►
From YouTube: Layer5 Community Meeting (Aug 27th, 2021)
Description
Layer5 Community Meeting - Aug 27th, 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
All
right
yeah,
so
let's
get
started
so
we
have
had
a
lot
of
members
joining
in
this
week
and
I
think
some
of
them
so
this
community,
we
have
a
tradition
where
newcomers
get
to
introduce
themselves
on
the
call
so
that
we
get
to
know
each
other
better
and
everyone
in
the
community
can
get
to
know
you.
A
So
look
rahul
biology.
Would
you
like
to
start.
B
Hello:
everyone,
I'm
rahul,
I'm
a
student
from
india,
I'm
in
my
third
year
pursuing
a
degree
in
computer
science.
My
goals
are
to
learn
more
about
layer
5
and
to
contribute
to
the
community
for
now
and
then
eventually
appear
for
an
internship.
B
A
Purvi,
I
think
you
are
new
here.
Would
you
like
to
go
next.
D
D
C
Jared
in
the
house,
good
nice,
what
I've
got,
I
don't
have
my
voice,
so
I
sound
like
well,
I
sound
like
a
grumpy
old
man.
I
guess,
but
but
anyways
excited
that
jared's
here
excited
that
rahul
is
here
as
well.
It's
nice
to
have!
Well,
we
don't
have
a
record
number
of
attendees
today,
but
bring
on
what.
What
do
we
have?
Was
it
36?
Is
that?
C
Where
we're
at
so
we're
invite
your
friends,
I
guess
we
it's
been
a
while,
since
we
broke
the
record,
but
nice
good
I'll
try
to
be
quiet
on
the
call.
So
I
don't
sound
like
it's
a
nut
so
yeah,
so
I
don't
spread
the
grump
around.
I
guess.
A
Claire,
charles,
I
don't
think
you
have
introduced
yourself.
Would
you
like
to
introduce
now.
A
Let's
go
all
right
if
anyone
else
is
new
on
the
call
and
didn't
get
the
chance
to
introduce.
Please
do
that
on
the
chat.
Moving
on
to
the
announcements,
we
have
a
couple
of
graduating
interns
who
just
finished
their
gsoc
internship
with
cmcf
and
at
layer
5..
A
I
bet
like
most
of
you
guys
are
aware
of
the
work
they
have
been
doing
and
pious
has
written
a
blog
on
his
experience.
A
A
So
an
upcoming
one
is
the
lfx.
I
guess
we
will
announce
the
main
piece
next
week.
A
So
if
you
want
to
do
an
internship
or
if
you
are
thinking
or
something
like
that,
please
check
out
the
faqs
at
this
page.
So
congratulations
guys.
A
And
another
topic
is
that
we
are
kubecon
in
north
america
is
approaching
so
layer.
5's
are
two
projects.
Mesherie
and
snp
are
in
rcmcs
sandbox
projects,
so
we
will
be.
A
We
will
be
involving
quite
a
bit
there
so
we'll
we
will
have
what
do
they
call
it.
A
Yeah
I
have
project
office
hours
and
all
those
stuff
at
kubecon.
So
this
is
a
call
for
volunteers
to
do
some
social
media
engagement
and
even
speak
at
kubecon,
so
speak
at
our
virtual
booth
at
kubecon
dollar
stuff
leave.
Would
you
like
to
add
anything.
C
Yeah
yeah,
if
I
can
I'll
see
how
long
my
voice
holds
out,
but
there
are
a
number
of
you
who've
been
here
for
a
really
long
time
and
there's
some
of
you
who
have
just
showed
up
today,
maybe
and
kind
of
some
of
that
matters.
How
long
you've
been
here,
and
some
of
that
doesn't
really
matter.
I'm
not
even
sure
why
I
mentioned
it
just
now.
C
I
mentioned
it
to
say
that
some
of
you
have
been
here
before
the
community
had
produced
two
cncf
projects
with
a
third
project,
get
nighthawk
building
out
fairly
quickly.
C
Well,
you
know
at
this
point
like
we
have
more
there's
more
to
say
and
to
show
people
about
what
has
been
done
here
than
certainly
I
have
time
for
I
I've
I'll
say
that
I've
said
this
any
number
of
times
and
I'll
say
this
again,
and
that
is
that
there
are.
It
is
quite
frequently
that
we're
receiving
requests
to
talk
to
tell
people
about
service
messages
or
tell
people
about
cloud-native
application.
C
Networking
tell
them
about
what
we
do
here
and
it's
my
for
my
part
always
been
my
desire
that
I
get
to
shut.
My
yapper
and
other
people
are
getting
up
on
stage
and
talking.
That's
why
I
always
well
part
of
the
reason
why
I
always
ask
one
of
you
to
join
me
on
stage,
because
it's
not
about
from
again
I'm
being
very
selfish
in
this
line
of
thought.
But
like
it's
not
about
me,
it's
I
don't
need
the
stage
time.
I
don't.
C
Nobody
needs
to
listen
to
me
talk
anymore,
and
so
we've
got
two
projects,
which
means
we're
gonna
perpetually
have
spots
on
stage
or
you
know,
and
some
of
that
is
to
have
a
virtual
booth
to
have
recorded
like
materials
about
what
you've
done
about
how
smi
conformance
works.
