►
From YouTube: Meshery Development Meeting (Oct 13th, 2021)
Description
Meshery Development Meeting - October 13th, 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
mystery
development
meetings
today
is
13th
of
october
and
it's
the
kubecon
week.
So
the
first
topic
we
have
for
today
is
is
about
kubecon.
So,
as
we
mentioned
last
week,
mashari
is
participating
in
this
year's
bug
bash.
So
you
might
have
seen
a
couple
of
pr's
coming
from
the
cncf
bugwash.
So
this
is
basically
an
event
where
you
can
fix
bugs
in
measuring
and
earn
prices
or
win
prices,
so
do
check
it
out.
A
We
also
have
the
meshery
project
office
servers
depends
today
so
13th
october
and
at
2
30
p.m.
Pacific
time,
so
it
might
be
a
late
for
a
couple
of
people,
but
some
of
the
measuring
maintainers
will
be
joining
the
call
and
we'll
be
hanging
out
and
talking
about,
measuring
and
showcasing
measuring.
So
if
you
are
joining
at
kubecon
feel
free
to
jump
in
on
our
project,
office,
servers
and
yep.
A
Check
out
that
link
in
the
chat,
we
also
have
the
smp
project
office
servers.
That
is,
that
is
on
friday,
so
also
check
that
out
as
well
lee
you
have
do
you
have
something
to
say.
B
Yeah
well,
yeah
come
come
to
those
office
hours
come
show
your
support
come
ask
questions.
Take
a
look
at
the
road
map
we'll
be
talking
about
roadmap
later
today
for
what's
for
measuring
between
the
dot
six
release,
that's
happening
now
and
1.0.
We'll
also
be
telling
people
about
the
dot.
Six
release
like
the
fact
that
there's
three
new
beta
features
and
a
whole
bunch
of
other
things
included
so
yeah
join.
It
is
late
for
some
of
you
that
are
on
this
call.
B
The
call
on
friday
same
thing:
mesherie's
performance
management
functionality
is
based
upon
the
open
specification
service
mesh
performance.
That's
the
name
of
the
spec
service
mesh
performance,
there's
a
as
a
matter
of
fact.
A
couple
of
press
releases
went
out
today
just
a
little
bit
ago
about
service
mesh
performance
and
measuring
both
of
them,
and
so
there's
the
world
has
a
new
standard
smp.
B
B
Mesri
is
the
canonical
implementation
of
s
p,
but
jump
in
there
as
well
like
there
are.
There
are
a
number
of
things
to
do,
particularly
those
of
you
who
might
be
on
the
call
that
aren't
into
whose
focus
isn't
just
coding.
B
The
s
p
project
has
a
number
of
initiatives
around
interacting
with
labs,
getting
into
the
cncf
labs
running
a
number
of
different
tests,
understanding
different
benchmarks,
working
with
folks
from
red
hat
hashicorp
intel
publishing
you
know
and
studying
the
results,
publishing
them.
There's
an
ieee
paper
that
we're
well
that
it
publishes-
I
guess,
sometime
this
month,
but
but
the
the
group
that's
there
wrote
a
paper
we're
being
asked
to
write
a
second
one.
B
It
wasn't
the
hardest
thing
in
the
world
to
do
that,
first,
one
because
we've
already
done
most
of
the
work.
The
second
one,
though,
should
really
talk
about
things
that
we
haven't
done
yet
things
about
the
benchmarks
that
need
to
be
run.
B
So,
particularly
if
testing
is
your
thing,
that's
a
great
project
to
get
involved
in
the
project
itself.
We
don't
when
we
meet.
We
don't
just
talk
about
service
mesh
performance.
We
talk
about
get
nighthawk,
so
next
time
we
meet
is
actually
on
the
28th.
A
googler
is
presenting
he'll,
be
presenting
on
a
c
plus
project
on
the
nighthawk
project,
and
we
want
to
do
some
well
intriguing
things
inside
of
nighthawk.
We
want
to
do
some.
B
We
want
to
create
an
adaptive
load
controller,
so
I
won't
go
further
into
it
than
just
to
say:
there's
different
languages
being
used
and
a
bit
of
a
different
skill
set
needed
for
some
of
what
we're
looking
to
do,
those
graphs
right
there.
Those
were
produced
by
a
grad
student
at
the
university
of
austin
or
university
of
texas
austin
of
ut
austin.
B
We've
been
engaged
with
a
professor
and
a
couple
of
grad
students
to
do
examinations
like
this.
So
like
we
talk
about
web
assembly
in
smp
and
envoy
and
nighthawk,
and
so
yeah
it's
at
first
blush.
I
think
a
spec
project
is
to
be
to
be
candid.
Like
super
super
boring,
they
they
tend
to
be
if
they
really
become
a
standard.
B
Oh
god,
it's
like
every
apostrophe,
every
like
it
it
all
it.
It
drags
on
forever.
We're
not
there.
Yet
there's
a
lot
going
on
so
jump
in
even
this
site
get
nighthawk,
it
needs
some
love.
It
generally
captures
what
we're
trying
to
do,
but
like
some
of
the
images
on
the
site,
those
were
only
ever
placeholders.
Those
are
literally
placeholder
images
that
are
on
the
site
and
I
think
I've
seen
like
seven
eight
people
call
out
like.
Oh,
I
can
only
see
half
of
the
picture.
B
It's
like
yeah,
it's
just
a
placeholder
and
I'm
chuckling,
not
at
the
people
who
are
calling
it
out.
I'm
chuck
like
that
six
grid
thing.
I
don't
like
what
that
is.
I
don't
know
you
know
again
just
a
placeholder
like
we
should
get
something
else
up
there
and
by
we
I
mean
not
me
any
one
of
you,
nicole
udkarsh
has
taken
some
animated
gifs
of
how
how
performance
management
works
in
measuring
that
that's
great,
I
don't
know
if
a
nearby
is
on,
but
anyway
there's
there's
a
number
of
you.
B
Who've
done
works
around
this
monav
is
on
right.
Now,
he's
been
fixing
some
things
on
the
the
performance
management
functionality
at
mesri,
like
screenshots
of
the
new
warning
sign
that
might
be
bad
like
that
might
be
not
like
if
we
have,
that
would
be
better
than
having
placeholders
on
this
site.
So
so
there's
so
I
guess
yeah,
there's
a
few
things
for
any
of
you
who
are
either
didn't
find
your
place
in
measuring
smp
has
got
a
whole
other
set
of
concerns
and
activities
going
on.
B
So
there's
a
lot
of
room
for
people
to
step
up,
there's
a
lot
of
room
to
make
new
friends
as
well
across
the
organizations
that
I
was
just
mentioning.
So
as
a
matter
of
fact,
I've
some
of
the
folks
at
intel.
