►
From YouTube: Layer5 Community Meeting (Nov 20th, 2020)
Description
Post #KubeCon recap and @envoyproxy discussion.
A
Hello,
hello,
hello,
friday
friday,
the
20th
thanksgiving
next
week
for
some
of
us
or
for
any
of
us
that
I
guess
so
choose
to
have
thanksgiving
beneath,
is
your
hand
raised
in
order
to
say
something
or
just
playing
around
with
the
tools.
A
A
But
yeah
feel
free
to
to
interject
anytime,
though
cool
all
right.
Well,
there's
a
link
to
the
meeting
minutes
in
the
chat.
We
record
our
meetings
and
post
them
on
youtube
for
the
viewing
pleasure
of
the
masses.
Apparently
those
things
get
watched
I
swear.
Half
of
them
are
from
my
mom.
You
know
she's,
just
she's
trying
to
keep
up.
A
Cool
all
right,
I'm
waiting
for
long
to
his
audio
to
connect.
Actually,
I'm
way
long
is
your
audio
connected.
Can
you
can
you
hear
me?
A
Yes,
oh
nice,
well
good,
we're
just
about
to
kick
off
the
meeting
and
well
as
if
you've
met
sriti
before
she'll
tell
you
that
there's
a
tradition
in
the
community
first
time
anybody
joins
one
of
the
calls
they
get
to
just
give
a
short
introduction
to
say
hi
to
everybody
and
and
take
a
moment
to
introduce
yourself
to
say
where
you're
dialing.
In
from
how
you
got
here
that
way,
people
can
get
to
know
you
better
and
make
some
friends.
I
guess.
A
So
yeah,
if
you
would
say
hi
to
everybody,
introduce
yourself.
C
Okay,
I'm
leyland
I'm
from
malaysia.
So
currently
it's
12
a.m.
So
it's
about
going
to
sleep!
Then
I
saw
the
notification.
Then
I
just
click
inside
yeah.
A
Well,
nice
thanks
for
coming
on
there's
a
couple
of
people
who
join
from,
I
think
a
time
zone
just
earlier
than
you
occasionally
we'll
have
some
folks
in
australia
joining
and
they
are
particularly
unintelligible
at
one
am
in
the
morning
so
like
like
most
of
us
now.
Actually
I'm
gonna,
I'm
gonna
see.
If
I
can
do
this,
I'm
not
sure.
A
If
this
is
possible,
I
wanna
see
if
I
can
embarrass
a
little
and
has
anybody-
or
maybe
it's
just
me
so
far-
has
anyone
used
measuring
ctl
system
complete
completion?
A
If
you
haven't
you're
missing
out
and
if
you
haven't,
you
should
do
it
and
then
give
your
thanks
to
weilong
for
for
having
it
completed
this
command
here.
A
Which
is
really
nice,
because,
oh
because,
if
you
have
relatively
fat
fingers
like
mine,
then
you
mistype
all
the
time
and
you're
trying
to
breeze
along.
So
if
you,
if
you
execute
this
command
and
then
follow
the
instructions
based
on
your
particular
type
of
shell,
you'll,
be
able
to
do
fancy
things
like
this,
like
type
s
and
press
tab.
A
Oh
you
know
what
actually
hold
on
press
s,
hit
tab
and
it'll
fill
in
the
command
press
tab
again
and
it'll
tell
you
you
know
here's
what
here's!
What
else
you
could
do?
Maybe
it's
config
great
and
from
here
you
finish
off
the
command.
So
so
there
had
been
a
contributor
about
a
year
ago,
about
a
year
ago,
his
name's
mayank
shah.
He
is
currently
a
community
bridge
intern.
I
think,
with
the
linker
d
group
he
had
raised
up
this
particular
request
and
tried
to
implement
it
and
gotten
blocked.
A
C
A
And
actually,
what's
actually
there's
a
kubecon
talk
a
little
later?
Well,
I
guess
maybe
about
the
time
that
you're
waking
up
there's
a
cubecon
talk.
It
has
a
recording
of
some
meshri
ctl
command
line
and
I'd
have
to
go
check.
It
might
actually
show
off
your
work
so,
like
your
work
might
be
on
cubecon
stage
later
today.
I
I'll
go.
Look
because
that's
that's
pretty
neat!
A
That's
pretty
cool
cool,
okay,
good!
Hopefully,
we've
we've
either
entirely
scared
you
off
from
spending
any
more
time
here
or
or
you
recognize
that
that
your
efforts
are
appreciated
like
it's.
It
was
very
helpful.
A
A
Yes,
nice,
okay,
okay,
yeah,
very
good.
