►
From YouTube: Istio User Experience Working Group April 02, 2019
Description
Istio User Experience Working Group session April 02, 2019
A
A
Manager
so
before
we
get
started,
I'll
tell
you
what
I
plan
to
do
I
wanted
to
show
everyone,
some
slides
of
link
or
D,
which
I
consider
to
be
our
big
competition,
and
they
have
been
considered
doing
really
well
in
the
UX
area.
So
maybe
that
would
inspire
some
people
here
to
do
some
UX
things
for
is:
do
I
had
I
planned
on
showing
on
having
a
colleague
of
mine,
show
the
destination
rule
management
prototype
I
will
that
she
couldn't
make
it.
We
will
schedule
it
for
a
future
meeting.
A
Then
I
have
some
stuff
on
the
install
I
want
to
show
some
slides
I
made
on
the
difficulty
of
the
install
I'm.
Not
expert
on
the
operator
would
lengthy
here
some
thoughts
on
that,
but
we
can
briefly
discuss
it
and
then
maybe
skip
this
mesh
proposal,
and
we
have
awesome
talk
about
sto
control
to
do
some,
some
PRS
and
some
new
features
that
have
been
proposed.
A
A
Our
install
is
possible
for
the
default
path
and
anybody
wants
to
jump
in
as
I
talked
about
stuff,
but
the
next
thing
you
do
with
link
Rd
is
to
see
if
your
control
plan
is
healthy,
there's
a
sign
any
check,
link
or
D
check.
That
does
a
lot
of
the
stuff
that
we
do
with
our
new
verify
install
command,
but
it
does
sort
of
here.
A
Another
really
nice
thing
that
they
do
on
their
install
is,
if
you
will
tell
you
if
the
version
that
of
the
control
point
you're
running
is
the
latest
version.
So
I
didn't
install
the
back
little
version,
but
it
would
have
told
me
if
I
had
done
that.
So
this
is
a
very
nice
feature
that
we're
trying
to
move
our
verify
install
towards,
but
any
improvements
there
would
be
great.
A
A
I
have
a
live
version
of
that
running
right
now,
I'll
bring
up.
So
when
you
go
to
the
overview,
it
lists
all
of
the
namespaces
on
your
cluster,
and
if
that
namespace
is
spitting,
a
linker
D,
it
will
say
meshed
now,
I
didn't
actually
try
to
install
any
demos,
I,
just
just
typed
this
after
a
fresh
install.
So
the
only
thing
meshed
is
their
own
control,
plane,
namespace,
link
or
D,
but
automatically
it
gives
this
great
view
of
their
services.
A
With
some
of
the
things
that
you
know,
we
record
in
Prometheus,
but
they're
getting
here
sort
of
live,
so
this
is
being
literally
live,
updated
right
now
and
then
direct
links
to
the
Griffons
that
we
have,
whether
they
have
they
also
install
a
graph
on
it
under
control
flame,
but
with
a
link
that
to
take
you
right
to
a
particular
view
for
a
particular
app.
But
this
is
this:
is
nice
and
just
based
on
seeing
that
that's
why
we
have
this
baby
step
in
his
DL
1.2,
the
SEO
control
experimental
dashboard
command.
C
C
A
Oops
so
obviously,
two
intentions:
first,
if
they
have
some
good
stuff
that
we
feel
would
be
easy
for
us
to
add.
We
should
add
it.
The
second
is
that
we
hope
to
you
know,
be
able
to
capture
link
ready
users
who
need
more
power
from
our
richer
things,
so
it
would
be
good
for
us
to
at
least
know
sort
of
the
terminology
they
use.
So
we
can
say
you
know
when
you
have
a
meshed
namespace.
We
have
a
labeled
moon
space
and
that
kind
of
thing,
because
just
to
level
set
us
help
us
talk.
A
A
If
we
can
make
things
easier
for
ourselves,
let's,
let's
take
a
look
just
for
the
benefit
of
everyone,
but
the
idea
of
a
service
mesh
is
a
universal
concept
and
you
know
we
don't
want
to
recreate
all
the
difficulties
we
had
like
das
vs.
UNIX
line
endings
and
things
if
there's
some
real
similarities
that
could
be
adopted.