I'd
specifically
said
that
that
we
would
get
interns.
We
would
have
you
that
have
them
engaged
directly,
so
yeah
cool,
good,
sorry
to
interrupt
so
much.
I
forgot
that
we
had
measuring
office
hours
later
today,
there's
so
much
going
on
jump
in
there.
B
I
would
so
much
like
to
never
hear
my
voice
speaking
about
mesri,
but
hear
your
voices
speaking
about
it.
That
would
be
wonderful,
so
hopefully
this
is
the
last
kubecon
that
I
assist
with
office
hours.
I
guess
I
guess
that
gives
everyone
about
a
six-month
timer
from
here
until
cute
county.
U
is
the
project
itself
isn't
too
far
away
from
you
know,
moving
from
sandbox
to
incubation,
you
want
to
prepare
for
that
and
get
ready.
C
B
C
So
there
were
three
states
when
we
deal
with
fetching
data
from
charts.
One
is
when
the
chart
data
is
directly
available,
which
directly
fetches
after
every
say,
four
four
thousand
seconds.
Then
there
is
a
state
when
data
is
not
physically
available.
We
say
that
we
will
fetch
the
data
and
we
will
keep
fetching
three
three
arrows
states
have
rise
three
times
we
will
fetch.
If
we
can't
get
it
back,
then
we
will
stop
fighting.
C
So
what
was
initially
happening
was
when
we
were
fetching
data
data
might
not
return
back
after
three
repeats,
so
the
chart
data
was
paused
with
the
previous
state.
No
new
data
was
getting
fetched
and
the
error
message
was
significantly
displayed
because
no
three
requests
were
ended
and
we
could
not
fetch
it
again.
So
that
was
the
error
which
I
dealt
with.
So
what
happens
now
that,
when
the
error,
when
the
array
states
finishes
three
requests,
this
displays
the
error
message.
C
So
if
I
stop
the
grafana
and
it
will
now
no
don't
be
able
to
fetch
data,
so
the
loading
states
that
we
are
trying
to
reconnect
to
the
reconnective
grafana.
So
we
are
trying
to
reconnect
gravana
three
arrest
states
will
rise,
so
we
wait
for
the
fetch
request
to
finish.
C
Yes,
so
now,
thus,
if
three
phase
requests
are
finished
and
we
can't
connect,
so
it's
not
displaying
any
data.
Initially,
it
used
to
pause
from
the
previous
data
and
if
I
restart
the
graph
server.
B
And
I
I
wish
dhruv
patel
was
on
the
phone
right
now
to
see
this.
This
his
face
might
be
flush.
This
is
kind
of
embarrassing
mono.
I
think
this
is
the
best
update
that
I've
seen
to
this
area
for
a
really
long
time.
Okay,
it
it's
yeah,
that's
great
man!
Yeah
I
mean
I
don't
you
know.
This
is
just
my
comments,
but
others
are
welcome.
B
To
comment
like
the
the
ux
is
a
bit
sexy
where
you've
got
the
horizontal
kind
of,
like
you
know,
not
spinner,
but
the
horizontal
progress
spot,
yeah
and
then
red
like
actually
that
warning
sign.
We
might
want
to
change
it,
so
they
could.
There
should
be
an
orangish
in.
Hopefully
you
have
access
to
figma.
You
might
find
something
a
little
less
dangerous
like
a
little
more
of
a
warning
than
it
is
like
red,
hot
danger.
B
I
say
that
to
also
say
that
what
was
there
previously
was
like
a
small
bit
of
text
that
was
in
bold
and
in
red
and
really
like
any
time.
People
see
red
like
we're
all
taught
red
is
very
naughty
right.
It's
bad!
You
know
things
are
things
are
not
good,
whereas
green
is
sort
of
the
opposite
of
that,
and
so
it's
a
nice
improvement
that,
like
hey,
it's
obvious
that
people
like
in
the
disc
panel.
B
You
know
you're
not
obviously,
you're,
not
seeing
anything.
People
don't
need
like
a
red
ugly
thing
in
their
face,
reminding
them
that
they're
looking
at
a
broken
interface
like
that,
rather
what
you've
done
is
a
bit
more
subtle
and
which
is
nice.
It
doesn't
mean
that
we
couldn't,
we
could
potentially
say
something
in
there.
We
could
potentially
improve
upon
this
by
having
like
a
a
blurred
like
an
intentional
placeholder
image.
That's
like
a
blurred
chart
that
you
know
shows
them
that
something
would
be
there,
but
they're
currently
not
saying
anything.
B
So
it's
not
just
totally
blank,
but
like
I
don't
know
that
you
should
take
any
of
what
I'm
saying
and
put
it
into
action
just
yet
like
we
might
want
to
do
a
mock-up
or
something
like
the
word.
The
words
there
was
an
error
communicating
with
the
server
they
might
make
sense
to
like.
I
know
that
you
just
got
done
removing
them.
B
If
they
were
put
back
on
the
panel,
they
don't
have
to
be
read
anymore,
like
you're
already
giving
them
the
warning,
sign
and
they're
already
not
seeing
things
and
anyway,
I
guess
what
I'm
trying
to
articulate
is
you've
made
some
subtle,
some
changes
that
that
indicate
the
same
thing
to
the
user.
They're
done
in
a
more
subtle
way:
you've
added
functionality
around
the
progress
indicator,
which
is
a
lot
nicer,
you've
added
functionality
to
timeout
after
the
three
you
know
three
requests
and
then
the
individual
panel
refresh
that's
nice
yet
too
oft.
B
I
think
one
of
the
problems
we've
had
in
this
area
in
the
past
is
that,
like
especially
by
the
way
just
nomenclature
for
everyone.
That's
looking
at
this
is
that
we're
looking
right
now
at
four
panels,
cpu
disk
and
and
two
other
ones.
Those
panels
are
contained
within
a
board,
and
these
terms,
this
nomenclature.
This
is
source
from
grafana.
B
It
really
it
only
lets
you
look
at
one
of
those
boards
at
a
time.
So
it
has
this
accordion
right
now
is
how
you
nav,
and
so
there's
a
lot
of
power
in
here.
Most
users
never
get
here,
and
we've
struggled
to
kind
of
get
people
here,
we're
almost
there
we're
almost
to
the
point
where
mesri
should
just
provision
of
prometheus
and
start
to
collect
its
own
metrics
without
waiting
for
someone
to
provision
a
grafana
and
a
prometheus.
B
Anyway,
just
a
bunch
of
random
feedback
just
to
say,
like
one,
you
know
great
great
job,
like
some
good
steps.
One
thing
that,
like
yeah,
I
don't
know
I
you
know
you
we
timeout
after
three
fetches
and
on
one
hand
that's
like
good,
so
we
don't
just
sit
there
burning
the
user's
cpu
with
a
bunch
of
fetch
requests
that
aren't
going
to
go
anywhere
because
the
data
source
isn't
available.