Does
that
put
you
you
help
me
with
me
my
poor
geography.
Does
that
keep
you
in
the
same
time
zone
or
does
that.
A
Got
it
cool
all
right?
Well,
hopefully,
everybody
on
the
call
has
access
to
the
meeting
minute
so
feel
free
to
help.
Take
some
notes
and
slap.
Your
name
into
the
meeting
minutes.
We've
had
a
few
people
join
this
week,
a
little
bit
of
a
different
crowd
of
folks
who've
joined
this
week
than
some
of
the
other
weeks,
and
that's
because
that's
because
cubecon
is
going
on
and
thanks
to
the
thanks
to
the
works
that
each
of
you
have
been
doing.
A
We
we
have
the
stage
three
times
and
so
sriti
has
actually
highlighted
the
three
different
talks
and
where
the
works
are
being
shown
off
so
of
the
constituency
of
people
that
have
joined
this
week.
It's
been
a
little
more,
maybe
on
the
user
side,
which
is
which
is
really
nice
in
part,
because
we
have,
we
typically
have
so
many
contributors
joining
and
here
to
learn
things
and
impact
the
projects,
and
so
so
anyway,
so
nice
to
have
some
more
users.
A
A
Good,
so
so
some
newcomers
other
other
newcomers
that
other
people
that
have
come
to
the
community
and
are
now
now
starting
internships
with
lair
five.
So
one
of
them
is
utkarsh,
and
I
don't
know
that
utkarsh
is
on
at
the
moment,
but
he
has
just
rece.
Just
this
week
come
in
through
community
bridge
or
if
you're,
not
if
you're
unaware
I,
this
is
community
bridge
or
rather
there's
a
there's,
a
linux
foundation
program
called
community
bridge.
That
was
just
renamed
this
last
week.
I
guess
yeah.
A
Just
a
week
ago,
it's
now
called
lfx,
I'm
not
entirely
sure
what
the
x
stands
for,
but
but
so
the
branding
is
no
longer
community
bridge
lfx,
there's
a
few
different
resources
that
the
linux
foundation
puts
forth.
A
We
are
participating
in
part
through
this
program
to
for
intel
and
linux
foundation
to
sponsor
a
couple
of
interns,
so
there's
still
an
open
spot,
and
I
am
not
signed
in
at
the
moment
and
somehow
mushrik
is
the
first
listed,
which
I
think
I
shouldn't
say
somehow.
I
know
that
the
administrators,
I
think,
did
this
on
purpose,
which
is
nice.
It
also
means
that
about
once
a
day
we
get
someone
applying
so
so,
hopefully
we'll
have
another
person
coming
through
community
bridge
at
some
point.
A
If
by
the
way,
if
anyone
is
either
interested
or
if
anyone
has
people
asking
them,
you
know
how
to
best
position
for
such
an
internship
it.
The
answer
is:
join
the
community
and
get
engaged
and
then,
if
you're
engaged
and
you
like
it
and
you're
being
impactful,
then
you
should
you
know
you
should
be,
and
if
you
want-
and
if
that
person
wants
an
internship,
then
we
should
we
should
do
one
so.
A
All
right
that
was
kind
of
an
update
as
to
utkar
she's
coming
through,
I
guess
technically,
not
community
bridge
but
lfx.
A
A
A
So
austin
is
gonna
help
in
a
few
different
areas,
and
one
of
those
is
if
you've
used,
meshri
you're
familiar
with
well
you're,
familiar
with
the
fact
that
it's
a
it
feels
like
an
infrastructure
tool,
and
I
say
that
to
mean
that
its
user
interface
could
use
some
help.
A
Actually,
a
fair
bit
of
help
measury
is
more
powerful
than
people
than
users
generally
understand.
They
generally
don't
understand
that
because
it
the
the
flow
of
the
user
experience
in
the
measure.
Ui
is
not,
as
needs
to
be
refactored,
to
expose
some
of
those
capabilities,
much
more
apparently
to
the
user,
so
so
good.
So
austin
is
here
as
well.
A
Okay,
all
right
wait!
Great!
We
made
mention
of
a
few
of
the
talks
that
are
going
on
this
week.
Some
of
you
on
the
call
two
of
you
three
of
you
participated
in
these
talks.
So
by
the
way
the
call
for
papers
for
next
cubecon
is
already
open.
A
A
A
Okay,
so
for
today
we
didn't,
we
didn't,
have
a
lot
of
topics
lined
up
to
to
dig
into.
I
think
this
first
one
is
worth
spending
a
little
bit
of
time
on
anybody
running
who's,
running
mac,
os.
