►
From YouTube: Meshery Community Meeting - Oct. 4th 2019
Description
@NikhilLadha3 demo'ed the new in cluster and out of cluster (Kubernetes) deployment interface for Meshery. We're getting ready for #NetworkServiceMeshCon @nservicemesh!
A
B
A
C
A
D
It's
actually
semicolon
assumed
that
we
should
all
do
the
semicolon.
There's
a
the
come
on.
That
means
you
use
inside
the
UI
food
machine
folder
for
your
purpose,
and
it's
me
setup
your
lips
inside
that
come
on
this.
The
functions
angular
is
CD
CD,
space,
dot,
dot,
separated
by
simple,
then
three
commands
inside
then
I.
Don't
recommend
I,
give
up
I,
get
an
error
like
a
path
hot
specific,
but
when
I
copy
the
command
and
then
separately,
that
is
instead
of.
D
Actually,
I
try
to
I
try
to
come,
and
so
actually
I
have
a
topic
to
of
slack
two
days
ago.
At
that
time
we
talked
about
this
reducing
so
you
say
that
it
will
work
so
I
dragon
mm.
Well
then
come
on
from
it.
Isn't
one
so
I
try
to
run
PowerShell
then
semicolon
works
in
motion,
but
back
towards
physically
tightening
it
a
zipper,
but
when
I
have
accepted
eight
books,
that
was.
C
C
That
was
a
great
use
of
the
first
few
minutes
anyway,
because
we
just
need
to
give
people
a
couple
of
minutes
to
join,
and,
moreover,
this
excites
me,
because
I
speak
as
much
as
I
do
mostly
to
fill
in
the
air
and
provide
opportunities
for
others
to
have
ideas
and
thoughts,
and
then
chime
in
and-
and
so
this
is
great,
because
this
is
one
of
the
first
times
where
the
community
onto
its
own
is
just
doing.
Is
you
guys
are
doing
new
thing
which
yeah
tickles
me?
This
makes
me
really
excited.
C
F
C
Link
to
the
meeting
minutes,
putting
it
into
the
chat,
I
say
it
every
time
and
I
really
I
really
do
mean
it
just
like,
and
you
know
and
kill
we're
just
talking
and
and
using
the
airtime.
Please
each
of
you
go
in
and
use
the
paper
time.
If
you
will
in
Google
Docs
record
your
attendance
help
take
notes.
Last
time
we
actually
had
a
great
set
of
discussions.
I
think,
last
time
we
met
and
in
part
because
of
the
discussions
were
so
engaging.
C
We
have
not
so
great
meeting
minutes
so
you'll
help
me
where
you
can
please
and
grecian
and
everybody.
These
are
not
my
meeting
minutes.
This
is
not
my
project.
This
is
your
project.
These
are
your
meeting
minutes.
This
is
your
and
I
hope
she
super
dungeon,
Anand
and
mcil,
and
everyone
who's
on
the
call.
I
know
we
spent
a
lot
of
time
talking
this
this
week
in
our
Wednesday
call.
I
know
we
spent
a
lot
of
time
talking
about
what
it
is
that
I
hope
that
you're
taking
away
from
the
project,
and
so.
C
Developing
that
sense
of
ownership
over
the
project
is
very
healthy,
and
it's
something
that
I
hope
you
know
I
think
maybe
you're
the
only
one
on
the
call
that
hasn't
had
to
listen
to
me
go
on
and
maybe
fangy
who's
here,
which
is
awesome,
hasn't
had
to
heard
me
kind
of
pontificate
about
what
it
is
that
I'm
hopeful
for
you
guys
to
gain
and
glean
not
participating
here.
I
think
a
lot
of
that
stuff
is
probably
relatively
obvious,
but
but
I
want
it
to
be
as
much
of
a
two-way
street
as
possible.
C
Switching
tracks
going
back
to
the
meeting
minutes,
I
think
there's
I'm
trying
to
keep
the
general
announcements
that
we
have
relatively
short
because,
frankly,
there's
so
much
going
on
in
the
service
smash
ecosystem.
Would
you
sit
here
and
have
a
general
conversation
about
mushy
things,
but
I
want
to
try
to
get
to
some
specific
items.
I
measure
you
today.
If
we
can
I
think
the
general
announcements
were
last
week
we
said
Krrish
had
just
given
a
workshop.
