeah
as
soon
as
you
know,
as
soon
as
everybody
understands
that
it's
immediately
go
and
use
google.
You
know
because
everything,
because
you're
just
going
to
have
to
use
that
anyway,
to
do
any
sort
of
automation,
so
you
might
as
well
just
get
used
to
it
and
that's
what
everybody
got
trained
on
for
the
most
part.
H
Yeah,
just
even
from
my
from
a
learning
perspective,
there's
some
really
interesting
ironies
here
that
the
best
way
to
learn
is
often
by
mimicking
right
by
copying
and
pasting
someone
else's
and
you
get
copy
and
paste
into
a
graphical
console.
Sorry,
that's!
I
think
we.
I
really
wanted
to
get
substantiate
some
of
these
anecdotal
evidence,
so
we
can
actually
sort
of
drive
some
richer
discussions
with
these
console
teams
to
see
whether
I
don't
know
anyway.
I
think
that
investments
are
inordinate
with
the
return.
So
that's
personal
opinion,
but.
B
Natasha
reached
out
to
the
replicated
folks
who
own
cube,
cuddle,
I
o
and
customize.
I
o
so
we're
on
an
email
thread
with
them
and
they
are
going
to
be
transferring
those
to
the
cncf.
So
I'm
working
with
bob
to
get
the
service
desk
ticket
put
in,
but
we
will
soon
have
those
domains.
D
A
Okay,
thank
you
very
much
friends.
That
was
another
first
film
meeting
from
six
diy,
see
you
next
week
for
blocker
back.
If
I
remember
correctly,
eddie
is
running
the
blocker
path
next
week.
Is
that
so
yeah
that's
about
right
time?
Okay,
see
you
next
week!
Thank
you
very
much.
All.