B
B
Yeah-
and
it
doesn't
have
to
be,
it
doesn't
have
to
be
much.
It
might
just
be
like
one
additional,
a
small
additional
piece
of
logic
that
that
includes
like
hey
once
we've
given
up
on
those
first
three
then
find
every
minute
thereafter
make
an
attempt,
and
you
leave
it
at
that
or
something.
B
B
Anyway,
this
is
this
is
great:
okay,
I'd
like
to
catch
up
with
you
to
talk
about
to
explore
and
anyone
else
who
wants
to
catch
up,
but
just
to
explore
like
what
to
do
next,
because
this
is
really
powerful
stuff,
it's
buried
in
the
ui
more
and
more.
This
works
like
for
such
a
long
time.
It
had
been
a
troubled
area.
B
People
had
blamed
the
grafana
sdk,
and
I
think
that
that
was
probably
true,
but
yeah.
We
need
to
get
get
some
on
the
on
the
main
dashboard.
We
just
need
to
start
to
expose
a
few
things.
Some
of
those
things
it
doesn't
have
to
be
like
these
line.
Charts
are
pretty,
doesn't
have
to
be
this.
It
can
just
be
a
big
fat
number,
just
just
a
simple
digit
that
says.
B
I
don't
know
number
of
service
meshes
under
management
or
or
for
things
like
some
of
the
adapters,
like
the
nginx
adapter,
how
many
versions
of
engine
x
does
measury
interact
with
and
maybe
for
all
the
ada
and
anyway
there?
I
guess
I'm
just
so
randomly
thinking,
but
but
the
point
being
is
like
the
dashboard
itself,
the
main
dashboard
it
just.
It
needs
to
completely
change,
and
it's
time
that
we
get
some
mockups
going
you
all.
You
all
are
welcome
to
make
suggestions
every
which
way
jump
into
figma.
B
So
we
had
made
a
couple
of
color
tweaks
to
the
performance
test
results
and
the
chart
that
was
being
used
in
there.
He
was
bringing
the
coloring
of
the
charts
a
little
more
in
line
with
the
theme
that's
defined
in
figma
monavie
for
nothing
else
like
you
might
be
able
to
do
the
same
with
a
few
of
the
colors
here.
C
B
The
tooltips,
like
in
general,
we're
trying
to
we're
landing
right
now,
somewhere
between
roboto
and
open
sands,
as
the
fonts
that
are
being
used,
and
you
might
take
a
look
at
what's
being
used
in
those
tooltips,
the
tooltips
themselves.
Don't
have
rounded
edges.
We
might
want
to
round
those
corners
off
like
just
just
a
few.
B
Cool
okay:
I
was
just
kind
of
waiting
for
other
feedback,
but
so
also
mono.
The
the
we've
got.
We've
got
this
nasty
like
habit
of
building
up
layer
upon
layer
in
the
ui.
So
so,
if
you
hover
your
mouse
on
the
very
like
light
gray
background
of
the
yeah,
so
whatever
whatever
that
is
so
that's
so
whatever
that
is
layer
one,
then
we
go
to
layer
two,
which
is
the
first
card,
the
first
or
yeah
there,
and
then,
if
you
go
down,
we've
got
measuring
results
like
even
further
down
anyway.
A
All
right,
thank
you.
Mano.
A
All
right,
thank
you
for
the
demo
man.
Next
up,
we
have
push
giving
an
update
on
the
unit
test
coverage
in
measuring
so
piosha.
Could
you
talk
about.
A
Oh
hi,
okay,
so
there
is
an
update
about
message.
Ctl
wait!
A
second,
not
necessarily
my
solution,
tl
test
coverage.
There
is
a
link
attached
here.
If
you
click
it.
A
So
recently
we
kind
of
crossed
30
percent
mark
okay,
this
part,
mexico
and,
as
the
sub
point
of
this
agenda,
says
calls
for
participation
like
we
want
people
to
come
and
write
unit
tests
for
like
messages.
Okay.
So
how
do
you
write
it?
A
A
A
Let
me
get
back
there.
It
is
yeah,
it
is
thirty
percent,
okay
and
the
next
goal
that
we
want
to
reach.
That
is
40,
which
is
kind
of
attached
here,
with
the
version
0.7
total.
So
that's
the
next
step
we
have
to
do
and
like
now
the
question
is:
how
do
I
find?
Where
do
I
have
to
write
the
test
like
which
area
requires
the
testing?
Okay,
the
code
code
will
help
you
here.
You
need
to
log
in
by
the
way,
to
like
kind
of
go
to
the
files,
exactly
check.
A
A
A
Thank
you
for
taking
this
up
like
30,
is
a
long
way
we
have
come,
so
I
I'm
sure,
like
a
lot
of
people,
have
been
helping
to
get
this
number
up
so
yeah.
If
you
are
interested
in
this,
get
involved,
reach
out
to
pyush
and
help
get
the
numbers
up.
B
Does
it
assist
with
identifying
who
might
like
a
leaderboard
for
who
is
helping
get
the
code
covered.
B
B
I
see
I
see
okay,
yeah
yeah
it'd
be
interesting
to
identify
those
few
individuals.
B
And
uplift
them
in
their
works:
okay,
yeah!
If
something
comes
to
mind
about
how
to
something
non,
not
manual
that
comes
to
mind
about
how
to
identify
those
people
like
because,
like
it's
all
in
there,
the
usernames
that
you
know
completed
that
coverage,
the
way
that
codecov
is
creating
a
percentage
and
like
that
it
should
be.
There
should
be
mario,
just
randomly
or
jared
you
guys,
don't
are
either
the
two
of
you
very
familiar
with
codecov
as
a
tool.
B
B
To
writing
unit
tests
or
what,
specifically
regarding
to
identity,
to
figuring
out
to
getting
a
leader
board
for
who,
which
contributors
have
contributed.
The
most
code
coverage.
B
And
rodolfo
are
probably
the
two
individuals
to
see
if
credentials
are
needed.
I
think
for
the
the
free
account
that
the
project
has,
I
think
I
think,
there's
like
five
or
so
accounts
that
can
we
can
that
can
be
created,
and
if
we
spill
over,
we
can
go
request.
The
cncf
to
maybe
get
get
us
help
us
out,
but
yeah.
B
That's
on
the
wrong
the
wrong
side
of
100
or
of
a
much
higher
percentage,
but
piyush
has
been
championing
this
just
kind
of
beating
the
drum
and
helping
get
things
through,
which
is
great
and
there's
been
a
number
of
contributors
that
have
come
to
help
help
get
the
number
up
it's
great.
