►
From YouTube: Meshery Development Meeting (June 02nd, 2021)
Description
Meshery Development Meeting (June 02nd, 2021)
Join the community at https://layer5.io/community
Find Layer5 on:
GitHub: https://github.com/layer5io
Twitter: https://twitter.com/layer5
LinkedIn: https://www.linkedin.com/company/layer5
Docker Hub: https://hub.docker.com/u/layer5/
A
Welcome
everybody:
this
is
the
mashery
development
meeting
today
is
june
2nd
and
let's
get
started.
I
think
I
I'll
just
give
a
quick,
a
quick
overview
on
what
is
happening
in
measuring
ctl.
A
So
this
is
the
master
recital
command
tracker.
I
guess
some
of
you
might
be
familiar
with
this
because
I
have
been
showing
this
on
every
meeting.
A
A
There
are
a
lot
of
columns
and
that
needs
to
be
filled,
a
lot
of
boxes
that
needs
to
be
green,
color,
so
yeah,
and
we
also
have
bug
hunting
sessions
every
week,
and
I
guess
this
week's
will
be
this
weekend.
I
guess
most
probably
so,
yeah
that's
something
to
look
into
and.
A
We
and
there
are
also
some
open
issues
in
measuring
ctl.
There
are
a
lot
of
open
issues
in
measuring
serial,
but
most
of
these
are
good
first
issues
and
most
of
these
issues
are
related
to
jekyll
are
basically
measuring
dogs,
and
some
of
these
are
related
to
golang.
A
So
you
can
take
a
look
at
those
and
we
we
need
your
help
here
and
that's
it
with
that
said,
if
you,
if
you
like,
to
work
on
it
just
being
just
being
on
the
mercedes
channel
and
yeah,
we
will
assign
some
issues
to
you.
A
So
with
that
said,
let's
move
on
to
drew
who
would
be
demoing
mesh
map
logs
terminal
logs,
rude
yeah,
hey?
I
never
knew.
Thank
you,
hi
hi
there
do
you
want
to
share
your
screens.
B
Can
you
guys
see
my
screen
yeah
so
before
I
start
something?
Let
me
just
let
me
start
with
the
demo.
Then
I
will
tell
you
what
is
happening
so,
as
you
can
see,
I'm
currently
changing
few
of
the
conflicts
of
my
of
my
kubernetes
cluster
and
you
you
can
see
over
here.
There
are
some
things
happening
now,
okay,
so
let
me
get
into
what
we
are
doing,
so
we
wanted
to
add
a
capability
where
measuring,
could
and
act
as
a
terminal
in
which
you
can
interact
with
your
interact
with
your
kubernetes
cluster.
B
B
Thing
which
we
are
doing
is
focusing
on
the
log
part
of
it.
So
currently,
this
particular
terminal
is
only
like.
It
is
hooked
to
a
measuring
operator
over
here,
which
is
in
my
cluster.
B
B
Currently,
what
is
in
works
is
basically
anyone
would
be
able
to
choose
which
cluster
he
wants,
which
part
he
wants
to
get
his
locks
from
where
he
wants
multiple
pods
to
get
his
locks
from
which
container
from
which,
which
container
from
this
particular
pod.
He
wants
a
lot
from
so
now
to
make
that
selection
easier.
B
Inside
mesh
map
also,
so
that's
the
basically
the
idea
behind
this
and
our
efforts
are
going
towards
that.
So
currently,
the
thing
which
I
demoed
was
basically
we
are
working
on
the
capability
for
them
to
choose
from
mesh
map
itself,
but
for
now
I
had
set
this
terminal
to
get
locks
from
machine
operations
so
that
I
could
show
you
how
we.
B
So
that
was
my
demo
and
pretty
fast
explanation.
If,
if
someone
has
any
doubt,
I
can
clear
it
or
I
can
develop
more
into
it.
If
you
like,
like
this
particular
thing,
this
is
a
feature
which
we
add
in
michigan.
C
Yeah
always
kind
of
you
always
kind
of
wonder
that,
don't
you
if,
when
there's
no
questions
so
dhruv,
just
to
add
some
flavor
to
what
you're,
showing
there's
some
intriguing
aspects
of
what's
happening
here.
So
one
of
the
things
is
that
you're,
basically
showing
like
a
live
tale
of
a
container,
that's
running
in
yeah,
exactly
kubernetes
and
near
as
I
understand
like
you
put
while
you
might
not
be
showing
it
right
now
that,
like
the
intention
of
what
you
and
abhishek
are
working
on,
is
that
a
user
might
well?
C
They
might
you
know
they
might
want
to
see
logs
from
one
container,
maybe
multiple
containers
at
the
same
time
or
maybe
or
maybe
even
how
would
it
look
if
they
wanted
to
see
like,
and
I
don't
know,
multiple
containers
at
the
same
time,
but
not
in
the
same
terminal
session.
B
Okay
yeah,
so
that
is
also
a
workflow
which
we
have
thought
about
and
is
on
implementation.
I
couldn't
show
it
because
it's
half
of
work
done
and
half
not
done,
but
what
we
have
in
mind
is
basically
they
will
have.
You
can
think
it
as
a
different
instances,
like
you
have
in
vs,
where
they
will
be
able
to
switch
between
the
switch
between
the
groups
which
they
have
created
like
group.
One
would
be.
B
I
want
the
blocks
from
all
the
containers
which
are
in
for
measuring
operator,
so
that
might
be
one
the
other
group
might
be.
I
want
blocks
from
two
pods
every
container
in
this
two
particular
port,
so
that
might
be
group
two
and
they
will
be
able
to
switch
between
those
two
groups
more
like
there
would
be
a
drop
down,
most
probably
over
here
in
the
right
hand,
side
where
they
will
be
able
to
switch
between
groups,
and
they
would
be
named
appropriately
depending
on
the
port
name
or
something
operator.
D
Than
a
drop
down,
could
it
be
simple
buttons
like
we
have
tabs
that
are
very
easy
to
navigate.
D
Not
like
everything
is
showing
in
the
screen,
but
like
we
have
tabs
from
the
top
bar,
we
can
switch
it
in
our
browser
and
everything
like
this.
Oh
okay,
you're,
okay,
you're
talking.
B
About
this
sessions
that
might
be
doable
like
what
I
had
in
mind
was
something
like
in
vs
code,
we
are
able
to
switch
between
tabs,
they
have
wow.
Oh.
D
B
C
Oh,
hey,
wait!
A
second
look
at
that
that's
kind
of
an
interesting
concept.
I
guess
I
didn't
realize
that
before
so
I
think
what
ayush
was
you
know
suggesting
sort
of
to
to
explore
is
the
use
of
tabs,
as
you
might
find
them
in
your
terminal,
just
those
those
tabs
right
at
the
top,
but
then
also
it
looks
like
so
aligned
with
your
prior
thinking
or
your
current
thinking.
Dhruv
is
the
controls
in
vs
code
for
having
a
drop
down
to
switch
between
them?
That's
that's
one
thing,
but
it
looks
boy
I
don't.