For
example,
the
fact
that
there
are
a
couple
of
ge,
github
actions
now
the
fact
that
mesh
re
does
that
the
fact
that
s
p
produced
these
results.
C
The
fact
that
get
nighthawk
is
running,
optimization
routines
or
whatever
is
going
on
so
we're
going
to
have
a
couple
of
virtual
booths.
Those
are
going
to
need
to
be
staffed
with
folks
that
like
to
interact
with
other
people,
it's
a
virtual
booth,
so
we're
gonna
have.
C
Office
hours,
usually,
where
you
you
know,
you're
just
you're,
just
welcoming
people
who
are
stopping
by
and
telling
them
about
the
project,
answering
questions
for
both
service
mesh
performance
and
for
measuring,
then
we're
also,
and
then
you
know,
we
need
to
tell
people
about
what
you
all
are
doing
and
slap
your
name
on
the
front
of
things
like
push
just
did
he
just
did
like.
It
was
definitely
more
than
three
months
of
work
that
he
did
for
g
suck.
I
mean
he.
C
He
got
a
running
start
to
that
thing,
which
is
exactly
how
people
land
on
internships
around
here
as
they
show
up.
They
just
say
they
help
they
just
say.
Yes,
let
me
add
it
I'll
help.
Let
me
do
let
me
engage,
they
find
a
niche,
they
find
a
foothold
on
something
and
they're
encouraged
to
move
as
briskly
as
they
can
and
that's
what
happened
in
push's
case
and
the
point
of
calling
him
out
is
to
say
that
he
went
and
wrote
up.
C
C
5
io
is
standing
which
much
like
open
source
contributions,
though
they
stand
for
all
time
right
until
the
repo
is
deleted
or
until
the
site
goes
down
which
which,
as
long
as
I'm
breathing
probably
isn't
gonna
happen,
you
know
I'll
have
to
hand
it
off
to
all
of
you
and
you
can
carry
it
forward.
C
So
the
point
is
is
like
it's
about
push.
I
mean
to
me.
This
is
about
push
not
really
about
the
stuff
that
we're
doing
and
that's
the
same
thing
about
putting
you
any
of
you
on
stage
or
or
in
the
virtual
booth
or
in
the
office
hours
or
if
you
want
to
write
an
email
that
we
need
to
send
out
to
the
3
000
people
who
are
signed
up
to
the
newsletter
we've
got
three
twitter
handles
got
again.
C
No
one
wants
to
hear
my
corny
jokes
on
twitter,
and
so
there's
lots
of
engagement
to
do
these
are
your
projects
come
represent,
bring
good
english
but
come
represent
yeah.
So
if
those,
if
any
of
that
stuff
that
I
I
just
mentioned,
is
of
interest,
you
know
ping,
somebody
ping,
a
community
manager
or
me
or
vendor
or
any
of
the
maintainers
in
slack
and
there's
something
to
do.
You
know
this
kubecon's
a
super
busy
time.
Dockercon
is
usually
a
busy
time
for
us
as
well.
C
Yeah,
it's
not
just
that!
We
also
chair
the
cncf's
service
mesh
working
group.
We
also
chair
the
cncf
technical
advisory
group
for
networking
we
have.
Some
of
us
are
cncf
ambassadors,
like
there's
a
lot
of
opportunity
afforded
to
us
in
collectively
in
that
way,
to
publish
blogs
on
the
cncf.io
to
have
webinars.
We
just
had
one
you
can
have
as
many
of
those
as
we
like
yeah,
there's
just
ample
opportunity.
C
C
Even
if
you've
barely
done
anything,
you
don't
contribute
code
that
doesn't
matter
at
all.
It's
not
about
code
yeah.
I
just
anyway,
I'm
gonna
I'm
about
to
get
on
a
soapbox.
Now
I
won't
so
so
with
that.
Just
it's
a
good
opportunity
to
get
involved.
A
Yep,
thank
you
lee,
so
please
get
in
touch
with
the
community
managers
interested
in
this.
F
Yep
all
right,
can
you
see
my
screen.
F
Yeah,
it
was
worth
it,
it
was
worth
it
yeah
sure.
So,
basically,
on
the
first
topic,
so
I
have
made
something
called
the
name
paper
so
basically
earlier
in
our
in
our
dashboard
previously
was
the
name,
the
name
of
the
serviceman
service
mesh
pots.
It
was
in
the
original
developer,
friendly
kind
of
name
right,
but
what
what
I'm
doing
here
is
basically
map
this
name.
F
So
there
is
basically
some
hardcoded
values,
then
send
back
a
more
friendly
kind
of
like
more
general
user
name
in
some
sense,
so
so
yeah.
So
this
is
only
only
for
for
the
starters.
Maybe
so
I
saw
that
there
there
might
be
a
lot
of
space
for
contributions
here,
because
we
cannot
because
yeah,
maybe
you
can-
you
can
have
your
own
mesh
and
then
start
your
own
mess
and
then
saw
that
the
name
has
not
been
mapped
in
our
name
neighbor.