What's
unfortunate
is
that
much
of
what
we're
talking
about,
I
think,
is
primarily
within
measuring
ctl
and
not
necessarily
the
rest
of
measuring,
which
means
that
we've
got
a
long
way
to
go
so
yeah.
E
Just
from
a
brief
glance,
I
don't
see
anything
obvious
in
there.
I
can
keep
looking
a
bit,
but
I
don't
see
anything
obvious
in
the
documentations.
Okay,
dang.
B
Just
in
case
okay
yeah-
and
it
may
not
be
something
that
they
facilitate,
so
that's
too
bad.
Okay,.
A
Thanks
piers
next
up,
we
have
darren
darren,
has
been
working
on
updating
measuring
ctl
to
use
helm
charts
darren.
Do
you
like
to
talk
about
your
recent
pr.
F
Yep,
let
me
show
my
screen
first
sure.
F
So
I
have
two
prs
for
that:
migration
one
is
in
under
the
mesh
kit
ripple
and
then
the
other
one,
of
course
under
the
measuring
repo.
Here
first,
I
will
talk
about
the
pr
under
the
mesh
kit
repo,
because
that's
the
utility
functions
that
we
are
using
in
mastery
cto
to
do
the
installation
using
the
hem
charts
so.
F
Yeah,
so
basically
I
updated
this
library
so
that
we
support
the
upgrade
while
the
hem
upgrade
command.
Basically,
because
previously
we
did
not
support
the
upgrade
for
the
function
that
we
had
for
applying
the
hem
chart.
So
I
basically
did
that
and
then
to
do
that.
Basically
I
just
you
know,
let's
go
through
from
the
beginning
on
the
left,
is
what
we
have
right
now
in
the
master
branch
and
on
the
right
is
what
I
have
in
my
branch.
F
So
I
basically
added
like
three
constant
variables:
uninstall
that
stands
for
you
know
installing
time
shot
upgrade.
You
know
for
upgrading
hem
shot
install
for
installing
the
hem
chart,
so
those
are
used
for
checking
what
the
user
wants
to
do
and
then
I
basically
deleted
this
delete
boolean
variable,
because
previously
we
only
had
like
two
actions,
one
is
delete,
the
other
one
is
create
or
install,
but
now
since
we
have
three,
so
I
think
it's
better
to
not
have
the
boolean
variable
well
it
well.
F
If
we
want
to
keep
it,
we
will
need
to
have
other
boolean
as
well
for
create
and
upgrades,
because
there
are
like
three
cases
now.
So
I
basically
just
you
know,
replace
that
with
a
string
of
action
and
then
and
then
I
would
go
back
to
why
we
use
this
as
a
string.
F
So
oh
here's
one
example
in
line
254,
we
check.
If
this
action
is
in
and
then
it
is,
we
want
to
check
the
hem
chart
is
already
existed
in
the
cluster
or
not
because
ham
doesn't
really
support.
Well,
hemp
install
at
least
is
not
item
potent
I
I
remember
one
of
my
viewers
mentioned
that
and
then
so
to
make
it
kind
of
you
know.
Item
put
into
the
user
is
basically
we
check
if
the
action
is
installed
and
then
and
then
we
I
have
a
function
to
check.
F
You
know
if
the
user
well,
if
the
cluster
already
had
the
release
that
the
user
is
going
to
install
if
it
is
then
basically
just
update
action
from
from
install
to
upgrade
because
ham
does
support,
upgrading
a
hem
chart
from
a
to
b,
even
though
a
and
b
are
the
same.
So
that's
what
this
little
part
is
so
yeah.
That's
why
I
use
you
know:
string
variables
of
boolean,
otherwise
you
you
well,
you
have
it
just
like,
in
my
opinion,
just
a
little
bit
cleaner
or
more
cleaner.
Then
we
have
one.
F
You
know
boolean
variable,
to
try
to
figure
out
which
one
of
the
well,
which
cases
is
from
the
three
cases
and
then
also
updated.
I
think
that's,
oh
yeah.
I
also
updated
the
january
action,
so
this
is
the
function
that
so
when
it
comes
to
here,
we
actually
do
the
either
like
either
one
of
the
cases.
All
the
action
that
we
want
to
do
either
install
upgrade
or
the
default
is
the
install.
F
So
we
using
the
hem
api
called
new
install
to
do
the
new
uninstall
for
the
uninstallation
and
then
same
for
the
upgrade
and
then
for
the
install.
This
part
really
had
that
before
the
install
and
then
I
same
for
the
uninstall.
The
only
part
I
added
is
the
the
upgrade.
F
So
that's
what
I
have
for
the
pr
under
the
mesh
kit
and
yeah
one
of
my
rewards
ashes.
I
think
he
brought
up
the
backward
compatibility
issue
because
I
deleted
this
variable
and
then
I
added
another
string
to
do
the
condition
checking,
but
I
am
concerned
that
you
know
keeping
these
boolean
variables
along
with
the
new
fields,
while
the
action
string
that
I
added
is
confusing
to
the
you
know
to
the
viewers
of
the
code
and
then,
in
terms
of
the
barrel
compatibility,
I
think
we
do
need
to
sacrifice
that
you
know.
F
Sometimes
you
know
to
improve
the
existing
code
base,
plus
the
fact
that
you
know
all
the
adapter
installation
code
they
won't
break
until
they
upgrade
the
mesh
kit
to
the
newest
version
after
these
prs
merge.
So
I
don't
really
see
the
backward
compatibility
is
the
big
issue
here,
but
yeah
I'm
open
to
you
know
any
other
suggestions
if
there
are
any
either
either
in
a
call
or
you
know
offline
on
the
pr,
but
that's
all
I
have
in
this
pr
and
then
yeah.
This
is
just
like
this.
F
This
just
serves
as
a
utility
function
for
the
actual
work
in
the
system
saw
system,
update
and
system
stops.
So
in
here
that's
what
the
main
things
are
happening.
Let
me
do
compare,
oh
my
god
there's
so
many
windows
yeah,
so
the
one
on
the
right
again
is
my
branch
and
then
basically
I
deleted
a
lot
of
the
like.
All
these.
You
know
downloading,
manifests
and
then
and
then
overriding
some
values
in
the
manifest
and
then
applying
the
manifest
because
they
are
all
actually
done
in
one
api
from
the
match.
F
Kit
that
we
just
saw,
that
is
the
apply.
Hem
chart
function,
api.
So
what
we
need
to
just
pass
in
the
hem
chart
url
I
have.
I
have
the
status
a
constant
variable
and
then
that
points
to
the
the
hamstring
location
to
be
published,
and
then
the
hemshot
name,
of
course,
and
then
you
know
it's
mastery,
because
that's
what
the
hand
chart
is
called
in
that
location
and
then
this
is
just
the
hem
chart
version
that
we
specified
and
then
I
have
little.