C
C
It
should
should
you
know
abhishek
and
yourself,
and
whoever
else
is
working
on
this
arrive
at
a
point
wherein
you
know
if
the
if
the
workflow
for
a
user
is
that
they've
they've
selected,
maybe
a
handful
of
pods
or
a
handful
of
containers
and
they
start
to
get
logs
and
it
turns
out
either
there's
just
too
many
logs
coming
through.
They
can't
keep
up
or
they
accidentally
selected
a
container
that
they
actually
they
don't
need
logs
from
and
so
they've
established
a
session,
but
they'd
like
to
basically
remove.
C
You
know
like
eject
one
of
the
containers
from
the
current
active
sort
of
multiplexed
stream
of
logs
that
this
type
of
a
control
like
if
there's
a
boy
getting
emails.
If
there
was
a
session
here
and
then
configuration
of
that
session
like
if
there's
a
sub
list
of
the
names
of
the
pods,
they
might
have
a
check,
this
might
be
a
checkbox.
Then
they
can
like
uncheck,
that
part
on
check.
C
You
know
what
I'll
I
won't
bore
everyone.
I
will
show
you
what
that
later,
but
yeah
I'm
so
sorry.
I
can't
believe
it
by
the
way,
just
as
a
side
note,
because
if
I
don't
do
it
now,
I
might
forever
lose
my
opportunity.
Dhruv
is
that
is
that
a
covent
cut?
Is
that,
like
you
know
that
that's
too
lazy
to
go
out
but
okay,
cool
good,
I
just
okay,
good
good
and
anyway
do
other
people
have
thoughts
for
for
dhruv
like
the
there's
I'll
toss.
C
One
in
I
don't
know
if
everyone
is
necessary
and
drew.
If
you
don't
mind,
if
you
still
have
it
up,
if
you're
able
to
share
just
kind
of
look
at
it
like
one
of
the
cool
things
about
the
log
viewer
that
that's
being
worked
on
here
is
that
well
it
opens
up
precedence
for
more
real
time.
You
know
just
more
real-time
feedback
from
the
infrastructure
under
management,
so.
C
C
C
The
use
cases
that
I
just
the
use
case,
I
just
said,
is
related.
It's
not
exactly
the
same,
but
but
something
for
all
of
you
to
think
about.
The
other
thing
I
think
to
think
about
is
while
dhruv
was
showing
this
capability
in
context
of
the
mesh
map
plug-in.
I
think
the
in
the
intention.
C
I
think
the
intention
here
is
probably
to
like,
if
you
consider,
if
you've,
let
me
bring
up
mesher's
ui.
If
I
have
it,
if
you've
used
the
dashboards
like
the
control
plane
dashboard,
I
may
not
have
started
measuring
just
yet
then
you'll,
you
will
have
seen
in
your
system
when
you're
running
a
service
mesh,
it
will
tell
you
meshrey
will
tell
you
the
details
of
your
service
mesh,
I'm
starting
mystery.
Now.
C
A
So
I
guess
we
have
a
couple
of
newcomers
on
the
call
ankit,
manish
and
dabi
and
manish
is
having.
I
has
a
mic
problem,
but
if
you
would
like
to
introduce
yourself,
you
can
turn
your
mic
on
and
yeah
just
introduce
yourself
or
you
can
write
down
in
chat.
So
anybody
want
to
introduce
themselves.
E
F
G
Hello,
hello,
everyone,
my
name
is
abhishek.
For
short,
you
can
call
me
avi
and
I
am
again
a
junior
from
new
delhi,
india
and
I
want
to
learn
more
about
service
meshes
and
distributed
system
and
micro
services
in
all.
So
I
just
find
this
project
and
layer,
5
organization,
and
it's
it
really
sounds
amazing
to
me
so
that
I
can
learn
and
at
the
same
time
I
can
contribute
in
a
meaningful
way,
so
yeah.
A
Thank
you.
Thank
you
guys
welcome
to
the
community.
We
also
have
a
newcomers
call
that
is
taking
place
tomorrow.
I
guess
you,
you
will
probably
have
access
to
the
community
calendar,
if
not
just
ping
ping,
on
the
chat
and
someone
will
get
that
sorted
out.
So
we
have
a
new
commerce
call
tomorrow
and
that's
a
good
place
to
get
started
and
get
yourself
oriented
with
the
layer.
5
projects
awesome.
So,
let's
move
on
to
the
next
item
on
our
agenda.
F
H
Mention
you
I
was
not
supposed,
I
I
hope
no
one
saw
the
flash
the
flash-
that's
not
something
to
be
seen
as
of
now
that's
in
work.
Today's
demo
is,
unfortunately,
not
of
that.
It
is
about
being
able
to
configure
virtual
services
to
provision
virtual
services
and
control
virtual
services
using
pardon
files.
I've
removed
pattern
files
quite
a
lot
quite
a
lot
of
time.
This
time,
it's
different,
I'm
not
doing
the
same
thing
this
time.
H
It's
about
being
able
to
create
a
virtual
service
and
completely
configured
using
pattern
files.
You
would
be
able
to
access
almost
everything
that
is
configurable
in
a
virtual
service
using
patterns,
so
I'm
demoing
a
very,
very
basic
functionality
here.
Basically,
I'm
setting
up
a
http
route,
it's
named
route
and
I'm
setting
up
the
destination.
This
is
stable.
As
you
see,
these
services
does
not
exist,
but
right
now
the
demo
is
about
being
able
to
create
virtual
service
and
configuration
using
buttons.
So
I
have
my
seo
adapter
running.
H
It
is
the
one
which
would
be
able
to
create
virtual
service
because
it's
a
construct
of
istio
and
I
have
my
mystery
server
running-
I'm
going
to
apply
the
pattern
and
wait
for
it
to
be
applied
while
it's
getting
upside.
Let
me
explain
what
the
pattern
is.
So
here
I'm
trying
to
provision
sd
service
mesh,
which
I
think
is
already
present
in
my
cluster,
but
yeah,
I'm
trying
to
provision
seo
service
mesh
of
version
1.9.1
and
then
I'm
trying
to
provision
a
virtual
service
create
a
virtual
service.
H
It
depends
on
instance,
so
basically
it
says
that
hey
once
you
create
once
you're
sure
that
it
still
exists
on
my
system.
If
it
didn't
create
it,
if
it
does
it
don't
do
anything
then
go
ahead
and
create
a
virtual
service
for
me,
so
it
has
given
me
logs
and
it
has
successfully
created
the
things
that
I
asked
it
to
create
and
it
okay.
It
was
already
running,
but
let's
see
if.
H
If
this
has
been
created
a
virtual
server,
so
here
I'm
what
I'm
doing
is.
I
am
trying
to
see
if
a
versus
service,
which
is
in
a
networking
api
istio's
networking
api
group
has
been
created
in
or
not
in
test
name
space,
so
yeah,
I
think
it
is.
It
is
the
same
configuration
that
we
actually
set
up
there
most
of
the
time
this
configuration,
the
configuration
that
we
are
doing
here
would
be
exactly
shown
here,
but
that
would
be
the
case
99
percent
of
time.
H
Sometimes
it
will
be
possible
that
we
tweak
a
few
things
that
we
provide
in
here
and
then
we
actually
pass
it
on,
but
the
basic
idea
is
that
you
should
be
able
to
create
with
the
services
using
patents
and
we
should
be
able
to
completely
configure
configuration
and
manage
its
life
cycle
using
patterns.