So
you
can
see
on
the
source
code
basically.
F
F
F
And
then
so
there
is
the
first
topic
and
then
the
second
topic
will
be
the
data
planes.
So
earlier
in
our
dashboard,
we
only
show
the
user
control
plains
and
then
we
didn't
show
the
user,
the
the
data
planes
so
in
the
service
mod
service
mesh.
Basically
there
there
are
two
components:
there
are
the
control
planes
and
the
data
points
so
database
is
managing
how
the
how
the
communication
between
each
box
corresponds
to
basically
and
then
so
here.
F
What
I'm
doing
is
that,
so
all
the
data
is
actually
there
already
inside
our
messenger,
and
what
we
are
trying
to
do
here
is
this.
We
just
resurface
some
of
the
data
that
is
inside
the
message
into
the
mystery
ui
yeah.
So
what
I
do
here
is
we
just
add
new
column,
called
data
planes
and
then
show
the
user
that
what
is
the
number
of
database
data
planes
that
is
running
on
the
on
the
cloud,
basically
yeah.
So,
for
example,
there
is
still
pilot.
F
There
is
a
one
data
planes
and
then,
if
you
hover
on
the
data
planes
value,
you
can
see
the
details
of
the
data
planes.
So
here
the
name
of
the
data
planes
is
still
proxy
and
the
status
is
ready.
F
So
you
can
see
that
the
it
is
using
a
how
many
amount
of
resources
inside
those
data
planes
and
then,
if
the
data
planes
are
not
ready,
it
will
show
and
occur
accordingly
an
error
message
something
so
you
can
see
that
there
are
new,
an
additional
columns
which
is
called
the
last
state.
It
is
terminated
and
then
the
error
is
2.5
and
the
last
thing
that
it
is
running
is
at
sometimes
finish.
D
I
had
a
question
about
the
the
first
thing
you
demoed,
so
you
said,
you're
hard
coding,
the
mapping
of
the
pod
names.
D
Doesn't
this
add
a
layer
of
of
of
overhead
to
maintenance?
Should
these
names
ever
change,
would
it
be
possible
to
parse
out
the
names
define.
F
Yeah,
so
thank
you
jared
for
the
feedback,
so
I
I
was
thinking
about
that.
But
the
thing
the
thing
is
we
yeah.
There
is
actually
a
good
suggestion,
and
maybe
for
that
we
can.
We
can
do
that
in
the
in
the
default
case.
So,
for
example,
we
for
example,
we
have
the
mapping
already
and
then
we
can.
We
can
definitely
do
that
in
the
default
case.
F
For
for
why
we
do
the
mapping,
I
guess
we
just
want
to
like.
F
Just
have
a
sense
of
reliability,
I
think
in
terms
of
like
what
is
it?
What
is
this
exactly
and
then
we
can
just
hard
code
like
what
is
that
yeah
accordingly,
but
definitely
this
is
a
good
suggestion.
G
C
Yeah,
but
it's
it's
worth
it's
worth
noting
something
like
jared
points
out,
a
really
a
great
point,
and
it's
one
of
the
perpetual
challenges
of
management
software
that
well,
you
know,
how
do
you
provide
a
highly
curated
like
service
mesh,
specific
experience
and
do
so
across
like
all
these
meshes
without
having
much
of
fragile
code
a
bunch
of
hard-coded
things,
there's
a
balance
in
there
somewhere.
C
One
of
the
things
that's
been
done
is
inside
of
the
service
mesh
performance
specification
is
there's
a
proto
that
defines
that
sort
of
sets.
The
well
standard
is
the
wrong
word,
but
it
sets
a
common
convention
for
what
names
to
use
for
each
service
mesh
such
that
then
any
component
in
measuring
or
elsewhere.
That's
using
that
would
have
a
common.
You
know
common
moniker,
a
common
way
of
identifying
the
name
for
that
surface
mesh.
In
this
case.
This
is
about
pods
and
actually
like.
C
If
we
can
crack
this
nut,
it
would
be
fantastic
like
if
you,
if
you
look
at
and
I'll
just
kind
of
play
it
through
out
loud
as
I
were,
to
potentially
try
to
solve
this.
So
if
if,
if
I'm
looking
at
pod
names
and
and
they're
broken
down
like
this,
it's
like-
oh
hey-
you
know
these
first
few
or
that
first
one
with
istio
galley.
It's
like
look!
Why
don't
we
capitalize
the
first
letter
on
things
there
are.
C
You
know,
there's
libraries
for
that:
okay,
good,
why
don't
we
split
it
on
hyphen,
great
there's,
a
you
know
like
that
seems
pretty
possible
okay
and
then
there's
there's
a
collection
that
fall
into
that
category.
Good,
okay!
Well,
then,
that
just
leaves
us
with
ooh
ingress
gateway.
Egress
gateway
is
like
if
the
g
had
been
capitalized.
That
would
be
a
way
of
you
could
identify
that
programmatically
and
say
insert
white
space
there,
but
without
it
like
okay,
then
how
are
you
gonna?
