►
From YouTube: SIG Usability 20190910
Description
See our SIG page for more information: https://github.com/kubernetes/community/tree/master/sig-usability
A
A
So
I
have
a
bunch
of
things
here,
because
I've
been
digging
their
stuff
and
doing
a
bunch
of
thinking.
The
first
one
is
I
want
to
kind
of
bring
up
if
we
want
to
put
secured
defaulting
and
like
friendly
security
as
something
on
our
radar,
because
that's
something
that's
a
very
high
impact
for
the
average
user.
For
example,
if
you
don't
have
a
cloud
security
policy
set
up,
which
is
not
something
it's
easy
to
set
up,
anyone
can
just
take
take
over
your
cluster,
so
that's
kind
of
an
area.
B
B
B
Basically,
the
way
I
was
looking
at
it
was
I
originally
came
at
this
from
a
multi-tenancy
perspective,
so
it
was.
If
you
have
a
secure
cluster
and
you're
shipping
that
to
people
it
fails
conformance
tests,
because
the
conformance
tests
require
that
your
that
your
kubernetes
cluster
be
pretty
insecure
in
order
to
pass
conformance.
And
so
what
I
was
proposing
was
that
we
put
together
a
set
of
profiles
for
how
you
want
to
run
kubernetes
clusters,
and
so
we
could
have
like
the
regular
conformance
profile
and
then
we
could
have
additional
profiles.
B
So
one
could
be
a
secure
single
tenant
cluster
profile
and
then
one
could
be
a
secure,
multi
tenant
cluster
profile,
and
these
would
be.
You
know
just
basically
like
profiles
that
we
circulate
among
the
SIG's
and
working
groups,
and
then
we
have
the
opportunity
as
part
of
the
work
in
the
multi-tenancy
group
and
anything
that
we
did
here
as
well.
To
work
with
the
security
researchers
so
then
pass
over
our
proposed
secure
defaults
to
them
so
that
they
could
penetrate
tense
them
and
then
give
us
the
results.
B
And
then
we
could
iterate
on
those
profiles
and
so
I.
Those
are
kind
of
the
steps
that
we
had
lined
up
as
something
out
of
the
multi-tenancy
group,
and
we
have
to
sort
of
very
rough
drafts
that
we
would
love
people
that
poke
holes
in
and
suggest
how
to
improve.
As
we
make
those
initial
steps.
But
it
would
be
awesome
if
we
wanted
to
then
extend
that's
like
coo
baby
M,
so
that
everything,
but.
A
Like
step,
one
of
making
sure
that
there's
a
way
to
get
conformance
passes
would
be
good
but
I
think
longer
term.
If
we
go
down
this
road,
we
need
to
look
at
changing
like
the
base
conformance
stuff
like
resemble.
It
would
be
awesome
if
pod
security
policy
was
default,
applied
like
network
security
policy
default
applied,
and
that
was
the
expectation.
A
B
Yeah
I
totally
agree.
There
has
been
some
suggestion
that
people
might
not
be
open
to
having
different
conformance
profiles,
but
I
think
that
I
feel
like
there
is
demand
for
this
in
the
community
to
have
secure
by
default
setup,
and
the
thing
that
comes
up
is
that
there's
a
lot
of
different
ways.
You
could
approach
this,
but
I.
Don't
think
any
of
us
are
saying
there's
only
one
way.
We're
just
saying
here
is
a
way
and
then,
if
you
want
to
suggest
a
different
way
than
let's
add
that
to
the
list,
you
know
yeah.
A
B
B
A
Yeah
so
next
thing
I
had
on
the
agenda
is
I
wanted
to
look
at
like
a
very
forming
a
very
alpha
version
of
some
kind
of
usability
review.
So
this
is
something
where
we
could,
like
maybe
jump
in
at
the
provisional
cap
stage,
and
offer
some
advice.
So
this
isn't
something
that
we're
going
to
be
formally
doing
but
I'd
like
to
kind
of
have
one
or
two
ideas
of.
If
we
did
it,
what
would
it
look
like,
so
we
could
potentially
narrow
it
down
more
than
actually
propose
something.
A
A
A
A
I'm
not
sure
like
coming
from
more
the
engineering
side
of
things,
I
think
my
concern
is
sometimes
lack
of
bird's-eye
view
when
it
comes
to
system
design.
But
it's
a
little
bit
funny
because
that
falls
partially
under
the
purview
of
a
couple
groups
like
cigar
ch
and
in
particular,
like
API
conformance
yeah,.
