►
From YouTube: Meshery Development Meeting (Aug 04th 2021)
Description
Meshery Development Meeting - Aug 04th 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
So,
let's
get
started
as
people
join
in
so
welcome.
Everyone
to
the
measuring
development
core
today
is
fourth
of
august,
and
we
have
some
items
listed
here
today.
So
it's
going
to
be
a
good
meeting.
So
before
we
get
started,
is
there
someone
new
on
the
call.
B
Okay,
so
hello,
everyone
and
my
name
is
anshuman,
and
I
am
currently
pursuing
my
b
tech
from
enterprise
engineering
college
and
I
am
working.
I
am
learning
front-end
and
back-end
development
and
I
I
want
to
explore
in
this
field
more
so
I
joined
this
open
source
platform
so
that
I
can
enhance
my
skills.
A
Sounds
good
amazing,
thank
you
for
introducing
manchester.
I'm
sure
you'll
find
a
lot
of
interesting
projects
to
work.
Work
on
here.
C
C
A
Yeah,
actually,
it's
anshuman
the
double
at
the
end
of
mine,
but
it's
having.
A
C
E
E
I
don't
know
that
there
are
any
other
newcomers,
so
schumann,
the
man
is
the
winner
today,
nice
to
nice,
to
meet
you
shuman
thanks
for
jumping
in
thank
you
yeah.
It
was
nice
of
you
to
introduce
in
slack
too
that's
always
it's
really
nice.
This
is
some
more
people
read
that
than
you
know.
D
B
A
All
right,
so,
let's
move
to
today's
agenda.
First
up,
we
have
a
small
announcement
that,
as
you
might
be,
knowing
measuring
recently
became
adopted
into
the
cncf.
A
So
measurey
is
a
cncf
project
right
now,
so
in
cnc
cncf
cloud
native
community
groups,
merceri
has
a
community,
so
this
is
pretty
new,
so
we'll
have
all
the
event
details
and
all
the
other
stuff
that
comes
up
with
this
community
and
you
can
rsvp
for
events,
learn
about
upcoming
events
and
all
those
stuff.
So
we
are
still
exploring
this
tool,
so
we
will
be
adding
more
more
stuff
here,
so
you
can
join
this
community
by
going
through
this
link
and
joining
so
yeah.
That's
it.
A
We
also
have
one
for
service
mesh
performance,
which
is
one
of
the
other
projects
that
we
have
been
working
on.
So
you
can
also
find
that
in
the
cncf
community
groups,
all
right
so-
and
we
also
have
sessions-
have
planned
a
session
this
month
with
the
cncf
student
groups.
So
it's
also
a
cncf
community
group
and.
E
But
but
we'll
send
out
somethings.
A
Yeah
yeah
so
they'll
be
having
a
scheduling
a
session
with
the
cnc
students
group.
So
most
of
us
we
have
a
lot
of
student
representatives
in
the
life
community,
so
it
will
be,
should
be
a
good
session,
so
stay
tuned
for
that.
G
Oh
yeah
sure
let
me
share
my
screen.
G
G
You
guys
can
see
my
screen
right,
yeah
yep,
so
hello,
everyone
good
morning
or
good
good
night,
so
my
name
is
marik
and
I
am
going
to
show
you
about
the
issue
that
I've
been
working
on
last
week.
Actually,
so
it
is
about
the
looking
for
data
fetching
for
grafana
or
prometheus.
G
So
so
here's
the
thing.
So
you
can
start
your
measuring
server
and
then
go
to
go
to
a
settings
file
and
if
your,
if
your
graphana
or
prometheus
are
not
configured,
it
will
keep
on
catching
the
data.
So
you
see
in
my
computer,
I
I
haven't
set
up
my
group
funnel
or
prometheus,
but
every
time
I
I
reload
the
page
it
will.
It
will
keep
on
fetching
the
data
from
from
them
yeah.
Basically,
so
what
we?
G
So
that's
one
so
yeah
so
and
then
moving
on
to
the
next
issue
that
I
want
to
talk
about
is
that
I've
been
working
on
the
subscription
for
performance
result
just
to
try
and
dip
my
hands
dirty
in
the
graphql
environment,
in
the
measuring.
So
again,
this.
G
Yeah
so
so,
originally
the
the
issue
was
about
creating
the
subscription
resolver
for
performance
result.
But
then,
when
I
looked
into
it
more,
it
was
a
bigger
issue
than
it
originally
was
so
instead,
instead
of
creating
it
directly
into
subscription,
I
do
I
firstly
do
the
query.
Graphql
query
for
the
performance
result
yeah.
G
So
there
was
the
initial
stage
and
then
there
will
be
a
later
more
stage
to
it,
for
example,
creating
the
creating
the
subscription
and
then
and
then
after
that,
after
finishing
all
with
the
back
and
stuff,
and
then
you
can,
you
can
move
into
the
front
end
and
create
the
create
the
integration
between
the
front
end
and
then
back
and
using
react
really.
Oh
basically
that
so
this
issue
consists
of
multiple
stages
that
it
has
not
been
taught
before.
G
G
Yep
so
yeah,
so
this
is
the
data
that
we
are
fetching
from
the
mesh
sync
and
yeah.
So
this
is
the
the
issue
that
I
want
to
talk
about.
