►
From YouTube: Meshery Development Meeting (June 8th, 2022)
Description
Agenda-
-Shivay and Meshmate Adithya Krishna is talking about Meshery in global conferences.
-[Abhi/Ashish] Multi-context Support For Meshery and Extension
-[Uzair] Fix NATS connectivity test #5583
-[Uzair] updates on Settings page.
-[Aaditya Narayan Subedy] SMP Dashboard
-[Ruturaj] App Mesh adapter with Viper #97
-[Pranav] Progress on user profiles
-[Adi] - New Event - Composing Cloud Native Infrastructure with Docker Desktop and Meshery - In-person Docker Developer Community Meetup - Bengaluru 🎉
-[Gaurav] Meshery v0.7.0 Planning
-Refactoring Meshery UI: meshery-ui-restructuring | How to contribute
-Volunteer Needed: v0.7.0 Roadmap feature champion assignment refresh
-[Nithish] Managing multiple channels in golang can get messy
-[Debopriya] New Blog posts:
-Validating Meshery CLI Functionality
-Evolution of Meshery CLI Command Reference
-LFX Experience - Pranav Singh
-SCA Contributhon
-PerfBytes: Ask Me Anything
-[Rohith] Copy to clipboard not aligned #715
Join the community at https://layer5.io/community
Find Meshery on:
GitHub: https://github.com/meshery
Twitter: https://twitter.com/mesheryio
LinkedIn: https://www.linkedin.com/showcase/meshery/
Docker Hub: https://hub.docker.com/u/layer5/
B
C
C
D
E
B
Okay,
hi
hi,
everyone
and
yeah.
My
second
call
I
started
from
the
last
week
itself
so
yeah.
I
am
a
freelance
devops
and
cloud
infrastructure
engineer
so
yeah.
I
work
with
kubernetes
throughout
the
day,
so
yeah
yeah,
like
you're
right,
mr
lee,
I
know
flood
city,
so
yeah
yeah.
I
would
be
happy
to
assist
in
anything
with
flux,
cd
or
cloud
native
infrastructure.
B
B
F
A
quick
note,
if
I
could
just
to
insert
one
is
oh
deborah
priya
may
not
be
on.
I
will
hunt
her
down.
She
has
oh,
is
she
not
available
she's
on
the
call?
Oh
I,
this
is
the
second
time
I've
done
that
I
will
hunt
her
down
anyway,
so
anyways,
there's
a
few
community
events
and
blog
posts
and
things
that
you
all
are
doing.
F
You
know,
promote
promote
those
things.
So
I
want
to
make
sure
that
we
highlight
them
here.
The
real
reason
why
I
interrupted
was
just
actually
to
comment
on
what
tapon
had
said,
which
is
which
is
just
to
make
a
quick
note
that
we
talk
a
lot
about
like
particularly
on
this
call.
We
talk
a
lot
about.
F
Well
often,
we
talk
about
coding.
Sometimes
we
just
like
argue
and
people
prove
me
wrong
about
various
features
and
functions
and
system
behaviors,
and
so
we
do
design
spec
reviews.
Sometimes
we
do
planning
for
the
next
release.
There's
a
lot
that
goes
on
in
this
call,
and
one
of
the
things
I
wanted
to
highlight
was
that,
like
some
of
the
most
important
contributions
that
people
make
in
the
community
are
not
about
are
not
code.
Specifically,
they
are
not
code,
and
so
it
might
seem
like
that's
the
core
focus
and
that's
the
only
thing.
F
That
means
anything
and
it's
like
couldn't
be
further
from
the
truth.
There
are
many
many
people
here,
actually
there's,
there's
a
guy
who'd
been
joining
here
for
a
year
and
a
half
very
consistently,
never
said
a
one
word
on
the
calls.
Never
I'd
had
discussions
with
him
in
conversations
separately,
and
he
was
just
here
just
to
learn
just
absorb
like
and
that's
actually
very
welcomed.
It's
like
great
to
just
see
another
person
on
the
call
have
them
encourage
others
to
join
anyway.
There's
a
ton
of
things.
F
F
F
As
a
matter
of
fact,
I
was
just
talking
to
pranav
singh
who's
on
the
call
who's
been
making
waves
in
the
code.
One
of
the
challenges
that
he's
trying
to
help
out
with
is
a
permissions
issue
in
the
cncf
lab
we're
running
mesherie,
we
run
in
the
cncf
lab.
It's
actually
an
equidix
data
center,
we're
running
performance
tests
across
all
the
different
types
of
service
meshes
a
googler
just
reached
out
to
me
last
week
asking
to
participate
in
this
anyway.
