►
From YouTube: Istio UX Working Group Meeting July 28, 2020
Description
Review UX items still scheduled for completion in 1.7
UX-owned Istio documentation automated test status
Hongyi on Unifying Istioctl Command Output
Should istioctl command like dashboard support “type/name” notation like kubectl ?
A
A
I
think
that
should
be
a
one-line
change
event.
If
that
there's
three
with
environments,
martin
are
you
on.
Can
you
tell
me
if
any
of
these
three
are
done.
B
Yes,
so
first
is
done
the
second
one.
I
think
we
decided
that
we
weren't
going
to
do
anything
further
for
it,
because
the
first
item
sufficiently
covers
what
we
need.
A
B
And
then
the
last
one
that
one
I'm
not
sure
I
know
somebody
is
working
on
it,
but
I
I
don't
know
the
latest
status.
A
B
I'm
happy
to
close
that
one
out
with
some
comments.
If,
if
you
agree,
I
I
think
that's
what
we
were
leaning
into
last
time
that
we
met
with
with
genon
who
was
doing
the
uninstall
work,
and
I
can
update
the
issue
to
reflect
that.
A
Perfect
then,
thank
you
very
much
sure,
so
I
wanted
to
talk
about
testing
first,
I
want
to
talk
about
the
automated
tests
and
then
I
want
to
remind
everyone
that
testing
day
is
tomorrow.
Anyone
who
can
should
test.
So
let
me
tell
you
about
the
automated
tests.
A
Frank
from
documentation
has
taken
every
page
in
the
istio
io
doc
and
assigned
it
to
a
working
group.
So
security
got
17..
We
got
33
of
these
documents
and
I
went
through
them
and
labeled
each
one
with
its
title
and
then
tried
to
figure
out
what
its
sort
of
priority
would
be.
So
what
frank
thinks
that
we
should
do
is
write
is
review
all
those
and
write
automated
tests
for
the
ones
that
we
can.
A
So
I
worked
on
automated
tests
for
analyze.
I
did
this
test
and
if
you
look
it
now
has
a
check
mark
on
it.
It
took
me
several
days
to
learn
the
framework
and
get
it
all
done,
but
all
of
these
commands
are
sort
of
run
automatically
when
the
documentation
is
built,
and
so,
if,
when
that
happens,
it's
great
for
us
because
there's
a
spreadsheet
that
we
normally
do
for
every
release
testing
day
and
we
are
trying
to
do
those
tests.
A
So
here's
the
tests
that
we
have
anything
that
says
needs
automation,
we're
supposed
to
be
sort
of
adding
automation
if
it's
possible
and
we're
supposed
to
be
doing
as
much
as
we
can
to
test
the
pages,
especially
from
our
working
group
and
especially
the
p0s
as
part
of
those
testing
days.
A
So
anyone
who
can
and
has
time
please
write
these
tests.
If
you
need
help
read
this
read
me
or
ping
me
and
I
will
try
to
get
you
started.
A
few
of
these
items
looked
like
they
needed
some
editing.
The
automatic
sidecar
injection
page
screwed
up.
A
A
A
A
If
anyone
has
somebody
and
their
team
was
trying
to
learn
this
to
you
or
something
maybe
having
them
try,
some
of
these
tasks
would
be
a
good
way
to
get
them
on
board
and
get
this
stuff
going.
Okay,
hong
yi
has
prepared
a
document
for
us
about
the
output
of
istio
cuddle
ani.
Can
you
walk
us
through
this
document
that
you
want
to
review.
C
Yeah
sure
so
this
document
in
particular,
will
focus
on
staccato,
pre-check,
verify,
install
and
analyze
so
currently
estimate
analyze
has
a
pretty
good
output
control.
C
However,
the
pre-check
and
verify
install
commands
are
kind
of
are
pretty
unstructured
in
terms
of
their
output,
so
this
dock
is
trying
to
unify
their
output
to
support
basically
three
output
formats.