Actually
so
some
of
the
issues,
so
the
model
that
generated
from
the
graph
scale
does
not
really
connect
with
the
photo
buff
from
the
messenger
so
so
be
because
we
define
the
mess
basic
model
in
the
in
the
messy
repository
using
the
using
the
protobuf,
and
then
we
recreate
this
model
using
graphql
schema
in
for
ourselves.
G
E
Yeah
nice,
do
you
so
nice
that
does
great.
You
covered
two
issues.
The
first
issue
it
just
to
clarify
the
user
experience
improvement
that
you
made
there
that
didn't
have
anything
to
do
with
graphql.
Is
that
right,
like
how
you
how
you
enhanced
it?
That's
all
fun
and
stuff?
E
Okay,
good,
and
maybe
you
know
what
actually
looking
at
the
long
list
of
topics.
Maybe
I
was
gonna
take
people
through
what
you've
done,
but
but
since
we
have
a
lot
of
topics,
maybe
I
won't,
but
I'm
good
with
zubarik.
E
Yesterday
we
were
on
with
well,
let's
see
we
were
on
with
an
engineer
from
intel
who
is
using
measuring
and
they
are
quite
into
our
the
initiatives
here:
they're
quite
interested
in
into
service
mesh
performance
as
a
specification
and
mesherie
as
a
tool
to
help
them
study
the
behavior
of
cloud
native
of
microservices
of
cloud
native
software
on
their
hardware,
not
just
their
chips
but
but
other
things,
so
so
they're
really
interested,
and
so
we
spent
quite
a
bit
of
time
on
a
call
yesterday
trying
to
help
make
sure
that
measure
is
running
and
doing
what
they
wanted
to
do.
E
So
that's
that
like,
even
though
in
concept,
it's
a
fairly
small
change.
It's
impactful
to
the
experience,
because
as
a
user,
if
all
of
you
have
used
the
mastery
or
if
you
haven't
you'll,
eventually
see
that
it'll
use
toaster
notifications
to
let
you
know
what's
going
on
and
when
you
get
a
red
colored
poster
notification.
E
You
know
it's
unnerving
because
something
there's
an
issue
somewhere
and,
like
you
know,
half
the
time
you
can't
figure
out
how
to
fix
it.
So
that's
great!
That
was
an
aberrant
notification,
something
that
we
didn't
need
the
second
one,
the
graphql
resolver.
I
think
you're
right
that
do
you.
Let
me
pull
up.
E
This
screen
real,
quick
to
give
everyone
context,
because
I
think
this
context
well
helps
for
newcomers.
It
also
helps
to
keep
everyone
else
here
on
the
same
page,
so
so
your
you're
looking
at
another
user
experience
enhancement,
but
one
that
requires
a
much
more
significant
investment
in
terms
of
code
and
use
of
graphql
as
a
technology.
So
today,
if
someone
goes
in
and
runs
a
performance
test
when
they're
done
running
the
test,.
E
E
That's
not
facilitated
through
a
subscription,
though
through
a
graphql
subscription
to
just
like
listen
for
the
total
count
and
be
notified.
You
know
be
pushed
that
total
account.
This
is
in
part,
what
you're
working
toward
to
make
sure
that
on
this
ui
and
a
couple
of
other
uis
that
that
just
happens.
E
My
understanding,
if
I
recall
like
there's
just
there's
a
callback
that
has
been
written
into
the
ui
here
to
instead
of
forcing
the
user
to
refresh
to
get
that
new
number
that
we're
doing
a
call
back
and
so
we're
sort
of
doing
a
poor
man's
subscription.
By
occasionally,
you
know,
requesting
more
data.
E
Let's
turn
that
into
a
subscription,
and
now
we
can,
just
you
know,
have
the
ui
listening
for
that
new
info
and,
as
you
dig
into
it
even
further,
the
thing
that
you're
identifying
is
that
there's
the
the
way
that
the
information,
the
performance
information,
is
characterized
and
captured.
There's
a
fair
bit
of
detail.
There's
analysis
that
goes
on.
There's
information
about
how
long
you
were
running
the
test
and
how
intensely
and
what
was
the
statistical
averages
etc.
E
E
All
of
this
information,
they
define
those
protos
to
find
a
way
of
capturing
that
info.
So
what
you're,
acknowledging
and
kind
of
identifying
is
that
there
are
discrepancies
between
how
the
protos
define
it
and
how
it's
being
captured
in
and
persisted
in,
measuring
you're
right
that
we
want
to
rectify
that.
We
want
to
use
service,
mesh
performance
protos
as
the
source
of
truth
or
like
as
the
thing
the
way
to
implement.
E
And
so
just
for
familiarity
with
everyone
else
for
everyone
else,
the
other
project
that
we've
just
donated
to
the
cncf
service
mesh
performance.
It's
the
specification,
it's
not
just
the
protos
there's
some
other
activities.
As
a
matter
of
fact,
you
you
all
should
go
to
that
community
meeting.
You'll,
there's,
there's
a
number
of
things
to
be
involved
with
and
work
with
intel
with
work
with
red
hat
with
work
with
there's
a
bunch
of
people
involved.
E
E
You
don't
have
to
swallow.
All
of
that
in
one
go
bark,
but
did
what
I
said.