A
Well,
what's
that
okay,
I
couldn't
everyone
spoke
up
all
at
once,
so
I
couldn't
I
couldn't
hear
anyway.
Some
of
you
are
probably
running
mac
os.
Some
of
you
maybe
have
upgraded
to
big
sur.
I
did
and
I'm
enjoying
some
new
features,
but
I'm
also
not
enjoying
some
kerfuffle
around
homebrew
and
it's
sort
of
between,
I
think
homebrew
and
a
new
version
of
xcode,
which
I
think
I
now
have
resolved
on
my
system,
which
really
wasn't
much
of
a
challenge
to
resolve.
A
I
just
needed
to
understand
that
that's
what
needed
to
be
done
anyway,
the
the
and
that
was
to
upgrade
to
the
latest
xcode,
having
not
done
that.
I
was
having
a
real
difficulty
using
mastery
ctl.
A
A
What
made
this
more
confusing
for
me
is
that
recently,
we've
been
having
some
challenges
with
builds
completing
successfully,
and
it
just
started
right
around
the
same
time
that
I
had
upgraded
to
big
sur,
and
so
is
it
just
a
coincidence.
I'm
sure
kush
has
been
looking
into
some
of
these
issues.
I
think
this
is
some
of
his
work
trying
to
to
fix
the
issues,
but
for
those
of
you
who
don't
know,
we
use
github
actions
throughout
most
of
the
projects
or
not,
I
mean
throughout
all
the
projects.
A
This
is
a
little
curious.
This
makes
me
think
that
maybe
we're
referencing
the
wrong
secrets,
because
every
repository
has
secrets-
and
those
are
those
just
need
to
be
referenced
in
the
workflow,
not
not
sure.
A
So
one
of
the
things-
that's
that's
a
bit
of
an
issue
for
mastery
ctl-
is
that
we
use
go
releaser
as
part
of
the
build
process
to
really
to
compile
mastery
ctl
to
a
few
different
target.
Operating
systems,
target,
architectures
and
and
that's
convenient
to
use,
go
release
or
it's
convenient
for
us
as
a
project
to
just
use,
go
release
or
to
go.
Make
these
various
builds
and
then
have
them
released
from
here.
Go
releaser
will
also
facilitate.
A
A
This
tap
is
is
what
is
used
in
part
for
brew
homebrew
as
a
tool
to
be
able
to
install
mastery
ctl.
A
Now
here's
the
thing
with-
and
this
is
actually
what
one
of
the
users
this
today
stumbled
into.
I
don't
think
our
instructions
were
necessarily
clear
that
they
that,
as
a
user,
you
first
need
to
add
this
repository
as
a
tap
to
your
brew.
Configuration
after
you've
done
that,
then
you
can
install
mastery
ctl,
otherwise,
brew
doesn't
really
know
where
mastery
ctl
is
and
where
to
pick
it
up
from
and-
and
this
is
an
artifact
of
using
go
releaser
so
I'll.
A
Try
to
make
this
a
little
more
evident
what's
going
on
here
is
that
if
anyone,
if
you've
used
brew
or
yum
or
apt-get
or
scoop
or
chocolatey
or
like
I
don't
know,
there's
a
bunch
of
package
managers,
you
occasionally,
you
know
you
occasionally
have
to
configure
your
package
manager
to
pull
from
different
repositories.
A
Measuring
ctl,
since
we're
publishing
with
go
releaser,
go
releaser
actually
in
their
documentation.
It
has
this
note
right
here
which
is
basically
saying
that
go
releaser.
Doesn't
it
doesn't
generate
a
core
formula,
a
valid
core
formula
for
homebrew,
and
because
of
that,
our
users
first
need
to
run
this
command
to
configure
their
local
brew
tool
to
point
to
this
layer,
5
project,
to
point
to
this
repo,
this
repo
calls
out
formulas
for
ashrae
ctl
and
where
to
go,
get
measuring
ctl
and
the
latest
release
is
423
and
it
points
over
to
where
to
go.
A
A
So
we
have
an
open
request,
actually
so
so
way
way
long.
Maybe
this
is,
I
don't
know
if
this
is
a
an
area
of
interest
for
you
or
not,
but
or
or
for
anyone
on
the
call
is,
it
would
be
a
lot
better,
a
lot
nicer
if
we
weren't
using
go
well,
we
weren't
using
go
releaser
for
publishing
measuring
ctl,
but
instead,
instead
we're
able
to
allow
users
to
just
do
one
command,
so
something
for
people
to
think
about
I'll.
Put
a
link
to
that
issue
in
the
meeting.
A
A
A
Okay,
anybody
have
any
comment
on
that.