C
C
C
C
Okay,
well,
then,
just
moving
on
so
the
by
the
way
semester
II
will
be
presented
at
NSM.
Con
I
think
we
mentioned
that
last
time,
but
as
part
of
that
presentation
is
the
fact
that
we
would
be
demonstrating
the
NSM,
adapter
and
so
her.she,
hopefully
give
listening
to
this
later.
We
need
to
move
from
beta
version
of
the
adapter
to
much
more
of
a
stable
version,
because
there's
gonna
be
people
poking
at
it
in
a
very
good
way,
and
so
she
says
she's
not
on
maybe
we'll
move
forward.
C
There's
one
thing
that
there's
a
new
community
member
that
I
don't
know
will
be
on
the
call
today
his
name's
Anton
and
some
of
you
have
probably
seen
him
he's
hoping
that
he
would
be
on
because
I
think
he's
got
a
bit
of
an
interest
around
SMI
and
I
mentioned
last
week
that
SMI.
That
community
is
asking
for
measure
e
to
be
a
conformance
tool
again
quickly.
C
Smi
as
the
service
mesh
interface
they'd,
like
for
measure
e
to
be
a
conformist
tool
to
validate
that
various
service
meshes
have
implemented
to
spec,
have
implemented
service,
mesh
interface
specification
and
that's
the
same
mask
that
open
SMI
is
asking
for
as
well,
and
we
need
volunteers
and
I,
suspect,
Anton
or
aunt
might
be
the
guy
I,
don't
know
between
him
and
mr.
Nick
Jackson
okay.
C
So
that's
one
activity,
the
other
one
is
to
just
a
healthy
reminder
that,
as
you
open
a
pull
request
to
in
the
description
of
the
pull
request,
not
in
the
pull
request
title
but
in
the
description
itself
link
but
put
in
the
the
issue
number
that
you're
addressing.
So
let's
see
if
this
first
one
has
it
just
as
an
example.
It
may
not.
C
C
C
That's
great,
although
the
Anand,
the
suggestion
would
be
as
opposed
to
putting
the
issue
number
in
the
pull
request
title
if
you
would
put
it
into
the
description
two
reasons,
one
is
what
I
just
said
about
github
automatically
closing
your
issue
when
the
PR
is
merged.
So
if
280
is
mentioned
in
the
description,
then
it
will
be
Auto
close
when
PR
306
is
merged.
The
other
reason
not
to
include
it
in
the
title
is
because
it
can
be
a
bit.
C
It
takes
a
moment
to
figure
out
what
280
is
versus,
what
306
is,
which
one
is
the
PR
number,
which
one
is
the
issue
number.
So
just
a
housekeeping
out
upset
we're
going
over.
We
talked
about
signing
of
your
commits
in
the
past
and
I
think
everyone
on
the
call
has
gotten
into
this
habit
of
signing
your
commits,
which
is
great
because
we
enforce
a
developer's
certificate
of
origin
as
we
enforce
that,
and
actually
an
ons
example
here
is
great.
C
He's
got
his
sign-off
going
for
his
commits
Suman
had
been
helping
and
dungeon
as
well
been
helping
with
enhancing
the
contributing
guide,
which,
which
does
probably
need
more
explanation
and
expansion.
At
some
point.
Maybe
we'll
move
this
contribution
guide
into
something
like
the
mesh
redox
to
just
make
it
easier
to
go
through,
because
I
think
there's
going
to
be
a
number
of
sections
that
a
number
of
different
concerns
and
questions
that
people
have.
But
gungeon.
C
That's
good
and
assume
had
added.
Oh
this
I
think
though
Superman
was
Ashish
or
I
might
have
been
I.
She
should
I,
don't
remember,
but
it
had
added.
Just
if
you're
using
Visual
Studio
code,
which
you
know
it's
a
relatively
popular
IDE,
then
you
can
have
Visual
Studio
code
automatically
sign
all
of
your
commits
for
you,
so
you
don't
even
have
to
really
think
about
it
if
you're
doing,
if
you're
using
command
line.
C
C
So
we'll
postpone
the
discussion
on
Anton's
question
about
always
deploying
SMI
when
you
deploy
sto
will
postpone
that
because
Anton
isn't
here
so
okay
with
that,
let
me
switch
back
and
say
you
know.
Fangio
are
you?