F
You
know
part
that
figure
out
the
like
which
chart
version
we
want
to
do.
Basically,
you
know
because
we
got
this
mastery
image
version
is
basically
just
the
version
storing
the
contest
and
then,
if
it
is
not
latest,
then
we
know
it's
either
like
one
of
the
older
versions
called.
Let's
say:
we've
0.5.5,
something
like
that.
That's
and
now
we
pass
that
to
this
api
and
then
this
api.
We
put
that
version
of
hand
chart
down,
and
then
here
is
the
valued
override.
F
This
is
equivalent
to
to
to
what
we
had
in
the
manifest
that
overriding
the
hem
shots
in
the
well,
not
the
hem
charge
the
plane
manifest
in
the
file
system,
but
here
instead
we
don't
override
the
values,
but
we
do.
We
use
the
hem
the
existing
features
by
provided
by
him
by
basically
setting
the
override
value.
If
you
do
it
in
the
terminal,
basically
just
like
him
in
store
blah
blah
blah
and
then
that's
that's
set,
and
then
you
know
just
do
whatever
value
you
want
to
override.
F
F
So
that's
basically
what
this
little
you
know
field
is
about,
and
then
this
has
to
match
the
the
hem
charts
that
I
basically
also
change
in
the
pr,
because
I
not
here-
that's
I'm
trying
yeah
here
so
basically,
I
split
this
repository
into
two
fields.
One
is
called
about
repository
and
then
the
other
one
is
the
tag
you
know,
because
that's
the
image
tag.
So
in
this
code
I
basically
override
this
tag
into
whatever
version
that
the
user
is
using
for
the
measuring
ctl.
F
So
that's
what
that
is
about
for
the
start
and
then
it's
the
same
logic
for
system,
stop
and
update.
You
know,
instead
of
overriding
all
the
well
all
particular
values
in
the
manifest
we
just
pass
in
the
override
values
into
these
applied
time.
Chart
functions
that
we
had
in
the
mesh
kit
and
I
specified
the
actions
here.
If
you
know,
if
you
want
to
do,
install
or
upgrade
or
the
uninstallation
and
yeah,
that's
basically
what
I
have.
F
I
believe.
Oh
one
thing
to
mention
in
the
stop
command.
Is
that
let
me
do
the
compare
with
branch.
Oh
no
one
thing
I
wanted
to
bring
up
in
the
stop
is
that
so
hem
natively,
if
you
do
ham
uninstall,
there's
a
caveat
that
it
doesn't
delete
the
crs
and
the
crds
that
you
have
in
the
hem
chart.
So
in
order
to
do
a
clean
uninstallation
we
have
to
do.
You
know
invoke
the
well.
I
had
to
create
a
function.
You
know
to
delete
the
crs
and
then
delete
the
crds.
F
That's
what
this
is
about.
So
inbox
delicious
basically
calls
the
wrapper
function
here
that
deletes
the
cr
instance
that
we
have
in
the
cluster,
which
is
the
broker
and
then
about
the
mastery
broker.
That's
the
actual
instance
name,
and
then
the
mastery
mesh
sync
and
then
this
guy,
you
know
it
just
calls
the
function
called
delicio
that
actually
do
the
dirty
job
for
us.
F
So
this
is
just
like
a
wrapper
for
calling
this
function,
and
then
this
function
will
delete
whatever
resource
you
specify
what
you
asked
me
to
do.
You
know
and
then
the
same
for
the
crds.
I
have
like
a
wrapper
function
and
then
that
calls
the
other
function
that
do
the
actual
job
of
deleting
the
resources
that
we
specify
in
this
wrapper
function
and
then
yeah.
So
enough,
for
you
know
talking,
let's
do
the
quick
demo
first,
so
I
just
do
a
make
to
update
my
mastery
cto
in
my
local
machine.
F
So
yeah
so
from
the
users
per
step
perspective.
This
actually
works
the
same
as
before.
So
as
you
can
see
here,
you
know
we
have
all
the
we
have
mastery
and
then
all
the
adapters
installing
and
then
we
have
the
broker
coming
up,
and
then
we
have
a
mesh
thing
coming
up
as
well,
but
both
of
them
they
take.
They
will
need
a
little
bit
of
time
to
to
get
up
and
running.
F
Oh
one
thing
I
forgot
to
mention
is
that
so
previously
we
download
the
the
manifest
to
a
folder
called
address
that
call
a
tutor,
slash,
dot
mastery,
but
now
the
mesh
could
actually
download
that
to
this
slash
temp,
but
if
we
really
wanted
to
download
to
the
same
location,
I
can
easily
change
that
one
in
the
dimension
kit,
but
then
so
for
these
four
points
that
no
one
do
just
well
mentioned
in
my
pr.
I
think
the
only
thing
that's
not
resolved
is
the
second
point.
F
Sorry
it's
the
second
point,
because
we
don't
actually
override
values
in
either
start
or
update
with
the
utility
function
in
the
mask
kit,
but
we
did
that
previously
in
the
manifest.
I'm
not
that's
the
only
thing
that
I'm
not
really
sure
about,
like
I
don't
know
if
we
really
want
to
override
the
value
in
the
headshot
while.
B
We
can
so
darren.
I
don't
know
what
the
suggestion
is
on
the
first
one,
but
but
I'll
suggest.
Maybe
this
is
the
same
thing
that's
suggested,
but
that
yeah
we
we
that
we
would
not
use
the
slash
temp
for
anyone.
F
B
Yeah
yeah
hussein,
as
we
talk
about
that,
do
you
recall
if
the
bash
script
that
installs
measuring
it
is
using
a
slash,
temp
or
or
not.
G
I
Actually
so
two
things
from
my
side,
so
one
thing
that
we
were
also
doing
in
manifest
based
installation
was
that
we
were
selectively
passing
a
few
environmental
variables.
So
if
you
would
pass
an
environmental
variable,
called
mystery,
underscore
server
underscore
custom
host
or
something
I
don't
exactly
remember.
The
environmental
variable
that
particular
environmental
variable
will
be
passed
on
to
the
mystery
server
also.
So
it
would
be
if
it
was
at
least
being
appended
to
the
environment
of
the
container
of
machine
server.
I
So
are
we
and
I
think,
we'll
be
doing
that
with
overrides
over
it?
So
is
that
thing
being
covered,
because
I
couldn't
actually
see
it
in
the
code.
F
I
don't
think
so
is
that
in
okay?
Can
you
point
it
out
in
the
pr,
because
I
probably
missed
that
part
in
the
yeah
yeah?
I
actually
pointed
out.
I
I
Yeah-
and
second
is
so-
we
are
right
now
deleting
custom
resources
in
mystery,
and
I
think
probably
we
we
can't
move
this
into
meshkit
the
meshkits,
so
now
that
we
have
action
names.