Did
it
just
regurgitate
what
you
were
saying
does
that
help
reinforce
your
understanding
a
bit?
So
it's
not
exact
so
mesh
sync
is
sort
of
an
aside
like
mesh
sync
is,
you
know,
facilitates
discovery
of
things
that
are
going
on
in
kubernetes,
but
really
it's
the
service
mesh
performance
protos
that
you're
that
we
need
to
ultimately
align
to
and
that
may
take
any
number
of
pr's
to
get
there.
E
Yeah
yeah
good.
I
hope
that
you're
not
placating
me
with
you
yeah
yeah,
that's
good!
That
helps
thanks,
but
hopefully
it
helps
reinforce
it
like
it.
There's
a
couple
of
there's
a
red
hat
or
a
principal
engineer
who
is
looking
at
those
specs
and
trying
to
make
sure
that
they're
correct
so
I'll
have
to
get
you
introduced
as
they
do
that
so.
H
And
one
more
thing
is
like
you
know,
currently
what
we
are
installing
prometheus
in
the
adam,
which
is
there
that
is
not
a
full-fledged
prometheus,
so
that
doesn't
have
a
node
exporter
so,
and
I
just
discovered
that
if
you
install
a
standalone
prometheus
with
their
prometheus
chat,
chiali
can
able
to
access
it
with
the
required
field.
So
it
is
not
really
necessary
to
install
it
from
the
add-on
of
isd.
E
So
not
that
we
can't
give
a
rip
about
chiali
here,
but
but
to
your
point,
I'm
samir.
If
we
don't
have
the
node
exporter,
there
are
a
number
of
additional
metrics
that
mastery
would
like
to
collect
that
it,
wouldn't
it
won't
be
able
to.
H
Yeah,
so
I
why
I'm
telling
is
like
in
the
ui
now
we
are
focused
on
installing
it
from
istio
adam,
but,
like
you
know,
if
you
want
to
leverage
the
other
add-ons
from
the
stu
also,
instead
of
installing
the
prometheus
addon
from
istio,
we
can
also
use
that.
So
that
is
why
I
gave
the
galley
example.
E
All
right,
let's
take
a
note
on
that
as
samir
you're
correct
and
we
need
to
support
the
ability
to
either
connect
to
or
maybe
deploy
a
standalone
prometheus
that
has
a
node
exporter
like
it's
it's
important
and
and
the
when
someone
runs
a
performance
test.
Today,
mescheri
will
look
for
a
node
exporter.
It
wants
to
pull
in
that
data.
H
A
All
right,
ashish
or
rudraksha
would
any
any
of
you
like
to
present
this
next
item.
I
Okay,
I'll
talk
about
it,
so
basically
to
give
a
little
context.
We
are
working
on
incorporating
a
functionality
inside
of
mesh
kit
that
can
help
us
to
generate
schemas
and
definitions
which
would
be
used
by
different
machine
adapters.
So
I'll
just
quickly
share
my
screen
and
it's
still
a
work
in
progress.
I
Okay,
so
basically
there
are
there,
there
will
be
two
functions,
one
would
be
get
from
manifest
and
one
would
be
get
from
helm.
I
haven't
yet
written
the
get
from
function,
but
the
get
from
manifest
function
is
working,
I'm
still
confused
about
the
naming
convention.
So
currently
I've
named
the
package
is
manifest
so
that
that
is
that's
a
different
issue.
We
can
change
the
names
later.
I
So
basically,
this
get
from
manifest
function
is
the
one
which
would
be
called
via
different
adapters
when
they
would
be,
they
would
want
schemas
and
definitions
for
particular
resources.
So
as
parameters
they'll
have
to
give
the
url
it's
the
responsibility
of
the
adapter
to
give
the
url
from
where
we
have
to
fetch
the
manifest
from
which
we
would
generate
the
schemas
and
definitions.
It
also
has
to
give
the
resource.
I
There
are
currently
three
resource
types
service,
mesh,
kubernetes
and
measuring
and
different
definitions
would
be
produced
for
different
resource
types,
and
then
the
user
also
have
to
give
the
user
in
this
case
is
measured
after
machine
adapters.
So
it
also
has
to
give
the
config,
which
would
contain
the
name
of
the
service
mesh
and
the
service
mesh
version,
and
in
return
it
will
just
return
the
component
struct,
which
is
just
a
slice
of
schemas
and
definitions.
I
So
currently
I
have
I'm
just
calling
this
function
from
the
main
dot
booth
to
just
a
demo
and
I'm
just
hitting
this
url
for
a
specific
version.
So
if
I.
I
Okay,
so
a
few
fields
are
missing
here
I
I
was
actually
I
had
doubts
and
I'll
clear
these
out.
So,
for
example,
creation
timestamp
is
null
and
the
api
version
is
empty
right
now,
I'll
clear
up
with
version
fill
these
up
other
than
that
we
have
definitions
and
we
can
also
have
schemas.
I
Yeah
this
is
the
schema,
so
yeah,
basically-
and
currently,
I
have
only
one
function
for
get
from
manifest
if
the
user
wants
to
give
you
produce
it
from
helen
chart
so
that
the
user
can
also
give
url.
For
that,
I
will.
I
haven't
written
this
function
yet,
but
I'll
get
it
done
by
tomorrow.