F
The
challenge
in
that
effort
right
now
is
just
it's
a
permissions
issue
for
meshri
to
be
able
to
reach
out
grab
the
cube
config.
That's
on
the
bare
metal
machine
that
that
automation
of
provisioning
has
been
done
for
it's
like
that
type
of
automation,
a
type
of
a
devops
centric
sysadmin-centric
knowledge
is
like,
is
critically
missing,
and
it's
blocking.
F
I
was
just
presenting
to
the
cncftoc
yesterday
on
initiatives
that
on
some
of
these
initiatives,
actually
on
that
very
initiative,
it's
about
service,
mesh
performance
and
the
public-facing
dashboard
that
this
community
has
produced,
and
the
holdup
for
me
being
able
to
show
them
a
really
great
looking
performance
dashboard.
Is
this
permissions
issue?
It's
like
there's
a
lot
of
those
kinds
of
things
and
so
dude.
I
just
wanted
to
take
a
moment
to
let
everybody
know
whether
you're
writing
a
blog
post
speaking
somewhere
telling
somebody
else
just
using
the
thing
complaining
about
it.
F
B
Oh
okay,
I'm
speaking
on
moot
all
right.
So
let's
get
started
off
with
the
agenda
items
now
then
so
lee
briefly
spoke
about
shivai,
giving
a
talk
about
mesh
mates
in
in
austin.
So
that's
it
about
this
topic
right
or
is
it
would
like?
Would
I
think
aditya
krishna
is
on
the
call?
So
if
any
of
you
would
like
to
talk
a
little
bit
more
about
this.
F
B
F
F
B
A
So
basically,
this
context
feature
has
been
added
here,
and
this
has
two
primary
purposes,
one
of
which
is
to
allow
you
to
perform
operations
across
any
number
of
your
clusters
at
once,
uniformly
and
parallely,
so
that,
if
you
have
some
kind
of
bootstrapping
to
do
like
you
have
nine
of
your
clusters
and
all
of
those
start
with
the
same
configuration
you
don't
have
to
you
know,
do
it
one
by
one.
Just
you
have
to
select
any
number.
A
So
you
can
imagine
that,
as
you
select
these
clusters
and
the
any
number
of
clusters
that
you
select
operation
would
be
applied.
So
that's
one.
The
other
primary
purpose
is
that
we
have
a
bunch
of
ways
to
get
an
idea
about
what
is
going
on
in
your
cluster,
one
of
which
is,
as
you
can
see,
on
the
dashboard.
So
it
displays
what
service
meshes
you
have
installed
here.
A
So
this
is
a
pretty
basic
stuff
and
if
you
will
dig
in
a
little
deeper
you'll
find
mesh
map
visualizer,
which
actually
gives
you
a
much
more
fine
grained
view
from
port
to
services
to
which
I'll
go
over
in
a
minute
in
inside
the
demo.
So
there
as
well,
when
you
will,
you
can
use
the
same
context,
switcher
and
the
number
of
the
cluster
that
you
would
take.
A
That
would
be
the
cluster
from
where
you
would
be
seeing
the
data
so
now
that
we
have
got
the
purpose
out
of
the
way,
let's
first
let
me
first
demo
the
first
primary
purpose
and
then
we'll
jump
on
to
the
second.
So
right
now
I
have
my
seo
and
linked
adapter
running
and
for
deploying
any
service
measure
performing
any
service
mesh
specific
operations.
We
have
adapters
for
each
of
the
service
meshes.
We
have
a
seven
eight
nine
adapters.
A
So
I
can
be
pretty
sure
that
I
have
my
operator
deployed
across
all
my
mini
cube1
and
mini
cube2.
I
just
have
two
clusters
running
because
I
wanted
to
have
mercy
on
my
system.
It
does
not
have
that
much
ram
to
support
five
or
ten
clusters
so
and
yeah
I
didn't.
I
could
have
demoed
with
the
case
and
all,
but
let's
just
go
with
mini
cube
for
now
so
yeah.
So
I
have
my
seo
adapter
running
my
linkedin
after
running,
I
can
connect
to
two
of
my
clusters.
So,
let's
start
so.
A
First
of
all,
what
I'm
going
to
do
is
I'm
going
to
select
only
my
mini
cube,
cluster
and
I'll
be
deploying
an
istio
service
mesh.
This
is
something
that
I
did
last
week
as
well,
but
there
would
be
minor
change
as
we
have
improved
graphical
subscription,
so
we
will
also
be
seeing
the
data.