That's
just
my
two
cents:
it's
not
audience
official
position,
but
I
think
we
should.
We
should
take
a
look.
D
Is
it
near
on
setting
the
garage
at
7:00?
That's
a
hell,
it's
a
good
hip
reflection,
I,
agree,
I,
don't
know
this
point
of
conversation.
We
had
had
that
ServiceMaster
day
last
week
was
one
thing
to
your
point:
EDM
one
thing
that
is,
you
know
would
be
beneficial
to
all
is
like
it's
to
the
extent
that
the
gnomon
clip
that
we
end
up
plumbing
to
some
of
the
same
human
creature.
D
A
Thank
You
Lee,
so,
let's
play
through
the
other
features
that
we
have
in
linker,
DS
UI,
there's
a
top
feature,
choose
sort
of
the
namespace,
you
want
to
click
start
and
it
will
start
reporting
all
of
the
traffic.
That's
happening
there.
So
it's
a
very
nice
skin
on
the
important
things
you
can
get
up
in
telemetry
and
you
can
stop
that
top
routes,
which
is
a
similar
tool
to
take
a
look
at
various
namespaces
and
how
they're
communicating
what's
happening
in
they're
filtered
for
things.
A
A
A
Shows
it's
just
it's
just
boilerplate,
so
I
thought
that
was
interesting.
Ui
choice
that
instead
of
sort
of
asking
you
about
all
of
your
paths
and
HTTP
methods
on
a
complicated
form,
they
just
sort
of
give
you
like
here's.
What
sort
of
looks
like
I
meant
the
things
you
want,
any
things
that
you
need
and
then
you're
done,
I
thought
that
was
kind
of
inspiring.
A
So
so
what's
the
service
profile,
so
it
seems
to
be
similar
to
our
virtual
service.
It
has
you
know,
rules
for
path,
regular
expressions,
oops
and
status
conditions
and
retries
sort
of
the
stuff
that
we
do,
and
this
is
how
they
expose
it
to
a
young
document
same
as
we
do,
but
just
instead
of
pointing
people
at
the
documentation
they
just
give
you
you
have
to
commented
out
one.
A
So
I
thought
that
was
interesting:
they
have
a
sort
of
service
mesh
command.
That
sort
of
tells
you
it's
a
summary
and
it
sort
of
prompts
you
to
connect
namespaces
to
it.
So
this
is
sort
of
setting
up
their
injection,
giving
you
the
command
on
the
CLI
to
do
it,
no
link
or
D
inject
and
then
some
namespace
options
oops
and
it
sort
of
tells
you
how
many
pods
are
there,
what
sorts
of
things
you're
doing.
So
that's
pretty
simple.
A
F
A
F
The
reason
I
ask
is,
from
the
surface
mesh
conference,
there
are
users,
I,
think
multiple
users.
They
did
indicate
that
it's
interesting
for
their
developers
like
the
handshake
between
the
dots
and
the
operators.
It's
typically
just
a
page
of
configuration
and
that's
all
they
need.
So
basically
the
micro
service,
the
developer,
would
say
these
are
the
recommended
configuration
I
would
like
and
it
doesn't
have
to
be
in
the
ISTE
of
format
it
could
be.
F
You
know
this
is
the
retries
I
would
like,
and
this
is
my
the
who,
which
services
I'm
talking
to
you,
which
services
depends
on
me,
so
it
could
be
a
list
of
configuration
and
then
the
ops
team
would
take
that
configuration
and
then
using
their
SEO
knowledge,
which
the
dev
team
doesn't
have
to
have
the
SQ
knowledge
and
then
translate
that
into
it's
still,
yellow
because
there's
a
learning
curve
involved.
But
then
the
nice
thing
is
show
team
I've
gone
through
the
learning
of
complexity,
oh
it
still,
but
the
devs
doesn't
have
to
learn.
A
G
Sorry,
you
heard
me
yes,
I
have
question
regarding
today.
We
have
like
some
of
this
feature
provided
by
Kali
and
we
want
to
add
such
kind
of
easy
to
ball.
So
the
question
is
like
we
have
different
feature,
separated
between
two
tools
or
2.4.