So
if
anyone
has
any
comments
on
how
I
should
rename
the
package
or
if
anyone
wants
to
see
how
it's
working
functionally
or
anything
else,
I
would
love
to
have
comments.
I
Yeah,
that's
what
that's!
What
I'm
you're
talking
about!
This
read
file
source
function,
yeah
yeah,
I'm
already.
E
Using
oh,
I
see
that
is
from
oh,
okay,
oh!
Is
that
what
the
utils
dot
means?
Yeah,
okay,
good
good
yeah!
Do
we
know
yeah
yeah,
there's
someone
might
want
to
share
the
spreadsheet
where
you
guys
are
tracking
the
mechanisms
I
was
just
about
to
ask
like
which
adapters
we
think
we're
gonna,
which
service
meshes.
E
We
will
collect
manifest
from
versus
which
ones
we
will
collect
helm
charts
from,
and
then
I
realize
that's
in
the
spreadsheet.
So
maybe
you
could
share
it
in
chat
afterwards.
Should
I
beat
you
to
it.
E
I
The
jq
thing
actually
rudraksha
was
right.
It
was
too
dumb
of
me
to
use
jq
when
I
can
just
unmarshall
in
marshall
from
start.
So
that's
why.
That
day
was
speechless
when
he
said.
E
Yeah
I
mean
yeah,
that's
why
it's
why
other
people
are
there
to
point
out
point
out
things
like
that,
but
but
actually
I
wouldn't
necessarily
phrase
it
like
that.
I
would
say:
hey
if
you
were
looking
at
transition
systems,
the
current
implementation
is
using
jq,
hey.
Is
there
a?
Is
there
a
jq
available,
natively
like
it
makes
sense
that
you
arrived
at
that
anyway,
I'm
good
last
question.
E
I
Like
the
do
you
adapt
the
adapters
themselves
right
now,
I
haven't
added
anything
in
the
adapters
because
adapters
are
just
going
to
consume
this,
so
once
we
are
sure
about
this
is
how
we're
going
to
do
it.
Inside
of
the
mesh
kit,
we
can
go
ahead
and
start
working
on
the
adapters.
A
I
Okay,
so
actually
the
first
thing
that
I
wanted
to
discuss
was
like
right
now:
the
configuration
for
our
while
developing
machinery
ui
like
we
are
using
node
one.
We
were
actually
using
nodemon,
for
you
know.
Basically
what
it
does
is
that
we
actually
have
a
proxy
on
top
of
the
next
normal
next
year,
so
next
development
server,
so
that,
like
whenever
the,
whenever
a
request,
is
it's
supposed
to
be
routed
to
the
back
end?
I
Then
it
just
like,
if
you
have,
if
there
is
a
request
that
starting
with
the
slash
api,
then
it's
supposed
to
be
routed
to
9081
in
case
of
development.
So
for
that
we
were
actually
using
a
proxy,
but
while
actually
implementing
the
proxy,
like
we
were
actually
using
node
mode
on
top
of
next
year
server.
So
what
it
was
actually
doing
was
for
each
and
every
change
that
has
been
made
to
the
any
js
files.
I
In
the
context
it
will
actually
rebuild
the
whole
ui,
but
actually,
if
we
are
using
next
year,
server
like
we
all
know
that,
like
there
is
a
it
gives
us
instant
feedback
like
yeah,
that
hot
module
replacement
thing.
So
that
was
not
like
we
were
not
able
to
leverage
it
because
we
were
using
this
notebook
on
top
of
it.
So
actually
I
have
just
removed
like
I
believe
that
was
I
mean
like.
I
I
don't
think,
I'm
not
sure
why
people
have
done
that
like
it's
been
that
changes
like
that
thing
was
built
two
years
before
so
not
sure
why
that
was
written
that
way,
but
I
have
right
now
changed
it
to
node.
So
as
a
result
of
that
like
right
now,
if
we
just
make
a
change
in
any
ui.
I
So
we
would
actually
get
the
instant
feedback
as
we
would
actually
get
when
we
are
using
the
normal.
I
mean
normal
next
year
server.
So
that
is
one
thing
that
I
have
change
right
now,
yeah,
and
apart
from
this
like
from
this,
I
will.
I
Move
on
to
the
updates
that
I've
made
to
connection
wizard
so
basically
like
from
the
last
time
like
I
haven't,
made
much
of
the
updates.
In
this
case,
I've
just
been
working
on
the
stylings
and
trying
to
refactor
the
code
so
that
people
can
contribute
to
start
contributing
to
it.
So
something
like
the
changes
that
I've
made
right
now
is
that
I
have
added
the
button
over
here
for
basically,
we
were
actually
experimenting
with
the
what
which
one
to
go
with,
but
right
now.
I
This
is
what
we
have,
but
definitely
we
could
use
some
help
or
suggestions
like
if,
if
you
have
a
better
icon
or
something
like
that,
but
yeah,
and
so
actually
some
alignment
issues
were
fixed
like
I
don't
think
you
would
be
able
to
notice
it
like
yeah,
so
those
things
and
yeah
over
here
like
when,
if
if
people
were
there
last
time,
I
demoed
this
thing,
they
would
have
noticed
that
the
styling
of
the
adapter
space
would
happen.