The
same
the
same
thing
would
be
also
accessible
in
the
ui
very
soon.
That
is,
you
should
be
able
to
have
a
dynamic
form
in
there
where
you
can
create
or
destroy
or
basically
manage
water
service,
then,
basically,
all
the
resources.
H
That's
something
I'm
not
going
to
remember
today,
because
it's
very
ugly
right
now
but
yeah.
If
anyone
has
comments
or
suggestions
of
feedback.
C
H
Yeah
yeah,
so
it
will
be
picked
up
by
mesh
sync
also
so
mesh
think
is
right
now
capable
of
completely
capable
of
detecting
a
lot
of
resources
in
the
humanities
cluster.
So
when
we
actually
basically
provisions
something
using
patents,
so
you
can
see
me
actually
it's
not
yeah
basic.
H
Basically,
the
idea
is
that
we
we
will
place
some
annotations
in
the
metadata
which
will
help
a
mesh
think
able
to
figure
out
that
what
all
right
so,
let's
say
that
we
are
creating
this
virtual
service
and
which
is
services
single
resource,
not
a
very
good
idea,
but
not
a
very
good
example.
But
let's
say
we
are
creating
a
virtual
service
here.
So
in
order
to
be
able
to
get
that
from
mesh
sync,
what
we
actually
do
is
we
will
be
placing
some
annotations
in
the
resources
that
we
are
creating.
H
So
it
is
very
much
possible
that
one
service
that
is
listed
here
is
actually
going
to
create
a
lot
of
native
humanities
resources
right
now
for
virtual
service.
As
I
was
saying,
it's
not
a
very
good
example
because
in
this
case
it's
going
to
be
creating
just
one
virtual
service,
but
that
virtual
service
will
have
some
annotations
so
that
measuring
can
identify
that.
Okay,
these
10
resources
are
actually
supposed
to
represent
one
application.
H
So
basically
it's
a
way
that
we
are
synchronizing
what
resources
patents
are
creating
with
what
mishink
and
and
basically,
ultimately,
mishma
can
basically
understand
it's
a
way
to
synchronize
patents,
resources
with
native
keys.
If
that
made
sense,
I
I
I
I
said
everything
in
here:
I'm
not
sure
if
actually.
I
C
Well,
between
that
and
the
spinning
cursor
like
it's
a
little
too
fancy
for
me,
I'm
not
a
hard
time
keeping
up
so
on
line
nine.
Well,
what
was
formerly
12,
which
you
know
at
this
point
like
I'm
just
totally
confused-
is
that
does
that
vsc
stand
for
virtual
service,
something
something
or
is
that
a
misspelling
of
svc?
I.
H
Know
it's,
it
doesn't
stand
for
anything.
So
the
thing
is
it's
the
name
of
the
resource
that
I.
C
H
Just
okay
got
it,
so
you
can't
name
it
anything.
It's
not
it's,
not
a
part
of
syntax.
You
can
call
it
anything.
C
So
is
samir
on,
or
was
that
smear
just
now?
No,
it
was
okay.
Well,
we'll
have
to
send
samir.
The
snippet
of
this
recording
this
is
is
great,
is
just
what
he
was
looking
for.
Are
there
any
aspects
of
the
ways
that
you
can
configure?
C
C
Yeah,
let
me
rephrase
like
so
in
in
virtual
services,
you're
able
to
users
are
able
to
configure
virtual
services
to
you,
know
somewhat
complexly
or
more
complex
than
what
we
see
here
by
by
way
of.
C
Well,
either:
well,
I
I
guess
not
not
going
through
the
examples,
but
there
are
other
options
that
people
can
add
in.
Are
there
any
ways
that
people
can
configure
a
virtual
service
that
you're
aware
of
that
isn't
being
supported
here
like.
H
Okay,
so
so
it's
it's
a
question
like
that
there
so
okay,
so
I
hope
the
question
is:
is
that
is
right
now
the
schema
that
we
are
capturing,
that
is,
that
is
actually
1.9.1.
The
scheme
virtual
service
schema
that
1.9.1
actually
supports.
So
it's
a
question
like
that.
What
if
someone
is
using
seo
dot
eight
and
it
has
a
construct
of
virtual
service,
so
will
we
be
able
to
configure
that
or
was
the
question
that
if
a
virtual
service
already
exists
and
can
can
patent
configure
that
one.
C
Both
of
those
are
good
questions.
I
would
like
to
know
the
answer
to
both
of
those,
although
what
I
so,
I
think
the
second,
the
second
question
is
that
you
would
ask,
is
so
all
right,
so
three,
three
questions,
one.
Is
this
a
complete
implementation?
H
No
yeah
so
as
of
1.9.1,
I
I'm
not
sure
if
there
is
a
latest
like
more
latest
version,
so
everything
that
it's
a
crd
would
support.
That
is
everything
that
you
can
actually
go
in
the
virtual
service
and
basically
configure
it.
Everything
is
actually
exposed
here.
C
Nice
and
then
for
the
other
two
questions
yeah.
I
guess
you
know
it.
These
might
lead
to
longer
conversations
than
we
should
given
the
agenda
that
we
have.
But
it
sounds
like
you
know,
so
so
I'll
make
a
couple
of
statements
and
then
you
can
clarify
so
current.
You
know
so
measury
adapters
service,
mesh
adapters
that
are
made.
You
know
that
for
measuring
they
are
capable
of
supporting
multiple
versions
of
a
given
service
mesh,
so
letting
someone
provision
different
versions
of
those
meshes.
C
They
inherently
they
don't.
I
think
we've
done
probably
it
hasn't
been
a
strong
use
case
and
we've
probably
done
a
poor
job
in
the
past
of
allowing
someone
to
manage
multiple
instances
of
the
same
type
of
service
mesh
that
might
be
running
at
different
versions
and
actually,
in
fact,
whether
or
not
they're
running
at
different
versions,
we've
probably
done
a
poor
job
of
letting
them
identify
which
of
the
two
different
instances
of
the
same
service
mesh.
C
Do
they
want
to
send
subsequent
operations
to
there's
probably
some
room
for
improvement
here,
which
means
there's
probably
room
for
improvement
with
respect
to
kind
of
the
first
question
that
you
said:
karsh,
which
is
yeah
if,
for
example,
you're
the
user,
is
configuring
a
virtual
service
and
maybe
they're
using
a
specification
that
only
works
with
istio
1.9.1?
H
Yeah,
so
for
actually
the
answer
varies
for
different
resources,
but
if
we
talk
about
only
virtual
service
and
one
dot
seventh,
so
I
get
the
question
it's
like
what,
if
they're
incompatible
between
crds?
In
that
case
this
will,
we
will
get
an
error,
because
the
schema
would
kind
of
let
it
pass,
but
it
would
notice.
The
ics
controller
would
basically
complain
that
it's
not
valid.
H
I
think
that
mission
control
would
come
in,
but
yeah
we
will
get
an
error
for
1.7
or
for
anything
that
is
being
provisioned
with
istio
with
the
istio
adapter
mysteriously
adapter.
That
will
work
just
fine.