So
is
there
any
idea
to
have
a
single
place
where
we
have
all
the
artificial
design
pictures
within
the
same
tool?
So
we
we.
A
Want
that
so,
as
a
like
I
say
as
a
baby
step
in
is
to12
I've
merged
this
week,
disability,
you
can
type
sto
control,
experimental,
dashboard,
key
ally,
and
it
just
brings
up
t
ali
dashboard.
So
at
least
users
will
find
it
if
they've
installed
it,
and
they
won't
need
to
know
what
port
it
is,
how
to
get
there.
A
The
I
know
there's
some
people
at
IBM
trying
to
look
at
integrating
with
the
open
source
cooee
shell.
There
are
I'm
sure
all
their
open
source
efforts,
I,
don't
know
what
the
best
way
to
sort
of
bring
all
these
tools
together
is
hoping
that
we
can
have.
You
know,
discussion
as
part
of
this
workgroup
and
how
we
get
just.
G
Yeah,
because
of
what
we
have
actually
likely
the
last
version
we
can.
We
got
cherry
with
confirmation
and
then
from
Charlie
we
can
go
to
a
distribute
tracing.
We
can
go
to
pro
meters,
so
I'm,
seeing
that
like
okay,
this
is
like
the
center
point.
One
I
should
go
the
first
time
which
is
Kali
and
then,
if
I
wanted
to
deep
dive,
I
need
to
go
like
distribute,
placing
from
atop
stuff
like
that
I
go
from
there.
G
G
A
G
Yeah,
how
do
you
prefer
both
of
them
but
yeah
I
prefer,
like
okay,
I
have
a
such
kind
of
comma,
so
I
can't
use
it
on
an
operator,
so
I
can't
use
it
to
check
when
I'm
debating
my
my
service
match
and
if
I
go
like
Clara
or
ECU
dashboard.
If
something
goes
down,
they
just
like
a
red
or
is
it
a
red
spot
on
centimeters?
A
F
G
A
A
Because
one
thing
we're
going
to
need
is
to
sort
of
direct
people
like
when
people
have
a
problem
with
validation
or
something
our
our
documentation
could
say.
You
know,
go
to
key
Ollie
and
look
at
this
page
to
see.
If
you
have
any
problems,
or
it
could
say,
you
know,
run
if
you'll
wait
or
it
could
say
just
look
for
poorly
named
service
ports.
What
so
we
want
to?
We
want
to
guide
people
to
the
right
documentation,
I,
think
yeah.
F
D
H
Yeah-
and
we
also
need
to
I
think
with
what
what
Rick
is
trying
to
say,
we
need
to
agree
on
a
formal,
dashboard
or
never
or
car
equivalent.
As
for
sudden
that
you
just
showed
four
linker
D.
What
is
that
4-h
do
and
I
think
if
we
can
agree
that
you
know
we're
interested
in
key
ally
being
the
the
one
dashboard
where
you
can
from
there.
You
know
launch
all
the
other
ones,
then
we
can
focus
on
whether
we
should
add
features
to
city
Ally
or
do
something
else
in
parallel
and
as
a
community.
H
If
we
can
agree
on
that,
then
we
can
invite
the
the
key
ally
community
to
get
on
call
and
talk
about
their
road
map
and
see,
if
they're
interested
and
then
and
being
the
dashboard
for
it
co.
So
if
you
look
at
key
ally
website
and
read
all
their
features,
it
seems
like
a
lot
of
the
screenshots
that
they're,
using
or
directly
from
from
SPO,
so
they're
kind
of
positioning
themselves
as
the
dashboard
for
SEO.
E
I
think
that's
exactly
the
point.
If
they're
sort
of
you
know
suitable
and
want
to
be
the
main
entry
point,
that's
the
dashboard
that
you
start
at
and
then
yeah
I
think
ed
pointed
out
that
we
need
CLI
kind
of
equivalents
of
a
lot
of
the
functions.
But
but
again
the
central
point
from
a
UI
perspective
could
be
key
ally
if
yep
so
I
think
we
need
to
make
that
decision
soon.
I
I
E
Think
the
point
the
ad
was
making
so
there's
got
to
be
another
CLI
kind
of
equivalent
with
so
you
could
do
this
without
the
fancy.
Gui
part.