Did
anyone
struggle
with
that
the
first
time
that
they
installed
mastery.
A
Ctl,
I
know
a
number
of
you
end
up
just
using
the
bash
script
to
install
mastery
ctl,
not
a
package
manager
like
brew,
so
you
wouldn't
have
bumped
into
it.
I
know
there's
a
number
of
you
that
just
compile
mastery
ctl
itself,
you
don't
you
don't
use
a
package
manager
or
you
don't
use
the
released
binary.
A
So,
okay,
other
topics-
this
is
probably
just
more
of
an
announcement.
I
suppose
there
are
we
we
myself
and
then
it's
been.
I've
done
this
with
a
few
different
people
in
the
community,
we'll
occasionally
give
work
service,
mesh
workshops
through
o'reilly,
and
I
don't,
if
you're
an
o'reilly
subscriber.
A
Linker
d,
we're
giving
two
different
workshops
or
three,
I'm
sorry,
three
different
workshops,
one
is
on
an
introduction
to
linker
d.
Another
one
is
an
introduction
to
istio
and
another
one
is
an
advanced
and
so
here's
the
thing,
here's
the
secret,
I
suspect-
maybe
most
of
you
aren't
subscribers
to
o'reilly
the
the
secret
is
the
content,
those
labs
that
we
walk.
People
through
they
are
they're,
actually
open,
source
and
available
to
you.
So
if
you
go
to
layer,
five
io,
slash
workshops.
A
A
A
All
right,
so
we
we
might
end.
Today's
call
early
depends
on
things
to
talk
about,
so
I'm
just
looking
over
the
list
of
people
who
are
on
the
call
we
might
want
to
skip
or
or
push
get
nighthawk
to
the
bottom,
but
anyone
who's.
So
it's
a
new
project,
that's
forming.
It's
called
get
nighthawk
and
I'll
talk
about
it
later,
but
it's
open
to
any
of
you
that
want
to
participate
kind
of
a
good
time
to
get
in
on
a
new
project.
A
It's
fairly
small
there's
going
to
be
a
couple
of
technical
hurdles.
For
the
first
time
we
will
engage
with
a
couple
of
data
scientists
because
there's
a
couple
of
data
science
things
that
we
want
to
do
in
that
project
and
not
for
the
first
time
we
will.
A
You
know
what
I
guess:
I'm
just
talking
about
get
nighthawk
now,
so
I
so
in
this
project,
it's
if
you're
familiar
with
envoy
as
a
proxy
or
even
if
you're,
not
if
you're,
familiar
with
nginx
or
h.a
proxy
or
envoy
or
others
envoy,
is
a
very
is
very
popularly
used
as
the
data
plane
component
in
service
meshes
and
envoy
has
a
load
generator.
A
It's
called
nighthawk.
The
nighthawk
project
was
created
as
a
subproject
of
envoy,
just
like
fortio
was
created
as
a
subproject
of
istio.
A
The
nighthawk
project
is
gaining
in
popularity,
it's
more
capable
than
fortio
it's
one
that
we
would
like
to
invest
more
and
more
into
in
for
a
couple
different
reasons.
We
want
to
invest
into
it.
So
one
of
those
reasons
is
that
soon
soon
fortio,
I'm
sorry
soon
nighthawk
you'll,
be
able
to
take
and
deploy
multiple
instances
of
nighthawk
and
tel.
Tell
those
instances
to
collaborate
amongst
themselves
to
generate
load
in
a
distributed
way
and
then
to
send
back
their
results
from
those
tests
to
a
central
location.
A
Do
statistical
analysis
on
those
results
and
and
given
that
capability,
I
expect
that
we're
able
to
ask
ourselves
and
ask
the
industry
a
few
new
questions
that
maybe
we
haven't
either
thought
of
or
been
able
to
answer
very
easily
before
and
so
we're
engaged
it
isn't
new
that
we
would
engage
with
universities.
A
A
K
kangapur
something
yeah
anyway,
anyway,
with
a
professor.
There,
who's
done
lots
of
research
in
this
area.
I
think
he
wants
to
bring
in
a
few
of
his
students
to
kind
of,
engage
and
learn
some
things
and
help
on
some
things,
so
the
project
itself
that
we're
looking
to
do
it's
fairly.
To
give
you
a
quick
essence
of
what
the
get
nighthawk
project
is.
A
A
So
there's
envoy
has
a
proxy
that
you
can
download
and
run
in
your
data
planes,
and
it
didn't
really
have
a
lot
of
it.
Wasn't
really
didn't
have
a
lot
of
dis
distributions.