Are
you
in
a
place
where
you
might
be
able
to
speak
for
a
moment?
Oh
you
can't
do
the
intro
today,
no
problem,
okay,
I
got
it
great.
I
also
just
want
to
say
Pavan,
it's
great
to
see
you
back
on
the
call
man,
it's
good
to
have
you.
Oh.
E
C
Okay,
okay,
good
good,
okay,
let's
get
into
the
meat
of
it
so
demos
for
today
yeah,
let's,
let's
it's
been
a
while
since
we've
had
a
demo.
So
let's
let
me
stop
presenting
and
say-
and
you
know,
I
asked
if
either
gireesh
Nikhil
Anand,
both
all
of
you
I
know,
have
been
working
on
an
additional
functionality
in
ma
cherie.
It's
a
grecian.
G
G
C
C
D
That
is
the
Sun
actually
currently,
the
we
have
the
neighbors
or
kind
of
switch
on
the
environment,
which
we
can
switch
between
interest
upon
submission
and
out
of
pleasure
contradiction.
So
we
see
we
the
issuer's
to
change
it,
OH
taps
in
which,
which
we
can
switch
from
one.
After
that
we
can
configure
where
we
can
see
the.
What
is
the
consideration
currently
in
our
page
and
come
back
to
the
predicate
to
our
philosophical
tradition,
fanaticism.
G
C
I
can't
help
myself,
but
explain
it
a
little
bit
and
that's
just
because
guys
it
is
most
of
you
of
deployed
imagery.
Most
of
you
run
measure
eat
locally.
You
know
in
on
on
docker,
probably
on
your
laptop
or
something
like
that,
and
when
you
do
it
I
think
most
of
you
have
probably
experienced
measuring
the
configuration
of
memory
by
giving
it
your
cube,
convenient
giving
it
your
cute
config
file
and
that's
a
very
valid
way
of
deploying
mystery,
and
that's
the
one
that
we
present
most
prominently
in
the
documentation.
C
Having
said
that,
there
is
another
way
of
deploying
measuring
and
that
is
to
take
in
Ranma
Sri
as
an
application
inside
of
your
kubernetes
cluster.
Some
of
you
have
probably
done
that
to
date,
but
probably
most
of
you
haven't
I'm
just
just
guessing,
and
the
thing
is,
is
measure
e
needs
to
communicate
with
kubernetes
differently.
If
measure
is
running
outside
the
cluster
versus
inside
the
cluster
and
the
kills
work
here
is
helping
improve
the
user
experience
of
those
two
deployment
models,
so
something
to
kill
back
to
you,
yeah.
D
We
can
switch
it
where
this
part,
so
the
third
complication
is
that
if
the
conviction
amount
is
already-
and
if
you
change
them,
you
find
you
add
that
when
you
find
friends
you
find
it
correct,
it
is
not
completely
submit
button,
will
not
be
there
if
you
want
to,
but
if
you
detect
thank
you,
configuration
and
I'm
very
the
confirmation
videos
over
this
insight.
So
when
we
are
done
without
all
the
flesh
apart,
there'll
be
additional
interspace
and
see
what
is
the
configuration
if
in
case,
we
don't
want
to
see
change
anything.
D
Otherwise,
if
we
want
to
change
anything,
we
can
switch
to
the
outer
place
to
turn
and
we
get
uploaded.
You
find
to
see
Jesus
ready,
the
only
change
that
is
their
parents
whenever
we
don't
need
a
submit'
by
the
vector
so
that
whenever
we
upload
a
file,
it
automatically
takes
the
to
file
any
contention
and
will
be
automatically
refresh
the
page
to
show
the
connection
away.
C
A
C
That
okay,
thank
you
for
that
question,
because
I'm
willing
to
bet
you
11
out
of
12
of
you
on
the
call
that
same
question
and
and
and
shame
on
all
of
you
for
not
having
asked
it.
So
it's
a
here's,
here's
how
I
would
describe
it
as
the
the
way
that
memory
is
presented
to
most
or
the
way
that
it's
sort
of
written
up
when
people
say
I'd
like
to
run
ma
cherie
or
rather
than
they'd
like
to
understand.
My
shrink
is
that
we
didn't
really
present
them
this
architecture
and
we
try
to
say.
C
Oh,
hey
and
I
know.