The
reason
is,
the
reason
is
that
mystery
allows
three
latest
version
by
default
and
those
three,
because
we
are
the
right
now,
the
schema
that
we
have
belongs
to
v1
alpha
three
virtual
service
schema
that
we
are
using
belongs
to
v
one
alpha
three
seos.
We
went
for
three
viagra
and
the
three
latest
versions.
H
They
support
that.
That,
basically,
is
crd.
It
is
recently
promoted
to
v1
vr1,
I
suppose,
but
even
after
three
hasn't
been
duplicated,
so
yeah.
Basically,
if
it's
being
provisioned
for
via
mystery,
it
won't
be
an
issue,
but
definitely
it
would
be
an
issue
in
case
someone
is
running
a
really
old
version
of
istio
and
its
schema
varies.
Its
virtual
schema
varies
with
the
one
that
we
have.
The
admission
control
will
complain.
C
I'll
I'll
withhold
my
questions
we
can
get
to
other
stuff
drew
fab.
Has
a
follow-up
question.
H
We
don't
do
version
control.
There
is
a
construct
for
version.
That
is,
you
can
place
an
version
in
there,
but
it's
more
like
you
are
using
a
version
version
control
system
that
you
are,
you
have
a
vcs
and
because
you
are
going
to
change
it
here,
so
vcs
is
going
to
appear.
So
basically,
it's
like
mesri
does
not
have
a
version
control
system
over
this.
So
if
you,
if
you
change
it
to
1.0.2,
it's
same
as
1.0.1,
at
least
for
now,.
H
Hello
yeah,
I
think
my
internet
connection
is.
C
Just
yeah
briefly,
the
this
new
capability
that
utkarsh
is
demoing
around
the
use
of
pattern
files.
So
this
yaml
file
here
it's
really
concise.
It's
simp,
a
simple,
yet
powerful
way
to
configure
to
describe
your
service
mesh
configuration.
C
Your
application
configuration
the
behavior
of
the
mesh
that
you
would
like
to
like
if
you'd
like,
to
tell
it
to
do
something
like
do
a
roll
out
and
canary
things
or
or
have
a
traffic
test
like
a
performance
test
run.
All
of
those
things
can
be
described
inside
of
a
pattern
file.
C
It's
the
istio
adapter.
That
really
has
this
support
today
and
there's
a
bit
of
mostly
copy
and
paste
that
needs
to
happen
between
the
implementation
done
in
the
measure
istio
adapter
to
other
adapters.
So
if
you're
on
the
call
and
you're
into
golang,
then
this
is
a
great
way
to
help
propagate
this
new
pattern.
Service
mesh
pattern
capability
across
other
service
mesh
adapters,
because
it's
largely
a.
J
C
And
paste
thing,
but
it'll
get
you
in
there
and
learning
service
meshes
measuring
and
service
mesh
patterns,
so
we'll
just
it's
just
a
call
out
for
a
volunteer
or
three
on.
A
Next
yeah,
actually,
I
guess
we
can
move
on
to
the
next
item.
We
have.
J
A
It's
showing
that
we
are
viewing
your
screen,
but
I
guess,
could
you
like,
stop
the
share
and
try
it
again.
A
A
To
give
you
some
context
on
what
our
has
been
working
on,
we
have
both
rest
apis
and
graphql
apis
for
measuring
and
for
rest
apis.
We
are
generating,
we
are
automatically
generating
docs
using
swagger
and
push
has
been
working
on
it,
and
so
basically,
what
anishka
is
trying
to
do
is
to
redesign
the
automatically
generated
docks
to
fit
into
the
measuring
docks.
J
J
Yeah
so
as
suggested,
I
went
through
the
available
style
guides
that
were
available,
so
I
have
like
used
the
font
and
everything
according
to
that,
and
the
other
thing
that
was
suggested
by
lee
was
like
having
the
colors
that
are
according
to
like
the
brand
kit
we
have
already
so
I
use
those
colors
only
and
I
have
like
made
them
into
a
little
lighter
tone
and
we
did
not
have
any
color
that
was
in
red,
so
I
used
something
that
was
like
that
I
thought
might
fit
with
this
like
design
so
yeah.
D
The
documentation
is
to
be
implemented
in
the
machinery
ui.
I
am
not
sure
about
that.
C
C
C
Sure
yeah,
so
I
usually
just
politely
pointing
out
that
the
template
that
was
used
so
that
yeah
an
anushka
you
are
using
the
screen
template
for
the
meshery
ui
in
the
application
application
and
that
the
and
it
looks
good.
Although
he's
just
highlighting
that
the
fact
that
the
rest
api
reference
itself,
it'll
it'll,
end
up
being
inside
of
mesh
redox.
And
so
I
think
we
can
sort
of
safely
ignore
the
yeah
yeah.
C
Some
feedback
from
me,
which
is
one
we
don't
often
get
to
see
and
review
things
in
figma
on
this
call.
This
is
nice.
It's
a
hi!
It's
you've
put
together
a
high
quality
design
like
you,
there
aren't
small
errors
that
you
have
sitting
around
in
here.
C
Like
you,
don't
all
of
the
everything
that
you
have
in
here,
you
did
on
purpose,
which
is
nice
and
that's
different
than
which
is
means
that
you've
done
a
quality
design
which
I'm
really
thankful
for,
because
not
everyone
does
that
so
kudos
on
that.
A
couple
of
items
that
for
you
to
consider
one
is
that
it's
not
it's
unobvious
and
it's
actually.
C
I
don't
think
it's
there's
two
red
colors
and
I
don't
know
that
they're
in
the
mesh
reui
figma
there's
a
layer,
five
ui
sigma
project
that
does
have
two
colors
and
I'll
point
those
out
to
you
and
you
can
test
to
see
if
they
work
both
of
them,
both
of
them
kind
of,
like
the
greens
kind
of,
like
the
other,
the
teal
blue
and
the
caribbean
green.
C
That
are
two.
I'm
gonna
use
the
wrong
color
term
here,
so
everyone
bear
with
me
but
they're,
two
deep
in
color
like
they're
too,
like
you've,
you've
faded
them
out
here,
which
makes
things
legible.
So
that's
that's
great,
like
it's
the
same
color,
but
it's
not
as
intense
and
so
nice.
C
The
effect
of
that
doing
so
on
the
green
means
that
the
white
text
and
the
light
green
have
a
fairly
low
contrast
and
so
give
it
some
thought,
like
maybe
we've
hit
the
inflection
point
by
which
the
contrast
is
too
it's
not
a.
Maybe
that
needs
to
be
black
text,
maybe
maybe
not
last
point
of
there's.
Okay,
two
other
points
of
feedback
under
the.
C
C
Other
eye-
and
these
are
just
like
hey-
I
look
like
if
we
went
forward
with
just
this,
like
it's
pretty
good.
These
things
that
I'm
getting
to
are,
you
know
you've
done
a
good
job
if
I'm
having
to
get,
if
I'm
not
really
having
to
reach
for
to
give
feedback,
so
so
take
that
with
a
grain
of
salt.
That
is
that
I
think
I
would
contend
that
the
current
table
that
you,
that
and
by
the
way,
everyone
just
again
like
novendou,
was
giving
context
there's
a
version
of
this
implementation.
C
That's
in
docs.measure.io
today
that
anushka
is
improving
on,
and
this
is
an
improvement,
but
the
table
in
there,
the
name
format
type.
