►
From YouTube: Istio User Experience WG meeting 2019-03-19
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
Okay,
so
I
wanted
to
have
a
mission
state
to
make
sure
everybody
understood
what
I'm
trying
to
do
and
what
we
want
it
to
happen,
and
so
everyone
feels
that
they're,
a
part
of
everything
here.
So
the
mission
of
the
user
experience
working
group
to
make
sto
easily
learn
about
understandable
and
just
configure
it
easily
fixable.
This,
too
should
be
configurable
and
must
benefit
operators
and
developers
of
all
skill
levels.
A
People
have
given
me
feedback
and
focused
on
the
advanced
operators
and
I
want
to
make
sure
we
we
hit
all
of
our
stakeholders.
The
UX
working
group
establishes
relationships
with
third
parties
who
produce
tools
in
the
asou
ecosystem
to
help
them
deliver
better
integration
with
Cory
Steel
features.
I
want
us
to
reach
out
to
people
like
us
to
mesh
who
are
making
their
like
Castillo
validation
tools,
so
that
we
don't
try
to
take
everything
on
ourselves.
I
want
to
play
well
with
others.
The
UX
working
group
contributes
Kubota
cores.
A
Do
we
want
to
improve
this,
to
control,
to
help
install
option
and
the
other
tools
to
give
better
visibility
into
what
this
deal
is
doing
and
what
it
is
not
doing?
We
campaign
for
good
installation
defaults,
reduce
unnecessary
configuration
options
and
against
surprising
behavior.
We
improve
error
messages
to
make
them
better
more
visible
and
closer
to
the
problem.
We
improve
the
configuration
troubleshooting
and
debugging
experience
for
is
to
users
of
all
skill
levels,
so
you
know
I'm
really
going
to
get
on
people
food.
Just
the
compare
to
improve
the
error
messages.
A
It
would
be
great
they're,
confusing
New
York's
working
group
cards
with
the
documentation
working
group
as
part
of
the
tools
and
got
into
god
users
to
the
documentation.
We
also
create
documentation
for
the
sto
tool
experience.
The
UX
working
group
collaborates
with
the
configuration
working
group
to
ensure
that
the
tools
render
the
configuration
correctly
and
in
an
understandable
way,
is
to
ships
with
55
CR
DS.
That
configure
is
do
our
mission
is
to
help
developers
know
which
they
are
DS
to.
Why
we
also
make
tools
to
show
operators
which
CR
DS
affect
their
microservice.
A
We
to
improve
visit
amount
by
improving
the
networking
performance,
the
UX
group.
Our
goal
is
to
reduce
the
time
it
takes
developers
to
learn
to
deploy
in
the
scale
for
products
agility
of
deployment.
Uptime
is
more
important
than
latency.
Time
must
be
easy
to
find
and
remediated
steal.
Miss
configurations
and
I
I
got
early
on.
I
got
some
feedback
about
this.
We
if
Co
cares
about
performance
just
I
want
us
to
focus
on,
learn,
ability
and
uptime
remediation
time.
That's
our
key
metric
New
York's
working
group
widens
the
abstraction
level.
A
C
E
E
F
G
B
A
F
A
A
F
F
The
same
thing
that
Martin
was
just
talking
about
so
I
was
just
gonna,
bring
up
that
we
had
talked
about
it
in
steering,
but
I
don't
know
people
I
mentioned
it.
Rahm
had
volunteered
to
help
work
on
the
survey
he
couldn't
be
here
today
at
this
meeting,
but
he
wanted
to
work
with
you
to
figure
out
how
to
do
that.
Okay,.
H
A
That
makes
sense,
I
know
I'm
one
of
some
questions
to
find
the
role
of
the
person
who
is
sort
of
answering
the
survey,
because
it
may
be
that
everyone
who
sees
themselves
as
an
operator
likes
it.
Everyone
sees
themselves
as
too
developer
does
not
or
everyone
who
sees
themselves
as
a
beginner
likes
it
or
advanced,
does
not,
or
vice
versa.
But
yes,
I
think
this
would
be
great.
I
will
take
this
up
with
round
yeah.
H
F
H
I
A
A
I
A
A
A
I
B
I
I
sing
at
the
clerk
she
bad,
but
I
will
try
to
speak
clearly
for
an
extra
sang
for
new
stream
and
ID
for
inviting
me
here,
and
my
name
is
Joaquin
doll.