A
It
wasn't
packaged
in
very
many
ways,
so
a
separate
project
spun
up
separate
but
highly
related
projects
spun
up
to
help
facilitate
people
getting
getting
envoy,
and
so
they
created
a
cli
called
get
envoy,
and
then
they
do
a
bunch
of
builds
of
envoy
every
night
or
every
time
that
there's
a
release.
It's
a
fairly
small
project.
A
Much
of
the
you
know.
The
value
here
is
between
these
two
things.
So
as
we
look
at
get
nighthawk,
which
would
be
like,
we
would
be
creating
a
number
of
packages
for
different
target
platforms
for
for
nighthawk,
the
load
generator
envoys
load,
generator
we'll
be
working
on
that
and
at
the
same
time,
we'll
provide
some.
A
We'll
provide
some
cli
support
for
retrieving
those
we're
not
going
to
go
off
and
create
a
new
cli,
rather
we're
going
to
build
on
top
of.
What's
in,
I'm
sorry
not
build
on
top
of
we're
just
going
to
extend
measuring
ctl
to
do
to
facilitate
some
of
this
really.
That
was
already
in
scope
for
mastery
ctl
if
you
use
messy
ctl
to
deploy
a
service
or
I'm
sorry
to
invoke
a
ted,
a
performance
test
which
I'll
show
you
now
by
the
way.
A
But
today
and
again
thanks
to
weilong,
you
can
do
a
measuring
ctl
perf
and
this
sub
command
I'll
set
docker
to
start
so
so
mastery
ctl
is
actually
starting
docker
on
my
system,
there's
a.
A
I
think
this.
This
will
say
that
it
failed
in
a
moment,
but
actually
I
can
see
docker
starting
on
my
system.
Now
it
just
takes
a
little
while
longer.
I
think
then
this
command
is
willing
to
wait.
So
we
don't
have
the
best
logic
in
this
command
about
determining
whether
or
not
docker
is
ready.
So
I
think
it
failed
saying
that
docker
has
failed
to
start,
but
that's
not
true.
It's
still
starting
it's
just
taking
too
long.
A
So,
okay,
so
docker
has
started
now.
So
if
I
run
this
command
again,
what
it
will
tell
us
is
okay,
great.
If
you
want
to
use
mastery
ctl
to
invoke
a
performance
test.
Here's
the
flags
that
you
can
provide
like
what
load
generator
did
you
want
to
use?
What
endpoint
did
you
want
to
hit
for?
How
long
did
you
want
to
do
it?
How
how
hard
did
you
want
to
hammer
on
it
and
that
kind
of
thing
you
can
also
just
you
don't
have
to
type
out
all
these
parameters
on
the
command
line.
A
A
It
doesn't
have
to
be,
but
you
know
a
handful
of
pages,
probably
probably
have
a
blog
post
or
two
on
it
or
what
have
you?
But
it's
just
it's
fairly
straightforward.
A
I
that'll
be
a
little
bit
difficult.
I
expect
more
help
is
needed.
We're
engaged
these
are
engi.
This
is
an
engineer
from
google
and
there's
another
one
engaged
from
red
hat,
which
is
good
that
we're
getting
good
that
we're
getting
that
that
type
of
engagement
from
the
nighthawk
maintainers.
So
these
are
nighthawk
maintainers
and
they're.
A
A
Nope
she's,
not,
I
think
we
bored
her
to
death
with
the
other
stuff
either
that
or
I
know,
she's
attending
kubecon,
so
she's
been
focused
there.
So
abhishek.
Let
me
let
me
ask
you,
I
don't
know
if
talking
about
smi
conformances.
A
B
It's
just
that
a
couple
of
commands
are
being
tested
out
like
install
installation
of
the
service
mesh
and
sample
applications
or
running
conformance
test
on
all
the
commands
that
we
have
that
those
have
been
tested
out
currently
and
when,
specifically,
if
you
look
at
kuma,
we've
been
able
to
install
kuma
as
an
installation
of
kumas
and
running
conformance
test.
Both
of
these
are
being
are
in
a
running
state
right
now.
D
Yes,
I
well
console
runs
on
the
on
measuring
adapter
library,
just
not
the
current
version,
so
the
latest
release
of
console
already
uses
the
measuring
adapter
library
offered.
The
latest
release
of
the
console
adapter
already
uses
the
measuring
apps
library.
So
I
plan
to
just
update
to
the
latest
version
this
weekend.
D
B
D
B
A
We
yeah
actually
not
now.
Is
I
don't?
We
don't
have
any
other
topics
it
I'd,
love
to
see
it
advanced.
Can
we
go
through
what
that
looks
like
now,.
