►
From YouTube: Meshery Development Meeting (Nov 10th, 2021)
Description
Meshery Development Meeting - November 10th, 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
everyone
to
the
machine
development
meeting
today
is
the
10th
of
november
and
if
you
have
any
topics
to
add
to
the
meeting,
minutes
feel
free
to
answer
so
that
we
can
discuss
it
on
the
call,
as
well
as
add
your
name
to
the
attendees
list
in
the
meeting
minutes.
A
B
C
D
A
E
So
yep,
so
there
have
been
an
initiative
going
on
to
refactor
the
mississippi
ui
and
bring
in
some
of
bringing
redux
and
reductible
observable
and
litex
tank
intervention
away
because
of
the
state
management
concerns
that
we
had
and
yeah.
So
this
is
actually
the
proposal.
So
if
you
want
to
look
into
what
problems
it's
trying
to
address
and
how
we're
trying
to
approach
it,
then
yeah
you
can
look
into
this
particular
document
and
also
like
yeah.
This
is
the
branch
that
we'll
be
working
on.
That
is.
E
So
this
is
where
we'll
be
working
on
the
refactoring
of
measuring
ui
and
also
we're
also
tracking
the
progress
of
of
our
contributions
as
well
in
spreadsheet
over
here.
So
if
you
are
into
ui-
and
you
want
to
contribute,
then
you
can
just
come
over
here
and
put
your
name
into
like
these
are
the
components
that
we're
trying
to
build
first
and
then
like
this,
will
be
the
globally
reusable
components
and
we'll
be
building
other
components.
On
top
of
this.
E
So
and
that's
what
that
is
what
is
going
on
right
now
and
you
can
see
that
people
are
already
asking
some
of
the
tasks
and
they
are
working
on
it.
So
you
can
just
come
over
here
and
put
your
name
and
start
working
on
some
of
the
on
these
tasks.
E
So
if
you
are
like,
if
you
don't
know
how
to
get
started,
then
we
have
this
readme
file
inside
the
branch
itself.
There's
also
a
link
to
this
particular
readme
file
in
the
meeting
minute.
So
you
can
go
over
here
and
read
about
what
is
the
workflow?
What
is
the
flow
that
you
should
follow
in
order
to
contribute
to
the
ui
refactoring
process
and
yeah?
That's
it
so
yeah
there's
a
lot
of
work
to
be
done
and,
if
you're
into
ui,
this
might
be
a
huge
opportunity
for
you
to
learn
and
yeah.
D
D
Let
me
share
my
skin,
so
I
have
just
raised
up
here
like
30
minutes
ago,
so
I
have,
I
think
in
the
spreadsheet
I
I
have
completed
this
app,
but
loading
spinner
and
the
data
table-
and
I
have
created
this
pr
that
this
app
part
and
if
you
want
to
see
how
this
thing
is
actually
globally
visible.
I
have
an
example
for
you,
so,
basically
right
now
you
can
see
in
my
app.js.
I
have
created
this
one
right
now.
I
have
hard
coded
this
background.
D
Color
I
will
remove
it
and,
and
you
can
see,
and
in
the
in
in
the
header.js
I
have
used
that
custom
apple
wrapper
and
you
can
see
this
actually.
This
is
actually
coming
from
that
background.
If
you
want
to
be
sure
about
that,
I
can
show
you
so
basically,
if
you
just
change
this
to
red,
I
think
it
will
affect
my
background
yeah.
So
basically,
it's
coming
from
there.
So
basically,
this
is
how
we
are
creating
these
components
reusable
and
the
other
two
components
I
have
raised
in
my
pr
and
also
so
yeah.
E
Okay,
so
yeah
that's
great
and
also
like
regarding
like
we
we
were
like,
I
mean
I
heard
that
we
would
be
doing
some
design
changes
on
into
the
measure
ui
as
well
like
we
are
going
to
rewrite
a
lot
of
things
and
we
thought
that
we
might
as
well
do
some
design
changes
to
make
it
look
better.
So
so
lee
like
do
you
think
this
is
some?
This
is
the
area
where
we
might
change
the
structure.