I'm
a
software
engineer
of
the
equation.
I
work
in
the
OSS
Department
I'm
responsible
for
driving
the
Escalade
equation
with
our
file,
gos
assisted
in
my
company.
I
Hers
at
the
agenda
first
I'd
like
to
briefly
introduce
how
we
integrate
East
sto
with
our
OS
s6,
and
tune
a
legit
service
communication
management
capability.
Then
I
will
talk
about
the
obstacles
of
a
healthy
impatient
when
we
try
to
integrate
sto
and
how
we
try,
mm
and
I
hope.
I
could
get
some
help
and
a
suggestion
from
the
community.
With
this
problem.
I
So
for
the
control
plan
component
also
SMED
solution.
We
pick
the
pattern
from
a
skill
for
traffic
management
and
a
mixture
for
elementary
collection.
We
use
a
console
for
the
sauce
registry
beside
that
we
create
our
own.
Some
temple
feel
component,
like
religious
traitor,
transparent
re-register,
a
service
on
behave
of
the
application.
Our
configuration
server
to
provide
this
for
a
child
for
the
sto
lutein
Lord
and
colorist
able
to
provide
lists
for
API
for
kuroda
book
deployment,
because
it's
a
love,
our
main
use
case
and
also
our
use
of
Ranieri
UI.
I
I
We
don't
use
the
pistol
egress
to
come
toward
outbound
traffic.
Actually,
we
think
it's
a
little
bit
hard
to
tell
which
act
or
services
and
it
would
be
accessed
by
the
intro
services
in
other
ones,
especially
the
large
the
network,
for
example,
some
Victoria
services
of
register
to
our
system
the
long
time.
So
so
we
just
we
just
the
regular,
actually
request
of
the
of
this
service
in
our
own
system.
We
just
leave
the
explosives
and
bypass.
I
I
I
A
I
I
So
this
figure
shows
how
we
combined
a
Piccadilly
and
a
away
all
the
traffic
engines
and
for
our
system,
a
big
getaway,
just
some
simply
upload
the
service
discover
and
a
traffic
management
functionality
to
the
noise
at
the
car
and
the
folks
on
the
API
management
feature.
So
for
us
I
think
it's
clear
a
picture.
They
took
component
law
server,
Marcel
worm
to
different
requirement,
but
also
cooperate
with
each
other,
and
I
was
saying
that
the
you
know
a
set.
I
I
So
this
Lee-
this
is
actually
this
poster
problem
way.
We
faced
when
we
integrated
steel,
because
in
our
use
case
and
in
telecom
use
cases,
we
often
get
to
supporting
both
port
network
interface
for
some
reason
like
functional
separation,
our
control
and
the
trap
network
plan,
Niq
aggregation
and
found
him
for
the
dominance
of
network
supported
for
implementation
of
different
network
service.
I
I
But
we
part
with
that,
given
a
type
she'd
are
an
outbound
requests,
so
actually
I
mean
it's
an
employment
liquid,
so
otherwise
out.
The
quest,
even
with
lady
liked
into
the
way
you
can
Terry
and
enjoy
will
clutch
at
the
end
event.
So
so
we
create
an
issue
on
the
underfoot
japchae,
which
our
request
to
the
a
pistol,
and
it
was
merger
to
the
1.0
to
fix
it.
I
I
The
problem
is
that
had
a
complaint
or
an
additional
conflict
when
there
are,
there
are
HTTP
and
headed
east
ECB
service.
Listening
on
the
symport
or
multiple
editors
at
Eastham
service,
seasoning
on
the
same
port
and
the
conflict,
p2p
traffic
are
simply
drop
by
anyway,
without
you
mean
application
areas.
I
So
there
is
a
assumption
here
by
law
in
the
steel
that
PCB
services
should
have
a
Utica
and
you,
the
iWeb
G.
Otherwise,
service
port
must
be
deliberate
allocated
to
avoid
conflicting
with
each
other
with
which
is
not
always
practical
import
action,
because
sometimes
we
can't
just
advantage
a
port
for
larger
micro-services.
I
D
Can
make
a
quick
question?
We
generally
example
of
this
case
to
see.
Is
this
for
redundancy
between
what
really
the
same
in
was
really
the
same
services
running
multiple
slices
are
literally
two
different
two
different
services
performing
different
actions,
but
you're
wanting
them
to
listen
on
the
same
court.