This
is
kind
of
small
I
think
if
I
go
full
screen
and
you'll
lose,
but
but
each
of
you
have
seen
this
before.
This
is
just
that.
What
we
say
is
all
right.
Ma
cherie
is
a
management
plane.
It
is
intended
to
help
you
run
surface
finishes
and
interface
with
service
meshes
a
number
of
them.
It
is
comprised
of
a
set
of
containers,
it's
written
n
go
and
uses,
react
and
and
here's
what
that
architecture
looks
like
it.
C
It
generates
load
and
it
speaks
to
excuse
to
kubernetes
and
it
speaks
to
meshes.
It
has
adapters
one
one
adaptor
per
mesh,
and
so
it
has
one
container
and
each
adaptor
is
deployed
inside
of
a
container
and
measure
e
as
a
program
is
deployed
inside
its
own
containers.
So
there's
measure
e,
plus
one
adaptor
for
each
type
of
mesh
that
you
interface
with
and
the
way
we
sort
of
say.
You
know
you
when
you're
doing
that
you
can
deploy
that
just
it's
a
doctor.
C
Okay,
it
sort
of
assumes
that
you're
going
to
be
using
docker
locally,
and
so,
if
you're
running
docker,
then
just
run
these
commands,
and
you
can
start
mush
me.
That
said
in
a
smaller
way,
it
says,
or
if
you
want
to
install
mush
round
kubernetes,
you
can
do
so
by
using
by
using
manifest
files
and
deploying
missery
on
top
of
kubernetes
there
inside
of
kubernetes.
So
so
the
diagram
here
would
be
that
this
is
kubernetes
over
here.
C
You're
running
a
service
fashion
on
top
of
kubernetes
and
instead
of
measuring
being
outside
of
running
in
a
different
docker
environment.
Over
here
measure
e
would
be
running
adjacent
to
your
service
fashion.
We're
running
inside
of
kubernetes
and
I.
Don't
know
that
we're
using
the
most
easily
comprehensible
or
comprehended
terminology
inside
of
the
Metro
UI,
we're
saying
that
we're
using
the
terms
in
cluster
and
out
of
cluster,
and
that
assumes
a
lot
I,
think
that
that
terminology-
and
it's
confusing
to
me
honestly
and
then
that's
what
Nikhil
is
trying
to
help
clarify
right
now.
C
The
version
of
mystery
that
everyone
has
available
to
them
I
think
has
this
type
of
an
interface
where
you
can
flip
the
flag
between
I'm
telling
mutant
configuring,
mysteries,
saying,
hey
I'm,
running
it
inside
of
a
cluster
or
no
I'm
running
it
outside
of
the
cluster.
If
I'm
running
it
outside
of
the
cluster
I
need
to
give
you
my
cube,
config.
Well,
this
isn't
very
good
user
experience
and
akil's
trying
to
improve
okay.
Thank
you
very
much.
A
I
kind
have
that
may
be
an
issue
not
issue,
but
let's
say
why
I
didn't
understand
bucks.
For
me,
it's
like
the
path
to
reach
to
the
in
cluster.
We
are
already
running
the
sri,
so
that
was
confusing
for
me
like
what
you
explained
right
now
for
me
makes
total
sense,
but
it
means
that
measure
is
already
pre-installed
in
a
kubernetes
cluster
if
we
want
to
deploy
it
in
cooperative
right.
So
when
you
launch
the
front
page,
you
arrive,
and
he
says
like.
A
Okay,
do
you
want
like
out
of
or
in
but
this
page
for
me
like
as
a
user
right,
is
something
that
I
because
I
run
my
sri
already
I,
don't
know
if
make
sense,
you
know
like
there's
a
yeah
I,
don't
know
how
to
say
like
misplaced
or
something
everybody's
like
confusing.
To
ask
me
after
I'm
running
mystery,
I
know
what
I
want
to
do
without
yeah
totally
yep,
so
it
makes
sense.
I
mean
then
I
understand
that
that's
where
we
will
install
the
connectors
right.
C
There's
a
yes
party,
they're,
already
I'm
missing
all
the
converter.
No,
no
Kevin
you're
not
can
everyone.
C
C
C
This
status
information
should
be
moved
over
to
the
settings
page
to
say
this
is
how
this
is
the
status
of
your
current
configuration.
That's
not
really
what
we're
wanting
to
present
on
the
dashboard.