I
I
mean
like
this
was
actually
in
the
center
of
the
page,
but
right
now
I've
moved
it
to
the
left
side,
but
I
actually
think
like
it
has
to
be
much
closer
to
this
thing
so
that
it
looks
like
they
both
are
actually
a
single
component
yeah.
So
that
is
something
and
I've
also
added
the
advanced
settings
button
which
the
functionality
of
this
is
that
right,
it
would
actually
take
it.
I
I
Apart
from
that,
I
don't
think
I
have
made
any
other
change,
so
actually
we
would
actually
be
releasing
this
sooner
and
we
would
be
waiting
for
the
feedback
from
the
users
and
they
would
be
building
like
going
from
there
yeah.
This
is
the
update
that
I
wanted
to
share
yeah
and
still
my
I'm
not
able
to.
I
think
I
haven't
troubleshooted
this
machine,
so
it's
not
working
right
now.
E
Any
any
feedback
here
and
people,
so
this
is
a
new
user
interface
that
has
yet
to
be
exposed,
but
I
expect
this
week
today
tomorrow
it
will
start
to
pop
up
in
the
edge
release.
One
thing
to
note
is
that
abi
who's
on
the
phone.
C
I
I
I
E
Yeah,
that's
a
that's
exactly
what
I
was
going
to
clarify.
That's
a
great
question!
So
abi
had
added
the
dash.
Can
you
hover
your
mouse
over
the
dashboard?
E
I
can't
make
sure
yeah
the
avi
had
added
that
as
a
menu
item
and
if
you're
running
the
current
version
of
messagery,
if
you're
running
it,
it
will
have
a
dashboard
the
ability
to
navigate
to
the
dashboard
there
and
it'll
also
have
it
in
the
nav
bar
up
on
the
right-hand
side,
and
so
it
is
not
so
actually
we're
just
going
to
deduplicate
and
remove
the
dashboard
navigator.
E
I'm
sorry
remove
the
dashboard
icon
from
the
navbar
and
just
have
it
in
one
spot.
It's
actually
in
two
spots,
because
if
nafish
clicks
on
dashboard
he'll
be
taken
to
the
main
dashboard,
if
he
clicks
on
the
meschery
logo,
he'll
also
be
taken
to
the
main
dashboard
and
then
what
and
so
we're
just
no
is
the
answer
to
your
question.
The
connectivity,
wizard
or
the
connection
wizard
is
not
replacing
this
main
dashboard.
E
Now
there
is
yeah,
there
is
work
to
be
done
on
the
dashboard
itself.
Barak
has
a
use
case
for
enhancing
the
dashboard
here.
On
the
left
hand,
side.
One
thing
that
needs
to
be
done
for
the
project
is
to
basically
take
that
right
hand,
side
the
connection
status
and
like
move
those
out
of
the
way,
those
are
really
important
for
having
meshri
up
and
running
and
connected
to
all
the
systems
great.
E
But
that
is
not
really
what
people
that
are
using
meshri
are
here
to
look
at
they're,
not
using
meshri,
to
look
at
those
connections
they're
using
measuring
to
manage
a
service
mesh.
So
there's
lots
of
information
about
the
service
meshes
that
mesherie
is
managing
that
information
needs
to
bubble
up
into
a
ui
is
aditya
krishna
on
the
call?
E
No,
but
there
are
let's
get
fancy
like
this.
Is
there
there's
room
for
some
fairly
neat
designs
on
charts,
charts
and
statistics
and
big
numbers,
red
and
green
colors?
H
Are
we
planning
to
remove
those
multiple
service
messages
in
the
left,
yeah.
I
No,
we
are
not
gonna,
remove
those
service
measures
actually,
but
we
want
to
collapse.
Those
things
like
the
life
cycle
is
a
parent
when
you
and
inside
that
is
multiple
child
menu
that
depicts
different
service
meshes
right.
So
we
are
gonna,
incorporate
a
feature
that
will
you
know
collapse
that
a
life
cycle
item
and
all
its
children
and
then
we're
gonna
see
the
next
parent
item
like
configuration,
that's
what
we're
gonna
we
are
thinking
about.
Adding
a
accordion
or
collapsible
menu
is
a
better
one.
I
I
So
adding
that
will
also
add
additional
benefit
like
showing
the
inform
important
information
that
is.
C
That
really
is
valid
and
advantageous
to
user.
So
that's
one
thing
and
yes.
I
Actually,
since
we
are
on
the
like,
we
are
in
the
same
topic
like
I
was
thinking
that
I
believe
someone
has
created
an
issue
that
this
icon
was
not
being
displayed
for.
I
You
know
collapsing
the
whole
thing,
and
I
think
that
is
an
issue
that
we
have
to
think
about,
because
right
now,
if
we,
if
we
are
here
and
the
user,
doesn't
know
that
we
can
actually
scroll
down
here
and
you
know
collapse
it,
then
it
might
not
be
very
intuitive
in
that
case,
so
I
think
we
can
actually,
you
know,
increase
the
index
of
it
and
keep
it
floating
over
here.
That
would
be
a
better
option.
I
guess.
I
Longer
skin,
I
could
not
see
it
but
yeah.
That's
that's
really
makes
sense
to
have
that
collapsible
button
fixed
onto
the
bottom
of
the
drawer
so
that
people
can
just
collapse.