C
Well,
okay,
maybe
you
can
reduce
the
number
of
hard-coded
things
to
just.
If
you
see
the
word
gateway
split
it
out
or
if
you
see
the
word
ingress
break
it
apart,
because
it's
considered
its
own
word,
it's
like
okay,
all
right
and
then,
and
you
might
be
able
to
and
it
might,
the
the
question
becomes
whether
which
one's
more
fragile
is
kind
of
what
it
ends
up
coming
down
to
and
yeah.
You
end
up
in
this
particular
case
having
yeah
an
extra
well.
This
isn't.
C
C
Yeah
is
the
is
the
thought
process
is
the
like
where
where
yeah
anyway,
I'm
just,
I
don't
have
an
answer.
I
was
just
kind
of
playing
it
out
loud,
it's
like
because,
because
if
people
can
identify-
and
I'm
not
I'm
not
directing
this
at
jared
michael,
you
brought
it
up,
so
you
have
to
have
a
solution.
That's
not
that,
but
yeah
yeah
yeah!
This
is
a
great
thing
to
call
out
this.
This
makes
some
of
us
uncomfortable.
C
I
think
it
made
barack
uncomfortable
he's
like
oh
okay,
well,
and
so
I
don't
know
like
is,
it
is
another
way
of
making
like
because
what?
What?
C
If
what
happens
here
is
if
you're
going
to
display
the
content
of
a
service,
mesh,
you're,
saying
here's,
your
control,
plane,
components
and
somebody's
looking-
and
you
know
it's
one-
it's
istio
1.12
or
something
and
they're
looking
they're
like
yeah,
I'm
pretty
sure
I've
got
a
healthy
issue
or
deployment
they're
like
yeah,
but
where's
my
where's,
my
pilot
where's,
my
pilot,
like
you
know
it's
pretty
important
like
I
don't
it's
not
here
in
this
widget,
it's
not
in
the
dashboard,
but
I
see
it
deployed
in
kubernetes
like
what
oh,
my
shari's
got
an
issue
like
what
anyway,
and
it
could
be
like.
C
Oh
because
the
language
here
is
so
barrack.
That's
one
thing
to
confirm
is
like
if
the
language
here
doesn't
match,
like
let's
say
that
pilot
was
renamed
to
captain
or
whatever.
C
F
Yeah
for
that,
if
we
didn't
cover
the
case,
then
it
will
return
to
the
original
name.
C
Okay,
so
there's
a
catch-all
that,
like
okay,
so
at
least
okay,
that's
good.
That
makes
it
slightly
less
fragile
right
because,
because
what
would
be
really
a
really
wrong?
Turn
is
like
in
trying
to
provide
the
right
ux
and
if
you
didn't,
if
you
didn't,
have
the
catch-all,
then
you'd
be
in
a
bad
spot,
because
yeah
so
anyway.
Just
thinking
that
just
thinking
that
through
because
it's
this
is
really.
I
know
we
just
spent
like
10
minutes
on
like
a
white
space
or
whatever.
C
But
you
know
it's
it's
what
makes
this
is
like
why
someone
would
use
a
tool
like
mesherie,
because
it
feels
good
and
it
feels
nicer
than
using
istio,
ctl
or
or
whatever
the
alternative
is,
and
this
is
in
part,
what
makes
it
feel
good.
So.
H
Yeah
sure
so
it
is
about
service
mesh,
interactive
labs,
so
we
have
interactive
labs
which
have
side-by-side
instructions
which
guides
user
on
how
we
can
use
labs
and
so
right
now
we
are
improving
the
existing
labs
and
creating
the
new
ones,
and
many
of
you
like
makna,
surya,
shankar,
aditya
and
shantan,
had
I
mean,
are
contributing
towards
this
one,
so
they
are
helping
in
improving
the
existing
labs
and
creating
the
new
ones.
So
so
there
were
some
questions
regarding
like
how
we
can
create
labs.
H
So
there
is
a
source
mentioned
in
the
docs,
so
yeah.
So
if
you
can
see
the
mastery
adapters,
so
there
are
different
scenarios,
so
you
can
see
how
how
they
are
written
and
and
yeah
and
and
can
see
how
it
is
in
the
ui
part,
so
so
yeah.
This
will
give
you
an
idea
how
you
can
create
your
own
adapters,
so
you
just
have
to
it's
like
you
just
have
to
replace
the
existing
adapter
names,
and
I
mean
there
is
much
more
but
it
is.
H
It
is
more
like
that,
and
there
was
another
question
like
what
image
we
can
use
so
so
all
the
images
are
all
the
supported.
Images
are
available
in
the
counter
code,
docs,
so
yeah.
These
are
the
images.
So
right
now
we
are
using
mini
cube
as
image,
but
we
have
to
start
it
by
ourselves.
So
if
you
can
find
our
image,
which
which
will
start
the
mini
cube,
then
that
will
be
great
and
the
next
question
was
regarding
how
we
can
test
the
labs
locally.
H
So
yeah,
I'm
still
learning
about
this
one,
and
I
have
used
a
long
way
for
this.
So
there
is
a
create
author
profile
link
in
the
window.