To
give
an
example
of
a
surface
like
that,.
I
So
the
bureau
figure
below
bigger
shows
an
example,
for
example,
where
I
have
so
HTTP
services,
which
maybe
provide
some
little
child
and
will
I
also
have
another
TCP
services.
Maybe
we
have
some
priority
protocol
based
on
p2p,
which
you
simply
still
can't
understand,
or
maybe
just
some
like
FTP
Celsius,
my
god,
but
for
some
reason,
because
of
this
to
service,
are
developed
by
two
teen,
so
for
some
reason
they
they
have
to
choose
this
import
like
a110,
true,
when
we
try
to
integrate
their
associates
into
our
sip.
I
This
is
the
assistant
and
the
Treader
will
build
long
long
or
we
don't
lead
to
this
to
some
type
of
vessel.
They
still
will
complain
that,
okay,
there
are
come
catholic
between
these
two
service
because
we
love
we
we
don't
use.
We
have
she
I
mean
in
our
cup
corporate
castro.
We
don't
use
the
we.
Actually,
we
just
used
a
all
IP,
so
the
p2p
service
will
be
listening
on
the.
I
A
F
I
C
I
think
having
generic
problem
presented
here
is
really
helpful
for
this
group
too,
but
specific
network
problem,
especially
a
solution
portion,
the
network
work,
will
definitely
would
the
sun's
the
right
place
because
these
s,
at
the
end
of
the
day
you
have
to
push
the
changes
through.
The
network
working
group
are
for
networking
specific
changes,
yeah.
F
I
A
I
So
this
is
a
some
our
issues
we
have
experienced
some
fixed
or
partial
assault
problem,
including
support
support,
stick
session
for
weight
based
approaching,
we
created
the
issue
and
it's
what
takes
five
they
enlisted
and
I'm
not
so
was
made,
should
be
used
to
download
Nate
a
focus
on
the
history.
Okay,
already
fixed
and
regular
three
steel.
They
had
high
CPU
usage,
who
will
work
with
council
history
and
included
in
sto
1.1
and
a
fixed
nor
fork
too
many
times.
Wait.
Soft
soft
cost
by
particles
for
the
history
fixing
or
folk.
I
Established
hit
the
collection
destroyed
by
removal,
the
sale,
configuration
from
pared
I.
Think
it's
to
issue
pass
partially
fixed
and
appear
in
marketing
to
distill
below
is
the
same.
We
we
need
help
first,
wine,
the
issue,
one
nine,
seven
one:
the
establishment
to
election
should
be
shouldn't,
be
destroyed,
buying
away
and
it's
the
absent
service
calls
reflection
itself
and
otherwise
the
listener
Catholic
I
mentioned
before
and
when
I
were
working
late,
IP
working
six
support
our
own
pork,
okay,
I.
Think
of
that
at
all
Thanks
you'll
see.
A
A
If
you
do
that,
it
will
list
the
dashboards
that
I
have
implemented
I've
also
on
this
particular
build
gotten
rid
of
all
of
the
log
rotation
stuff,
which
is
those
Scrolls
away.
The
stuff
I
want
to
see
when
I
use
this
to
yo
control
them
have
a
idea
for
that
as
well.
So
if
I
do
dashboard
seiki
ollie
when
I
do
that,
it
does
all
the
stuff
that
we
tell
people
to
do
to
run
key
ally,
does
a
port
forward
and
sets
up
if
it
opens
the
browser
for
that
place
and
just
brings
it
up.
A
A
A
It
brings
up
this
interface
and
if
you
look
at
memory
usage,
it
gives
you
the
live.
Remember
usage.
A
lot
of
people
tell
me
they
have
been
trying
to
look
at
these
kind
of
things
for
is
do
and
they
didn't
they
weren't
aware
this
was
here
so
I
just
threw
this
together,
but
I
was
wondering
if
anyone
liked
this
idea
of
having
a
dashboard
control
I,
you
know,
I
saw
that
link.
Rd
has
one
so
I
thought
we
should
have
one.
But
of
course
we
don't
have
an
integrated
experience.
A
A
I
saw
I,
don't
have
a
zip
can
installed
right
now.
So,
when
I
do
that,
oh
I
do
have
zipped
install
it
says,
there's
a
customize
error
message
for
each
one:
it
says:
there's
no
color
couldn't
find
Zipkin
pod
or
something
like
that.