The
second
part
or
the
second
primary
purpose
that
I
talked
about.
So
when
I
have
selected
mini
cube,
mini
cube
is
thinkable
okay.
So
now,
if
I
come
to
my
life
cycle,
so
this
needs
a
little
improvement.
A
So
if
I
just
click
on
life
cycle,
it
just
shows
me
this-
maybe
some
improvement
over
here.
Okay.
So
if
I
go
to
my
istio
dashboard
so
from
here
I
can
select
my
namespace.
Let's
say
I
want
to
deploy,
is
to
your
service
machine.
My
sd
system
name
space
in
my
mini,
cube
cluster
right
now,
when
I'll
click,
one
plus
on
when
I'll
click
on
this
button,
there
would
be
no
modal,
although
was
working
on
something
and
by
this
week
will
have
a
model.
A
A
A
There
is
one
little
caviar
to
point
out
for
the
people
with
kenais
is
that
right
now
the
lock
says
that
the
not
the
log,
the
text
that
you
see
here
says
no
service
message
detected
in
mini
cube,
2,
comma
mini
cube
clusters,
so
it
should
only
it
basically
by
default,
comma
separated
names
of
all
the
cluster.
It
shouldn't
take
the
name
of
all
the
clusters
over
whichever
the
selected
clusters
are.
It
should
only
so,
for
example,
right
now
the
circuit
cluster
is
mini
q2.
A
So
it
should
only
say
that
no
service
message
is
detected
in
mini
cube
to
cluster
okay,
a
little
bit
of
nitpick
moving
on,
so
I
can
also
bring
my
link
id
adapter,
and
if
I
go
to
my
linkedin
page,
I
can
select
my
linkedin
system
name
space.
I
can
select
my
mini
cube.
Okay,
so
my
mini
q2
cluster
is
already
selected
now,
if
I
click
on
plus
wait
3
seconds,
maybe
I
come
back
to
the
page.
A
Wait
three
seconds:
it's
sometimes.
Sometimes
it
takes
a
little
while
to
deploy
okay
yeah,
so
I
can
see
so
by
default.
If
I
have
select
all
so
when
I
reloaded
the
page
as
you
could
see
so
when
you
reload
the
page,
the
select
all
gets
selected
so
by
default.
The
view
is
all
the
clusters,
not
some
of
them,
not
one
of
them,
and
now
you
are
seeing
because
I
have
selected
all
you
are
seeing
data
from
both
the
clusters.
A
Now
again,
that
was
the
second
primary
purpose.
If
you
you
switch
back
and
forth
like
you,
can
you
would
be
seeing
data
from
different
clusters
here?
Okay,
so
that
was
first
purpose.
There
was
a
second
purpose.
There
is
a
the
next
part
of
this
demo
is
to
go
over
to
meshmap
and
before
deploying
a
book
info
app
in
one
of
these
clusters.
Let
me
go
back
to
the
visualizer
and
toggle
the
view
there,
and
these
are
some
issues
that
I
think.
Obviously
I
don't
have.
The
latest
pull
of
the
ui,
so
that's.
A
Why
I'm?
I
was
getting
these
little
snap
bars
over
here,
so
just
to
remove
the
clutter,
I'll
select
only
pod.
So
right
now,
if
I
click
on
select
all
and
I'm
seeing
my
istio
over
here
in
linkedin
over
here
some
parts
from
mystery
or
somewhat
from
linkedin,
and
if
I
deselect
mini
cube2,
I
would
be
only
seeing
my
hto
ports
if
I
deselect
mini
cube,
and
if
I
select
me
only
mini
cube
2,
I
would
be
seeing
stuff
from
linkedin,
so
pretty
basic
stuff.
A
You
can
use
the
same
exact
context
feature
in
your
visualizer
to
toggle
between
the
views
of
your
cluster.
So
by
default,
if
you
select
all
you
get
the
entire
view,
all
the
clusters,
although
it's
pretty
deceptive
too,
because
there
is
a
view
selector
button
over
here,
which
says
single
cluster
and
which
basically
means
that
you're
only
seeing
one
cluster,
but
it
kind
of
is
there
because
you
know
for
conventionally
not
for
conventional
reasons.
It
was
here
before
before
the
multi-context
things
came
in
so
it
still
says
single
cluster.
A
But,
as
you
can
see,
I'm
seeing
data
from
multiple
testers
so
yeah
this.
This
would
need
a
little
bit
change,
but
again
a
small
nitpick,
so
yeah
now
let
me
go
back
to
the
designer
and
I
have
a
sample
up
here
just
to
deploy
for
fun,
so
we
have
a
designer
mode
as
well.