If
you
can
click
on
this
one
create
author
profile
on
the
top
okay,
so
I
have
created
my
own
profile
to
test
the
changes
so
yeah.
If
you
can
find
another
way,
then
please
do
share
in
the
slack
channel
or
in
the
community
so
yeah,
that's
it
from
my
site.
You
want
to
add
something.
C
Yeah
we
were
I'm
going
through
this
yesterday
on
the
measuring
continuous
integration
meeting,
the
one
that's
held
twice
a
week
or
I'm
sorry
once
every
two
weeks,
and
you
know
and
surya
shankar
was
going
through
it
and
and
was
stumbling
upon
random
challenges,
which
seemed
to
be
on
catecota's
side
of
the
house.
C
But
I,
but
it
had
also
been
a
long-standing
ask
of
mine
that
we
try
to
use
an
image
that
already
has
kubernetes
running
and-
and
I
don't
think
we
are
yet-
and
I'm
really
hoping
that
someone
can
do
that
to
change
it
to
kubernetes
cluster
running.
If
we
can
try
that
image,
that
would
be
lovely.
C
C
H
C
There
it
is
right
there,
let
me
I'm
pointing
it
out
kubernetes
cluster
running
that
one
we
should.
We
should
definitely
try
that
just
to
the
right,
the
image
id
is
what
gets
used
and
it's
in
the
index
dot.
C
I
think
in
the
labs
we
should
just
make
like
by
the
way
in
order
to
test
things,
it
might
just
be
to
make
a
copy
of
an
existing
lab.
Try
out
that
image
id
see
how.
A
It
goes
but
but
it
still
requires
us
to
start.
Like
start
it
manually
when
the
scenario
starts.
C
Okay
run
launch
when
the
scenario
starts
to
ensure
the
cluster
is
running
cool.
Let's
do
run
launch
you
know,
but
like
it
should
be
further
along
anyway.
You
know,
I
don't
know,
I
haven't
run
it,
but
what
I
do
know
is
we
haven't
tried
it
and
people
still
have
issues
with
mini
cube,
running
and
so
yeah.
C
C
C
C
C
But
let
me
get,
let
me
hear
that
name
again
for
who's
going
to
try
out
that
image,
because
because
I
swear
I'm
going
to
punch
my
my
display,
if,
if
we
don't
try
out
that
image,
I
would
show
sorry.
Mr
das,
is
that
it
yep.
D
C
Okay,
all
righty
nice.
Thank
you
so
much
for
that
honey.
I
put
you
on
the
spot
as
well
with
respect
to
like
I'm
just
even
covering
all
this
today.
So
that's
wonderful
and
suhani
is
tracking
a
number
of
requests
for
new
labs
and
things.
So
please
please
chat
with
suhani.
If,
if
you
dig
in
it's
great.
A
Thank
you
soni,
so
I
have
shared
the
links
to
the
images
in
the
chat,
so
you
can
check
it
out.
I
I
So
a
little
background,
just
yesterday
we
were
on
a
call
with
a
person
named
paula
and
he
has
a
pretty
dope
product
and
related
to
va
and
he
wants
to
do
heavy
networking
stuff
with
it.
Currently
he
is
using
istio
and
it's
a
little
bit
messy
to
manage
all
his
services
using
that
so
he's
looking
forward
to
using
misery
for
doing
that
and
he
reached
us
he
reached
out
to
us
asking
for
help
on
deploying
misery
on
his
eks
cluster.
I
It
was
quite
a
debugging
session,
so
his
setup
was
a
windows
machine
and
he
used
powershell.
Also,
he
tried
using
a
wsn2
for
doing
stuff.
Basically,
he
told
us
that
he
went
to
this
page
indicating
the
star
page.
If
you
go
here
live
site
and
there
is
all
this,
then
there
is
this
information
about
what
platform
you
want
to
deploy
mystery
on,
and
basically
this
isn't
much
detailed
about
a
distinction
between
operating
system
or
platform.
So
he
might
have
gone
here.
I
I
I
I
He
ended
up
missing
it
somehow,
and
he
wasn't
aware
about
this,
so
this
was
a
little
bit
surprising
because
it
should
be
available
in
the
place
in
the
first
place
and
it
should
make
it
easier
to
deploy
stuff.
He
also
said
that
I
tried
him
and
with
him
this
version
was
pretty
old,
so
even
helm
installation
didn't
work
fine
for
him,
so
there
are
some
changes
that
need
to
be
done
here
as
well
as
this
page.
It's
it's
kind
of
a
similar
ux
like
this
on
this
page
as
well.
I
I
I
Now
this
is
a
bit
of
odd
because
we
have
multiple
ways
of
installing
and
some
of
them
directly
install
messaging
like
this
pass
script.
They
would
take
you
an
install
misery
for
you
without
being
aware
about
whether
you
are
trying
to
install
it
on
gk
or
aws,
whatever
it
tries
to
find
your
default
config
and
install
on
it.
So,
oh,
okay,
that's
that
that's
a
deal
I
mean
it
should
not
happen
other
than
that
we
have
brew
users.