I
would
contend
that
that
table
may
not
need
the
coloring
of
that
table,
I
think
is,
I
would
suggest,
is
probably
good
to
go,
as
is
it
might
need
some
font
changes.
I
don't
recall,
but
but
what?
But,
I
wouldn't
necessarily
add,
advocate
for
a
gray
background
in
the,
if
you're
going
to
use
a
gray
background
on
the
table
that
it
would
just
be
alternating
rows.
So
the
people.
D
So
that's
one
thing:
please
try
removing
the
lines
border
lines,
let
them
let
them
be
transparent
and
have
a
circular
sense.
That
would
look
nice.
J
That
we
have
here
for
these
get
enclosed
right.
J
A
Yeah,
thank
you
anishka.
Next,
we
have
gabriel
who
will
be
demoing
measuring
filters,
filter,
ui,
so
filters
basically
means
awesome
filters.
Web
assembly
filters,
so
we
we
use
that
to
yeah.
We
we
we
apply
those
filters
to
ny
proxies
and
we
can
use
it
to
configure
configure
that
so
gabra
has
been
working
on
adding
a
user
interface
in
measuring
for
for
applying
filters.
F
Yeah
so
hope
I
am
audible
properly
and
my
internet
connection
remains
stable
for
the
time
yeah.
So
recently
we
have
added
support
for
the
filters.
That
is
wasn't
filters
and
mischief
into
just
an
overview
of
that.
So
today
I'll
be
demonstrating
the
crud
operations
on
the
filters,
and
we
will
be
also
doing
a
difference
from
the
patterns
one
so.
F
C
F
See
the
filter
file
is
uploading
and
deleting
successfully
and
the
major
difference
what
we
have
in
filters
from
the
pattern.
One
is
that
in
the
patterns
ui
here
we
can
see
we
can
edit
the
patterns
from
here
the
ui
itself,
so
using
that
update
button
over
here,
but
in
the
filters
ui,
we
don't
allow
the
users
to
edit
it
from
over
here
so
later
on.
F
We
are
thinking
of
applying
a
major
form
of
react,
json
schema
to
which
a
user
can
create
a
filter
over
from
the
machine,
ui
itself
filter,
ui
itself.
So
that's
another
quick
demo
for
the
filter.
One.
A
To
add
to
this
like
gaurav-
and
I
are
working
on
like
the
deploy
part
like
currently,
it
doesn't
deploy
these
filters,
so
you
can
just
only
perform
the
crud
operation.
So
I
have
some
questions
regarding
that,
but
I
I
guess
I
will
post
that
in
slack
to
save
everyone's
time
here.
A
C
Garage
in
terms
of
an
item
that
you
might
get
well
a
couple
of
things,
one
if
you
would
consider
consider
these
things,
maybe
the
answer
is
no.
These
are
stupid
suggestions.
So
again,
if
that's
the
case,
you
can
provide
that
feedback.
Privately,
that's
fine,
I'm
joking!
Actually,
you
should
really
provide
that
publicly.
As
a
matter
of
fact,
just
as
a
reminder
to
have
everyone
on
the
call,
please
do
as
much
as
you
can
publicly
for
a
few
reasons.
C
One
everybody
learns
at
the
same
time
when
you
do
it,
it
gives
everyone
an
opportunity
to
engage
and
express
opinion
or
help
or
do
things,
and
it
also
helps
the
few
maintainers
that
are
here
not
avoid
repeating
themselves
like
answering
it.
You
know
the
same
question
a
bunch
of
different
times
to
different
people
and
so
think
think
twice
some.
Some
conversations
have
to
be
private,
like
if
you're
gonna
tell
me
that
his
suggestion
is
stupid,
then
I
suggest
that
you
do
that
publicly.
C
Really
I
do,
but
anyway
I
recognize
some
things
need
to
be
private.
Fine,
maybe
you
start
private
but
try
to
bring
it
public
as
often
as
you
can.
It's
just
it's
just
it's
nice.
A
couple
things
you
might
consider
is
that
in
the
table
itself,
at
some
point,
users
might
want
to
do
bulk
actions
like
they
want
to
delete
two
filters
at
the
same
time,
and
so
on.
Yeah.
F
And
one
more
thing,
I
would
like
to
point
out
that
once
in
the
patterns
also
and
in
the
filters
also
so
when
the
user
is
uploading,
the
patterns
of
the
filters
file
and
we
click
on
delete,
so
it
gets
deleted.
But
we
need
to
click
over
here
to
close
that
pop-up
window.
So
I
think
another
feature
should
be
added
that
it
should
be
closed
automatically.
C
Maybe
some
bulk
actions
for
things
like
delete
most
notably
and
then,
if
you
would
there's
an
api,
there
should
be
a
mastery
server
api
that
and
the
vendor
can
not
have
said
that
this
is
true
or
deploying
the
filter.
So
if
a
filter
has
been
uploaded
and
it's
time
to
like
apply
it,
there
should
be
a
deploy
endpoint.
C
The
deploy,
endpoint
may
or
may
not
do
what
it's
supposed
to
do,
but
from
the
uis.
From
your
perspective,
that's
not
your
that's
somebody
else's
problem,
so
if
you
would-
and
since
this
is
a
a
beta,
an
experimental
area,
if
you
would
go
ahead
and
take
that,
I
suggest
like
take
that
next
step
and
put
in
the
action
to
do
a
deployment
and,
let's
start
to
that'll,
help
us
think
through
what
that
user
interaction
needs
to
be
because
it
may
not.
It
may
be
that,
like.
Oh,
that
is
a
stupid
suggestion.
C
A
Yeah,
I
was
just
about
to
mention
that,
like
do,
we
need
to
add
a
deploy
endpoint
like
we,
we
can
just
pass
passing
this
filter
file
directly
to
the
adapter
right
and
the
adapter
handles
the
rest
of
the
stuff.
A
That
is
that
that
is
what
I
was
thinking
like
fine
for
that.
A
I
I
Yeah,
so
I
am,
this
is
like
the
first
term
attempt
to
get
some
interactivity
into
mystery
and
so
I'll
just
term
we're,
starting
with
the
context
delete
of
function.
So
I'll
just
show
my
config.yaml.
I
I
So
we
want
to
give
the
user
a
choice
to
set
to
set
a
context
which
will
be
the
next
current
context
once
they
delete
the
existing
one.
I
I
Now,
let
me
let
me
try
to
show
you
how
the
flag
works
so
because
this
has
to
be
auto
scriptable
so
that
it
can
be
automated.
I
have
added
the
set
flag.
The
set
flag
was
chosen
because
it
was,
it
was
already
part
of
the
context
command.
I
think,
and
a
similar
user
experience
is
always
good,
so
the
set
flag
and.
I
Yeah,
okay,
this
this
is
a
bug
I
have
to
fix
the
silent
flag
but
yeah
other
than
that
this
is
it
it's
a
pretty
short
demo.
A
F
C
With
a
command
like
measuring
ctl
system
context
view
that
you
know,
I
refer
to
that
one
fairly,
often
to
figure
out
like
what
context
am
I
currently
in
and
and
and
you
know,
and
then
that
command
works
quite
well
for
telling
me
the
name
of
my
context
and
it
also
spits
out
with
you
know,
like
the
the
full
text
of
that
view.