D
Yeah
I
think
then
it
definitely
like
that's
definitely
like
a
really
good
idea
and
that,
if
and
kind
of
like
as
a
prerequisite
to
that
it
would
almost
be
like
do
we
know
what
journeys
were
focusing
on
that
or
maybe
like
particularly
common
or
or
you
know,
you
know,
deployment
like
the
the
journey
to
deploy
something
to
communities
is
I
know
that
came
up
a
lot
during
the
last
coop
con
that
it
was
something
that
people
found
challenging
to
do
so.
It's
kind
of
like
yeah,
like
kind
of
like
which
one
do
we
choose.
D
D
It
was
like
the
tough
questions
on
Stack
Overflow
and
they
they
came
out
with
like
here,
a
comment,
app
developer,
related
questions,
and
here
a
common
administrator,
related
questions
and
kind
of
like
the
top
four
buckets
under
each
one,
and
just
like
off
the
top
of
my
head.
I
think
app
developer
was
like
deployment
communication
so
like
how
do
I
expose
an
app
and
then
also
dev
tools
and
persistence
so
like
how
do
I
hook
up
storage
kind
of
like
with
the
little
asterisk
there
that
these
are
all
related
see
getting
started?
D
You
know,
so
it's
a
it
is
kind
of
leaning
towards
that,
but
that
it
was
an
interesting
quantitative
analysis
that
if
we
wanted
to
kind
of
like
say
we
wanted
to
do
a
usability
review
of
this
particular
journey
like
deployment
like
that.
That
could
give
us
something
to
get
started.
I.
E
E
Tasha
I,
like
Gabi
what
you're
saying
about
like
using
the
stories
to
drive
the
usability
reviews,
perhaps
another
way,
I,
don't
know
if
it
makes
it
easier.
This
complicates
everything
but
another
way,
maybe
to
look
across
stories
and
think
about
evaluation
of
the
taxonomy
language
used
if
that
works
really
for
the
API
or
if
it's
more
for
the
for
the
tools
that
they're
crazes,
like
you
cuttle,
or
something
like
that.
But
this
this
you
know
any
I
think
it
could
cuddle
just
because
it's
a
very,
very
popular
tool.
E
A
I
know
that
six
Eli
is
working
on
getting
some
kind
of
telemetry
from
coops
e-tail,
but
I
don't
know
how
well,
especially
with.
However,
they
format
it
or
anonymize.
It
would
translate
to
having
user
stories
like
internally
we're
trying
to
work
on
something
similar
where
we
kind
of
get
like
a
history
of
someone's,
more
or
less
good
detail
commands
internally.
We
don't
have
like
the
same
security
concerns,
because
that
way
we
can
kind
of
easily
see
like
people
are
trying
to
do
one
functional
action,
but
it's
split
over
end
commands
that
kind
of
thing.
A
Yeah,
that
would
be
a
good
thing
to
follow
up
with
sixty
likes
I.
Think
oopsie
tail
is
probably
like
a
very
good
point
to
look
at
because
it's
a
very
discreet
surface
and
it's
because
usually
it's
humans
using
it.
So
we
don't
necessarily
have
to
filter
out
the
signal
of
bots
or
something
if
that
would
be
a
data
source
that
we
want
to
look
out.
D
Was
just
any
questions
that
were
attacked
kubernetes
on
Stack
Overflow,
so
you
know
that
the
talk
didn't
really
dive
into
like
here,
like
the
raw
questions
it
was
kind
of
like
here
are
just
the
general
categories
and
maybe
like
a
little
bit
about
them
like
where,
like
the
specific
questions
about
deployment
that
they
scraped
on
Stack
Overflow,
but
it
didn't
go
into
like
that.
Much
detail.
A
A
A
D
Share
the
Stack
Overflow
talk,
he
has
his
slides,
the
guy
that
gave
it
slides
in
the
in
the
talk.
I
think
that
that
would
be
a
really
interesting
because
I'm,
just
remembering.
Like
the
developer
questions,
there
were
also
the
administrator
questions,
but
that
might
be
interesting
to
look
at
and
to
see
if
there's
anything
worth
pulling
out
from
there,
where
we're
like
we,
we
could
do
a
usability.
We
could
pick
one
of
the
buckets
and
try
and
do
a
usability
review
of
one
of
those,
for
example,
but
yeah
I'm
happy
to
share
that
afterwards.