Okay,
again,
we
can
do
this,
so
basically
we
should
I
I
propose.
I
They
should
be
able
to
fetch
misery,
ctl
on
it
and
accordingly
deploy
it
on
platform
of
their
choice,
whether
it
be
any
of
these
or
yeah
what
else
so
now
more
things
can
be
added
on
this
and
it's
an
open
point
of
discussion
lee.
Would
you
like
to
add
more.
C
Yeah,
it's
yeah,
it's
not
the
simplest
of
things
to
do
so.
The
goal
is
to
one
step
into
one.
It's
it's
two
steps.
It's
choose
your
platform
step.
Two
is
then
manage
your
mesh.
If
you
run
the
bash
script,
it's
really
pleasant.
You,
you
will
run
the
bash
script,
it
will
download
the
software
run.
The
software
meshery
connect
to
kubernetes
and
it'll
auto
pull
up
the
ui
for
you
as
well
as
long
as
you're,
not
on
a
headless
system.
C
C
As
you
know,
people
who
run
software,
let's
simmer
on
that,
a
little
bit.
It's
it's
like
we
gotta
try
to
keep
it
to
one
step,
give
people
choice.
There
is
choice.
One
of
the
caveats,
one
of
the
something
that's
happening
to
people
today
is
it
root
rocks.
If
you
revisit
like
windows,.
C
C
I
don't
know
it's:
it's
really
tough
to
people.
Everyone,
people
us
humans,
particularly
myself,
super
lazy.
Oh
god.
I
don't
want
to
read
words.
Oh,
are
you
kidding
with
me?
I
don't
want
to
type
that
stuff
in
either.
I
just
want
to
click
that
copy
to
clipboard
button
like
and
anyway,
that
happens
to
a
lot
of
people,
and
so
the
challenge
is
it's
this
system
of
trade-offs.
Again,
it's
like.
We
can
write
a
script
that
accounts
for
any
platform
you
might
deploy
on
we'll
figure
that
script
will
figure
it
out.
C
It'll
it'll
do
what
it
needs
to
do,
which
also
means
that
if
we
do
that,
that's
a
lot
to
sustain
and
to
it'll
be
and
a
lot
more
bugs
to
potentially
happen,
and
so,
if
you
don't
do
that,
if
you
don't
give
people
choice-
and
you
say
well,
the
only
way
is
to
run
a
bash
script
and
not
these
other
ways
that
that
might
not
fly
and
so
yeah
it
something
for
all
of
you
to
simmer
on
a
little
bit.
C
It
is
a
challenging
requirement
to
basically
keep
it
to
a
single
step
now
granted
in
this
single
step.
There's
a
couple
of
lines
and
that's
fine
still
one
step,
you
click
to
copy
to
clipboard
run
it
you
know,
so
I
don't
think
we
can
figure
it
out
necessarily
on
the
call.
But
it's
something
that
rude
rocks
wanted
to
bring
up
to
everyone
here,
because
everybody's
got
an
opinion.
C
So
yeah,
as
you
think
about
it,
please
drop
your
comments
into
slack
or
in
to
discuss
the
discuss
forum
cool.
A
A
J
So
basically,
like
one
important
aspect
in
creating
a
successful
web
app
is
how
we
maintain
the
state.
J
So
in
that
context,
like
the
current
version
of
mystery
ui,
the
state
is
being
managed
by
react,
mostly,
which
is
okay,
but
if
you
want
our
app
to
be
scalable
like
we
should
have
a
more
robust
architecture,
so
we
recently
had
put
together
a
proposal
which
has
been
accepted
and
in
the
implement
implementation
stage
right
now,
the
link
will
be
in
the
docs
or
it
will
also
be
put
in
the
chat
as
well
or
in
a
slack
channel
also.
J
So,
if
you
guys
want
to
learn,
know
more
about
it,
you
can
do
so
by
going
through
going
to
the
link
and
looking
at
what
are
the
changes
that
will
be
made
and
those
things
yeah
and
also
like.
If
you
are
interested
in
contributing
to
measure,
then
it
would
be
helpful
if
you
make
yourself
comfortable
with
the
frameworks
or
libraries
that
are
going
to
be
used,
because
there
are
some
new
things
that
are
going
to
be
introduced
like
redux,
really
and
all
those
things
yeah
so
yeah.
J
A
But
thank
you
disha
lee.
You
were
able
to
say
something.
C
I
was
just
gonna
use
this
as
an
example
to
say
that
nitish
had
suggested
that
maybe
something
should
change
here.
So
per
usual
people
said
that
good
make
a
proposal.
He
grabbed
the
template
for
design,
specs
and
started
writing,
and
then
people
started
commenting
and
it's
about
a
week
and
a
half
later
now.
I
think
it's
about
a
week
and
a
half
ago
the
document
was
empty.
C
Now
it's
full
and
now
there's
a
massive
change
being
planned
for
actuaries
ui
in
a
much
more
much
better
way,
a
more
sophisticated
way
and
so
yeah.
If
natish
can
do
it.
So
so
can
everybody
else.
C
Cool,
so
please
get
involved
because
there's
there's
a
lot
to
do
specifically
here
this
doc.