So
in
visualizer
mode
you
can
visualize
stuff
in
your
designer
mode.
You
can
design
stuff.
So
I
have
my
so
these
are.
This
is
my
seo
gateway?
This
is
my
virtual
service.
A
This
is
my
community
service
and
some
of
my
mystery
applications.
So
basically
a
lot
of
times.
You
have
an
application.
That
is
just
a
amalgamation
of
your
deployment
in
a
service
and
for
the
reason
that
you
don't
have
to
separately
design
deployment
and
service.
We
give
out
this
application
component
out
of
the
box
so
yeah,
so
we
have
a
few
applications
over
here
service
and
some.
So
this
is
in
the
book
info
app
that
you
would
find
on
istio's
studios
documentation.
A
So
let
me
select
so
let's
say
that
I
want
to
deploy
it
in
only
one
of
my
clusters,
because
I
don't
want
my
system
to
hang
so.
I've
selected
mini
cube
too,
and
if
I
click
on
deploy-
and
I
hope
that
there
are
no
errors-
okay,
yeah
deployed
book,
so
my
design
is
successfully
deployed.
I
go
back
to
my
visualizer
and
since
I
have
selected
only
mini
q2,
I
would
be
seeing
data
from
mini
q2,
so
I
should
be
seeing
my
book
in
forever
now.
A
Yeah
and
let
me
also
filter
out
the
name
space,
because
my
book
info
app
is
deployed
in
the
default
namespace.
So,
as
you
can
see
your
product
page
reviews
and
ratings
details,
your
basic
booking
for
app
that
has
been
deployed,
and
if
I
go
back
to
my
mini
cube,
there
should
be
nothing
over
there
because
yeah.
A
So
in
there
there
are
no
parts
in
my
default
name,
space
of
minicube
because
I
didn't
deploy
it
there,
so
that
was
the
overall
use
of
this
context
feature
the
two
primary
applications
that
I
told,
one
of
which
is
to
manage
the
cluster
on
which
you
want
to
perform
an
operation.
The
other
one
is
to
manage
the
view
of
the
clusters,
the
data
from
which
you
want
to
see
and
that
pretty
much
sums
it
up
and
yeah.
That
was
the
demo.
A
Yeah
yeah,
thank
you
for
pointing
that
out.
I
think
basically,
user
has
already
worked
on
it
and
I
was
going
to
commit
those
changes
since
to
show
you
the
settings
page.
There
is
a
sweet
little
table
that
I
can
show
you,
but
I
didn't
merge
it
because
you
know
half
an
hour
before
a
demo.
You
shouldn't
merge,
commits
because
things
are
working
now
and
I'm
not.
A
No
yeah,
don't
worry
so
he's
also
working
on
some
other
stuff
as
well.
So
it's
fine.
He
has
changes
made
and
yeah.
That's
a
very
good
question
that
setting
speech
will
have
a
table
and
I
swear
to
god.
I
really
wanted
to
show
that
table
because
I
don't
have,
but
I
don't
have
it
right
now
here,
but
yeah
there
would
be
a
table
and
all
of
the
clusters
that
you
see
right
now.
You
would
be
able
to
configure
the
operator
of
each
of
these
clusters
individually
as
the
rules
of
that
table.
A
B
B
So
first
item
is
this:
nat's
connectivity
test
was
not
giving
the
correct
result.
Despite
messing
was
machine
was
sending
data
and
broker
was
broker
was
able
to
send
data
through
the
missionary
servers.
The
connection
were
all
right,
but
then
also
in
the
ui,
we
are
seeing
that
nats
broker
is
not
connected,
so
the
issue
I
found
out
was
this:.
B
F
F
B
Is
this
settings
page
that
mentioning
the
settings
pages
will
look
like
this
so
here
I
have
currently
have
only
one
cluster,
so
it
is
only
one
row
and
if
you
expand
this
row,
you
will
get
the
information
about
the
communities
cluster
and
here
the
information
about
the
operator
nesting
and
the
nets.
This
the
usual
connectivity
test
and
the
detail
about
this,
and
if
you
have
multiple
cluster
and
all
this
cluster
will
pile
up
in
a
row
and
the
specific
context
operations
you
can
perform
by
clicking
on
this
icon.
B
A
A
B
A
F
Anything
any
feedback
on
this
settings,
page
yeah.