C
One
is
log
format
which
is
intended
to
be
human,
readable
and
the
other
two
are
json
format
and
yaml
format,
and
so
we
would
like
to
implement
these
output
for
pre-check
and
verify
install
to
be
as
close
to
as
close
to
easter
color
analyze
as
possible.
So
basically
the
output
might
look
like
may.
Would
you
mind
scrolling
down
to
the
section
on
proposed
output.
C
So
for
proposed
output,
maybe
a
little
bit
a
little
bit
more
scroll
a
little
bit
so
basically
yeah
right
here.
So
we
would
imagine
that
the
output
might
take
four
parameters
which
will
basically
show
the
level
of
an
output
message,
the
code
of
a
message,
the
origin,
which
is
where
the
message
origin
is
from,
and
also
the
detailed
message
of
why
this
message
is
emitted.
C
So
this
is,
I
also
put
an
example
here
like
in
istio
analyze.
You
would
have
error,
followed
by
the
code
like
isd,
0,
101,
0101,
that
kind
of
stuff
and
a
followed
by
origin
of
that
message.
So,
basically,
in
this
meeting
we
would
like
to
discuss
the
message
code
thing
because
for
pre-check
and
verify
install,
we
would
also
like
to
add
some
message:
codes
for
them
and
we're
we're
proposing
some
of
the
message
codes
for
these
two
commands
and
json.
Do
you
want
to
take
over
from
here.
D
Yeah
sure,
do
you
mind
scrolling
down
a
little
bit.
I
believe
it's
on
right
here.
Yep
there
you
go
yeah.
I
think
we
would
like
to
propose
that
yeah.
If
you
look
at
the
comment
on
the
right,
so
currently
we
have
0
to
100
for
internal
usage
and
100
to
200
for
config
analysis
which
we're
currently
using,
and
we
would
like
to
propose
that
200
to
300
for
control,
plane
resource
analysis.
D
D
Yeah,
so
so
just
wanna
wanna
emphasize
that
this
is
we
would
like
to.
This
is
part
of
like
kind
of
improving
how
users
see
the
error
messages.
D
The
the
the
command
line
may
not
be
still
like.
The
command
line
may
still
be.
The
separate
commands
is
still
cuddle,
but
the
we're
aiming
at
the
the
unified.
A
D
Yes,
so
I
think,
based
on
the
format
yes
well
well,
documentation.
Url
is
part
of
the
the
analysis,
output
message.
A
A
We
have
been
trying
to
plan
to
divide
the
istio
commands
so
that
some
of
them
that
are
related
to
installing
a
control,
plane
or
checking
out
the
control
plan
are
separate
from
user
facing
commands
to
check
the
status
of
your
istio
resources.
A
D
Right
so
I
think
that's
that's
the
main
like
why
we're
proposing
the
range,
so
I
guess
yes,
they
are,
they
might
be
different,
like
the
the
ranges
are
are
for
different
user
targets.
I
think
I
guess
0
to
100,
probably
right
now
it's,
although
we
don't.
I
don't
believe
we
have
error
messages
right
now
for
for
those,
but
those
are
mostly
for
like
easter
internal
stuff
and
targeted
for
easter
developers
and
then
100
to
200.
D
It's
mostly
for
easter
users
and
basic
the
api
users,
config
users,
and
then,
if
we're
the
two
200
300
will
be
the
the
actual
actually
the
mesh
operator
or
the
let's
say,
the
overall
mesh
operator,
but
more
more
for
the
who
who
installed
the
easter
control
plane
and
maintaining
still
control
play.
A
C
Test
plan,
I
don't
think
I
have
anything
to
add.
A
So
that
may
help
or
interfere
with
this
in
some
way.
D
Yeah,
I
think
it
should
be
fine
right
now
to
just
to
add
some
initial,
because
I
I
think
we
can
just
start
to
add
some
initial
integration
tests
for
these
commands
and
if
we
need
to
converge
later,
we
can
do
that.