You
know
if
you
have
a
GUI,
that
it's
Kali,
but
but
maybe
you're,
right
Martin.
It's
still
with
probably
if
it's
that
tightly
linked
to
to
sto
should
probably
part
of
our
code
base
rather
than
a
separate
repository
yeah.
I
A
She'll
vet
is
just
a
validation
tool
from
Aspen
match
that
was
presented
here
a
few
weeks
ago
and
tells
you
if
your
ports
have
poor
names
or
if
your
virtual
services
are
not
hooked
up
to
gateways
properly.
It's
just
a
validation
in
the
large
stos
validation
web
hook
and
the
command
line.
Validation
only
tell
you
about
a
single
item
being
added
where's
that
vet
tool
wasn't
all
of
your.
Is
your
resources
and
some
of
your
kubernetes
things
like
services
and
sort
of
apply
some
additional
ad
hoc
type
validation,
okay,.
D
C
C
I
mean
I
think
that
was
his
proposal
is,
but
I
am
Not
sure.
Additionally,
in
the
short
term,
I
think
we
can
also
benefit
from
just
linking
in
the
HT
or
dogs
out.
We
steer
it
to
say
you
can
try
this
out
to
know
quickly.
If
you
have
misconfigurations,
you
should
do
that
and
just
to
emphasize
on
Ed's
point,
we
are
not
replacing
the
functionality
in
HT
of
CTL
or
galley,
which
does
a
single
resource,
static
analysis
that
still
remains
there.
D
A
I
That's
high
honor
request
list,
but
for
1.2
were
not
we're.
Just
focusing
on
our
engineering
processes,
most
gotcha,
okay
yeah
we
have.
We
have
people
like
a
native
K
native,
wants
to
be
able
to
deploy
this
TOS.
It's
your
resources
and
know
when
they're
available
and
right
now
they
just
have
to
kind
of
send
fake
traffic
until
the
traffic
works,
and
so
we
really
need
the
status
to
be
recorded.
It's.
It
is
a
big,
a
complicated
proposition,
though,
because
one
CRD
results
in
configuration
for
a
thousand
different
pods
to
be
updated.
A
A
So
I've
been
I
want
to
make
it
easier,
so
we
I
looked
at
linker
D
was
you
know
this?
One
command
I
have
been
using
some
IBM
products.
That
sort
of
give
you
a
graphical
front-end
for
helm
charts,
and
this
turned
out
to
be
extremely
difficult
experience
with
Sto.
So
there's
so
many
settings
and
those
settings
are
either
not
documented
well
on
our
on
our
page
or
they
there's
just
too
many
of
them.
A
So
I
started
taking
screenshots
of
the
complete
screen
for
this
graphical
form
front
end
for
our
options,
pause
security
parameters,
global
settings,
the
ingress
name
do
I
want
the
HTTPS.
What's
the
name
of
this
proxy,
so
the
people
who
put
this
together
just
took
everything
in
values,
diamo
and
then
a
forum
for
it.
A
When
I
talked
to
Frank
about
the
configuration
profiles
and
how
nice
they
were.
But
but
then
I
I
like
to
you,
know,
I
install
like
this,
with
the
helm,
template
install
and
when
I
want
to
redo
things,
I,
just
sort
of
run
it
again
and
I
I
make
heavy
use
of
this
install
page,
which
lists
all
of
our
options,
many
of
which
have
no
description.
So
you
know
I,
don't
really
need
a
description
for
the
cert
manager's
thing,
but
for
a
lot
of
the
things
that
we
do
like
this.
A
Outbound
traffic
policy
mode,
which
has
been
getting
a
lot
of
discussion,
it
just
says,
allow
any,
doesn't
show
you
the
other
choice.
We
have
to
figure
this
out
and
I
guess
the
operator
or
the
SU
operator
that
we've
been
talking
about
is
a
solution
for
that.
Although
I
haven't
really
researched
the
operator
well,
a.
I
Partial
solution,
because
so
the
operator
is
going
to
provide
a
nicer,
more
structured
way
to
to
manage
this
with
a
single
CRD
that
we
can
sympathize
heavily
comment
and
that
kind
of
stuff.