We
want
to
make
sure
that
we
don't
rely
on
us
like
parsing,
is
parsing
a
string
like
if
we,
if
you
are
going
to
designate
between
in
cluster
versus
out
of
cluster,
it
needs
to
be
a
first-class
attribute,
which
I
so
yeah,
don't
otherwise
just
don't
designate
it.
It
is
something
good
to
designate.
It
is
something
that
that
there
is
logic
to
figure
out
whether
kubernetes
met
whether
mastery
is
deployed
in
or
out
and
so
yeah.
F
F
I
think
it
says
yeah
think
do
think
about
that
samara
if
you
would
ibm
they're,
both
potentially
destructive
destructive
actions
and
yeah
yeah.
Give
it
some
more
thought.
If
you
would
I'm
not
saying
yeah,
there's
an
improvement
needs
to
happen.
They
look
kind
of
awkward
hanging
out
in
that,
and
so
one
other
item
is
the
reset
menu.
Is
that
a
new
tab,
or
is
that
just
a
button.
F
Anyone
on
the
call
can
go
assess
the
intuitiveness
of
that
interface
and
if
it,
if
you
stumble
or
if
you
stub
your
toe
on
something,
that's
a
great
moment
in
time
to
just
kind
of
write
down
what
you
were
doing
and
what
you
couldn't
figure
out
and
why
it
took
you
15
clicks
to
do
that.
You
know,
and
so
that's
all.
Those
are
all
good
things
to
reflect
on
cool
who's.
Next.
B
So
so
I've
been
working
on
the
dashboard.
This
dashboard
basically
shows
the
results
of
cncf
cluster
labs.
It
basically
shows
charts
and
other
details
of
the
test
runs
there.
So
as
of
now,
the
all
the
designs
that
was
made
in
figure
is
complete.
We
can
like
change
pages
and
we
can
like
filter
according
to
that
test
that
we
have,
and
the
the
three
pairs
layout
was,
is
also
complete.
F
Nice,
so
this
is
live
data
yeah
yeah.
Everything
is
like
nothing,
hardcoded,
oh
nice.
I
don't
know
that.
I
have
the
best
suggestion,
but
I
will
note
that
it's
a
lot
of
yellow
on
the
page
yeah
I
you
know-
I
don't
know.
Maybe
that's
maybe
that's
the
thing
to
do.
One
thing
that
you
might
want
to
lean
into
like
one
is
the
we
talked
about
this
before
the
chevron
on
the
menu.
Probably
the
coloring
of
that
needs
to
styling
needs
to
change
a
little
bit.
F
Just
the
other
thing
is
in
the
results
page
where
you
have
environment
details
and
is
maybe
just
lean
into
these
things,
and
I
think
you
are
you're
you're
more
or
less
presenting
these
two
buttons
as
a
tabbed
menu,
which
is
you
know,
which
is
great,
like
maybe
lean
into
that
and
like
really
try
to
make
it
a
tap
the
menu
experience.
Where
one
looks
like
it's
active,
the
other
one
is
inactive,
and
then
you
know
you.
B
Actually,
what
I
want
to
say
here
is
had
done
this.
You
know
test
specification
and
environment
details
into
a
tabular
view,
because
we
were
showing
all
the
nine
ten
graphs
at
the
same
point
of
time
in
in
the
first
dictation
right.
So
since
it
is
not
required-
and
we
have
only
one
graph,
so
I
think
we
can
totally
show
all
the
details
at
the
single
page.
I
yeah
I
mean
everything
should
be
in
expanded
view.
B
People
may
you
know,
click
on
it
and
to
collapse
so
that
they
can
get
a
focused
view
of
it.
But
what
I
think
is
the
thing
should
be
expanded
view
rather
than
the
collapse
view
or
whatever
yeah
yeah
yeah.
That
seemed
fair
because
we
are
just
showing
one
chart
and
yeah
that
seems
familiar
yeah
other
than
that.
I
saw
like
when
you
go
to
page
one
yeah.
You
know
when
you
clicked
on
a
test
result
then
go
to
the
results
of
those
pages.
So
yeah
this
one.
B
So
you
are
seeing
empty
page
like
the
footer
and
the
header.
Is
you
know
meeting
each
other?
So
that
should
not
happen.
You
know
when
the
state
is
loading,
so
while
the
application
is
in,
you
know
casing
you
know,
while
the
operation
is
why
you're
fetching
the
result
and
your
result
is
not
yet
live,
and
you
can
just
use
this
any
loader
or
something
like
loading
component.
That
tells
you
that
something
is
being
loaded
right.
So.
F
Is
this
is
terrible?
This
is
horrible.
I
don't
know
just
no,
no
great.
This
is
such
a
what
an
accomplishment.