D
And
and
how
the
so
we're
using
a
fork
of
couple
right
to
where
we're
cuddle,
which
is
the
I
can't
remember
what
it
stands
for,
but
it's
a
testing
framework
to
test
kubernetes
operators
and
it
basically
works
by
providing
a
number
of
yaml
files
and
then
cottle
deploys
these
yaml
files
and
runs
some
tests,
which
are
also
specified
as
custom
resource
definitions
on
the
deployments.
For
instance,
you
can
assert
that
all
your
resources
have
been
deployed
with
a
given
set
of
properties.
D
So
you
define
a
dummy
file
in
this
sequence
of
manifests
and
then
the
capital
harness
or
the
costal
test
execution
framework
has
been
configured
to
run
a
custom
test
implementing
as
go
code
when
this
manifest
file
shows
up
in
the
sequence
of
test
steps.
In
this
framework-
and
this
is
the
mechanism
that
is
used
to
run
these
conformance
tests.
D
The
deployment
of
the
sample
application
or
the
deployment
descriptor
of
the
sample
application
is
one
of
these
files
of
this
manifest
files
that
kotlin
runs
through
now.
What,
in
the
case
of
console,
we
need
to
add
annotations
to
both
specifications
that
make
console
to
inject
the
agent
and
the
site
current
set
of
this
service
as
a
service
control
by
console
service
mesh.
D
B
D
Think
these
three
possibilities,
I
so
for
myself
to
be
able
to
to
implement
that
and
that
then
the
adapter
would
select
a
console.
Adapter
would
send
these
either
the
specific
deployment
into
the
function
or
the
service
call
to
the
when
when
it
invokes
the
smi
test
run,
which
will,
if
you
understand
it
correctly,
mean
that
we
actually
have
to
change
the
api,
because
the
conformance
test
tool
is
a
service
that
runs
in
the
cluster.
A
From
the
adapter,
so
so
in
that
case,
like
changing,
taking
advantage
of
the
fact
that
we've
forked
cuddle
and
making
some
changes
to
its
api
to
be
able
to
receive
the
additional.
D
Yes,
something
like
this
isn't
isn't,
don't
we
isn't
the
smi.
D
Conformance
tool
it
as
a
service
into
the
cluster,
I
think
so
and
then
the
the
adapter
actually
calls
the
api
of
our
servers,
smi
conformance
testing
tool,
which
is
running
as
a
service
as
an
adapter
in
the
adapter
and
that
api.
We
need
to
change
if
we
want
to
send
in
additional
information.
D
Yeah,
yes
totally
and
then
so,
and
this
is
when
you
run
this-
and
you
see
just
correct
me:
if
I'm
wrong
you,
then
you
you
automatically
a
namespace
called
measure
is
created.
D
Changed:
okay,
good
and
then
the
cuddle
tool
is
actually
creating
namespaces
for
each
of
the
test
suits
or
whatever
they're
called
to
yes
and
then
deletes
everything
after
the
rom.
D
Yes,
yeah,
so
yes,
so
so
whatever
we
do,
it
takes.
It
takes
some
effort
right.
So
that's
why
I
think
we
need
to
discuss
this
what's
the
best
way,
because
as
soon
as
we
start
changing
api,
then
that
has
some
consequences.
A
Yeah,
if
yeah
do
you
think
we
want
if
we've
waited
or
or
if
we
focused
on
other
work
in
the
meantime,
while
console
while
before
and
waited
for
console
110
to
come
out?
A
Is
that
our
only
issue
or
is
that
the
only
reason
to
potentially
take
on
this
work.
D
That's
also
something
I
wondered
about
isn't:
can
we
because
the
the
assumption
is
that
all
service
meshes
can
handle
exactly
the
same
application
deployment
manifest,
and
so
do
we?
Is
that
something
we
expect
to
be
true
or
more
or
truer
more
true
in
the
future
right
or
is
it
like
in
in
istio,
where
you
have
like
you
know,
you
have
to
define
virtual
services
which
are
part
of
eastern.
Maybe
that
will
be
a
part
of
the
thing
we
have
to
set
up
and
that's
very
istio
specific,
and
then
you
would
anyway.
A
So
the
thing
is
is
like
we
have
future
use
cases
for
examining
well
for
facilitating
service
onboarding
onto
a
mesh.
So
because
it'll,
be
you
know,
with
high
frequency
or
it'll,
be
a
use
case
that
many
many
people
have
when
they
they
go
to
use
a
service
mesh.
They
have
existing
services,
they
want
to
migrate
them
onto
the
mesh.