C
Yeah,
you
know
I
did
I've,
which
is
it
the
dashboards
that
sayethon
is
showing
yeah
yeah
for
sure,
although
since
you've
already
done
it
like
great,
like
let's
move
forward
and
and
that
way
it's
we're
converted
right
where
we're
on
to
the
new.
C
The
new
system,
the
new
ui,
the
new
architecture,
redux
state
management,
whatever
we're
calling
you
know
like
you,
were
there
and
yes
had
you
asked
that
question
before
cyatin
had
done
satan
had
done
this
then
we
should
have
taken
a
moment
to.
But
for
now
I
would
say
until
we
have
a
design-
and
you
know
on
file
that
we
would
just
go
forth
with.
C
I'll
use
and
I'll
use
and
well
an
indian
friendly
term
and
say
we'll
just
go
with
the
same
over
same
here.
C
C
Let's
just
go
with
the
same
thing:
for
now,
until
we
get
a
new
design
there,
there
is
so
abi.
Kumar
has
taken
it
upon
himself
to
try
to
enhance
the
way
that
adapters
are
shown,
and
we
might
take
a
quick
look
at
that
if
he
comes
to
the
call
later,
the
briefly,
the
connection
status,
the
right
side
set
of
dashboards,
that's
a
great
dashboard
for
mesherie's
health
and
health
of
all
its
components,
which
can
be
relocated,
probably
to
settings.
C
C
Geez,
quite
a
few
and
so
beautiful.
C
At
some
point,
how
many
like
did
you
have
a
sense
in
the
dish
of
the
number
of
other
rows
like
how
many
rows
we
would
have
in
here
for
before
we're
sort
of
fully
transitioned.
E
Cool
or
any
other
updates
on
this
yeah,
so
you
can
also,
you
should
feel
free
to
add
the
reviews
on
the
ps
that
people
are
making
so
yeah.
F
C
Nice
you're
saying
that
in
general
or
like
specifically
me
or
just
in
general,
no
in
general,
yeah
yeah
yeah,
that's
a
great
it's
a
great
point
like
so
I
don't
know
that
there's
a
bunch
of
it
might
be
pranav,
potentially
or
vedant,
as
you
two
gents
and
maybe
ayushman
as
well
as
you
three
gentlemen
potentially
go
to
contribute
and
measure
ui,
which
I
don't
think
you
have
yet
it's
not
as
it's
not
as
complex
as
you
might
think.
C
I
mean
yeah
the
let's
just
react
right
now:
it
is
going
to
get
more
complex,
so
if
you're
ever
going
to
now
is
the
time
to
get
in
because
we're
going
through
a
an
architectural
shift
and
so
there's
a
focus
on
the
contributor
guide
and
and
helping
people
understand
how
to
contribute.
C
So,
okay,
next.
A
Yep,
thank
you
natasha
and
santan.
Next
up
we
have
an
update
on
the
mashery
docs
versioning
from
aditya.
G
So
I
haven't
pushed
these
changes
so
I'll
just
show
them
locally
right
now.
So
right
now
I
have
taken
a
different
approach
that
we
spoke
about
so
now.
Basically,
the
like
the
latest
version
is
in
the
docs
folder
and
then
there
are
five
other
versions
which
each
have
the
config
like,
which
each
have
like
a
static
html
site.
So
now
now
after
we
merge
it,
we
will
have
five
different
like
we'll.
G
Have
five
different
sites
with
different
urls
like
the
the
latest
url
will
be
docs.measure.io
latest
right
now
I
have
version
one
running
so
that
will
become
dogs,
dot,
mesh
dot,
io,
slash,
v,
zero
point,
one
so
yeah
yeah,
I
have
done.
G
I
have
done
it
for
two
versions
and
the
latest
need
to
complete
it
for
the
other
versions,
and
we
will
actually,
I
like,
I'm
actually
showing
this
to
you
locally,
because
we
need
to
rewrite
the
build
and
preview,
also
the
build
and
preview
for
for
the
netlify
workflow,
because
that's
not
working
right
now,
so
we'll
have
to
update
that
so
that
it
can
show
us
what
yeah.
G
C
G
Yeah,
if
they've
chosen
to
stay
in
a
different
version,
they
would
stay
within
that
version.
I
could
make
it
happen
that
they
would
stay
in
dogs
that
measured
at
I
o
and
slash
latest.