C
However,
I
could
see
that
changing
into
an
interactive,
like
I
didn't
specify,
I
said
message,
ctl
system
context
view,
and
I
didn't
specify
what
context
I
want
to
view
the
command
itself
just
says.
Well,
if
the
user
didn't
provide
it,
then
we'll
just
show
them
the
the
current
one.
But
you
can
imagine
that
you
know
there.
We
will
just
show
them
the
details
of
their
current
context.
C
You
could
imagine
that
changing
to
oh
well,
not
only
will
we
give
them
the
name
of
their
current
context,
but
maybe
also
we
would
give
them
an
interactive
list
so
that
they
could
pick
which
one
they
wanted
to
see
details
of
by
the
way
adida
or
others.
C
If
you
go
to
change,
if
you
go
to
take
on
that
particular
use
case,
if
you
don't
mind,
give
consideration
to
the
fact
that
it
is
a
fairly
frequent
use
case,
at
least
for
me
that
I
don't
know
what
context
I'm
in
and
I'd
like
to
see
that,
and
so,
if
you
change
the
system
context
view
command
to
an
interactive
prompt.
C
C
It's
you
know
mastery
ctl
system,
context,
you
know
view
current
or
or
you
know
I
don't
know,
but
it
would
be
a
shame
to
change
that
around
enter
the
user
into
an
interactive
mode
that
they
didn't
really
want
to
go
into,
because
you
could
see
we
would
go
into
I'm,
hopefully
everyone's
tracking,
what
I'm
saying,
but
you
you
could
go
into
the
interactive
mode
and
it
would
probably
have
an
asterisk
next
to
the
current
context,
and
then
that
would
tell
me
I
would
cancel
out
of
that
to
answer
my
question,
which
is
what
is
my
current
context:
I'll
write
this
down.
A
Awesome
yeah,
yeah
yeah,
that's
something
we
should
consider
about.
Yes,
we
are
over
time.
Should
we
like?
Do
you
have
a
quick
update
on
mercedes,
github
action.
K
Yeah,
I
do
actually
okay,
so
small
recap:
I'm
working
on
a
github
action
to
perform
service
mesh
performance
test
and
over
the
last
few
weeks
I
finally
got
around
to
making
it
an
actual
action,
because
initially
it
was
just
a
workflow,
but
now
I'm
starting
to
port
this
over
to
natural
action,
and
I
would
like
to
show
what
that
looks
like
today.
K
Sure
is
my
screen
visible,
yeah,
okay,
so,
but
so
basically
one
of
the
requirements
is
the
user
should
be
able
to
pass.
In
a
token-
and
that's
what
I
have
right
here
and
it's
a
small
description
so
when
this
action
is
to
run,
it
needs
a
docker
file,
which
is
what
I
have
here,
but
so
this
would
deploy
measuring
and
by
the
time
the
container
starts
up.
It's
going
to
call
this
shell
script,
which
I
have
here.
This
will
actually
perform
the
tests
and
get
the
user's
token,
which
I'm
doing
up
here.
K
C
That's
great,
this
is,
you
know,
certainly
further
or
you're
you're
continuing
to
advance
here.
This
is
good.
I
just
I
I
I
haven't
seen
some
of
this,
which
is
nice.
No
doubt
you've
you've
been
considering
this
as
you've
gone
to
sort
of
shift
from
kind
of
a
statically
configured
workflow
more
toward
an
action
in
which
you're
abstracting
some
things
away,
you're,
requiring
inputs
and
and
kind
of
and
taking
those
variables.
C
If
you
would
you
in
and
consider
this
more
deeply
or
if
your
thought
is
totally
finished
and
fine,
but
the
question
is:
do
you
think
that
we
can
arrive
at
a
single
github
action
called
you
know
the
the
meschery
ctl
github
action,
which
is
intended
to
really
take
it's?
It's
not
it's
not
specif
like
it
is
not
one
specific
for
just
smi
conformance,
there's,
not
one
specific
for
just
performance
tests,
but
there's
a
generic
one
that
you
know
conveniently
allows
people
to
use
it
for
whatever
measuring
ctl
action.
C
They're
looking
to
do,
or
is
that
too
much
is
that
does
that
make
it
unwieldy
like?
Does
it
make
the
action
too
hard
too
difficult
to
use
and
too
difficult
to
create.
K
Oh
no,
I
definitely
think
it's
possible
if
I
was
an
action
user.
I
would
actually
like
that
such
that
I
could
I
could
specify
if
I
want
performance
test
or
conformance
tests,
and
then
the
action
just
takes
care
of
all
of
that.
So
it's
something
I
hope
to
work
towards,
but
I'm
trying
to
get
this
done
first
and
update
this
as
I
go.
C
And
you
build
for
my
part
if
it
greases
the
wheels
for
your
efforts,
I'd
be
happy
to
meet
separately
to
just
provide.
You
know,
try
to
help.
You
advance
the
effort.
C
C
Nice
good,
if
you
guys,
have
feedback
for
gibraltar.
Please
go
see
him
in
the
slack
channel
and
we
had
one
more
item
today
and
so
and
navindu
as
you
consider
entertaining
that
other
items
that
you
might
want
to
say:
hey,
no
one
is
obligated
to
hang
out
for
longer.
Some
of
us
will
talk
about
measuring
all
day
long.
C
I'm
I'm
game
for
that,
but
not
all
if
you
have
that
luxury
or
are
that
interested
and
so
navindy
if
I
use
speaks
next,
maybe
just
you
know
thank
everybody
for
the
time
here
and
and
yeah
get
them
off
the
hook.
If
they
need
to
go.
A
Yeah
so
guys,
thank
you.
Thank
you
for
joining
the
call,
so
for
newcomers
we
have
a
call
tomorrow
and
for
others
I
guess
we
will
meet
friday
on
the
community,
call
bye,
guys
and
over
to
you
eyes.
D
Yeah
yeah,
so
I
was
just
asking
when
we
go
to
this
route,
and
that
is
a
management
we
have
to
select
which
one
we
have
to
see
which
adapter
we
need
to
access
right
now.
D
D
I
was
trying
to
use
figma
for
the
first
time,
something
like
the
we
have
buttons
on
the
top
so
that
it's
easy
to
access
and
rather
than
having
these
cards
as
every
time,
we
won't
be
needing
these
descriptions,
and
these
tags
are
very
big
to
cover
the
whole
screen,
something
like
it.
We
could
have
something
some
lists
list
like
these,
and
this
could
be
expanded.
D
Also,
I
was
there
was
question
just
I
don't
know
how
why,
when
we
have,
we
can
access
access
the
different
type
different
adapters
from
this
url
then
why
we
are
having
separate
urls
for
each
and
every
one
like
like
for
coma
hto.
D
Could
we
just
remove
these
and
have
only
this,
and
this
thing.
C
Yeah
yeah:
this
is
you
guys,
are
on
onto
something
pisses
me
off.
No,
this
is
correct.
Yeah,
you
guys
are
onto
something
I'm
the
so
by
the
way,
there's
a
factoid
floating
around
somewhere
out
there
in
user
experience
like
user
design,
land
that
says
a
human
really
only
likes
to
reason
over
or
can
reason
over
about
seven
inputs
at
a
time.