The
dashboard
will
ultimately
present
the
fact
that
you're
connected
to
octarine
and
octarine
instance,
and
an
sto
instance
and
there's
five
services
inside
of
octarine
and
two
inside
of
Sto,
and
it
would
tell
you
more
about
your
applications
more
about
your
the
health
of
your
service
mesh's,
not
about
the
health
of
measure
e.
C
C
C
The
goal
here-
and
this
will
be
accomplished
very
shortly-
is
that
when
you
deploy
measure
e
inside
of
kubernetes
measure,
he's
just
gonna
know
that,
and
you
don't
have
to
tell
mr.
e
that
it
says
so
that
yeah
as
a
user,
you
don't
have
to
come
over
and
in
a
confusing
way
like
to
tell
mr.
e
something
that
it
should
already
know
to.
When
you
deploy
missery
on
your
laptop,
you
know
outside
of
kubernetes,
just
in
a
simple
docker
environment.
C
C
If
you
will
make
it
such
that
you
as
a
user
really
shouldn't,
have
to
come
over
and
answer
these
questions
in
the
screen
here
and
but
if
you
do,
if
you
want
to
use
a
different
cute,
config
file
or
if
you're
deployed
in
cluster
and
for
some
reason,
Mestre
didn't
understand
that
correctly
or
you're
deployed
in
cluster
and
Mesurier
sims
that
it
should
be
using
that
you
want
to
interface
with
the
default
namespace
or
something
like
that.
But
you
can
still
go
to
the
these
screens
and
and
change
part
of
where
mystery
is
pointed.
A
But
I
think
I
think
you
you,
you
say
the
phrase
here
and
I
think
that's
where
we
are
not
aligned
and
I
understand
for
me:
docker
desktop
because
it
contains
today
and
kubernetes
language
server
with
it.
That's
why
I
quote
like
already
publish
measure
e
to
kubernetes,
even
though
I'm
running
on
the
laptop
so
for
me,
I'm,
starting
with
it
before
I
reach.
This
page
right.
A
So
for
me
like
it's
like
the
out
of
out
of
cluster,
we'll
always
be
the
preferred
solution,
title
frame
when
we
are
running
doctor
for
doctor
desktop,
because
then
we
will
pick
up
the
cube
config
file
that
points
to
the
running
kubernetes
also
on
my
laptop,
has
added
water.
But
then,
if
you
are
running
a
real.
F
A
Or
a
real
kubernetes
cluster
may
be
on
edge
AWS.
Whatever
on
your
own
premises,
then
you
are
pointing
you
that
this
out
of
cluster
is
when
we
are
started
measuring
through
dr.
desktop,
let's
say,
but
then
you
are
pointing
out
to
your
kubernetes
cluster,
so
you
are
making
like
the
separation
between
doctor
each
cells
and
the
kubernetes,
where
I
as
a
user.
A
Again,
really
from
a
point
of
view
of
pure
user,
oh
yeah,
so
so
it
was.
It
was
just
funny
because
for
me
it
was
like
in
closer
and
I
was
like
what
marina
was
his
own
Pig
I,
don't
know
teacher
yes,
cluster
embedded,
and
so
that's
why
I
can
like
do
in
cluster
things,
and
actually
that's
not
the
case.
So,
okay,
thank
you
very
much.
I
mean
separation.
The
migraine
tactics
is.
C
Anyone
and
I'll
put
this
out,
for
everyone
is
I.
Don't
in
this.
This
is
confusing.
To
me
is
why
I
was
asking
for
it
to
be
changed
and
I.
Don't
know
that
the
terminology
that
we're
even
moving
to,
which
is
terminology
like
this
out
of
cluster
in
cluster,
that
still
is
confusing.
I
think
this
is
a
point
in
time.
Iteration
of
an
experience
that
needs
to
be
continues
to
continue
to
be
enhanced
and
may
need
to
be
enhanced
with
visuals,
because
our
because
the
world
in
which
we
live
in
is
very
flexible.
C
Awesome:
okay,
that's
that
was
thinking
to
kill
any
other
feedback
from
the
kill.
It's
not
very
sure
of
that.
G
Yes,
so
they
have
been
humming,
so
they
awesome.
Yeah
I
mean
like
my
headset
has
been
like.
You
know
it's,
it's
I
think
my
computer,
you
know
it
just
keeps
dropping.