I
think
that
is
possible
to
do.
G
C
I'll
increase
about
70
megabytes
and
we
have
what
a
gigabyte
offered
for
free.
It's
not.
G
A
concern
yeah
yeah.
We
have
plenty
of
space
to
use.
This
is
actually
the
best
possible
way.
I
think
there's
no
other
way
that
we
will
be
able
to
do.
C
And
actually
to
be
technical
like
technically
there's
like
10
other
ways
to
do
it.
It's
just
that
this
is
the
cheapest,
the
simplest
it
it
like.
There's
a
bunch
of
reasons
why
you
won't
find
broken
links,
you
won't
find
you'll
be
able
to
change
the
structure
of
the
site
and
you'd.
Still,
that
would
be
fine.
You
wouldn't
break
old
versions,
be
perpetually
backwards
compatible.
C
You
have
a
single
url
that
you
always
refer
to.
You
want
to
mess
with
dns.
You
want
to
have
to
be
running
multiple
sites,
so
just
run
one
you'll
snapshot
the
cut
like
there's
there's
like
a
time
you
you
don't
have
to
mess
with
multiple
branches
like
I
have
no
idea.
Why
other
projects
that
wrong
jekyll
and
that
have
the
same
conundrum
don't
take
the
same
approach,
but
we
should.
G
C
Thank
you,
but
I
mean
just
to
reinforce
what
the
vendor
had
said
is
that
it
is
a
consideration
that
you
wouldn't
want
to
make
sure
that
you
aren't
you
don't
blow
past
the
storage
offered
for
a
given
repo.
You
know,
of
course
you
could
break
out
docks
from
the
main
repo.
That's
less
desirable.
C
It's
a
solution.
Another
one
is
to
just
prune
the
number
of
versions
that
you
have
out
there
in
the
like
it.
You
know
it
might
be
painful
for
some
of
us,
but
I
think
we
could
all
agree
that
no
one
is
running
v0.1
or
2,
or
dot
3
or
dot
4..
So
no
one
cares
like
at
some
point.
We
can
prune
those
back
as
well,
so.
A
A
A
Before
the
v
0.6
release,
we
have
been
trying
to
move,
we've
been
trying
to
move
away
from
kubernetes
manifest
and
using
helm
charts
instead,
so
that
we
have
a
more
defined
way.
Okay,
but
yeah,
are
you
guys
able
to
see
my
screen
because
I'm
having
some
sort
of
flag,
I'm
not
sure
we
can
see
your
browser.
That
is
at
the
pull
request.
A
Anyways,
so
the
thing
is
that,
right
now
we
have,
we
have
helm
charts
in
missouri
ctl
and
when
mystery,
cpl,
actually
installs
misery
in
your
cluster
or
otherwise.
It
uses
the
helm
charts
to
install
that.
But
there
are
some
caveats
around
him,
mainly
that
those
cavets
are
happening
because
how
helm
works
helm.
Has
these
two
things
it
I
mean
on
every
object
that
helm
creates
in
your
cluster.
It
sets
these
two
attributes
or
in
more
kubernetes
specific
language
annotations.
A
A
Those
are
happening
because,
right
now,
the
current
like
the
current
flow
is
that
if
messy
server
doesn't
find
the
operator
in
the
cluster,
it
installs
it
by
itself
using
manifest,
which
doesn't
obviously
have
these
two
properties
which
helms
looking
for
and
that's
why
some
people
are
facing
the
bug
when
they
switch
from
local
deployments
or
topper
deployments
to
in
cluster
deployments.
That
message
ctl
refuses
to
install
measuring
on
the
cluster
because
they
have
some
crs
already
present.
A
A
Asuco
recently
created
a
new
and
published
a
new
chart
for
missouri
operator,
which
is
having
mystery
operator
a
simian
chart
and
mystery
server
is
a
sub
chart,
giving
you
the
ability
to.
I
mean
control.
If
you
want
to
install
message
server
or
not,
there
are
some
things
with
this
I'll
come
to
them
later,
just
a
little
go
through
of
what
is
happening.
We
first
like
the
revised
approach
we
first
detect
if
misery
is
running
in
cluster
or
not.