It
done
it.
It
doesn't
give
you
anything
like
here's,
how
to
use
helm
to
turn
on
Zipkin.
If
you
want
it,
which
would
be
what
I'd
be
looking
for
from
you
Frank
suggestions
on
how
to
handle
those
messages,
and
it
is
the
dashboard
active
as
long
as
you
haven't
under
control,
see
yeah.
C
A
Port
forward
is
just
is
just
running
and
there
it's
the
same
thing
we
use
it's.
The
code
is
the
same
code.
We
use
for
the
proxy
config
manned,
which
does
the
same
thing
as
these
envoy
dashboard.
It
just
sets
up
the
port
forward.
All
the
user
does
control,
see
it's
not
telling
anything
out
like
that.
It
is
up
its
API
machinery.
Api
is
I,
think
to
start
the
port
forward,
I.
F
A
Don't
think
this
is
everything
and
I
think
longer
charming.
I
want
to
go
to
something
where
there
is
a
wrapper,
so
I'm
a
there's
been
some
people
ask
me
to
show
you
this
I
wasn't
really
ready
to
show
it,
but
if
you
ever
have
used
linker
D,
this
is
its
user
interface,
which
I've
sort
of
made
screen
captures
for
so
I.
Don't
have
to
run
liquidy
right
now,
but
maybe
at
the
next
meeting
I
could
show
you
like
sort
of
competitive,
saying
what
is
linker
D.
Have
we
don't
have?
How
do
you
use
it?
A
A
A
But
I
could
I,
could
imagine
like
a
100
pay,
a
hundred
line
UI
that
just
does
sort
of
what
we
saw
for
ki
Ollie
difficult
for
they
need
to
bring
it
up
and
sort
of
points.
People
to
things
like
there
is
a
few
things
in
T
Ali.
The
UI
is
interesting,
but
mostly
it's
extremely
simple
stuff
that
we
already
could
do
you
know
documentation
just
takes
you
to
their
homepage.
A
A
Don't
have
a
good
screenshot
of
it.
I
told
her
I
wanted
to
make
a
new
service,
and
it
asked
me
a
question
about
the
namespace
and
the
service
name.
Then
it
sort
of
popped
up
this
instead
of
having
a
wizard
with
all
these
questions.
For
you
know
what
URL
pass
do
you
do
it
just
hooked
up
a
huge
thing
of
commented
out,
yellow
boilerplate,
so
their
user
interface
with
trivial?
It
just
fills
out
a
template.
You
could
easily
do
something
the
same
way
and
it
may
be
that
easy.
We
easy.
A
A
G
A
So
putting
all
these
dashboards
here,
it's
nice
because
it
lets
people
know
what
in
a
bar,
it's
hard
to
find
them
walking
around
our
dashboard,
and
it
also
lets
us
sort
of
test
them.
But
we
everybody's.
We
have
to
go
through
all
I
to
go
through
all
the
sto
control,
documentation
and
test
everything.
So
it's
sort
of
nice
to
have
them
here
to
remind
you
that
we
have
all
these
things
and
then
maybe
people
just
kiala
or
Jaeger
something
I,
don't
know
your
your
call.
F
G
A
C
A
So
I
agree,
and
for
for
these
are
global,
just
bring
them
up.
These
two
require
a
pod
to
be
selected.
So
then
we
need
a
browser
of
all
the
pods
in
the
users
namespace.
So
if
anyone
has
such
a
tool,
we'd
be
happy
to
you
know:
let
it
collaborate
with
them
on
how
to
how
to
put
this
best
into
the
tool.
There's
also
the.
A
Don't
know
how
to
add
to
the
gy,
but
it's
big
buzz
and
they
could
make
a
prototype
of
how
to
add
some
SEO
stuff
into
that
dashboard.
To
show
cool
things
that
would
be
awesome.
I
would
give
you
a
lot
of
space
at
the
next
one
of
these
meetings
to
demo.
Your
CRM
is
that
that's
that
sub-command,
so
I
have
submitted
I
have
a
couple
other
experimental
ones.
We
could
also
discuss
I
I.
J
Have
a
Christian
yeah,
so
this
is
not
going
to
be
like
the
kubernetes
bore
ride,
because
if
you
use
a
Kubek
at
a
dashboard,
it
actually
asks
you
to
deploy
a
polling,
the
cartridge
issues,
a
dashboard
are
we
are
you
planning
to
do
something
like
that
or
or
you
something
or
you
this
only
in
you
stoke
a
day.