D
B
A
D
C
A
A
A
D
Yeah
so
I
can
I
can
share
my
screen
real,
quick,
just
to
show
the
proposal
that
mentioned
last
time.
D
So
this
proposal
about
the
survey,
this
is
kind
of
a
cool.
What
was
pitched
at
our
last
meeting
I
want
to
say
that
almost
maybe,
like
a
couple
weeks
afterwards,
I
got
an
email
about
a
about
a
survey
that
CNCs
put
out.
That
was
actually
a
very
it
was
very
similar.
D
It
was
a
survey
about
taking
the
pulse
of
the
community.
I
am
actually
newer
to
the
current
ideas
open
source
community,
so
maybe
some
of
y'all
are
kind
of
like
already
familiar
with
the
survey
and
are
like
yeah.
Of
course,
that's
like
the
thing
that
goes
out
every
year,
but
for
me
at
least
it
was
like
super
it.
D
It
kind
of
made
me
wonder
if
there
was
a
way
to
maybe
yes
like
partner
a
little
bit
more
with
CNCs,
because
I
thought
it
was
a
pretty
in-depth
survey
and
they
linked
to
like
previous
surveys
as
well
as
I
know
like
one
of
the
inspirations
for
for
putting
out
like
this
initial
user
survey
proposal
was
the
the
Stack
Overflow
developer
survey
that
was
linked
to
here,
which
has
like
demographic
information
about
developers
and
just
kind
of
a
behavioral
stuff,
but
also
just
kind
of
also
it's
also
similar
posts
of
the
community
and
with
the
CN
CF
survey.
D
Almost
like
a
little
bit
of
a
bigger
question
like
what
would
we
be
interested
in
learning
and
kind
of
like
what
has
already
been
learned
through
this
CNCs
survey
as
well,
so
yeah
and
then
the
other
thing
that
I
that
I
did
was
there's
a
Thursday
firm,
called
strategy
that
has
kind
of
an
approach
to
their
own
approach,
to
a
research
called
jobs
to
be
done,
and
so
I
had
reached
out
to
them
and
they
had
also
kind
of
recommended.
D
A
D
A
F
One
thing
I
want
to
try,
and
that
is
I
think
what
Gabby
said
before
about
user
stories
and
focusing
on
like
specific
use
cases
and
then
kind
of
looking
into
this
survey.
So
I
think
those
two
can
go
hand-in-hand.
If
we
identify
what
you
want
to
focus
on
in
terms
of
usability,
we
can
use
that
as
a
platform
to
like
go
for
a
first
survey,
and
we
can
make
this
like
more
of
it.
Shouldn't
be
a
one,
odd
thing:
it
should
be
more
about
like
collecting
users.
F
You
know
that
we
can
communicate
with
regularly
on,
like
you
know,
anything
related
to
kubernetes,
or
you
know,
use
cases
that
we
might
be
building
on.
As
you
know,
different
teams,
and
if
we
identify
what
are
the
common
different
like
workflows
across
all
of
us,
we
can
like
pick
one
and
then
you
know
kind
of
start
dissecting
that
a
little
bit
more
like
for
us
like
say
develop
like
we
kind
of
divide
our
use
cases
into
two
pieces.
F
One
is
the
admin
infrastructure,
you
know
persona
that
is
looking
at
managing
the
infrastructure,
and
then
we
have
the
developer
persona.
That
is
looking
at
deployments
and
running
the
applications
and
for
us
both
of
them
are
really
important
and
and
specially
monitoring
troubleshooting
is
so
important,
so
I
went
to
keep
con
la
sio2
and
I
was,
and
what
I
did
was
I
focused
on
the
monitoring
and
troubleshooting
story,
because
I
wanted
to
kind
of
dig
deeper
into
that
and
kind
of
realize.
F
There's
so
much
out
there
and
you
know,
was
able
to
not
like
it
was.
It
was
so
much
information
but
at
the
same
time
I
was
like
okay,
where
do
I
start
from
like
what
do
I
focus
on
and
I'm
still
trying
to
figure
that
out
and
there's
where
I'm
you
know
kind
of
looking
at
the
community
like
this
Tuukka
Rask,
say:
okay,
let's
focus
here
for
the
12
and
let's
focus
here
for
the
admin.
F
A
Thanks
yeah
you're
you're
kind
of
leading
nicely
into
the
last
thing
I
had
tabled,
which
was
for
those
of
you
who
were
on
like
the
mega
thread
around
whether
or
not
we
want
to
deprecate
coop
CTL
cpu.