This
is
because
to
think
that
under
duress
you
just
started
iterating,
it's
like
hey.
We
need
to
have
a
dashboard.
For
my
part,
it's
like.
I
don't
know
what
should
go
there.
I
know
you
know,
I
don't
know
just
create
a
page
called
dashboard
like
put
something
on
there
and
then
you
just
started
iterating
from
there,
and
so
that's
great.
So
it's
great
it's
yeah,
I
mean
it'll,
be
highly
referenced.
F
E
Yeah,
so
I
will
not
take
a
lot
of
time,
so
just
the
context
is
that
earlier
app
measure
adapter
was
using
in
memory.
It
was
storing
its
config
in
memory.
Instead
of
that,
we
I
ported
it
to
use
wiper,
which
is
basically
file
based
config,
which
is
what
istio
was
using,
and
this
actually
also
solves
the
bug
where
the
cube
conflict
was
being
reset
every
time
you
start
up
the
adapter,
so
yeah
just
a
little
bug
that
was
fixed
and
while
I
was
at
it,
I
ported.
A
Yeah,
just
thank
you
because
we
are.
We
are
basically
lagging
behind
on
having
uniformity
across
adapters,
and
these
full
requests
make
me
happy
because
they
just
mean
that
we
are
iteratively
getting
closer
and
closer
to
have
more
and
more
uniformity
across
the
adapters
and
plus
this
little
boy
app
meshes.
You
know,
no
one
gives
much
attention
to
it.
So,
thanks
for
giving
it
attention
this
adapter,
so
everyone
talk.
We
mostly
use
this
theo
and
kuma
kumar.
B
So,
a
little
bit
of
context
of
what
I'm
working
on
is
like.
Whenever,
whenever
we
log
into
mystic,
we
don't
we
get
connected
to
mystery
through
an
authorization
through
a
remote
provider,
and
then
we,
the
remote
product,
gives
our
information
to
machine.
Then
we
get
logged
in,
but
we
don't
have
a
way
through
which
we
can
see
our
profile
data
or
we
can
edit
data.
So
currently,
I'm
working
on
a
user
profile
yeah
in
terms
of
ui,
it's
almost
complete
like
now.
B
F
Just
to
make
sure
I
understand
correctly,
if
you're,
if
you
use
meshri
and
if
you've
signed
in
to
the
mesri
provider
or
signed
into
you,
know
otherwise
known
as
mesh
free
cloud,
you
can
just
anybody
can
go
hit
that
url
and
log
in
and
then
see
both
a
summary
of
things
that
they've
managed
as
well
as
like
it
looks
like
they
can
do,
manage
their
api
tokens
as
well.
Okay,.
B
F
Previously,
people
would,
when
they
log
into
mesh
or
ui,
and
I
guess
it's
probably
still
there.
If
you
log
into
mesher
ui,
you
can
go
to
your
avatar
to
get
token
okay,
so
you
get
the
token
for
that
current
session
and
that
token
will
expire
after
24
hours
or
so
it'll
expire
after
some
time,
and
so
is
that
the
same
behavior
for
the
tokens
that
you
were
just
showing?
B
B
B
A
B
B
So
for
that
we
have
a
separate
branch
named
machine
ui
restructuring,
so
you
all
are
welcome
to
participate
in
the
restructuring
of
machine
ui,
and
here
is
the
tracker
for
the
component
that
we
are
using
in
machine
ui.
So
it's
a
call
for
volunteer
for
everyone,
so
you
can
take
any
of
the
component
from
the
current
machine,
ui
and
re-implement,
using
the
new
set
of
principles
for
dot
seven,
so
I'll
share
the
link
in
chat.
B
B
B
So
yeah
and
we
are
in
the
in
the
back
end
like
we
are
actually
using
a
library
called
gpl
gen
to
accompany
that
and
yeah.
So
the
way
these
things
work
like
in
in
go
like
you
it.
It
is
natural
for
us
to
use
golang
channels
to
actually
you
know,
make
those
subscriptions
work
or
actually
stream
the
data
to
the
ui.
So
that
is
what
we
are
doing
as
well.
So
yeah.
B
B
The
data
that,
like
let's
say,
for
example,
you
created
a
new
pod
or
you
created
something
like
in
your
infrastructure,
then
mesh
sync
would
actually
stream
that
delta
or
stream
the
change
over
to
mystery
server,
where
we
would
actually
store
that
as
a
cache,
so
yeah.
This
is
something
that's
happening
and
yeah
meshing
is
the
component,
which
does
that.