It.
C
A
Yeah,
so
the
issue
that
we
have
mentioned
right
now,
like
it,
seemed
pretty
unintuitive
to
a
new
user.
So
that's
why
we
opened
up
the
issue.
A
A
Okay,
so
current
code
coverage
that
we
have
right
now
in
our
machine
sequel
is
not
necessarily
the
whole
project
measuring
that
is
21.60,
and
we
can
check
that
on
protocols
dashboard
and
the
files
we
are
actually
going
through
right
now
are
some
files
in
package
utils
and
some
files
and
cli
okay.
So
there
are
a
lot
of
what
should
I
say,
issues
open,
but
very
few
folks
are
showing
up
there.
A
I
don't
know,
what's
the
like
problem
there,
I
think
these
are
really
interesting
issues
and
some
challenging
issues
also
like
testing
for
messy
secure
commands.
A
There
are
thousands
like
some
thousands,
like
many
commands,
that
needs
to
be
added
for
writing
unit
tests
and
integration
tests
and,
if
you
guys,
are
not
like,
if
contributing
or
writing
tests
for
your
these
commands
are
hard
for
you,
then
you
can
go
for
writing
simple
unit
test
for
functions
that
have
been
written
in
utils
package.
A
I
mean
in
this
folder.
So
there
are,
I
guess,
more
than
five
files
here
and
there
are
tons
of
function
written
here.
This
file
is
the
biggest
one
that
has
a
line
of
codes
more
than
700,
and
many
functions
are
written
here
and
if
you
guys
are
not
interested
or
you
think
that
writing
tests
for
message
serial
commands
is
hard.
You
can
guys
you
guys,
can
write
a
smaller
unit
test
for
these
smaller
functions.
A
Okay
and
this
week
we
do
not
have
any
number
one
unit
test
writer
or
number
one
tester
unit
tester,
because
there
was
no
new
pr
merge
this
week.
So
let's
hope
there
are
at
least
two
people
next
week
on
this
document,
which
I
have
the
title
as
number
one
unit:
tester
people,
okay,
and
let's
move
on
to
the
second
issue.
That
is
issue
with
api
system
version.
This
is
an
url
which
should
return
this
thing,
but
it
might
not
be
returning
this
thing
to
you
and
you
are
facing.
A
A
So,
let's
see
what
am
I
doing,
I
am
doing
I
as
I
you
guys
can
see
the
channel.
I'm
using
right
now
is
edge
and
the
version
is
latest
and
I
knew
I
have
deployed
all
the
machete
on
kubernetes,
okay,
I
am
able
to
fetch
the
latest
version
here.
I'm
not
sure
why
it
is
stable,
it
should
be
edge,
but
it
is
here.
Okay,
if
you
guys,
are
not
having
this
here
or
facing
issues.
A
My
recommendation
would
be
updating
your
containers
and
deleting
the
site.
Caching,
because
that's
what
what
happened
to
me
when
this
change
like
this
endpoint
was
updated
very
recently.
Previously
it
was
api
server
version
and
we
replaced
server
with
system.
A
I
also
faced
issues
when
it
was
released
and
measuring-
I
don't
know
I
guess
one
or
two
three
release
before
and
I
also
got
panicked.
It
should
work,
it
is
not
working.
Then
I
updated
the
containers
and
eventually
it
was
working
there.
A
A
I
am
really
not
sure.
I
was
also
getting
saying
that
I
like
we
are
not
able
to
fetch
it
or
the
time
taken
to
proceed
process.
The
request
was
too
high
and
process.
The
request
was
then
declined.
Some
kind
of
that
no
idea
to
be
honest,.
E
Yeah,
there's
a
so
looking
at
the
other
client,
not
measure
ui,
but
mastery
ctl
and
running
mystery
ctl
version.
E
You
know
we'll
try
to
invoke
the
same
api,
and
so
I
just
deployed
the
latest
edge
release,
and
then
I
haven't
tried
to
see
if
there's
an
issue
here
or
not,
and
right
now
from
this
client
on
that
edge
release
like
it,
it
was
able
to
hit
that
api,
and
so
that's
that's
interesting.
I
I
assumed
that
it
was
going
to
fail,
but
that's
so
I'll
test
from
measure
ui
as
well
to
see
if
that
issue
is
still
present
in
the
edge
release.
E
A
Okay,
okay,
so
moving
on
to
the
next
item,
rudraksh
or
ugly,
would
you
like
to
talk
about
publishing
the
release
notes
automatically
publishing
these
notes.
F
F
Basically,
we
were
just
waiting
for
the
action
to
return
the
release
notes
earlier
and
that
was
causing
it
to
delete
those
release
notes
once
a
release
was
made.
So
in
that
case
we
needed
to
merge
a
pull
request
which
contained
the
release
notes
before
creating
a
new
release,
which
was
a
hassle,
but
this
is
least
brain
came
up
with
a
elegant
solution
for
this.
He
just
he
said
that
whenever
we
make
a
release,
just
pull
the
release,
notes
from
the
release,
notes,
page
itself
and
yeah
just
put
the
trigger
here.