However,
there
are
some
customers
that
will
never
want
to
use
the
operator.
I,
don't
continue
to
want
to
run
the
values
out
the
amyl,
so
we
really
need
to
figure
out
how
to
solve
values
out
the
amyl
on
a
film
with
proper
metadata,
proper
annotations
and
that
kind
of
stuff.
The.
E
Interesting
part
of
the
operator
is
sort
of
defining,
essentially
an
API
that
represents
some
composite
set
of
things
kind
of
like
a
profile,
but
it's
more
it's
more
like
a
number
of
settings
that
can
be
kind
of
turned
on
or
off
that
are
interesting
to
everybody.
Kind
of
thing
like
I,
want
to
turn
on
egress
or
somebody,
the
egress
gateway,
or
something
like
that,
and
then
there
might
be
a
few
things
that
get
tweaked,
so
it
would
be
a
etirsa
gateway,
kind
of
thing,
but
I
think
fundamentally
what
this
is
what's
interesting
about.
E
I
And
we've
we
actually
started
to
review
the
CR
D,
which
is
at
itself
quite
large,
and
it
became
clear
there's
we
were
muddying
the
distinction
between
what's
programming
a
feature
versus:
what's
slipping,
some
particular
options
for
deployment
of
a
particular
component
like
so
you
want
to
turn
on
Shane.
Well,
that's
that's
really
not
a
feature
that
that's
kind
of
programming
the
binary
so
we're
we're
separating
this
making
it
very
clear.
I
There
are
sub
stuff
that
is
in
values
that
the
animal
that
really
should
just
be
a
regular
CRD
that
deprotonate
user
interacts
with
at
the
CRD
level,
not
by
programming
not
through
get
to
helm.
So
hopefully
that
up
I
think
will
clean
up
the
values
as
well,
but
we
still
need
to
fully
document
and
structure
the
yam,
also
its
consumable.
The.
A
A
Don't
have
a
good
way
to
know
what
settings
that
were
chosen
right
so
independent
of
this
I
was
wondering
if
there
was
any
interest
in
doing
a
reverse
mapping
of
the
helm
so
that
you
could
sort
of
see
what
was
chosen
based
on
inspection
of
the
various
things
in
this.
Do
like
the
deployment
of
pilot
and
the
deployment
of
galley
and
settings.
So
the
idea
would
be
that
you
would
do
some
kind
of
thing.
Maybe
you
still
control,
maybe
some
other
tool.
It
would
sort
of
tell
you
everything
that
had
been
set.
I
A
D
I
A
E
Okay,
well,
I
think.
The
same
argument,
though,
applies
that
it's
not
that
interesting
to
be
able
to
you
know,
ask
oh,
you
know
what
one
of
these
150
or
500
settings
that
are
available
in
values.
Imo
are
turned
on
for
me,
it's
much
more
important
to
understand
at
a
higher
level
what
features
have
been
turned
on
and
off
so
you
that
sort
of
should
be
the
the
vocabulary
that
we
use
with
users
right.
A
We
want
the
vocabulary
here
so,
if
I'm
talking
about
traffic
more
policy,
I
want
to
be
able
to
sort
of
you
know,
run
some
command
grip
traffic
mode
policy
to
see
you
this
so
I
can
sort
of
say.
What's
having
my
policy
do
you
have,
because
if
the
answer
to
my
question
of
why
doesn't
my
pod
work?
Is
your
traffic
mode
policy
is
not
there
in
a
service
entry?
A
E
A
Vertically
I
heard
about
all
these
things
right,
but
most
of
these
are
nobody
cares
about
like
how
many
galleys
are
we
going
to
scale
up
to
maybe
not
interesting,
but
this
one
is
particularly
tricky
because
in
the
documentation,
maybe
it's
got
spaces.
Maybe
here
it's
got
the
word
global
and
mode
in
front
of
it
in
the
search.
Maybe
doesn't
want
to
see
it.
So
it's
a
little
tricky
so.
C
But
I
do
believe
I
thing,
I
agree
with
Frank,
I.
Think
just
a
good
organization
of
these
values.
Around
features
will
help
simplify
this
a
lot
and
you
are
highlighting
some
of
the
more
important
variables
or
values
which
might
not
be
reflected
in
one
or
two
features,
so
we
can
handle
them
separately.