So,
in
that
a
delete
case,
we
can
actually
probably
get
rid
of
the
custom
resources
also,
and
I
think
it
should
fix
some
of
the
characters
issue
also
because
one
of
the
issues
that
he
was
facing
was
that
the
custom
resources
were
not
getting
deleted
from
the
from
the
system,
and
that
is.
I
F
Yeah,
I
I
believe,
that's
what
happened.
You
know
in
the
master
branch
right
now,
because
you
know
that's
like
a
caveat
in
ham.
It
does
not
manage
the
crds.
Well
enough,
you
can
actually
read
about
it
in
the
I
don't
know.
I
thought
it
install
carey
at
if
I
spelled
it
correctly
yeah.
So
this
one.
F
If
we
go
to
this
page
and
then
it's
curious
that
yes,
some
caveats
an
explanation,
so
it
actually
talks
about
there's
no
support
at
a
time
for
upgrading
or
deleting
crds
using
him.
So,
even
though
you
know
you
are
trying
to
delete
the
entire
hem
charge
and
the
in
your
hammer
due
has
you
have
the
cr
and
crds
this
command?
You
have
uninstall,
it
does
not
delete
that,
for
you
is
yeah.
F
F
If
I
do
the
system
stop
hit.
Yes,.
F
First
of
all,
you
can
see
all
the
parts
are
getting
deleted
and
then,
if
I
go
to
crd
or
crds
are
gone
because
you
know,
I
have
those
two
functions
to
delete
the
cr
instance
and
the
crds
and
then
namespace,
I'm
not
sure
if
it
deletes
the
namespace
or
not,
it
does
not
delete
the
namespace,
but
I'm
not
sure
you
know
how
big
the
deal
is.
You
could
not
delete
the
name
space,
but
it
doesn't
really
affect
the
functionality
because
you
can
do
the
after
you
stop
that
you
can
do
the
start
again.
I
F
And
also
you
know
to
the
point
of
item
protein
is
so
now.
I
have
already
run
that
once
and
then
I
want
to
you
know,
run
it
twice.
I
can
do
that.
It's
not
going
to
report
any
errors
that
ham
reports,
because
if
you,
if
you
do
the
same
installation
twice
and
will
complain
that
you
already
have
that
release
installed
in
the
cluster,
but
since
I
have
that
changes
it
it
now,
it
does
not
do
the
well,
it
doesn't
show
the
errors.
F
F
I
think
that's
all
I
have
to
cover.
Oh
actually,
I
think
I
wanted
to
cover
one
thing:
I'm
not
sure
if
it
is
a
bug
or
not
so
or
is
specific
to
my
environment,
let
me
go
to
the
so
now
I
am
in
the
you
know
the
ui.
If
I
want
to
install
something
right,
the
ping
is
successfully
to
the
adapter
link
id
and
then,
if
I
want
to
install
that
it's
not
showing
up
now,
because
what
I
saw
was
the
I
was
seeing
this.
F
F
So
I
saw
during
my
testing
like
last
night,
so
now
it's
working
fine.
So
when
I
wanted
to
install
something,
let's
say
the
istio
right
and
then
it
paints
a
lot
of
paints
to
the
adapter
is
good.
But
then
I
wanted
to
install
that
the
ui
would
show
an
arrow
of
you
know,
anal
file
to
the
server
and
then
that
specific
adapter
will
crash
with
an
error
of
panic.
F
I
Yeah,
actually,
this
comes
in
probably
either
you
had
vpn
issue
or
so.
Okay
by
default
adapters,
try
to
install
the
latest
one
in
order
to
get
the
latest
version.
They
actually
query
the
query
use
githubia.
So
what
hap?
What
does
happen,
sometimes
with
with
mishops
version,
one
which
we
are
using
in
the
ui,
that
those
queries
fail,
and
that
is
why
you
won't
have
anything
in
the
area
and
it
will
try
to
find
that
the
first
one
and
actually
yeah.
F
I
So,
ideally
so
the
panic
we
know
the
reason
of
the
panic,
but
ideally
the
adapters
try
to
gracefully
recover
from
the
panic
and
they
do
I'm
not
sure
why,
this
time
it
crashed
and
failed,
it
probably
didn't
restart
either
for
you
but
yeah.
We.
So
one
of
the
things
that
we
should
do
immediately
probably
is
to
first
check
that
is
the
length
like
the
basic
checks,
but
even
after
that,
actually
we
have
in
all
of
the
raptors.
We
have
traceable
recovery
from
the
conics
by
default,
so.
F
F
B
B
F
F
B
This
is
great:
it's
a
bad
joke
on
my
part,
actually
to
be
serious
for
a
moment.
It
actually
just
hearing
you
go
through.
It
inspires
confidence
that
that
you're
wrapping
your
mind
around
you
know
that
you're
being
diligent
with
the
variety
of
aspects
of
concerns
of
like
you
know
what
happens
upon,
delete
what
happens
upon
update
what
happens
upon
installation,
as
as
I
was
going
through
and
experimenting
with
other
adapters.
Here's
because
you
were
using
this
system
enough.
B
My
github
api
rate
limit
occurred,
yeah,
hey,
there's
two
things
that
are
kind
of
an
issue
there
one
is.
We
should
probably
design
a
better
system
that,
like
doesn't
use
that
api
as
heavily
there's
a.
I
think,
a
design
spec
coming
forth
for
kind
of
a
for
caching
there,
so
that
we're
not
acquiring
so
much
and
then
two
understanding
that
whether
it
was
at
a
rate
limit
or
just
network
connectivity
yeah,
the
adapter
shouldn't
fall
over.
I
think
part
of
what
ukarsh
was
saying.
B
Is
that
there's
been
there's
a
number
of
components
in
measuring
that
have
undergone
a
re-architecture
there?
More
or
less
the
components
are
in
a
v2
architecture.
If
you
will
and
the
nsm
adapter
and
the
cpx
adapter,
I
believe,
are
the
well
in
the
octarine
adapter,
these
three,
the
octarine
adapter
is
essentially
going
away.
B
Will
would
that
one
will
be
sort
of
converted
into
the
vmware
tanzu
service
mesh
octarine
was
acquired
by
vmware,
and
so,
but
so
these
last
two
standing
adapters
have
yet
to
be
at
least
near,
as
I
recall
that
they
don't
leverage
mesh
kit
yet
and
so
they're
missing
some
of
the
more
robust
yeah.
B
Some
some
feature
I
mean
they
might
use
mesh
kit
for
for
certain
aspects,
but
but
they
really
they're
not
using
patterns,
pattern,
components
and,
and
we've
invested
less
there
because
in
part
those
two
adapters
were
created
by
those
surface
mesh
projects.