It
weighs
the
pros
and
cons
of
a
particular
approach,
identifies
the
fact
that
there
is
a
need
for
state
management,
particularly
across
modules
that
are
being
used,
components
that
are
being
used
and
good.
So
this
doc
gets
sort
of
set
aside
as
a
reference
and
the
new
one
needs
to
be
written,
which
is
what
nitish
just
said
about
an
implementation
spec
and
how
to
approach
particulars
of
redux
and
of
redux
observable.
C
So
it's
interesting
also,
I
guess
you
know,
relay
as
a
graphql
client
is
also
in
there
and
specifically
redux
thunk
as
well.
So
yeah
go
go,
go.
A
So
if
you
are
interested
in
working
on
ui
get
involved
in
this
project,
would
you
like
to
talk
about
the
national
cli
block.
E
Okay,
so
we
want
to
start
posting
many
kinds
of
images
and
short
videos
about
measuring
one
of
them
being
starting
point
being
this
image,
so
we
need
something
to
go
along
with
this,
so
a
small
price
to
pay
for
anyone
to
get
their
name
associated
with
this
image
forever
is
to
write
a
small
blog
post
about
messy
cli.
C
No,
that's
awesome
I'll.
Just
I'm
repeating
the
same
thing.
Tennis
just
said,
and
that
is
this
graphic
was
just
made
and
it's
sitting
idle
in
the
community
drive
unused
unloved.
It
doesn't
have
you
know
it
needs
to
be
tweeted
out
and
it
can
be
tweeted
out,
but
it'd
be
a
lot
better
if
it
was
tweeted
with
a
link
to
something
interesting
about
meshary,
cli
and
there's
a
lot
of
interesting
things
in
there.
C
Actually,
some
brand
new
commands
that
have
not
never
been
talked
about,
and
you
know
what
some
commands
you
can't
find
anywhere
else
in
the
world
in
the
world.
B
A
A
Oh
so
ashish
would
you
like
to
go
next
on
the
runtime
generation.
J
Yeah,
can
you
see
my
screen?
Yes
yeah,
so
currently
in
the
whenever
the
mastery
adapter
starts
there?
J
Is
this
go
routine,
which
spins
up
this
register
capabilities
and
it
registers
the
ohm
pattern,
related
capabilities
like
workloads
and
workloads
and
trade
components
to
the
machine
server
so
right,
currently
what
it
does
is
there
are
these
static
yaml
files
inside
the
temp
inside
the
templates
directory
inside
the
under
workloads,
and
basically
it
uses
these
static
workloads
and
and
and
sends
to
sends
these
converts
these
into
the
ordes
and
center
to
the
meshi
server.
J
So
what
I
was
working
on
is
basically
introducing
this
another
function,
register
dynamic
capabilities
or,
basically,
what
it
does
is
it
fetches
the
latest
version
for
whatever
service
mesh
this
adapter
is
and
based
on
that
version
it
basically
fetches
the
manifest,
generates
the
components
on
the
fly
and
then
registers
them
onto
the
mesh
server
and
does
this
after
every
some
seconds
or
hours
or
whatever
time
we
specify
right
now
specified
one
hour,
so
I
can
just
give
it
a
little
demo.
J
So
if
I
change
the
register
capabilities
with
the
register
dynamic
capabilities
right
now,
I
have
my
message:
server
running
so
locally.
What,
whenever
messages
for
service
prints
up?
There
are
some
workload
definitions
already
these
are
registered
locally.
These
are
basically
registered
locally.
So
if
I
spin
up
the
adapter
right
now,
the
adapter
will
dynamically.
J
Okay,
so
registering
latest
workload,
components
it
generates
all
the
components
and
then
it
registers
them
with
the
meshi
server.
So
yeah,
it's
done
so.
If
I
reload
this
page,
so
as
you
can
see,
all
of
the
all
of
the
components
are
all
of
the
components
have
been
registered
by
the
latest
version,
and
I
have
done
the
exact
same
thing
for
the
other
three
service
meshes
the
I've
done,
basically
in
total
for
four
istio
puma
osm
and
linker
d
have
a
problem
with
linkedin
because
they
don't
have.
J
They
have
very
poor
formatting
of
their
helm,
charts.
There
is
no
crd
directory
and
they
are
using
partials
and
so
on.
So
I
have
to
figure
that
out
and
for
the
other
three
adapters,
there
is
no
support
for
packing
components.
Right
now
so
and
currently
it
only
registers
the
only
registers
the
work,
basically
workloads
and
not
trades,
unlike
the
static
function
because
currently
in
the
spec
with
cursor,
there
is
no
basically
he
has
mentioned
the
he
has
mentioned
about
the
trades.
But
right
now
we
are
not
supporting
them.
J
So
once
we
once
we
get
to
cross
that
bridge,
I
can
incorporate
that
function
here
too
other
than
that
one
thing
can
be
asked.
I
want
to
ask.