I
would.
A
A
If
you
bring
it
up,
it
brings
it
up
locally.
It's
got
its
own
web
server,
that's
serving
it
right
out
of
this
process,
and
it's
got
this
very
nice.
Ui.
That's
live!
That's
presenting
all
this
stuff.
This
would
be
the
dream.
If
anyone
has
something
like
this
to
contribute,
I
don't
have
one
okay,
cuz
it
to
lead
people
to
the
stuff
that
we
did
have
really
easily,
while
we're
waiting
for
that
to
materialize
and
I,
don't
know
how
to
get
there.
F
A
You
so
it
looks
like
we
only
have
time,
maybe
to
discuss
one
of
these
that
ideas
I
had
for
sub
commands
were
add
to
mesh,
which
just
helps
the
user
manipulate
the
the
injection
settings
on
live
pods.
This
test
connectivity
and
gateways
gateways
is
something
I'd
be
curious
to
know.
People
think
of
it
was
just
because
I
noticed
that
in
the
book
info
documentation,
we
had
many
many
paragraphs
to
help
people
find
where
they
had
deployed
their
stuff.
A
The
idea
was:
wouldn't
it
be
great,
just
type
sto
commercial
gateways
and
haven't
caught
up
the
list
of
all
the
virtual
services
from
all
of
the
gateways
with
the
service
public,
IP
and
note
port
in
there,
so
I'm
just
going
to
start
talking
about
it,
encourage
people
to
read
it
and
make
comments
on
it.
For
the
next
meeting
test,
connectivity
was
a
debugging
thing
just
to
tell
why
the
networking
isn't
working.
Maybe
we
can
skip
talking
about
that.
One
add
to
mesh
with
something
that
I
sort
of
bought
feedback
on
before
I
implement
it
it's.
A
Then
I
would
use
Q
control
patch
to
inject
the
pods
one
at
a
time
to
see
which
which
pod,
when
I
injected
it
caused
things
to
stop
working,
and
it
did
stop
working
at
uninjected
that
one
and
keep
going.
It
gave
me
a
the
next
list
of
which
pods
played
well
with
this
do
and
which
pods
didn't
and
then
I
thought.
Well,
this
particular
command
is
really
easy.
I
could
just
tell
people
to
do
this,
but
everyone
I
talked
to
hate,
cube
control
patch.
They
can't
figure
out
the
syntax,
so
the
idea
was.
A
G
A
If
you,
if
you're
timid,
you
make
a
namespace
with
the
label
without
the
label
for
auto
injection,
you
make
a
new
name
space
and
it
deploy
into
that
namespace.
This
is
for
timid
people
and
then
you
run
add
to
mesh
service
reviews,
but
still
reload
still
works.
Add
to
mesh
product.
Page
reload
still
works
at
the
mesh
ratings.
A
We
deployed
ratings
with
my
sequel
instead
of
the
dummy
one,
it
stops
working,
something
like
that
or
we
add
to
mesh
each
sort
of
thing
to
see
which
thing
does
so
it's
more,
but
just
a
technique
for
people,
because
so
many
people
come
to
me
and
they
just
don't
know
how
to
debug,
and
they
don't
want
to.
This
is
this
is
a
way
of
just
dipping.
Your
toe
into
is
tio
rather
than
your
whole
namespace.
Let's
try
understand
the
difference.
G
A
F
G
A
A
G
A
A
A
So
one
thought
people
I
had
was
you
know
if
do
control
get
pods
that
would
sort
of
work
like
queue,
control
get
pods,
but
tell
you
more
things
like.
If
you
had
a
sidecar
or
what
version
it
was
or
if
the
sidecar
was
healthy,
would
we
want?
We
want
something
that
looked
like
like
that
or
we
want
something
that
looked
like
like
like
I
like
the
is
mesh,
enabled
what
what
did
it
should
it
say?
J
C
And
from
here
I
would
be
interesting
to
know
whether
the
service
has
been
your
mash
enabled
or
has
a
sidecar,
and
what
are
the
rules
will
route
you?
The
service
is
mutually
ours,
they
enabled
and
what
are
the
kind
of
like
the
rules
or
if
they
have
multiple
versions.
I
think
that
would
be
interesting
to
see.
Yes,.