C
I
don't
have
a
reference
for
you
for
that
where
that
comes
from,
but
but
how
many
inputs
are
here.
I
don't
know,
there's
like
nine
of
these
yeah.
Actually
it's
disappointing
because
there
should
be
10
because
we
have,
but
but
but
anyway,
so
why
well?
Okay,
and
understandably,
if
you
understand
that,
then
why
do
we
have?
You
know
nine,
not
only
those
nine
plus
the
other
nine
above
it
well
for
marketing?
That's
why.
C
Because
screenshots
then
show
the
fact
that
measure
is
quite
powerful
and
you
know
manages
a
lot.
Okay,
fine.
I
think
we
passed
that
point.
I
think
it
is
understood
by
most
that
when
they,
if
they're
touching
mesri
or
they're
looking
at
meshri,
that's
why
it's
called
meshri,
because
it
does
more
than
just
one
mesh
and
so
yeah
you
guys
are
you
guys
are
probably
about
right.
There's
a
couple
of
reasons
why
it's
probably
time
to
reduce
the
user
input
overload
a
how
many
of
you
are
running.
C
How
many
of
you
have
run
istio
two
of
you
or
four
of
you,
maybe
somewhere
in
there,
and
that's
totally
fine
like
as
a
matter
of
fact,
if
you
never
run
in
any
of
them,
you
can
still
do
a
lot
of
good
damage
or
good
contributions
to
the
project
with
never
a
running
am
doesn't
really
matter
anyway,
but
so
my
point
of
the
the
moral
of
the
story
is
most
users
most
of
the
time.
C
They'll
just
be
running
one
service
mesh,
like
that's,
that's
true,
like
and
again,
like
I'm
gonna
digress
for
a
second
and
say
well,
if
you're
and
asking
then
why
the
heck
do
you
support
10
of
them
when
most
people
aren't
going
to
run
one
it's
like?
Well,
that's
not
the
point.
That's
a
and
that's
that's
a
that's
cream
on
the
top.
It's
the
cherry
on
the
top.
It's
it's
the
mentionable.
After
the
fact
it's
meshri.
C
Does
these
awesome
things
about
managing
service
meshes
like
letting
you
apply
service
mesh
patterns,
letting
you
see
active
logs
streaming
from
multiple
containers.
At
the
same
time
lets
you
visualize
your
mess
and
your
mesh,
your
mess
and
your
infrastructure,
and
and
it's
like
wow,
that's
compelling
that's
great
and
it
lets
you
manage
filters
and
apply
filters
and
configure
your
filters
and
like
okay.
This
is
helpful,
like
istio
doesn't
come
with
a
ui
that
does
this
and
so
nice,
and
then
you
say:
oh
yeah
yeah.
C
Also
it
doesn't
matter
what
mesh
you're
running
it
does
it
for
any
of
them.
Oh
then,
you
just
like
you
drop
the
mic
and
walk
out
of
the
room.
So
you
we
don't
have
to
start
with.
The
mic
drop
first
is
kind
of
what
you
guys
are
getting
to
ayush
and
adi.
C
D
D
We
can
have
a
single
route
management
and
have
different
different
accessing
different
adapters
at
a
time
with
these,
with
this
kind
of
bar,
rather
than
only
through
this
or
and
changing
it.
According
to
the
adapter
we
are
using
like
we
don't
have
to
create,
we
should
not
be
having
two
different
layouts
like
this
one,
and
this
one.
C
Very
much
need
some
fresh
thoughts
and
something
new,
and
so,
if
you
go
back
to
the
your
messaging
instance
for
a
moment
when
you
click
on
management,
which
I
think
in
the
mock-ups,
it
says
life
cycle,
which
is
what
we
ultimately
want
to
start
to
call
these
things
like,
because
because
it's
it's
performance
management,
it's
filter
management.
It's
patterns
management,
it's
conformance
management,
it's
like
so
calling
this
management
is
probably
wrong.
It
needs
to
be
called
life
cycle,
and
so
so
good.
C
But
when
you
click
on
the
that
link,
it's
like
well,
you
want
to
manage
service
meshes,
but
the
user
at
this
point
hasn't
identified
explicitly
which
mesh
they
want
to
manage,
and
so
that's
why
they
get
that
extra
toolbar
up
top
saying.
Well,
you
know
from
this
picker,
which
one
did
you
want
to
do
and
and
what
happens
is
if
you
click
on
an
adapter
on
the
left
hand
menu
that
you
are
not
actively
running
like
a
citrix
service
mesh,
potentially
or
whatever
one
you're
not
actually
running
you
will
it
will.
C
It
will
show
you
that
top
toolbar
again,
because
that,
like,
if
you
click
on
octarine,
I
think
that
we
will
see
this
happen,
so
your
octarine
adapter
is
not
running.
So
what
it
does
is
takes
you
back
to
management.
D
Like
these
are
not
configured
or
not
configured
yet
or
we
are
not
using
yet
those
kind
of
things
like
main
purpose.
What
I'm
saying
I'm
trying
to
say
that,
like
we
have,
this
is
a
slash
management,
but
when
we
go
to
some
specific
one
it
is
here
telling
it
goes
like
it
may
create
a
whole
separate
round.
C
D
This
type
of
this
type
of
notification
should
be
stored
somewhere,
like
in
notification
software,
because
it's
executing,
if
it
disappears
it's
kind
of
busy
to
me
how,
when
it
is
going
to
have
a
complete
or
is
it
completed
or
not
these
new,
these
type
of
things
to
be
stored?
It
needs
to
be
stored
somewhere.
I
think
also,
when
we
like
add
some
add-on,
it
will
just
show
it
is
up
and
it
is
still
executing.
D
C
B
Oh,
I
always
small
feedback
like
if
you
go
back
to
your
film,
you
have
a
you,
have
a
way
to
show
all
the
all
the
things
that
all
the
options
which
one
particular
sub
section
has
in
this
particular
layout,
which
we
can't
currently
have
like
a
person,
can
see
all
the
things
that
that
particular
adapter
can
do
like
when
you
click
on
play
and
delete.
You
get
certain
options.
B
H
C
D
Like
my
basic
point,
was
these
disc
descriptions
are
not
always
required
if
a
person
gets
used
to
it
and
then
only
the
headings
would
be
enough.
C
Yeah,
you
know
both
are
achievable.
I
I
consider
it's
like
actually
to
reinforce
the
point
that
I
use
she's
currently
making
it
the
the
I
like,
if
you,
whenever
you're
using
a
piece
of
software,
there's
icons
and
buttons
to
click
and
just
like
in
the
mock-up
up
here,
you've
got.
You
know,
six,
seven,
eight
different,
nine
front
service,
mesh
icons
and
wait
a
second.
What's
that
I
don't
know
because
it
doesn't
have
a
label,
it's
like!
Well,
of
course,
that's
kuma.
I
work
with
kum
every
day.
C
I
don't
need
a
stupid
label
on
the
thing
I
don't
need.
I
don't
like
the
icon,
not
the
label,
and
it's
like,
like
kind
of
like
two.
You
should
point
over
time.