A
The
manifests
that
they
use
today,
you
know,
aren't,
aren't
going
to
contain
all
the
annotations
or
you
know,
containers
pod
specs
that
are
necessary
for
it
to
you
know
those
services
that
so
that
onboarding
use
case
is
is
an
area
of
investment
for
us
or
like
noting
that
that's
a
future
thing
but
like
so
that's
that's
in
some
respects.
That's
a
plus
one
to
cracking
open
the
architect,
potentially
investing
here,
there's
another
plus
one
to
potentially
investing
here,
and
that
is
that
it's
a
little.
It's
slightly
odd
that
meshery
itself
and
the
adapters.
A
We
you,
you
guys.
People
here
put
a
lot
of
time
into
making
sure
that
measuring
through
the
adapters
is
able
to
provision
and
manage
any
given
service
mesh
and
yet
for
this
set
of
smi
conformance
tests.
We're
using
a
different
mechanism
to
deploy
the
meshes
and
to
apply
the
esser
and
at
a
high
level
like
that.
That
seems
you
know
not
good.
You
know
that
seems
not
ideal
like.
Ideally,
you
would
use
the
same
mechanism
today.
That's
part
of
why
we're
using
measures
like
it
can
do
life
cycle
management
of
a
mesh.
A
So
let's,
let's
use
that
and
then
let's
add
on
the
ability
to
have
assertions
defined
and
apply
those.
So
that's
kind
of
a
that's
another
plus
one
to
potentially
invest
like
if
we're
going
to
take
on
work
like
it's.
It's
a
it's
a
fair
question
to
ask
to
question
whether
or
not
continued
use
of
kudo
is
the
is
the
right
thing
or
maybe
just
part
of
kudo,
like
the
the
value.
A
That
was
a
lot
of
the
value
that
kudo
was
going
to
bring
or
cuddle
kudo
karo
kudo
was
going
to
bring,
and
but
it
has
its
own
assumptions
as
its
own
project,
and
so
you
end
up,
you
know.
Hence
we
ended
up
forking
to
overcome
some
things,
and
now
that
project
has
advanced
since
we've
forked,
I
don't
know
what
the
changes
are
or
what
the
but
but
we're
yeah
in
some
respects.
We're
kind
of
at
this
crossroads
of
we
want
the
convenience
of
being
able
to
define
an
assertion,
verify
the
assertion,
return.
A
A
A
So
I'm
you
know
so
on
my
system,
I
have
istio
console
open
service
mesh
and
nsm
installed,
apparently
geez
and
for
for
istio.
This
is
the
only
adapter
currently
that
supports
this,
but
this
is
this
is
very
much
so
something
you
know
desired
and,
and
that
is
to
let
people
operate
their
service
mesh
with
confidence
to
empower
them
with
some
best
practices.
A
It's
one
of
the
validation
operations,
one
of
the
operations
that
sort
of
falls
into
that
category
is
one
of
those
is
running
a
this
set
of
tests
that
we're
talking
about
smi
conformant,
another
one
is
it
use,
it
doesn't
use
kudo,
it
doesn't,
but
is
to
go.
Take
a
look
at
your
running
configuration
the
configuration
of
istio
that's
running
now
and
against
a
set
of
like
10
or
so
tests
examine
your
configuration
basically
basically
do
in
some
respects
what
we're
doing
with
cuddle,
which
is
go.
Look
at
your
environment.
A
Given
this
rule,
a
rule
that
says
hey,
you
shouldn't,
you
should
have
prefixes
in
your
you
know.
Whatever
this
this
rule
is
your
config
should
look
like
this
now,
the
the
library
that
we're
using
to
perform.
This
is
also
not
our
own,
rather
borrowed
from
another
project.
A
That
project
was
very
happy
that
we
had
done
it
at
the
time
and
a
year
later,
I
think
that
they
consider
we're
we're
somewhat
competitive
and
so
we're
still
in
good
relationships,
but
we're
just
you
know
anyway,
they're
not
advancing
that
project
anymore
anyway,
but
this
is
this
capability
is
something
that
we
desire
for
every
adapter
to
have.
A
Every
adapter
should
be
able
to
take
a
look
at
the
environment
check
to
see
if
you're
doing
something
well
or
not,
and
so
this
particular
function
is
also
an
or
this
particular
functional
area
is
also
a
need
of
rewrite.
A
So
yeah
it
makes
me
wonder
if
so
so,
here's
what
I
would
say,
here's
or
from
my
perspective
is
this
coming
wednesday,
and
so
michael.
If
you're
able
to
make
it
wonderful,
the
the
next
smi
community
meeting
is
being
held
on
wednesday,
the
25th.
It's
on
our
commute.