Sometimes
it
is
so
one
of
the
things
we've
been
working
on
is
so.
There
are
two
issues
that
you
know.
One
is,
if
you
guys
know
we
have
this
capability
to
actually
pick
an
adapter.
That's
right
now
reckon
I
might
run
it.
I
have
no
adapters
connected.
That's
intentional!
It's
my
development
apartment
sometime
tending
to
run
the.
G
I'm
not
using
them,
so
so
people
knew
about
this
earlier,
but
then
mentioned.
We
don't
have
this
capability
for
kubernetes
like
when
you
hover
over
them
like
and
I
will
actually
show
you
the
server.
But
beyond
that
we
do
not
have
much.
So
then
one
of
the
things
that
you
know
I'm
working
on
is
actually
to
bring
in
the
capability
to
ping
a
kubernetes
adapter.
G
So
now
that,
like
no
sorry
pink
pink
over,
ladies
sorry
I'm
her
doctor
so
earlier,
the
way
it
was
was
measure.
Ii
was
kind
of
a
very
thin
layer.
It
was
this.
Was
all
these
relying
on
the
adapter
for
cooking
with
great
ease
now
in
order
to
go
one
step
further.
What
we
are
working
on
now
is
to
make
merry
talk
to
go
varieties,
so
we
can
actually
get
some
metadata
now.
This
has
two
advantages
like
I.
G
Don't
talk
to
each
them
on
today,
so
the
first
thing
is:
actually
they
capability
to
ping
and
ensure
that,
but
you
know
indeed
like
then
you're
able
to
talk
to
ponies
now
what
it
returns
is
actually
right
now,
I
can
what
I'm
doing
is
I'm
just
talking
to
the
ladies
to
actually
get
a
list
of
nodes
and
to
understand
the
underlying
infrastructure,
no
one,
if
that's
a
common
function,
I've
written
and
I'm
going
to
be
using
two
places.
This
is
one
eventually.
G
F
G
Hopefully,
they're
gonna
next
week.
You
know
this
feature
would
be
in
a
better
state
on.
It
will
be
landed
right
now.
It's
just
I,
it's
just
on
the
local
and
worked
on
it.
So
this
is
the
first
part
which
actually
talks
to
kubernetes
ensures
that
the
connectivity
is
fine
and
you
know
we're
all
good.
So
that's
the
first
part.
The
second
part
is
that
I
hope.
Like
you
know,
people
on
this
call
know
about
the
about
the
source.
Mesh.
Spec
are
the.
G
Let
me
just
try
to
bring
it
up
here.
I
forgot
the
right
name.
It's
called
a
github
lair,
it's
the
ax
steel
benchmark,
spec,
yes,
yes,
the
sub
mesh
benchmark,
spec
I'm,
not
sure
I'd,
like
everybody
on
this
call
know
about
this.
So
to
give
you
a
highlight,
so
here's
a
spec
you
can
find
this
under
layer,
5
I/o,
so
this
mesh
benchmark
spec,
it
is
open,
so
everybody
in
the
sense
are
gonna
be,
are,
you
know,
be
really
happy
to
collaborate?
Click
like
on
all
the
people
who
are
interested.
G
One
of
the
things
we're
working
on
is
measuring
is
conducting
low
tests
on
your
behalf
on
applications,
but
at
the
moment
it
is
only
recording
in
client,
site
metrics
and
if
you
have
connected
measuring
tool
for
media's
instance,
it'll
also
collect
some
server-side
metrics
just
more
at
a
high
level
cluster
wide
metrics
beyond
that,
like
no
we're
not
doing
much,
because
you
know
all
those
metrics
which
are
the
duration
of
the
test,
recognize
like
some
payload
by
itself.
Now,
apart
from
that
now
one
of
the
other
interesting
things
is
the
capability
to
record
or.
G
It
should
gain
ground
on
which
the
test
is
being
spoke.
What
alone,
like
you
know
you,
for
example,
when
you're
on
a
test?
Now,
where
is
this
many
run
another
test
sometime
in
future
or
maybe
sometimes
pass?
You
can
only
compare
the
results
and
see
how
the
two
applications
or
how
the
tours
of
the
applications
were
found
by
looking
at
like
I
know,
there
are
client
set
metrics,
which
also
latency
and
throughput.