So
this
is
a
function
which
actually
listens
to
the
events
that
mesh
sync
is
sending
and
then
and
then
they
it
accumulates.
It
and
stores
throws
it
in
our
database.
B
So,
as
you
can
see,
we
actually
have
a
bunch
of.
We
actually
have
a
bunch
of
channels
over
here
so
but
yeah,
it's
not
wrong,
like
we
actually
have
a
lot
of
it's
not
wrong
to
have.
I
wouldn't
say
that,
like
it
is
not
a
good
practice
to
have
multiple
a
lot
of
channels,
but
then
in
our
case,
like
the
something
weird
happened,
and
it
keeps
on
happening,
which
is
that,
let's
say
like
I
as
a
developer,
I
come
and
I
want
to
make
some.
I
want
to
make
a
functionality
work.
B
But
then
those
are
not.
Actually
you
wouldn't
be
able
to
see
that
in
the
ui,
but
still
those
are
issues
so
yeah.
So
these
kind
of
stuffs
are
prone
to
yeah.
They
would
happen
when
you
actually
don't
have
a
you
know,
a
structured
way
of
managing
the
golang
channels.
So
yeah
we
actually
tried
to
introduce
something
called
broadcaster.
B
It
is
it's
not
something
it
is
okay
like,
but
but
from
from
our
experience,
you
can
actually
see
that
like
this
is
this
hasn't
helped
us
in
in
a
lot
of
yeah
in
a
lot
of
ways
so
yeah
this,
so
we'll
actually
have
to
come
up
with
their
own
solution
to
make
sure
that
the
handles
the
golang
channels
that
we're
using
are
being
managed
properly.
So
when
you
say
like
we
want
to
manage
a
channel
like
what
I
mean
is
that.
B
B
They
might
actually,
you
know,
manipulate
the
channel
to
two
different
things
which
yeah,
which
have
which
might
cause
issues,
and
they
have
crossed
issues
to
us.
A
lot
of
issues
so
yeah.
There
have
been
multiple
attempts
to
yeah
get
rid
of
these
things
and
one
of
which
was
broadcast
using
a
broadcaster,
and
you
know,
as
I
said,
it
hasn't.
Actually
helped
us
in
a
lot
of
ways,
so
so
so
the
ideal
thing
that
we
would
want
to
do
is
to
I
mean
yeah.
We
haven't
one
thing
that
we
can
actually
do.
B
Is
you
can
actually
just
refactor
the
whole
code
base
and
make
sure
that,
like
we
actually
have
a
channels
which
actually,
which
do
one
thing
and
like
they're,
only
responsible
for
doing
one
thing
and
and
they
yeah
they
should
not
be
used
for
doing
multiple
other
things.
If,
if
you
want
to
do
different
thing,
then
you
actually,
I
have
to
create
a
new
channel
so
yeah.
So
there
are
it's
essentially
like
whenever
you're
writing
code.
You
actually
have
to
okay
yeah,
let's
not
go
into
that.
B
So
yeah
yesterday
is
actually
one
of
the
things
that
we
can
do
right
now
is
to
just
right
now,
if
you
observe
you
can
actually
take
a
look
at
the
code
base
as
well,
so
you
can
actually
see
that
we
are
not
directly
using
the
channels
to
stream
the
data
back
to
the
let's
say,
if
you
take
a
particular
example,
if
you
take
a
particular
handler,
let's
say
the
handler
for
a
particular
a
subscription,
and
then
you
have
a
handler
so
right.
What
we
are
doing
right
now
is.
B
We
are
not
actually
using
the
single
channel
to
actually
stream
the
data
back
to
the
back
to
the
subscription.
We
are
actually
getting
the
data
and
then
like
streaming
it
to
a
different
channel
and
then
from
the
channel.
We
are
actually
streaming
it
back
to
the
subscription,
so
that
is
yeah.
One
thing
that
we
can
do
right
away
is
as
me
and
artists
are
discussing
yesterday,
which
is
to
remove
that
and
just
have
a
single
channel
that
streams
directly
back
to
the
graphics
subscription.
B
We
have
a
messaging
frameworks
pic
which
kind
of
in
a
indirectly
actually
deals
with
these
kind
of
issues.
So
I
would
encourage
you
guys
to
look
into
this.
Stick
and
yeah
suggest.
If
you
have
some
ideas,
then
come
up
with
them
like
yeah,
we
can
actually
discuss
about
it
and
maybe,
if
it's
good,
then
we
can
implement
or
yeah.
So
it's
basically
like
yeah.
This
is
the
situation
that
we
have
and.