And
why
or
why
not?
Someone
raised
a
pretty
good
issue,
which
is
that
we
don't
seem
to
really
have
personas
prioritized
in
the
project
and
in
that
context
it
was
where
do
we
place
the
like
workload
and
flexibility
of
our
developers
upstream
against
features,
because
that
was
a
feature
that
is
very
useful
to
have
as
a
user.
A
B
A
D
B
This
usability
group
I
was
told
that
they
have
a
bunch
of
people
who
are
signed
up
to
give
feedback
on
kubernetes
who
haven't
really
been
used
and
who
would
be
very
eager
to
chat
with
us.
So
from
what
I
understand
that
there
is
a
group
who
have
signed
up
already
to
be
sort
of
interviewed
about
their
experience
with
kubernetes
and
haven't
necessarily
like
actually
gotten
to
do
that.
Yeah.
B
Yeah,
you
know
what
I
think
would
be.
Awesome
is,
if
you
know,
I,
feel
like
we
really
need
to
kind
of
like
show
some
forward
moments.
I'm
on
a
couple
of
these
things,
and
what
would
be
awesome
is
if
someone
wanted
to
volunteer
to
sketch
out
a
couple
personas
that
we
could
collaborate
on
and
also
if
somebody
wanted
to
volunteer
to.
B
Brendan
has
mentioned
that
we
could
have
access
to
funds
if
we
want
to
fund
a
usability
study,
but
I
feel
like
that's
kind
of
jumping
the
gun
a
little
bit,
and
it
would
be
better
for
us
to
thoughtfully
plan
out
what
we'd
like
to
get
out
of
it
and
then
see
what
we
can
do
on
our
own.
And
then,
if
there's
something.
That's
like
very
time-intensive
that
would
be
like
you
know.
We
already
all
have
day
jobs,
then
that
would
be
the
kind
of
thing
we
could
farm
out.
B
B
A
F
Don't
speak
to
somebody
on
my
team
and
see
if
they
are
ready
to
pick
this
up
and
I'll
get
back
to
you
guys
about
it
and
if
I
can
get
them
to
like
commit
and
help
out
a
bit.
You
know
we
could
try
sharing
something
next
week
and,
moreover,
this
is
again
our
understanding
of
our
persona,
but
just
kind
of
you
know
putting
out
there
to
you
know
kind
of
dissect
a
little
bit
more
from
everybody's
perspective,
so
I'll
be
chopped
to
somebody.
Okay,
awesome,
okay,.
A
F
Very
you
had
something
about
icons
that
you
shared
and
I
wanted
to
kind
of
show
something
I
just
kind
of
wanted
to
share
something
with
everybody,
because
at
VMware,
because
we
are
kind
of
exploring
different
products
on
kubernetes.
We
do
have
clarity,
which
is
like
a
design
system
that
we
have
out
there.
Let
me
just
like
put
that
in
chat
and
the
team
has
come
up
with
like
a
few
icons
that
are
open-source,
so
we're
happy
to
like
kind
of
explore
that
a
little
bit
and
I
kind
of
talked
through
that
with
the
team.
F
F
E
F
Think
that's
all
like.
We
have
containers,
pods,
namespace,
nodes,
yeah
and
I
think
as
a
team
you're
still
exploring
things
so
as
when,
like
you
know,
a
new
object
type
comes
up
like
you
know,
we
looking
at.
We
have
data
stores,
of
course,
but
you
know
like
services,
we
have
the
team
helping
us
out
with.
We
have
like
visual
design
of
the
helps
us
out
with
adding
icons
into
this
repository,
so
yeah
I.
A
F
E
So
you
like
panel,
said
I
think
it's
not
a
fully
fleshed
out
set
of
default,
kubernetes
resources-
or
even
maybe
we
should
be
thinking
at
a
high
level
of
maybe
there
should
be
service
icons
for
popular
services
across
the
kubernetes
and
co
system,
but
I
think
to
start
with.
You
know:
one
gap
is:
do
we
have
visual
assets
for
the
common
criminales
assets?
E
Resources?
Is
this
something
that
the
community
can
get
behind
there?
Other
I
know
there
are
there's
another
icon
set
the
news,
but
pretty
certain
other
companies
using
different
icons,
and
would
it
benefit
the
community
to
settle
on
something?
That's
more
centered
about
a
way
of
sharing
things.
Yeah.