We
are
still
working
on
person
off
the
sauce
Crick's,
but
then
the
other
thing
that's
missing
is
information.
G
Was
the
metadata
will
denote
legend.
So
here
is
an
example
looking
on
something
like,
so
you
know,
every
time
a
test
is
run,
we
are
recording
the
results.
We
also
want
to
know
the
worst
of
coronaries,
on
which
that's
what's
wrong.
I
had
some
node
information
like,
for
example,
here
is
an
example.
Again
we
are
talking
five
outs
and
each
node
is
of
type
like
an
232
and
it
has
32
cores
and
it
has
like
not
as
much
memory.
So
this
kind
of
information
is
definitely
good
because
you
know
when
you're,
comparing
multiple
test
results.
G
It
also,
you
know,
gives
some
more
visibility
to
the
test
as
to
okay,
what
kind
of
an
environment
was
this
one,
because
it
might
not
be
an
apples
to
apples
comparison
like
if
you
compare
test
results
when
one
is
run
on
a
production
quality
cluster,
while
the
other
one
is
on
somebody's
lap
top.
So
in
order
to
make
you
know,
you're,
not
a
brief
clarity
that
they
really
can.
We
are
now
entering
into
the
area
where
we
are
trying
to
collect
more
metadata
from
the
dog.
G
So
one
of
the
things
like
you
know
we
did
as
far
as
this.
This
week's
development
is
actually
to
collect
the
Nord
metrics.
So
here's
an
example
of
a
test
run
on
Tehran.
Please
ignore
look
at
what
it's
really
doing,
but
then,
like
you
know,
the
crux
of
it
is
this
is
a
part
looking
at
which
essentially
is
like
you
know
the
the
plants
that
mercury
flight
mode
shows.
G
You
know
how
many
connections
with
you
for
the
dance,
how
long
the
test
went
where
their
errors,
the
many
generation
average
duration,
be
50,
75,
99
and
so
on.
With
all
the
percentile
values
are
here,
I
mean
this
is
probably
not
the
final.
What
I
was
trying
to
do
is
let
you
know
so
after
the
test
was
wrong.
I
was
also
trying
to
get
the
node
information
and
and
take
care
so
I'm.
One
thing
there
are
two
sites
or
little
one.
G
The
node
information
is
going
to
persist
it
along
with
the
test
results
now
whether
to
show
it
here
or
not.
You
know
it's
I'm
still
debating
it
with.
You
know
why
you
know
after
I,
finalize
I
will
definitely
bring
open
it
up
for
discussion
for
sure,
but
here
this
is
just
a
place
for
me:
I'm
testing
it
so
I
thought
that
that's
a
good
place
for
me
to
see
it
later
on
before
I
get
that
food
on
anything
so
I'm
just
displaying
it
here.
G
For
now
you
can
see
if
I
go
on
so
that
node,
you
know
I'm
using
for
testing
here
is
actually
a
sorry
there.
The
kubernetes
I'm
using
for
testing
there's
a
single
node
kubernetes.
It
just
has
one
node,
and
here
is
a
host
name.
I,
hear
that
I'm
for
CPUs,
here's
a
memory,
the
architecture,
the
operating
system,
the
kubernetes
worship,
the
propeller
run
pen
use.
G
So
so
this
is
kind
of
our
first
step,
but
gonna
work
your
twenty
kind
of
record
some
information
of
some
node
metrics,
so
that
we
can
eventually
kind
of
you
know
populate
it
and
give
it
to
a
format
like
you
know,
which
is
kind
of
standard,
and
this
is
the
like
the
format
you
know
we
as
part
of
layer.
Five
are
actually
working
towards.
Of
course,
this
is
not
yet
finalized.
I
can
we
are
your?
G
We
are.
Oh,
it's
open
for
contributions
from
everybody,
so
please
feel
free
to
comment
whatever
you
feel
like,
and
please
suggest
like
no
changes
to
respect
so
eventually
water.
The
way
it'll
be.
Is
that
make
sure
you
manage,
recording
the
results
it
will
actually
be
able
to.
You
know
present
the
results
in
this
format,
so
people
can
actually
even
download
them
and
use
it
for
their
offline
comparison
like
if,
if
that
would
help
and
again
you
know,
I
think
we
have
already
been
talking
to
follow
the
use
of
Mestre.