But
if
you
just
make
some
intermediate
steps
to
cluster
these
in
terms
of
features
and
turning
them
on
or
off,
we
can
make
it
much
more
simpler,
rather
than
I
agree.
C
I
Well,
I,
don't
know
Martin
Martin
see
to
be
engaged
in
doing
this,
but
what
will
happen
soon?
Is
the
operator,
so
I
expect
that
most
customers
will
want
to
interact
with
the
operator,
and
then
you
do
get
that
experience
kind
of.
Naturally,
you
can
look
at
the
CRD,
see
what
the
state
is
at
see
if
it
will
be
deployed
or
still
in
the
process
of
being
deployed.
A
So
that
would
be
great
because
so
many
settings
like
this
much
networks
thing.
You
know
it's
so
important
to
know
what
other
networks
your
multi
clustering
with,
and
if
your
only
documentation
is
that
it's
this,
it's
tough,
okay,.
I
A
And
I
I
have
a
proposal
for
this
SEO
control
mesh
command,
which
I
don't
know
if
I
was
going
to
talk
about
today
or
if
I
was
going
to
try
to
bring
that
to
the
configure
networking
meeting,
but
it
would
be
so
great
to
be
able
to
view
sort
of
what
what
multi
cluster
stuff
you've
got
set
up.
So
do
we
have
anything
else
to
say
about
install
before
we
talk
about
the
sto
control,
sub
commands.
A
Great,
so
we
have
two
things:
probably
we
want
to
talk
about,
or
do
we
have
time
to
talk
about?
Lynn
has
a
list
of
some
of
the
reason
proposals
I'm
going
to
go
to
in
a
second,
and
then
we
have
the
pull
request
for
TLS
I.
Believe
the
authors
on
the
call-
and
it
was
we
want
to
take
a
look
at
this
as
the
right
sort
of
design
for
the
command
and
then
also
currently,
this
tool
has
no
tests,
and
the
question
was:
what's
the
bar
for
getting
things
into,
it's
do
control
experimental.
A
Should
we
let
people
just
put
stuff
in,
should
we
make
them
do
this
at
a
branch
and
or
in
a
new
ecosystem
project?
What
what
is
the
rule
for
sort
of
proof
of
concept
things
in
this
case?
It
was
for
this
Auerbach
debugging
tool
to
replace
something
that
was
in
one.
Oh,
that
didn't
make
it
into
one
one
couldn't
work,
so
we
pulled
it
out
on
one.
J
So,
on,
okay,
add
something
so
regarding
the
test.
I
I
think
most
of
the
logistical
T
commands.
We
only
have
unit
tests
and
the
are
simple
unit
is
like
we
providing
some
config
time
as
input
and
we
are
searching
the
command
output
for
some
specific
string.
That's
the
unit
is
we
have
today
and
I.
Don't
think
we
have
into
in
test.
Photos
is
to
take
command
and
I
feel
that
is
not
necessary
to
end
like
enjoined.
Here's
for
you,
Stoker
T,
so.
A
You're
completely
right
that
we
don't
another
thing
that
I'm
dealing
with
right
now
is
some
regressions
in
some
of
the
commands.
Like
this
proxy
config
thing
is
saying:
looked
at
the
difference,
this
cluster,
aesthetic
and
envoy
saying
it's
got
no
value
of
the
first
type,
because
there
were
no
and
then
test
for
it.
So
what
we're
seeing
is
like
things
were,
what
push
to
envoy
changes
or
unvoyage
and
changes
and
a
command-line
tool
that
is
asking
envoy,
config,
DUP
and
then
parsing
that
and
we're
using
static
file
for
for
unit
testing.
A
It
is
starting
to
break
as
we
go
to
new
envoy
versions,
and
that's
going
to
be
worse
when
we
start
doing
the
secret
discovery,
service,
stuff,
so
I'm,
hoping
that
we
can
and
that's
not
part
of
your
PR.
They
should
be
indented
done
by
somebody
else
completely,
but
we
do
need
to
know
how
much
test
is
going
to
be
needed.
Being
I
I.