F
E
And
I'll
do
a
quick
just
recap
that,
like
hey,
so
that
everyone
knows,
whenever
you're,
whenever
you
land
a
commit
or
a
pr
on
the
project,
then
there's
a
few
things
that
happen
automatically.
E
One
workflow
is
to
basically
build
the
documentation
and
publish
it
into
a
as
a
temporary
preview
of
like
what
the
site
will
look
like
as
and
when
his
pr
emerges.
That's
very
helpful
for
reviewers,
it's
very
helpful
for
him
to
check
it
out.
There's
a
number
of
other
tests
checks
that
go
on
there's
work
that
that's
been
done,
that
you
all
have
done
to
anyway.
There's
a
bunch
of
checks.
E
E
It
was
released
a
little
while
ago.
These
are
the
good-hearted
folks
that
contributed
to
it.
Well,
what
do
you
know?
Ashish
tewari,
who
just
demoed
oh
nervous,
just
demoed
like
great
these
guys
are
you.
The
contributors
are
getting
acknowledged,
I'm
good
we're
seeing
some
of
the
same
names
from
release
to
release.
This
is
great.
E
And
so
anyway,
there's
a
lot
of
notes
that
go
on
every
time,
so
this
happens
just
automatically
so
that
people
don't
have
to
sit
there
and
go
back
and
document
it
good.
That
automation
is
very
helpful.
E
E
This
page
will
persistently
show
these
details,
but
we
also
published
the
release
notes
in
mesherie's
documentation.
If
you
wanted
to
learn
about
the
project-
and
you
wanted
to
follow
like
what
were
the
changes
from
one
release
to
the
next
well,
you
can
go
out
to
documentation,
go
to
releases,
there's
a
long
list
of
what
those
are.
Okay.
Well,
then,
getting
those
same
notes,
because
it's
just
a
copy
of
those
notes
like
published
over
here.
J
Hey
guys
I'll
just
share
my
screen.
J
So
I
will
be
showcasing
two
things
today.
One
of
them
is
the
autosave
for
basically
the
a
mesh
map
and
the
other
would
be
node
inventory.
J
These
are
the
two
names
of
the
features
so
before
like
earlier,
what
used
to
happen
was
we
had
not
a
very
proper
way
to
save
the
design
that
we
created,
like
there
used
to
be
a
small
pop-up
that
or
small
tooltip
poster
that
came
from
this
bottom
screen,
which
asks
that
hey
you
have
made
some
changes,
do
you
want
to
save
and
there
was
a
really
bad,
safe
button,
which
many
of
you
would
have
ignored,
which
would
let
you
save
the
design
that
you
have
created
so
that
has
changed
now.
J
We
have
several
ways
in
which
you
can
save
your
design,
so
the
first
one
is
the
save
as
button.
So
what
I'll
do
is
I'll
just
quickly
create
a
dummy
design,
yeah,
and
I
am
not
sure
if
you
guys
were
able
to
see,
but
what
happens
is
as
as,
as
I
drag
anything
from
the
designer
pin
over
to
the
canvas
and
I
drop
it.
J
A
lot
of
things
happen
and
mesh
map
looks
that
okay,
there
are
some
changes,
so
it
triggers
some
methods
which
basically
save
your
design
automatically
just
like.
If
you
have
used
google
docs
before
you
might
find
a
lot
of
similarities
in
the
functionality.
J
J
And
if
you
go
to
pattern
files
over
here
you'll
be
able
to
see
that
oh,
my
design
has
been
saved
and
it
has
been
converted
into
a
yaml
file.
If
you
want
to
go
back,
if
you
want
to
load
the
same
design,
it
would
be
back
over
here,
the
name.
Now
I
guess
my.
Maybe
you
like
what
you
have
done,
but
you
want
to
create
a
copy,
so
you
can
just
go.
Click
save
as
and
write.
J
J
The
next
feature
that
there
is
is
node
inventory
so
as
a
lot
of
people
who
maybe
or
may
not
be
a
little
bit
familiar
with
mesh
map,
but
there's
one
thing
that
you
can
do
on
meshmap
is
you
can
drag
and
drop
filters
over
nodes?
So
if
I
just
go
and
drop
on
my
filter
over
to
my
current
steel
mesh
you'll
be
able
to
see
that
there
is
some
small
thing
that
pops
below,
which
is
basically
the
filter
or
a
tooltip.
That's
showcasing
that
now
there
is
a
filter
dropped
on
this
node.
J
Now,
if
I
have
multiple
filters
onto
a
single
node,
what
happens
is
you
have
a
list
of
like
you
have
a
stacked
user
interface
for
the
filters
and
you
can
just
hover
over
them
to
see
which
filter
you
are
now
in
future?
What
will
be
done
is
when
you
hover
over
any
of
these
filters.
J
Your
mouse
pointer
will
basically
change
to
this
like,
and
it
will
tell
the
user
that
it
is
clickable
and
as
soon
as
you
click
on
any
of
these
filters,
you
would
be
able
to
edit
the
configuration
that
you
just
entered.
For
example,
if
I
drop
another
filter,
you'll
see
that
there
is
a
pop-up
that
comes,
and
this
is
basically
the
configuration
of
the
filter.