I
just
reached
out
to
citrix
folks
a
week
ago
trying
to
get
a
sense
of
where
they're
at
with
their
service
mesh
and
how
we
should
continue
to
sustain
that
adapter
and
so
as
a
side
note
but
yeah.
This
is
great.
B
I'm
kind
of
surprised
like
that,
like
understanding,
helm
itself
went
through
a
v2
architecture,
kind
of
get,
you
know,
displacing
tiller
and
getting
rid
of
tiller
and
coming
to
a
better
architecture
that,
even
now
it
still
seems
to
smell,
like
poop,
to
me
like
how
that's
surprising
that
it
doesn't
take
care
of
some
of
these
of
things
like
custom
resources
or
deleting
of
namespaces
or
that
type.
F
Of
thing
well,
deleting
the
namespace
is
not
it's
probably
not
a
job
of
the
ham.
You
know
the
only
caveats
that
they
have
is
the
crds,
because
I
don't
know
why
they
have
been
having
trouble
in
managing
their
lifecycle
raciality.
I
remember
that
they
used
to
have
that.
You
know
the
installation
of
crd
and
then
you
know
the
website
says
after
some
heavy
discussion
in
the
community,
they
decided
not
to
touch
that
in
the
uninstallation.
F
So
I
do
not
know
what,
but
that
is
probably
information.
Implementation
specific
to
the
project.
B
F
B
Cool,
I
I'd
say
that,
guy
to
summarize
that
real
quick
sounded
like
we
were
gonna,
it
sounded
like
we
might
press
forth
with
that.
Breaking
change,
understanding
that
yep
sometimes
to
progress.
That's
you
know
necessary
to
improve
that.
We
would
kind
of
control
the
rollout
of
the
mesh
kit.
That
would
be
a
known
item
that,
as
we
bump
version
to
the
new
library
of
meshkit,
that
there
would
be
an
update
needed
in
each
of
the
adapters,
but
so
so
any
so
with
respect
to
v
zero,
six,
zero
and
shift
to
pound.
B
F
This
does
not
address
that
because
I
think
that
is
a
small
issue
that
lives
in
the
hem
chart
for
the
engineers.
So
I
do
yeah.
I
was
planning
to
take
care
of
that
after
you
know
all
these
measuring
ctl
changes,
but
that's
I
don't
expect
that
to
take
up
a
lot
of
my
time
to
fix
that
small
bug,
because
it's
just
a
permission
thing,
but
you
know
to
clarify
that
these
prs
well,
these
pr's
just
don't
address
that
servicer
confirmation
issue.
E
A
All
right,
thanks
darren,
I
guess
we'll
leave
some
comments
on
the
review
after
going
through
it
in
the
leave
reviews
in
the
pr
after
going
through
it
once
again,
we
are
out
of
time,
but
we
still
have
a
couple
of
topics
to
discuss.
So
if
you
have
to
be
anywhere
else,
feel
free
to
drop
out
barika.
Would
you
like
to
discuss
about
your
topic
and
your
open
pr.
H
Yeah
sure
this
should
be
a
quick
discussion
actually
so
I'll
share.
My
screen.
H
Yeah,
so
this
is
the
thing
that
I
want
to
discuss
about
so
so
this
is
mainly
on
the
front-end
side.
H
So
one
of
the
behavior
that
we
want
in
the
mesh
map
is
so,
if,
if
you,
if
you
are
aware
that
we
have
like
something
like
extension
in
measuring
and
when
we
click
so
if
you
are,
if
you're
enrolled
to
the
alpha
program,
you'll
have
a
new
addition,
no
tabs
on
your
left
side
of
the
major
ui.
H
So
here,
if
you
click
on
it,
it
won't
actually
give
you
like
a
collapse
thing,
because
so
we
want
it.
We
we
desire
to
be
collapsible,
because
in
this
case,
because
the
extra
gene
itself
is
a
is
like
a
canvas
and
we
want
to
have
like
we
want
to,
let
you
use
the
whole
screen.
H
We
want
to
use
the
whole
screen
and
in
this
case
we
want.
We
want
the
sidebar
to
be
collapsed
yeah.
So
this
is
just
a
simple
implementation,
but
I
think
so
in
this
pr.
I
I
did
the
implementation
on
the
front-end
side,
so
it
will,
if
you
can
see
on
the
record
itself.
H
B
In
my
mind,
a
plug-in
that
loads
and
measures
should,
as
it's
defining
the
menu
items
that
it
wants
to
have
shown
in
that
extension,
point
that
there
should
be
a
consideration
for
behavior
of
of
what
happened,
what
the
desired
behavior
that
the
plugin
has
for
an
item.
That's
clicked,
and
so
there
should
just
be
an
attribute
associated
to
each
menu
item.
That
is
optionally
supplied
by
a
plugin
that
in
this
case,
if
they
it
would
be
if
they
wanted
to
collapse,
that
they
would
specify
that.
B
H
I
I
think
so,
yes,
that's
we
should
we
should
let
the
providers
decide
what
the
behavior
should
be
and
there
could
be
two
ways
we
can
achieve
this
one
is,
as
I
was
mentioning
the
review.
Also,
that
is
one
of
the
way
could
be
that
we
define
some
some
constants
like
these
are
the
options
that
you
can't
select
from
like
collapsing,
the
drawer
or
something
because
that
would
be
easier
to
implement
on
the
measure
ui
side
measuring.
I
I
can
declare
pre-define
some
of
the
actions
and
up
and
in
the
capabilities
we
can
select.
We
can
mention
that
this
is
the
behavior
of
this
particular
extension
point.
So
now
the
provider
gets
to
choose
not
what
extension
gets
loaded
in
what
way,
because
there
could
be
multiple
extensions.
So
this
is
one
of
the
way
in
the
second
more
difficult,
but
far
more
far,
more
flexible
way
is
that
these
so
right
now
we
do
code
injections
to
load
these
plugins.
I
We
can
also
do
code
injection
for
for
the
sidebar
icon
and
the
like
in
in
the
entire
upper
entire
entry.
We
can
do
code
injection
for
that
also,
that
is
when
you
click
the
entire
code
that
gets
executed
that
comes
from
the
provider.
That
would
be.
That
would
be
difficult
to
achieve
not
on
the
provider
side,
but
on
the
ui
side,
because
we
will
have
to
again
create
a
new
sandbox
if
you
would
and
yeah,
and
then
we
will
have
to
load
the
stuff
and
and
execute
it.
I
So
probably
we
can,
because
because
the
requirement
is
extremely
simple,
at
least
at
the
moment
we
can
go
with
the
first
way
and
that
is
the
dui
defined
predefined
behaviors
like
collapse,
sidebar
or
maybe
zoom
something
or
something
a
predefined
set
of
actions,
and
if
the
provider
choose
that
this
is
the
behavior,
when
you
get
to
get
click
on
the.