J
Yeah
I
think
so
we
actually
we
have
sitting
multiple
times
that
are
we
API
changes
are
breaks
sto,
it's
not
it's
not
really
to
use
the
container
tree.
So
most
of
the
API
in
the
HTS
API
is
used
by
pirating
the
generation
of
the
awakened
V.
So
if
the
upstream
envoy
API
changes,
it
will
first
break
pirate
and
easter
cat
a
those
commands
using
the
config
down.
We
are.
J
I
A
If
no
one
else
wants
to
I'm
going
to
try
my
hand
at
making
some
end-to-end
tests
for
sto
control,
because
I
think
that's
going
to
be
needed,
because
I
want
more
people
to
contribute
to
sto
control
so
having
more
tests,
for
it
means
it'll
be
easier
to
sort
of
review
those
contributions.
As
for
the
particular
one
we're
talking
about
it,
is
this
one
I
don't
know.
If
we
want
to
discuss
that
on
this
meeting
or
it's
since
it's
more
of
a
networking
thing,
it's
designed
to
be
discussed
there
I.
J
A
J
J
Yeah
yep,
okay,
so
so
currently
we
already
have
a
command
in
Easter
Cote.
That
will
check
the
authentication
policy
and
Relenza
distinct
visual
to
tell
you
that
the
videos
sitting
on
both
client-side
and
server-side
and
what
I'm
doing
now
is
to
enhance
our
command
and
also
provide
support
for
jot
and
our
bank
rows.
J
The
investing
is
totally
awesome
and
it
works
by
talking
to
pirate
and
and
by
getting
those
authentication
policy
and
disenroll
applied
by
customer
and
and
I,
don't
eyes
aim
and
outputs
Lea
means
a
man
lying.
What
I
am
doing
here
is
that
it
be
different.
It
talks
to
the
our
way
directly
to
get
a
config
down
and
analyze
the
v2
TOS
jobs
are
bagged
based
on
the
Envoy
convict
orally.
So,
theoretically,
you
can
use
it
even
without
SEO
kind
of
but
allow
it.
It
relies
on
some
and
a
suction
to
form
the
perigee
region.
J
F
C
J
If
you
Bo
tell
so
country,
it
only
checks
the
listeners
listening
and
Lodi
T,
so
this
is
used
to
used
by
alpha
e
to
forward
traffic
accuser,
actual
application.
So
for
this
product
page
we
can
see
it
has
a
to
designer
listening
on
its
note,
IP
for
the
first
a
while
we
see
it
has
a
certificate
which
means
when
the
client
is
connecting
to
this
pot
and
this
IP.
J
How
do
I
represent
this
certificate
for
the
for
the
kind
and
we
can
see
the
materials
is
enabled
and
the
more
District
and
we
can
see
there
is
no
child
policy
and,
as
there
is
no
a
pen
policy,
this
is
the
config
and
an
inner.
We
also
have
the
config
and
the
outbound
clusters,
the
other
one
cluster
is
used
by
our
way
when
always
connecting
to
some
upstream
servers.
J
So
we
can
know
how
does
product
page
you
talks
to,
for
example,
details
dot
default
and
we
can
know
how
does
put
a
page
talks
to
HTTP
being
so
we
can
see
in
the
satirical
on
our
way.
We
will
present
this
certificate.
Wayne
is
talking
to
this
upstream
and
it
will
validate
to
the
upstream,
with
this
sand
being
their
certificate,
and
when-
and
we
know
when
are
we
talking
to
communities,
it
doesn't
present
any
certificate
and
it
doesn't
require
any
certificate.
J
So
we
can
know
it
will
use
like
paying
takes
to
talk
to
you
to
the
kubernetes
deport
if
we
apply
some
sample
policy.
For
example,
I
have
prepared
some
trial
policy
for
product
page
and
if
we
provide
a
event
policy
and
we
as
well
as
Caban
again,
okay,
you
can
see
we
will
present
some
brief
description
about.
We
have
been
able
to
George
and
the
issuer
should
be
testing
at
security
ohio
and
we
have
enabled
our
bank.
We
have
only
one
rules.
The
row
name
is
service
viewer,
so
this
is
a
very.
J
F
E
J
So
so
in
the
future
we
extend
easy
to
now.