F
Nice
so
go
jump
into
the
spec
yeah.
There's,
there's
some
use
of
gnats
within
the
project
and
to
me,
if
you
need
for
distributed
systems
to
talk
to
one
another
like
use
of
a
message,
bus
like
nats
is
a
good
way
to
do.
Channels
is
clearly
not
getting
the
job
done.
We
clearly
have
more
complex
things
than
single.
You
know
unicast
messages,
so
those
so
we
need.
We
need
something
else.
So
a
great
that's,
a
good
opportunity
for
anyone
else
to
jump
in
we've
got
about.
F
G
So
recently
we
have
a
lot
of
blog
posts
published
in
our
site.
So
those
of
you
who
do
not
know
we
maintain
our
own
library
of
blogs
so
very
quickly.
This
one
is
about
a
recent
podcast
that
we
had
with
birth
bites.
Lee
was
a
guest
there
along
with
henrik
and
mark,
and
we
talked
about
service
meshes
and
messaging
and
how
we
use
them.
So
this
is
a
brief
recap
of
sorts
for
that
we
have.
G
We
had
shecode
africa
as
an
event
which
started
in
april,
so
this
is
by
the
participants.
This
blog
is
by
the
two
participants
that
we
had
they
talk
about.
You
know
the
project.
The
project
was
the
discuss
leaderboard,
they
talk
about
everything
they
talk
about
the
challenges
they
face,
how
they
finally
manage
their
way
through.
So
this
is
a
blog
of
like
about
their
experience.
G
Similarly,
we
have
a
blog
by
pranav
who's
on
the
call
he
just
completed
his
lfx
internship,
and
he
talks
about
his
way
up
the
contributor
ladder
here
and
he
actually
describes
everything.
Not
everything
he's
worked
in
a
lot
of
things.
Some
of
the
main
highlights
of
what
he's
worked
on
some
mistakes
that
he
made.
You
know
it's
just
a
brief
account
of
his
experience
here.
G
G
So
looking
at
these
blogs,
like
just
a
brief
note,
that
blogs
can
range
for
a
variety
of
topics,
so
if
you're
working
on
messaging,
for
example-
and
you
come
up
with
a
new
feature-
you
are
most
welcome
to-
you
know-
write
that
write
that
down
as
a
blog.
You
know-
and
we
can
have
this
up
over
here-
any
new
feature,
anything
that
you've
worked
on
any
experience
that
you
had
anything
that
you
want
to
talk
about.
You
can
just
write
it
down.
We
have
a
blog
template
that
we
follow.
G
G
So
see
like
you
saw,
this
is
an
experienced
course.
This
is
like
for
this.
You
don't
need
any
technical.
You
know
knowledge
like
you,
don't
have
to
write
code.
You
can
talk
about
any
piece
of
information
that,
like
this
is
about
debugging
convoy
proxy.
So
you
don't
have
to
necessarily
write
code,
whereas
this
one
is
about
code.
This
one
is
about.
You
know
github
and
everything.
So,
oh
nice,
okay,.
F
G
This
is
where
we
store
our
like
source
collections.
Blog
is
a
collection
you
go
to
blog
and
following
the
md
like
this
is
the
blog
template
and
following
this
mdx
format,
it's
very
simple:
you
just
have
to
copy
paste.
This
make
a
folder,
but
if
you
think
that
this
is
something
like,
you
are
only
into
writing,
and
even
this
is
becoming
too
much
of
a
trouble
for
you.
Then
you
can.
Of
course
we
have
a
slack
channel
called
block
kitchen.
G
You
can
join
that
and
if
you
think
that
you
cannot,
you
know,
follow
this
template
share
your
google
doc
like,
if
you
think
you
cannot,
you
know,
convert
this
into
an
mdx
format
or
it's
difficult.
You
know
contributing
to
the
site,
then
you
can
share
your
template
and
we
can
create
issues
so
that
someone
else
can
take
up
the
job
of
publishing
into
like
the
website.
But
the
process
is
simple:
this
is
the
blog
template
source
collection,
blog
blog
template.
F
F
F
D
D
So
I
I
just
came
up
with
a
simple
solution
with
I
actually
no
solve
that.
So
now
it's
this
build
is
running
on
my
local
machine,
so
I
just
solved
the
first
part
of
the
copy
to
click
both.
So
now
it's
properly
aligned
and
as
you
can
see,
I
already-
I
also
saw
the
scroll
error,
so
the
extra
space
is
not
being
used,
but
that
only
works
if
there
is
a
small
amount
of
information
in
the
in
the
blog
in
the
card.