J
So
if
you
want
to
edit
that
configuration,
you
would
be
able
to
edit
that
in
the
future
also
right
now,
if
you
want
to
edit
the
configuration
of
the
mesh
that
you
just
added,
you
can
just
basically
left
click,
maybe
I'm
not
on
the
branch
which
has
that
feature
anyway,
so
you'll
be
able
to
edit
that
configuration
as
well.
So
this
is
basically
your
inventory
or
the
nodes
inventory.
J
So
that's
all
folks.
E
So
nice,
so
samir
is
asking
some
great
questions
and
I
think
he's
for
the
most
part
getting
those
answered.
But
there
is
specific
to
samir,
so
sameer
has
been
giving
great
feedback
use
cases,
usage,
blog
posts
contributions,
the
guy,
the
guy's,
making
the
full
tour
which
is,
and
but
one
thing
that
samir
has
has
noted
and
he
has
helped.
E
The
project
advance
is
how
it's
how
measury
is
managing
envoy
filters,
and
actually,
I
think
it
correct
me
if
I'm
wrong
so,
but
the
in
the
inventory,
the
node
inventory
that
you're,
showing
here
the
example
use
case
that
you're
showing
is
the
fact
that
any
given
service
might
have
more
than
more
than
one
filter
and
that
they
could
okay
potentially
be.
J
E
We
should
we
should
I'm
going
to
do
that,
because,
because
samir,
it
will
not
work
entirely,
but
there
are
certain
aspects
that
will
work
and
okay
and
it
will
only
take
a
minute
for
the
the
contributors
who
are
working
on
it
to
describe
what's
broken
so
that
you
don't,
you
know,
get
frustrated,
but
but
the
but
the
rest
is
like
yeah,
it's
time
to
start
cooking
at
it.
H
Okay,
so
one
thing
is:
does
it
source
the
istio,
let's
say
the
mess
for
istio?
We
have
virtual
services
and
also
we
have
services
in
kubernetes.
So
does
this
pattern
will
show
both
the
things
or
because
you
know
if
some
one
of
these
things
breaks
so
whole
things
breaks.
E
Like,
ultimately,
yes
is
the
answer,
I
don't.
I
don't
know
in
this
very
moment
that
the
way
that
you-
and
so
the
question
is
hey
samir
saying
well,
this
is
great.
I
can
you
know,
design
configure
my
service
mesh,
but
also
I'm
configuring,
my
service
mesh
in
context
of
a
larger
kubernetes
deployment,
and
so
there
are
sort
of
kubernetes
native
constructs
like
a
service.
Just
a
kubernetes
service,
a
kubernetes
deployment
of
kubernetes
damon
said
you
know
like
there's
a
list
of
those
constructs.
E
Are
that
can
those
be
represented
as
well
because
they
have
bearing
on
you
know
like
an
ingress
controller,
I
can.
Can
I
visualize
and
show
that
here,
because
it's
part
of
my
architecture
as
part
of
how
my
application
works
so.
H
J
K
You
raised
you
yeah,
so
actually
you
can
see
in
so
this
is
actually
categorized
in
the
sd
conference
and
there's
this
virtual
service.
So
this
is
actually
the
istio's
virtual
service.
You
can
be
the
we
have
some
annotations
which
actually
before,
if
in
case
in
future,
someone
else
also
starts
using
virtual
service
as
a
constant's
name.
Even
then,
you
will
have
a
nice
categorization.
E
C
E
E
Is,
I
don't
believe
it's
a
bug
in
mesh
sink,
necessarily
maybe,
but
it's
preventing
mesh
sync
from
working,
which
means
a
large
portion
of
mesherie
itself,
doesn't
necessarily
work,
so
we
do
need
to
keep
it.
We
do
need
to
give
attention
to
this
and
work
through
it.
We
need
to
hammer
it
out,
so
we
can't
do
that
on
this
call,
but
I
wanted
to
highlight
the
fact
that
I'll
put
the
bug
number
in
there
it
might
be.
We
should
make
sure
that
we
get
the
latest
error,
that's
being
seen
into
that
bug.
H
Sync
start
so,
once
we
download,
like
you
know,
once
we
install
it
with
message
like
help
chat,
so
all
the
applications
are
all
the
components
get
installed.
In
fact,
the
operator
also
get
installed,
but
when
we
start
the
mercedes
server
messing
doesn't
get
installed
in
the
namespace.
H
H
But
I
remember
like
kovisek
has
given
me
a
helm
chart
of,
like
you,
know
some
zip
file
last
time
that
time
it
did
work
with
those
helm
installation,
but
when
I
do
it
via
the
guide,
which
has
been
there
in
the
website
to
test
it
doesn't
work.
E
That
could
be
yeah,
that's
very
good
information,
but
if
you
still
have
a
copy
of
that
chart,
the
one
that
was
working
previously
it'd
be
good
for
us
to
take
a
look
at
the
difference.
E
A
Guys,
as
always,
we
are
late.
So
thank
you
all
for
joining
and
we'll
see
you
next
week
on
the
misrecord
and
we
have
a
couple
of
other
calls
tomorrow.
So
if
you
are
interested
in
working
on
service
mesh
performance,
we
have
a
community
meeting
tomorrow
and
on
anchorman.
There
is
a
newcomers
meeting
that's
happening
on
thursday,
so
you
can
join
the
as
well
to
get
oriented
in
the
community.
A
So
thank
you
all
for
joining.
Okay.