Let
me
click
on
the
sidebar.
H
Yeah,
so
just
to
summarize,
so
I
think
I
think
in
the
inside
inside
the
providers,
we
should
add
a
new
additional
field
called
callbacks
or
unclick,
something
and
then
inside
the
database.
You
can
just
like
provide
an
enumeration
or
or
like
an
index
numbers.
For
example,
zero
on
the
major
ui
docs
would
declare
that
this.
H
H
B
What
do
you
think
lee
yeah?
To
be
honest,
to
be
forthright
to
be
candid,
like
I
didn't,
listen
to
it,
dancing
like
all
in
part,
because
in
part,
because
I'm
an
ignoramus
on
on
some
of
the
specific
like
we
are
successful
so
long
as
so
long
as
in
an
extensive
so
so
to
to
define
an
extension
point.
We
have.
We
have
a
system
wherein
we
afford
for
an
integrator
to
modify
to
augment
the
overarching
system
behavior.
B
We
shouldn't
let
them
specify
weird
things
we
shouldn't,
let
them
both
try
to
collapse
and
and
uncla
like
expand
at
the
same
time
like
like
that,
these
are
principles
and
properties
of
what
any
extension
point
should
be.
They
should
be
agnostic
to
any
specific.
B
Any
explicit
or
specific
plugin,
like
this,
like
no
plug-in,
should
be
called
out
in
the
mesri
server
itself.
So
so
it
should
be
dynamic,
like
the
the
part
of
the
plug-in
is
like
it
is
a
communication
between
when
you
plug
something
in
it's
like
dynamic
communication
between
my
system
and
this
via
usb
c,
that
it's
saying
I'm
I'm
this
type
of
microphone.
You
should
represent
this
type
of
volume
or
whatever,
like
so
anyway,
but
the
same
thing
with
the
plug-ins
is
like
oh
dynamically.
B
There
should
be
an
exchange,
and
so
I
think
that
yeah
it
it
can
make
a
lot
of
sense
that
there
is
some
of
this
that
gets
tracked
in
go
as
a
matter
of
fact.
I
don't
know
that
it'd
be
kind
of
you
probably
have
to
go
out
of
your
way
to
to
avoid
I'm
having
consideration
for
this
in
go.
We
do
want
to
validate
like
to
the
extent
that
a
plug-in
misspells,
the
word
collapse
or
whatever
indicator
that
we're
using
like.
B
We
want
to
capture
that
and
not
just
have
not
let
some
one
else
break
measure
ui
that
the
tab
index-
I
I
heard
tab
index,
that's
about
all
that
I'd
heard,
but
but
yeah
that
there
isn't,
for
the
extension
point
to
be
robust.
B
There's
we
don't
have
to
go
too
far
here,
but
but
we
do
want
to
consider
that
a
given
plug-in
and
a
given
remote
provider
may
want
to
insert
you
know
a
few
different
menu
items
as
they
do.
They
might
also
have
a
child
level.
They
might
have
a
menu
with
like
sub
menu
items
that
that,
if
we're
considering
for
that
that
that's
a
bit
more
of
a
robust
plugin,
we
don't
have
to
spend
as
a
project.
B
We
don't
have
to
spend
a
ton
of
time
there
in
advance
of
in
you
know,
solution
providers
bringing
that
use
case.
B
But
if
it's
simple
enough
to
consider
for
now,
you
know
great
if
we
can
also
like
we
would
want
to
do
things
like
auto
sort
by
alphabetically,
but
also
potentially
allow
them
to
specify
that
tab
index
or
to
specify
the
not
literally
the
tab
index,
but
the
sequence.
B
You
know
stuff
stuff
like
that.
We
don't
have
to
go
banana.
We
don't
have
to
go
nuts
on,
you
know,
making
it
the
most
polished
api
that
you've
ever
seen
like
it
does
all
the
things
that
anybody
could
ever
want
to
do,
letting
them
colorize
and
extend.
You
know,
change
the
colors
of
the
menu
and
all
that
like,
but
anyway,
point
is
like.
I
think
I
think
the
mark
of
success
here
is
so
long
as
there
isn't
specific
consideration
in
measuring
server.
That
really
only
works
for
that
plug-in.
B
It's
good
there's
one
other,
I
I
know
we're
well
over
time.
Those
that
are
here
don't
have
anything
better
to
do
so.
Rude
rocks,
mostly
we've
been
talking
about
release,
drafter
and
auto
labeling,
and
I
think
a
lot
of
the
reason
that
we've
talked
about
as
much
as
we
have
is
mostly
like
my
distraction
on
not
paying
attention
to
the
conversation
at
hand.
I
think,
and
a
bit
of
confusion
about
access
and
personal
access
tokens
and
not
knowing
history
around
conflict
with
auto
labeler
bots
and
welcome
bots
and
stuff.
B
Did
you
send
some
messages?
I
haven't
read
them.
Do
we?
Are
you
ready
for
experimentation
like
near
as
I
could
tell
it?
Didn't
it
like
somewhat
related
but
orthogonal.
The
problem
that
we're
experiencing
right
now
is
that
one
of
our
workflows
is
failing
because
we're
referencing
a
github
token
in
a
deprecated
way.
D
D
D
We
need
to
probably
use
some
trick
around
creating
a
different
workflow
using
the
same
released
after
action,
which
doesn't
really
need
that
token,
because
we
don't
really
need
a
patent
pull
request
so
yeah.
Basically,
the
thing
around
access
token
is
what
is
causing
the
failure
and
pull
request
and
which
is
not
at
all
affecting
released
after
release
drafts
right
now.
I
just
think
they
are
up
to
date
with
the
latest
prs.
B
Yep
yeah
we
might
want
to.
We
might
want
to
take
a
risk
and
maybe
under
the
assumption
that
the
conflict
between
the
welcome
bot
and
the
label
or
bot
is
maybe
gone.
B
I
don't
know
and
then
or
or
to
put
a
pin
in
this
and
just
like
or
decide
which
one
is
more
important
like
categorized
and
clean
release,
no
or
even
cleaner
release,
notes
or
welcoming
of
new
users.
D
D
B
Good
deal
so
yeah
as
the
maintainer
there
you
have
please
just
make
the
executive
decision
or
I
know
you're
well
informed
you're
as
informed
as
I
am
on
the
problem
now
and
actually,
I
think,
better
informed
and
so
yeah
do
what
you.
B
Okay,
nice
thanks
and
I
think,
that's
a
wrap.
B
By
the
way,
if
you
guys
don't
know
vedant
on
the
call
just
got
his
his
bat,
his
measuring
badge
today
or
will
be
very
soon
so
congrats
if
it
on
pretty
cool,
tell
tell
us
the
name
of
the
tattoo
parlor
that
you're
going
to
send
a
reimbursement.