A
This
is
how
we're
doing
it
just
just
taking
an
environment
variable,
and
this
approach
might
seem
questionable
at
first,
but
I
confirmed
with
a
few
people
that
this
is
indeed
the
one
which
feel
have
to
use,
and
basically
it
just
takes
an
environment
variable
and
decides
that
if
it
is
in
cluster
or
not,
we
need
to
check
if
measuring
is
in
cluster
or
not.
Basically,
because
charts
are,
I
mean
we
are
using
two
charts
now.
A
One
is
for
the
mystery
operator
and
one
is
for
the
mystery
server
and
if
the
current
install
was
done
using
misery
server.
If
this
mystery
server
finds
the
finds
that
mystery
is
already
installed
in
your
cluster,
it
won't
actually
use
the
other
chart.
It
would
just
use
the
measuring
server
chart
and
set
an
upgrade
and
then
use
appropriate
overrides
to
either
disable
the
operator
or
enable
the
operator,
and
if
messy
server
is
not
present
in
cluster,
it
uses
the
other
chart.
That
is
the
mystery
operators
are
which
installs
without
any
problems.
A
C
C
Well,
or
just
like,
do
you
have
you
have
a
pull
request
open
now
that
displaces
the
manifest
based
deployment
with
a
home-based
deployment.
A
A
The
only
thing
is
crs,
so
basically
helm
installs
crs,
but
it
doesn't
like
take
responsibility
for
them
and
if
it
finds
them
already
in
cluster,
I
mean
if
it
doesn't
find
them
in
cluster
mystery
operator.
Installation
might
fail.
So
I
guess
the
final
update
to
this
would
be
fixing
that
how
we
install
crs,
maybe
a
dry,
run
for
that,
and
I
think
then
this
vr
would
be
good
to
go.
I
was
talking
about
this
pr
itself
like
there
is
no
other
pr
other
than
this.
C
Cool
good
good,
and
so
have
you
found
a
a
common
approach
to
managing
crs
with
helm
that
other
projects.
A
A
Skipping
crds-
or
I
mean
setting
up
like
our
charts,
currently
installs
here,
so
I
think
we
can
still
keep
them,
but
when
misery
server
installs,
it
can
have
a
dry
run
which
actually
checks.
If
helm
installation
would
fail,
it
would
fail
due
to
missing
crs,
and
if
it
fails,
then
it
would
include
those
crs
as
well.
A
C
Okay,
or
actually
I
mean
actually
that's
good-
your
clarification
makes
me
wonder
that
we
might
not
be
speaking
so
so
just
just
to
the
crds,
the
ds
being
the
the
definition
like
this
cut.
This
custom
object,
so
bro
broke
a
mesh
sync
meshery
broker.
C
C
Okay,
so
briefly
like
to
get
so,
I
just
want
to
I'm
going
to
kind
of
in
a
hurry
to
like
get
this
one
done,
because
and
and
most
of
the
folks
on
the
phone
we
generally
like.
We
generally
take
a
lot
of
time
to
explain
stuff
and
I
don't
think
anybody's
really
most
people
aren't
necessarily
grocking
this,
and
it's
probably
not
helping
that
much.
My
question
is,
we
don't
want
to
leave.
C
We
there
might
still
be
situations
where
objects
were
orphaned
for
one
or
another
reason,
or
someone
goes
to
like
turn
off
a
cluster
they're
running
mesh
re
there.
They
turn
off
a
cluster.
They
turn
it
back
on
and
maybe
the
way
that
they're
turning
it
back
on
is
like.
Maybe
the
running
measuring
ctl
system
start
again,
even
though
there
might
be
an
actively
installed
or
run,
and
we
don't
want
the
response
to
come
back.
A
A
A
Okay,
I
think
it
would
be
dynamic,
but
yeah
sure
we
can
look
into
it.
C
Yeah
there
might
be
common
practices
around
that
yep,
then,
on
the
other
challenge
we
were
talking
about
about
like
leaving
objects,
orphaned,
the
crs
or
the
crds,
or
that
might
be
the
same
logic.
We
might
have
the
same,
delete
logic
like
hey,
let's,
clean
up
what
helm
isn't
cleaning
up,
which
is
would
be
kind
of
bad
because
we
kind
of
be
still
stuck
with
our
own
manifests
in
some
respects,
helm,
plus
some
of
our
own
logic
so
curious.