J
I
mean
this
question
that,
should
we
register
both
the
static
I
mean,
should
we
spin
up
both
these
functions
register,
dynamic
capabilities
and
register
capabilities
on
the
fly,
or
I
I
mean
I'm
not
in
favor,
of
replacing
the
static
thing,
but
I
don't
know
I
mean
how
the
master
server
is
going
to
figure
out
if
by
some
deep
copy
checks
or
what
that,
so
that
two
copies
of
the
same
thing
do
not
get
registered
so
that
it
does
the
updating
and
not
the
creation,
so
we'll
have
to
figure
those
things
out.
J
Other
than
that
we
can
all
I
mean
we
can
basically
split
up
two
go
routines,
one
for
dynamic
capabilities
and
for
static
capabilities,
or
if
anyone
has
some
other
suggestions,
we
can
do
that
as
well
other.
Otherwise
there
are.
There
is
still
work
to
be
done
here
and
there
is
still
things
to
be
added
in
the
mesh
kit
because,
as
I
said,
there
is
no
such
support
for
trades
right
now.
J
So
yep
that
was
it
from
me.
If
anyone
has
any
other
before
I
before
I
stop
sharing
my
screen
just
one
more
small
thing.
I
had
written
this
spec
for
basically
we
were
encountering
this
rate
limiting
issue
and
I
wrote
this
back
and
I
too
forgot
about
it.
So
I
just
remembered
that
I
wrote
a
spec
so
basically
I'll
share
this
link
in
the
comments
we
were
hitting
the
api
rate
limit
in
github.
J
C
C
C
What's
the
right,
why
hesitate
to
say
devops,
but
but
some
of
you
have
joined
the
community
and
have
said
like
hey,
I
would
like
to
get
more
familiar
with
docker
and
kubernetes
and
like
devops,
and
there
are-
and
this
this
is
a
good
opportunity.
The
service
mesh
labs.
That
suhani
spoke
to
that's
a
great
opportunity
for
those
that
are
into
operations.
C
This
opportunity
is
another
good
one,
and
that
is
that
there
are
a
few
users
that
have
asked
for
instructions
on
how
to
deploy
measuring
in
an
air-gapped
environment,
and
we
don't
have
all
of
those
instructions
written
down,
there's
a
few
considerations
that
they
need
to
take
care
of,
and
these
users
to
start
with
it
can
be
a
really
long
list
of
like
here's
all
the
things
they
have
to
do
great.
We
just
need.
A
Well,
thank
you
thanks
silly.
This
pack
is
in
the
chat
and
you
can
find
the
other
links
in
the
community
drive
as
well.
So
we
are
a
little
over
time
and
we
have
this
one
item
remaining
that
which
is
related
to
night
talk
or
measuries
or
layer.
5's
project
get
night
talk,
so
we
have
a
couple
of
new
projects
that
are
opening
up
here.
A
So,
if
you
like
to
know
what
network
is
then,
like
my
talk
is
basically
a
layer,
7
load
generator
performance,
testing
tool
developed
by
onward,
and
what
get
night
talk
is,
is
it's
a
layer,
5
project
that
aims
to
make
it
make
night
talk
more
accessible,
so
there
are
a
couple
of
projects
coming
around
it,
one
of
which
is
grpc
support
for
network,
so
this
project.
A
De-Prioritize
this
use
case
in
night
talk,
so
they
they
are
willing
to
work
with
someone
from
our
team.
If
and
they
will
help
us,
they
will
help
us
add
this
feature
into
night
talk.
So
that's
one
area
and
you
can
find
all
these
things
in
the
in
the
meeting
minutes.
A
Another
project
is
distributed
low
generation
which
we'll
use
for
distributed
performance
testing.
So
we
are
still
waiting
on
a
pr
in
night
talk
to
make
sure
that
we
can
do
this
quite
easily,
so
that
is
also
in
one
area
so
and
and
yeah,
and
then
another
area
is
adaptive.
Load
control.
So
netapp
has
has
this
feature
of
called
adaptive
load
control,
which
basically.
A
Which
basically
creates
load
according
to
a
feedback
loop
or
according
to
your
configuration
of
your
mesh
or
your
congregation?
So
basically,
what
we
are
trying
to
do
is
use
this
feature
in
measuring
and
possibly
work
towards
building
meshmark
as
a
benchmark
for
service
meshes.
A
So
these
are
a
couple
of
areas,
so
we
will
be
discussing
more
of
this
in
the
service
mesh
performance
community
meeting,
which
happens
every
other
thursday,
so
it
will
happen
on
next
thursday,
so
join
that
meeting
to
get
involved
in
this
project.
I
know
many
of
you
have
already
signaled
their
interest
in
working
on
this
project,
so
you'll
be
working
with
c
plus
and
go
so
if
you
are
interested,
please
join
the
next
week's
call.
A
And
we
also
have,
we
are
also
planning
for
the
next
release,
so
we
are
writing
a
release
blog.
So
if
anyone
else
would
like
to
volunteer
for
this
as
well,
please
signal
you
can
find
an
old
release
blog
in
the
missouri
website
or
on
the
layer
5
website.
So
please
signal
that
as
well
all
right,
so
anything
else.
A
All
right,
hello,
thank
you
thanks.
Everyone
for
joining
see
you
next
week.