A
The
current
tests
for
valid
or
analyze
a
bunch
of
incorrect
crs
that
they
check
over
one
reason
we
haven't
done
a
lot
of
integration
tests
for
pre-check
and
verify
install
is
that
they
we
don't
have
a
good
mocked
up
situation
of
a
cluster
say
that
doesn't
have
support
for
web
hooks
and
things
like
that.
Do
you
think
we
need
mocking
of
the
sort
of
unusual,
problematic
clusters,
or
do
you
think
it's
fine
just
to
have
tests
for
correctness.
D
So
so
in
analyze,
we
do
two
things
right
like
for
specific,
like
user
error
scenario,
we
do
unit
test
so
unit
tests
cover
those
scenarios
and
then
in
integration
tests
we
actually
it
overall
just
tests
out
the
general
functionality
of
the
the
command
line.
D
So,
like
I'm,
not
specif,
not
specific
to
user
scenario
like
we
test
out
if
the
json
output
is
valid,
if
if
the
command
line
can
output
correctly
in
certain
ways,
but
we
don't
test
specific
on
user
scenarios,
so
I
think
we
can
also
stick
to
that
kind
of
separation.
I
haven't
looked
at
the
unit
tests
we
have
for
and
for
for
pre-track
and
verify
install,
but
I
I
believe
here
what
honey
is
talking
about.
D
It's
more
focusing
on
the
integration
test,
where
you
just
kind
of
tried
to
test
out
the
the
easter
cuddle
command.
A
Using
istio
cuddle
last
week
and
the
week
before
to
troubleshoot
some
things,
and
I
was
thinking
that
a
lot
of
the
kubernetes
commands,
let
you
use
deployment
or
service
slash
and
in
the
name
of
the
deployment
or
service,
and
it
picks
a
random
pod.
I'm
wondering
if
we
think
that
istio
cuddle
should
do
the
same
thing.
A
I
would
like
it
and
I
was
sort
of
missing
it.
Do
we
have
any
sort
of
objections
to
it.
A
That
sounds
good
to
me
all
right.
Do
you
think
we
need
to
design
doc
ford,
or
should
I
just
implement
it
and
have
people
come
in
on
the
pr.
A
And
is
there
anything
we
need
to
see
for
next
week?
One
thing
I
really
want
to
do
next
week
is
either
do
or
prepare
for
the
plan.
The
roadmap
for
one
eight,
which
means
we
need
to
identify
candidate
issues
for
one
eight
and
sort
of
prioritize
them.
So
if
you
have
features
that
you
want
for
one
eight,
please
make
an
issue
if
you
think
it
is,
should
be
p,
zero
or
p
one.
A
Definitely
for
one,
eight,
let
me
know,
and
I
will
triage
it
as
such
and
then
next
week
I
can
prepare
a
meeting
with
all
of
the
highly
rated
one
eight
items
and
try
to
see
if
there's
a
sort
of
agreement
on
them.
Everyone
knows
how
zenhub
works
right
with
the
all
right.
Have
people
use,
design
hub
or
people
mostly
using
the
github
stuff.
A
Mix
match
yeah,
so
just
if
you've,
never
if
you're
not
familiar
with
how
we
do
it
in
istio.
There's
these
priorities,
new
issues
and
things
that
our
release
blockers
are
marked,
p0,
etc,
or
release,
blocker
or
or
p
zero.
If
I
don't
know
who
has
access
to
zen
hub
or
whatever
so
I
mean
I
can
move
stuff
around
to
set
priorities,
I
don't
know
who's
allowed
to
do
that.
A
So,
if
there's
items
that
when
you
bring
them
up,
you
feel
we
need
for
one
eight
and
they
have
p2
or
p
greater
than
two
in
them:
either
either
drag
them
to
the
p0
p1
or
ask
me
to
do
that,
and
I
will
try
to
use
that
to
prepare
next
week's
meeting.