C
A
So
basically,
helen
gives
the
explanation
that
we
don't
take
care
of
crs
people
have
their
own
way
of
doing
that,
and
crs
are
pretty
varied,
so
helm
is
still
to
come
up
with
a
I
mean
with
the
design,
which
also
takes
care
of
crs
right
now
it
would
only
install
crs
for
you.
It
won't
do
anything
else
other
than
that,
so
someone
else
need
to
take
care
of
deleting
those
crs.
C
C
I'm
I'm
saying
I'm
saying:
hey:
that's
that
that
doesn't
fly
like
that's,
not
a
good
answer,
press
on
that
and
what
pops
out
and
so
I'll
initiate
some
conversations
to
see
what
people
are
doing
best
practices
in
terms
of
ensuring
that
their
you
know,
artifacts
aren't
orphaned.
C
A
C
So
this
is
bad,
so
I
guess
I
didn't
realize
so
this
this
goaling
is.
This
is
scolding
that
you're
writing
that
so
so,
like
we're
getting
rid
of
manifest
to,
hopefully
be
in
a
better
place
with
helm,
but
now
we're
writing
any
amount
of
golang
to
basically
manage
the
install
anyway.
C
B
C
Yeah,
let's,
let's
go
I'll,
go,
initiate
a
couple
of
conversations,
pull
you
into
it
and
let's
see
if
we
can
get
some
things
figured
out
with
some
others
that
have
had
the
same
experience.
It
may
be
that
hey
look
we
don't
like
it
might
be
that
we
don't
use
helm
for
installing
through
mastery
server.
C
F
C
A
C
Good,
okay,
cool
anybody
else
have
comments
here.
No,
I
then
do
what's
next.
C
Oh
yeah
sure
yeah.
I
didn't
hear
what
kind
of
said
I
thought
it
was
yeah
hey.
So
the
next
topic
up
is
that
for
the
v070
release,
but
like
as
a
plan
for
that
next
minor
release,
the
measuring
needs
to
improve
in
the
way
in
which
it
allows
people
to
look
at
metrics
that
it's
tracking
metrics.
And
if
you
extrapolate
that
out,
it's
not
just
metrics,
it's
some
telemetry,
so
it's
potentially
distributed
traces,
it's
logs,
but
for
the
time
being
we're
focused
on
metrics.
C
Gosh
grafana
boards
and
then
c
metrics
in
mesh
roots.
It's
quite
helpful.
We
expose
metrics
in
three
places
in
in
measuring
today
and
there's
not
any
there's,
there's
a
little
bit
of
customization.
That
can
happen
probably
unbeknownst
to
most
of
you
did
anyone
have
measuring
up
but
unbeknownst
to
most
of
you,
you
can
incorporate
metrics,
you
can
bring
in
your
own
custom,
grafana
charts
with
your
own
metrics
and
define
them
and
measury
today
will
let
you
query,
run
a
prom
ql
query
for
metrics.
C
To
do
so
right
now,
so
let
me
let
me
share
a
couple
of
docs
real
quick
by
the
way.
There's.
Let
me
pause
from
it.
So
hey,
there's
a
couple
of
new
folks
that
are
joined
so
naprenda
and
pc.
Welcome
to
the
call
nice
to
have
you.
If
you
don't
mind,
do
you
want
introducing
yourself
in
the
chat
and
pc?
If
you
don't
mind
renaming
to
spell
out
your
first
and
last
name
for
us
that
would
be
really
nice
all
right.
So.
C
So
in
inside
of
measuring
today
the
everybody's
got
access
to
the
mastery
architecture
deck.
Let's
put
a
paste,
a
link
in
there.
C
That's
your
architecture,
deck!
If
you
look
at
the
architecture
today,
there's
there's
a
few
different
ways
of
looking
at
the
architecture,
but
this
particular
slide
slide
five.