You
can
give
you
two
other
names,
so
you
can
analyze
the
relationship
between
the
two
points
like
the
first
one
is
a
kind
pause
it.
How
I
talk
to
you
other
parts
like
the
second
part,
so
this
kind
of
like
a
static
analyzer
and
we
config-
and
it
gives
you
the
information
about
that.
Currently
I'm
fixing
windows,
security
CDs.
This.
E
A
J
A
A
So
the
first
thing
was
this
geo
control
version,
which
actually
turned
out
I
was
mistaken.
I
thought
that
we
didn't
have
a
good
steal,
control
version,
because
at
one
point
in
the
distant
past,
when
you
typed
it,
given
you
both
the
clinton
server
version,
it
turns
out,
we
actually
have
a
client
and
server
and
a
nice
summary
that
comes
if
you
do
is
to
control
version
remote
money
short,
so
it'd
been
lost
in
all
of
the
you
know
of
the
testing.
So
if
I
literally
do
the
two
control
version,
remote
short
is
very
nice.
A
Tells
you
all
of
the
versions
of
all
of
your
tools,
but
we
people
are
going
to
find
that
because
it,
the
default,
is
so
different
yeah.
That
seems.
A
Didn't
I
didn't
discover
that
until
I
went
through
the
testing,
when
we
had
testing
day
what
is
the
output
of
this
version
by
itself?
Oh,
if
you
do
it
by
itself,
it
looks
like
Q
control
version,
but
it
only
gives
you
the
client
side
by
and
if
you
just
do
remote,
it
gives
you
a
cube.
Control
version.
Look
like
thing,
but
for
all
of
the
SEO
control
playing.
A
F
A
Yes,
so
another
thing
is
been
interesting
for
people
is
that
when
you
install
when
you
download
this
deal,
it
brings
down
this
big
directory
full
of
stuff
the
command
line,
the
samples
all
of
the
ml
and
helm
stuff
that
you
need
it
turns
out.
95%
of
the
download
is
sto
control,
which
is.
A
A
I
A
A
So
yeah,
so
it's
do
control
check,
which
is
so
we
have
in
progress
this.
This
version
check
that
iris
is
doing.
A
F
A
F
A
His
tío
control,
all
you
get,
you
get
a
lot
of
details
for
things
like
virtual
service,
like
the
number
of
ports
and
interesting
things.
The
subsets
are
shown
to
a
destination
rule.
If
you
do
cube
control,
get
gateway
to
a
destination
rule
virtual
service,
you
don't
get
the
subsets
now
I
want
it
to
put
the
subsets
in,
but
it
turns
out
that
due
to
unfinished
ers
in
kubernetes
itself,
you
add
a
custom
column
like
the
one
I
did
for
host
or
gateways.
A
J
A
J
A
Something
like
it
is
to
control
what
are
my
types
you
could
type
in
to
keep
control
gate
or
something,
but
that's
very
amusing.
So
everyone
put
your
thoughts
on
that.
The
sto
controlled
dashboard,
which
I
I
put
in
one
too,
if
you
don't
so,
try
it
out
if
you're,
using
master
and
add
or
make
PRS.
If
you
want
to
change
it,
the
mesh,
if
I,
which
I
just
mentioned
them
that
that
looks
you
with
Auto
inject.
A
A
B
F
Yes
and
those
are
the
road
maps
will
carry
it
so
be
you
probably
heard
some
of
the
discussion
we
had
earlier
is
Chi
alley
position,
as
the
is
still
dashboard
for
everything
that
user
wants
to
see
within
the
ister
project
as
far
as
dashboard
lives,
I
think
that
would
be
good
discussion
to
have,
because
that's
going
to
influence
whether
we
can
build
another
dashboard
or
whether
we
can
you
know,
work
with
Sonia
to
get
that
done.
Okay,
sort.
E
Of
a
I
think
we
need
sort
of
summary
of
what
are
the
high
level
dashboard
functions
that
we
need
to
see
and
then
understand
which
subset
of
those
are
in
catania
already,
which
ones
could
be
added,
and
do
we
need
a
separate
dashboard
or
not?
That's
so
getting
those
answers
is
kind
of
an
important
part
of
our
key
allottee
piece
of
this.