A
the
more
intuitive
the
ui
is
the
more
familiar
it
is
compared
to
prior
experiences.
You've
had
the
less
explanation
that
the
user
needs
and
kind
of
going
back
to
my
earlier
remark.
If
you
have
to
explain
your
ui,
it's
like
a
bad
joke,
don't
want
to
explain
the
you
had
to
explain
your
joke.
It's
not
funny!
C
C
It's
another
thing
that
my
wife
hates
she
thinks
that
she
when
I
when
I
think
I'm
funny,
then
she
definitely
thinks
it's
not
funny
all
right.
We
need
some
ideation
here,
the
what
I
was
going
to
say
about.
I
think
that
you
can
achieve
both
that
sometimes
it's
done
through
user
preferences.
Like
users,
you
know
you've
seen
software,
that,
let
you
they
ask
you
do
you
want
to
see
the
label
or
not
see
the
label.
Do
you
want
to
see
a
really
tiny
icon
or
a
big?
C
You
know
fat
one
and
then
okay,
like
you
know
those
are
things
that
it
might
be,
that
what
you
do
is
start
off
with
over
explaining
the
ui
to
the
user,
where
you
have
labels
and
you
sort
of
and
then
over
time
as
they
become
a
power
user,
they
they
become
familiar
with
the
various
user
preferences
and
they
can,
you
know,
disable
some
things
the
to
the
term
to
reinforce
the
suggestion
by
dhruv.
Is
that
there's
some
amount
of?
Oh?
I
mean
it?
Wasn't
it's
not
entirely.
C
The
fact
that,
like,
oh,
I
didn't
know,
citrix
had
a
service
mesh,
oh
interesting,
and
then
people
you
know
like
because
they
can
see
it
they
go
into
it,
there's,
obviously
a
line
in
there
somewhere.
C
Where
know
there's
a
you
know
like
some
of
this
is
under
refining
and
understanding
some
user.
You
know
users
are
on
a
journey,
some
it's
the
first
time
ever
so
you
know
maybe
leaning
into
a
little
harder
into
discoverability.
Okay
for
others.
It's
like
you
know,
get
out
of
my
way.
I
just
I
want
to
do
like
as
few
clicks
as
possible
to
just
use
this
utility-
it's
not
about
sitting
here
in
the
ui
and
reading
everything
through
it's
just
like
I
want
to.
C
You
know,
I'm
just
going
to
go
use
the
github
action
now
like
I
didn't
want
to
bring
up
the
ui.
I
just
want
to
do
that.
Yeah,
and
so
part
of
this
is
actually
you
and
specific.
I
don't
know
that
and
for
a
number
of
the
others
of
you
ritesh
and
dayatan,
and
and
rishabh
I
don't
and
and
others,
and
I
don't
know
that
everyone
has
seen
the
work
that
has
been
done
around
the
configuration
wizard.
C
But
it's
not
a
part
of
an
a
current
release.
There's
about
three
or
four
community
members
who've
been
working
on
this
initial
configuration
wizards
such
that
when
someone
launches
mestre
ui
for
the
first
time
or
they
launch
a
new
mystery
deployment,
they
can.
They
can
go
through
this
installation
stepper.
You
know
like
next
next
next
next
next
day
they
configure,
meshri
and-
and-
and
the
point
of
that
is
like
it's
one
of
those
things
that
you
can
also
dismiss-
it's
like.
C
More
of
that
type
of
thinking
is
needed.
Hopefully,
if
you're
I
mean,
if
you're
on
the
call
now
or
if
you're
in
the
community,
you,
if
you
don't
have
access
to
these
mock-ups,
then
just
say
something
and
let's
get
you
in
there.
Let's
get
some
ideas
going
grab
some
screenshots
from
some
of
your
favorite
software,
and
let's
reflect
on
that.
Let's
think
about
that.
C
There's
the
you
know
like
one
of
the
things
that
you're
all
gmail
users,
except
for
someone
who
joined
recently,
who's
a
hotmail
user
and
I
feel
poorly
for
them,
not
because
hotmail's
lesser,
but
because
they
won't
have
access
to
the
google
drive
to
the
layer.
C
5
the
community
drive,
but
the
reason
I
bring
up
gmail
is
you've,
probably
all
seen
the
the
views
that
they
have,
the
I
forget
exact
name
of
what
it's
called,
but
there's
a
type
of
design
or
view
that
is
comfortable,
dense
and
super
large
or
whatever
the
other
thing
is
called,
and
you
can,
as
a
user
switch
between
those
and
you're
you're,
either
shown
a
lot
of
labels
and
a
lot
more
things
or
you're,
not.
F
C
Well,
they're,
probably
getting
that
fairly
right,
they're,
actually
showing
you
as
a
user.
It's
like
other
words
that
they
could
have
used
to
describe.
It
are
super
user,
which
is
probably
dense,
noob,
which
is
probably
you
know
your
grandma
that
you
can't
see
the
screen,
so
it's
huge
whatever
whatever.
C
So
those
are
the
kinds
of
things
that
we
could
consider
in
here,
as
well
like
for
my
part,
having
been
around
this
baby
for
a
while
meshri,
it's
okay
for
me
to
call
it
ugly,
which
is
yeah
those
uis
that
we
currently
have.
They
are
ugly,
I
mean
like
the
oh
yeah.
The
font
is
way
too
huge.
Why
are
we
describing?
Why
are
we
showing
people
to
sub
labels?
Why
is
it
a
plus
sign
and
a
trash
can
icon
and
like
what?
If
I
you
know,
there's
just
it's
well,
thank
goodness.
C
Ukarsh
is
bringing
really
this
highly
detailed
way
of
configuring.
The
mesh
that
now
we
can
begin
to
expose
all
those
nerd
knobs
into
the
ui
and
really
the
only
nerd
knobs
that
we
have
in
the
ui
right
now
are
those
little
switches
and
I
sometimes
go
through
and
click
on
them
just
for
fun,
because
that's
about
the
only
thing
that's
actually
fun
to
do
in
the
ui
and
it's
actually
really
responsive
and
and
to
ayush's.
C
Other
point
is
like
there's
a
plan
in
place
in
the
link
in
the
chat
to
overhaul
the
messaging,
the
notification
center.
So
right
now,
you've
got
three
errors
in
your
upper
right
hand,
corner
and
and
there's
a
start
to
something
that
could
be
glorious
in
terms
of
like
what
operations
did
I
invoke?
When
did
I
invoke
them?
Did
they
go
through?
Can
I
restart
them?
Can
I
replay
them?
Should
I
can
I
retract
them
like
what
happens
if
lee
was
signed
into
the
same
mystery
server,
which
he
can
he
can
bring
up?
C
He
can
hit
the
same
url.
If
it's,
you
know
publicly
facing
and
sign
in
and
what
happens
if
he
was
invoking
operations
and
like
are
those
my
failures
or
his
probably
probably
his,
but
you
know
yeah
yeah,
let's
get
into
this
and
let's,
let's
everybody
jump
into
figma.
Let's
everybody's
got
an
opinion:
everybody's,
a
user
everybody's
got
a
mouse,
and
so
most
of
you
are
gonna
be
wrong
by
the
way,
I'm
just
gonna
start
with
that
but
you're.