Later
five
community
calendar,
it's
normally
just
half
an
hour,
we're
extending
it
to
an
hour.
A
We're
gonna
invite
all
the
maintainers
of
the
various
service
mesh
projects
that
claim
to
be
compatible
with
smi,
we're
inviting
them
over
to
walk
them
through
the
program,
sort
of
walk
them
through
smi
conformance
and
here's
here's.
What
our
ask
is
of
them
and
our
ask,
will
there'll
be
a
couple
of
asks.
One
is
anyway.
Let
me
not
walk
through
all
that.
A
Let
me
just
say
that
I
think
the
answer
here
is
something
like
assuming
we
can
hobble
along
with
kudo,
as
it
is
today
that
we
really
want
to
get
to
the
other
side
with
at
least
one
of
these
service
meshes
saying
here.
A
Mescheri
is
the
tool
that
validates
this
mesh
and,
if
others
are
successful
as
well
great
and
we
can
claim
success
and
the
popularity
of
the
project
will
grow
and
the
re
the
reliance
on
the
project
will
grow,
support
for
it
will
grow,
and
it
would
be
after
that
that
we
would-
and
we
can
start
looking
at
that
now,
but
like
understanding
that
we
would
probably
leave
console
by
the
wayside
and
not
invest
in
rework
against
something
that,
I
think
what
we're
saying
or
what
I'm
saying
anyway,
is
that
I'm
not
sure
that
we
should
keep
kudo
long
term,
moreover,
that
we
should
keep
that
other
component
that
we're
using
for
configuration,
validation,
they're,
both
kind
of
doing
the
same
thing.
A
A
That
it
would
be
nice
if
we
that
we
were
able
to
land
some
of
the
momentum
that
we've
had
here
with
one
or
two
of
the
projects,
acknowledging
that
it
looks
like
right
now,
at
least
without
some
additional
work
that
we
might
be
income.
Our
approach
might
be
incompatible
with
console,
but
that
we
would
fix
that
in
a
different
way.
A
We
once
one
way:
oh
yeah,
very
good.
Well
I
mean
one
one
silly
way
would
be
I
just
this
just
occurred
to
me
now
is
that
well
we
could
have
manifests.
A
D
Yeah,
yes,
exactly
and
in
the
comment
in
this
document,
if
you
scroll
down
a
little
bit,
I
think,
on
the
right
hand,
side
I
sort
of
outlined-
and
this
is
sort
of
like
the
question
and
my
sort
of
suggestion,
what
I
talked
about
now
and
that
whatever
we
do,
if
we
start
from
the
inside
out,
so
that
if
we
start
defining
a
service,
very
specific
deployment
file
in
in
the.
D
That
would
also
be
a
good
way
to
start,
but
even
so
then
we
would
have
to
decide
where
to
try
to
swap
out
the
install
installation,
the
file
that
is
called
install
now
that
deploys
that
application
or
how
we
want
to
do
that
and
what's
sort
of
like
you
know,
how
do
we
achieve
that?
We
are
able
to
test
console
with
the
least
amount
of
work
and
risk.
A
Fair
enough
yeah,
it's
it
seems
like
we'd
wanna
yeah
for
advance.
Actually
something
you've
already
suggested,
like
sort
of
identify
in
advance.
If
it's,
if
a
few
switch
statements
are,
will
get
us
there
or
if
and
also
understand,
hey,
what's
the
release,
time
frame
for
1.10
and
then
actually
like
at
some
point
start
talking
about
what
it
means
to
get
rid
of
cuddle
and
to
get
rid
of
that
other
tool
as
well.
D
D
Well,
we
can't
catch
up
with
abhishek.
If
you
think
it's
easily
doable,
then
then
that
will
be
sort
of
like.
I
think,
a
good,
pragmatic
solution
to
be
able
to
run
this
soon
for
console
stopgap.
Oh
yeah,
yeah.
A
So,
just
a
quick
thumbs
up
or
thumbs
down
have
a
shack
like.
Are
you
generally
positive
on
the
idea
or
not
it's
it's
a
thumbs
up.
Okay,
with
some
additional
okay,
I
mean
good
yeah.
I
just
wanted
to.
I
want
to
make
sure
we
don't
we
we
get
something
out
of
the
time
spent
so.
A
Cool
good
all
right
thumbs
up
we're
just
a
few
minutes
over.
Some
of
us
have
another
call,
and
so
thanks
for
this
michael,
I'm
glad
that
we
made
progress
because,
hopefully
we'll
hopefully
next
week's
call
on
wednesday
will
be
fruitful.
I.
A
Okay,
talk
to
you
guys
next
week
see
everybody.