F
So,
basically,
the
whole
discussion
about
the
sages
ability
with
Tasha
started
off
with
these
icons,
because
you
were
like
trying
to
see
if
it
kind
of
matches
the
standard
of
what
the
Kuna
is.
You
know
community
wants
and
if
it's
something
that
we
could
open
source
and
like
work
with
the
team,
like
you
know,
other
teams
out
there
to
make
like
a
standard.
So
it
kind
of
started
off
with
that,
but
this
was
definitely
something
less
on
the
priority
list.
When
you
started
talking
about
usability
yeah.
B
If
you
know,
would
it
be
possible
to
suggest,
as
an
upstream
team
like
when
you're
describing
a
pod
or
a
namespace
or
like
all
of
these
different
concepts
in
a
UI?
Could
we
standardize
on
what
those
visually
look
like,
so
that,
at
a
glance,
a
user
knows
what
they're
dealing
with
instead
of
every
project
kind
of
having
a
radically
different
way
of
depicting
them?
Okay,.
B
And
so
the
what
Pamela
and
Alistair
we're
showing
is
an
example
of
an
open-source
site
that
they've
been
working
on,
which
is
just
basically
up
streaming,
a
bunch
of
icons.
If
anyone
wants
to
use
them
and
then
the
question
arose
like,
would
it
would?
Could
the
kubernetes
community
be
interested
in
these,
or
are
there
already
sort
of
standardized
icon
iconography
that
people
are
using
that
they
could
replace
in
there's
four
people
to
leverage
my.
A
B
F
B
A
B
Just
like
slide
it
in
there
yeah
cool,
okay,
well,
I,
think
kind
of
follow-up
action.
Items
from
this
meeting
are:
let's
see
if
we
can
do
a
call
to
just
kind
of
see
if
there's
appetite
for
a
wider
iconography
conversation
and
if
not,
we
could
talk
to
people
about
it
at
KU,
Khan
as
well,
since
I
think
a
bunch
of
us
are
going
to
be
there
and
we're
gonna
have
a
cig
usability
table.
B
A
D
B
And
then,
if
we
have
things
that
don't
fit
inside
that
initial
group,
we
could
start
kind
of
other
groups
of
questions
and
then
use
that
as
well.
But
I
just
think
it's
important
to
like
the
thing
I
run
into
all.
The
time
is
sometimes
you're
asking
for
feedback
from
people
who
aren't
the
right
people
to
be
asking
for
that
sort
of
feedback.
And
then
you
really
don't
get
the
information
like
you
can
get
pretty
skewed
results
that
can.
D
Yeah
I
guess
in
my
mind,
I
had
two
different
categories
of
questions.
One
were
questions
that
might
fit
into
a
survey
to
sent
to
end
users,
but
then
the
other
is
like
the
set
of
questions
that,
like
us
as
a
community,
have
about
end
users,
oh
cool,
so
that
was
kind
of
like
where
my
mind
was
burning
like
there's
those
two
categories:
okay,.
G
Do
you
mind
if
I
jump
in
you
guys
hear
me?
My
name
is
Ron
Norman
I
work
at
Google,
thanks
for,
let
me
join
is
my
first
time.
I
didn't
want
to
speak
up
too
much,
but
there's
a
few
areas
here
that
I
think
our
team
would
would
love
to
partner
in
on
especially
personas.
We've
done
some
persona
work,
as
well
as
some
partnership
with
the
CN
CF
on
the
survey
in
the
past.
G
As
far
as
some
research
questions,
so
I'd
love
to
plug
in
I
didn't
want
to
volunteer
too
much
to
drive,
but
I'd
love
to
partner
with
I
didn't
see
what
our
team
can
help
with
and
then
some
of
the
discussion
on
usability
review.
That
was
really
interesting
and
I'd
love
to
sort
of
compare
notes
and
maybe
give
a
few
thoughts
around
a
certain
method
called
a
heuristic
evaluation
that
some
of
you
are
probably
familiar
with.
G
We've
been
using
in
some
of
our
internal
processes
to
high
affect,
even
at
the
stage
of
reviewing
like
API
definition
and
CLI.
If
we
can
establish
a
set
of
good
UX
principles
and
then
validate
those
principles
with
heuristics,
it's
something
that
could
plug
into
almost
any
type
of
phase
of
a
project,
it
might
be
worth
considering
and
maybe
like
tabling
into
a
different
conversation.
But
is
another
thing
that
could
be
there
and
beneficial
as
another
UX
method.