C
So
it
highlights
most
of
the
logical,
the
high-level
logical
objects
that
mesh
retracts
and
is
aware
of,
and
it
tells
you
if
measuring,
creates
the
object
or
is
just
aware
of
it
or,
if
that's
an
extension
point
and,
and
so
today
there
are
these
dashboard
configurations
that
some
of
which
are
hard
coded
in
the
system
and
some
of
which
users
can
like,
I
said
users
can
add
their
own
custom
dashboards
just
in
a
very,
not
a
very
easy
way,
and
so
that
configuration
when
someone
does
do
a
configure
dashboards
that
will
be
persisted
into
their
user
preferences
on
the
on
the
back
end,
and
so
that
this
is
what
measuring
looks
like
today.
C
The
proposal
is
a
bit
more
like
this,
so
multiple
metrics
and
multiple
metric,
multiple
imagery
providers
and
custom
dashboards
or
custom
boards,
and
so
what
would
happen
is
we
would
begin
to
have.
You
know
new
objects,
new
nouns
in
the
mesherie
database,
one
for
a
dashboard
itself.
The
dashboard
might
have
multiple
panels,
panels
can
have
panels
are
basically
charts
or
or
they
they
show
metrics
and
panels
might
have
one
or
more
metrics
in
that
panel.
So
we
begin
to
have
a
new,
a
few
new
first
class
nouns
in
the
project
boards
panels.
C
C
Here's
the
dashboard,
here's
a
mock-up
of
what
that
dashboard
might
look
like
in
the
future.
C
Excuse
me
dude,
so
there
might
be
some.
It
might
be
something
like
this,
where
there
would
be
potentially
two
or
three
locations.
So
this
red
outline
here
is
intended
to
make
be
the
sort
of
the
concept
of
a
location.
C
So
in
this
location,
what
dashboards
would
you
like
to
show
and
within
and
this
being
kind
of
a
dashboard?
So
in
a
given
location,
you
might
be
able
to
show
multiple
dashboards.
So
one
up
here
and
one
down
here.
C
And
then
within
a
board,
you'd
have
one
or
more
panels,
and
so
we
try
to
give
the
users
the
ability
to
sequence
what
order
they'd
like
to
show
the
boards
in
sequence,
what
order
they'd
like
to
show
the
panels
in
and
then
give
them
a
little
bit
of
control
over
the
the
panels.
C
So,
if
you're
familiar
with
mesherie's
performance
profiles
and
the
cards
that
you
see
there,
you're
able
to
users
can
click
on
a
card
and
swivel
them,
like
flip
them
over
look
at
the
back
side
of
the
card.
So
if
a
user
is
looking
at
the
graphical
view
of
metrics
of
particular
metrics,
but
they
would
like
to
change
those
or
augment
them
if
they
could
potentially
click
on
the
this
gear.
Icon
basically
have
the
card
swivel
turn
over
on
the
back
side.
C
They
would
see
the
panel
json
the
configuration
that
makes
up
that
that
panel,
those
set
of
metrics,
and
so
these
set
of
metrics
are
compatible
with
grafana's
sdk
and
that's
in
part
how
we'll
import
and
potentially
export
panels.
C
People
can
then
save
them.
That
would
be
part
of
their
user
preferences
and
so
okay.
This
is,
and
so
in
other
locations
like
in
settings,
there
might
be
the
ability
there
might
be
other
sorry,
there's
multiple
places
to
expose
metrics
there
might
be
different
locations
that
people
can
place
their
boards.
C
C
Set
of
database
tables
that
are
designed
to
show
the
way
that
data
is
currently
stored
and
how
the
what
these
objects
currently
are
and
there's
a
proposal
for
what
that
would
change
to.
We
will.
We
will
take
this
out
of
this
drawing
and
put
it
into
an
architecture
slide
so
that
everyone
can
review
and
comment
and
ask
questions.
C
Aside
from
that,
there
will
also
be
an
adjoining
design
spec
in
which
we'll
talk
about
what
what
users
you
know,
what
are
the?
What
is
a
board?
What
is
a
panel?
What
is
a
telemetry
provider?
What
can
users
do?
You
know
have
various
you
know:
how
are
we
dealing
with
today?
We
have
these
dynamic
and
static
boards.
C
C
If
anybody's
interested
in
this
area,
great,
like
we'll
get
you
those
links
and
and
but
surely
by
this
time
next
week
or
well
before,
monav,
will
probably
send
out
a
design
spec
and
some
of
these
mock-ups
for
review
and
content.
C
Does
anybody
have
a
gut
reaction
or
you
know,
questions
on
the
scope
or
anybody
interested
in
participating
in
that
effort?.
A
All
right,
thank
you,
lee.
Thank
you
mano.
Next
up,
we
have
a
demo
from
abhishek
on
the
new
feature
in
in
the
dashboard,
so
abhishek.
B
H
Okay,
so
so
this
is
the
current
dashboard
that
we
have,
so
it's
just
a
very
minor
change
in
the
way
the
adapters
are
shown.
So
we
are
just
listing
the
active
adapters
which
are
running
and
if
the
active
and
the
number
of
active
adapters
are
same
as
the
number
of
available
adopters
and
there
will
not
be
any
added,
add
adapters.
H
So
currently
this
is
the
beer
like
they
have
environment.
That's
why
it's
taking
some
building,
otherwise
it
would
have
been
much
smaller.
H
Yeah
other
than
that,
the
other
change
that
you
can
see
here
is
the
elevation
on
the
on
this
on
these
cards
so
yeah.
So
it's
just
a
local
change
that
I
have
in
the
that
I
have
locally
only
so
it's
not
in
the
pr
currently
but
yeah.
It
looks
nice
to
me
actually.
So
that's
why
I
just
tried
to
show
it.
E
H
Okay,
this
one
yep,
okay
yeah.
It
was
just
an
experiment.
It
was
not
design,
so
we
can
just
replace
it
with
more
net
button,
so
it
just.
I
just
wanted
to
try
it
something.
Let's
try
something
like
this.
So
let's
get
it.
H
H
Definitely
actually,
this
is
coming
from
a
google
site
downloaded
in
the
public
directory
so
yeah.
This
color
can
be
changed
to
this.
So
I
yeah
since
I
had
to
add
avatar,
because
these
two
are
grouped
together,
so
it
can
only
contain
avatar.
So
that's
fine
right
now,
if
you're
thinking
to
just
replace
these
two,
so
only
one
advertise,
so
this
color
can
be
much
optimized
and
are
can
be
matched
to
this
button.
So
are
this
background,
so
nobody.
H
Okay,
so
any
comments
on
the
on
the
elevation
property
like
should
I
you
know,
push
it
to
the
branch
or
should
I
not
pronounce
this
say
one
more
time
about
the
elevation
I
missed.
So
if
you
see
the
previous
ui,
so
all
the
cards
are
elevated
and
yeah,
so
apparently
considering
it
that
these
are
all
on
the
same
page,
no
elevation
is
required
at
all
so
just
to
yeah.
So
there
is
slight
change
in
the
color
you
can
see.
H
Okay,
then,
to
item
1800,
so
just
change
this
I'll
change
this
and
add
plus
button.
Only
with
this.
C
But
the
concept
that
like
yeah,
I
think
we
should
do
some.
My
suggestion
would
be
to
do
a
little
bit
of
iteration
on
the
concept
that
you've
got
going
of
like
any
nut
like
any
number
of
additional
adapters.
I
think
that
it
historically
had
been
the
case
that,
in
some
respects,
we
wanted
to
facilitate
discovery
for
everyone,
every
user
to
come
in
and
they
would
see
like.
Oh
there's,
all
these
other
meshes
that
could
be
managed.
C
C
You
know
capabilities
the
less
about
it
becomes.
It
becomes
pretty
obvious
to
everyone
that
comes
to
trimester
that,
like
it,
doesn't
just
manage
one
service
mesh
and
so
optimizing.
The
user
experience
for
discovery
of
that
fact
is,
probably
you
know
not,
but
not
necessary
anymore,
and
so
anyway,
that
gives
credence
to
what
you're
doing,
which
is
compressing
the
user
interface,
avoiding
you
know
overloading
the
user
with
10
different
statuses
of
the
different.
C
C
So,
along
the
same
thinking,
though,
there's
a
much
more
significantly
important,
and
here
and
now
consideration
about
chips,
chips
are
those
elongated
ovals
that
we're
seeing
those
adapters
yeah
like
that's
what
you
call
that's
for
everyone
else,
that's
what
you
call
it
mesh
material
ui
calls
it
a
chip,
so
we're
using
these
chips
to
represent
connections,
connections
from
mesh
reservoir
to
other
measuring
components
like
adapters
operator
broker,
also
to
other
pieces
of
infrastructure
like
kubernetes
clusters,
grafana,
prometheus
jaeger,
other
things
well
here
and
now,
there's
a
consideration
for
the
fact
that
there's
a
pull
request.
C
Makes
bring,
you
know,
makes
us
forces
us
to
go,
build
out
an
enhanced
user
experience
around
the
fact
you
might
have
any
number
of
kubernetes
cluster
chips,
so
you
might
have
a
bunch
of
those
chips
and
those
chips
are
actually
bear
a
lot
more
weight
on
what
the
user
is
doing
in
all
of
the
rest
of
the
uis.
C
When
someone
invokes
an
operation
or
is
looking
at
the
statistics
of
a
given
set
of
services
or
like
they're
they're,
always
doing
it
in
context
of
at
least
one
cluster,
if
not
multiple
at
the
same
time
and
so
bringing
about
the
ability
to
the
easy
ability
to
switch
between
clusters
so
or
if
someone's
connected
to
10
kubernetes
clusters,
maybe
they
actively
want,
they
want
all
10
loaded,
but
for
a
given
operation
they
only
want
to
affect
five
of
the
clusters,
not
all
10..
Maybe
they
want
to
have
cluster
groups
where
those
five
are
for
tests.
C
So
there
are
some
more
pressing
concerns
around
kubernetes
clusters
chips
and
the
nice
thing
about
avi.
What
you're
exploring
here
is
that,
once
we
get
the
user,
the
ux
paradigm
figured
out
for
adding
a
chip
adding
a
connection
to
a
thing.
Well
that
will
be
applicable
to
kubernetes
cluster
to
all
of
the
other
types
of
chips.
So.
E
Yeah
yeah,
okay,
just
another
thing
that
came
to
my
mind
right
now,
so
so,
whenever
we,
whenever
user
clicks
the
adapter
button,
it
takes
us
to
the
page
where
we
can
yeah
in
this
page,
like
the
only
the
only
option
that
we
are
providing
to
the
user
is
to
select
the
available
list
of
from
the
available
list
of
adapters
and
connected
to
machine.
E
So
this
is
something
that,
if
you
can
think
about
like
we
can
actually
do
it
in
the
dashboard
itself
right
so
right,
right
right
where
we
place
the
adder
at
this
button,
we
can
actually
show
a
model
or
something
that
would
like.
We
need
not
redirect
a
user
to
a
different
page
in
order
to
do
this,
a
simple
task
of
adding
another
chip.
So
if
you
can
do
it
at
the
same
place,
then
it
might
be
more
convenient
for
the
user.
H
E
H
Yeah,
because,
actually
when
some
adapters
are
offline
or
also
or
they
are
online-
and
we
want
to
reconnect
with
so
what
we
are
doing
just
is
really
connecting
from
here.
So
the
same
thing
can
be
done
with
the
dashboard
itself
and
it
will
be
much
better
user
experience.
So
yeah
definitely.
E
So
yeah
one
thing
that
might
that
might
be
something
that
we
have
to
consider
is
that,
like
in
the
dashboard,
we
are
saying
that
we
are
only
showing
the
status
of
the
connection
and
if
we,
if
we
try
and
do
that
thing,
then
yeah
I
don't
know
like
it-
might
be
something
like
yeah.
It
might
not
strictly
be
that
be
like
that.
We
are
just
showing
the
status
of
the
connection
and
not
allowing
the
users
to
change
anything
or
to
something
here.
E
C
A
Yep
we
do
have
other
topics,
but
like
is
there
any
any
other
topic
that
needs
to
be
discussed
right
now
or
if
yes,
then,
we
can
stay
back
for
a
couple
of
minutes
to
get
those
discussed
as
well.
A
I
think
we
can
discuss
mine
on
friday
sounds
good
and
I
I
will
update
the
I'll
be
talking
about
the
smp
action
on
the
cncf
tag
called
that's
happening
next
week,
so
I'll
update
it.
There.
B
C
Nice
to
see
everybody.
Thank
you,
david.
A
Thank
you
guys
will
see
you
in
the
next
meetings.
Bye.