G
You
know
as
a
tool
in
several
places.
So
in
those
places
when
people
are
running
tests,
such
a
common
format
would
actually
be
enforcing,
so
we
are
marching
towards
that.
So
so
that's
the
two
parts
which
I
wanted
to
present
one
is
collecting
the
node
metrics
and
pretty
much
like
you
know,
using
that
to
ping
kubernetes,
like
you
know,
to
ensure
that
connectivity
from
less
readable,
Renesas,
the
okay,
that's
pretty
much
what
I
have
to
show
to
you
guys
today.
G
I
Hey,
hey,
sorry,
I'm
starting
on
the
video
and
I
could
not
get
enough
of
this
is
mute.
I
guess
it
was
from
the
previous
collection
li
were
spinning
water
in
and
out
of
cluster
things
Oh
when
you're
running
on
a
stock.
Currently
we
deployed
the
few
minutes
to
move,
make
sure
you
will
not
be
on
a
management
place
for
the
other
Cuban.
It
is
right.
C
Great
Ramesh
great
question:
I'm
gonna.
Let
sorry
does
it
measure
ease
it.
G
Yep
sorry
I
just
don't
put
my
answer
there,
so
my
sri
is
not
going
to
be
managing
we're
going
to
be
slick.
You
know.
Eventually,
the
plan
is
that
I
mean
even
know
like
you
can
actually
private
work
with
your
selfish,
like
sto
or
liquidy
like
the
front
ma
cherie.
So
so
it
is
going
to
be
the
management
plate
for
the
fish
sugar
sit
on
committees
or
not
is
actually
you
know
kind
of
I.
Think
it's
more
of
a
convenience.
G
G
I
mean
I'm
not
going
to
go
into
the
networking
specification
there,
but
that
our
network
parts
there,
but
then
the
point
I'm
trying
to
make
is
that,
like
you
know,
make
sure
he
gets
it
anywhere
as
long
as
it
can
talk
to
the
coordinators,
API
server
like
I
I,
think
like
I,
can
do
whatever
it
needs
to
do
like
you
know,
with
respect
to
managing
so
specialist
working
with
them.
You
know
talking
to
kubernetes.
G
I
G
Mean
so
fundamentally
so
so
this
question
is
not
actually
I
would
say:
it's
not
specifically
a
measuring
question.
It's
specifically
a
kubernetes
question.
So
so
sorry
I'll
just
talk
about
nice
release
capability
in
the
sense
that
if
codenys
can
do
it,
you
can
do
it
with
mushroom
I.
Think
that
probably
would
answer
your
question.
I
know
it's
not
the
complete
answer.
You
know,
injecting
injecting
stuff
into
a
running
container
I'm,
not
sure
who
do
that.
The
way,
for
example,
is
do
does
it
is
like
a
little
inject.
G
The
site
carjacker
will
actually
spin
up
another
container
in
the
safe
Park
when
it's
bringing
it
up
right
now,
and
it
has
its
way
of
doing
it.
Like
you
know,
with
with
a
lot
of
thought
things
like
I'm.
Just
talking
about
the
end
result,
I
mean
the
progress
to
get
that
money.
It's
still
like,
there's
been
quite
a
bit
like
I
missed
you,
that's
all
that
so
yeah
style
that
is
do
is
again
morning.
I
mean
at
least
today,
like
I
still
runs.
Well,
are
kubernetes,
so
yeah,
so
I
don't
agree.
G
I
C
C
D
C
Yeah
good
question
so
to
just
give
context,
the
issue
that
Gunjan
is
working
on
is
I.
Think
it's
really
it's
a
really
neat
issue,
and
it
is
the
notion
that
each
time
that
someone
runs
a
performance
test,
Kurush
was
talking
about
the
notion
that
there's
just
the
anonymous
statistics
that
are
collected
about
the
performance
test
and
and
part
of
the
hopefully
public
service
that
measure
will
perform
is
to
over
time
analyze.
Those
performance
test,
results
and
begin
to
inform
the
world
or
inform
those
interested
in
service
matches
about.
C
C
C
But
let
me
let
me
give
you
guys
the
right
link
so
sorry
wrong
wrong
with
the
the
second
link
in
the
chat
is
to
the
issue
that
comes
in
is
going
over.
So
there's
a
public-facing
endpoint.
It's
very
simple
in
nature.
It's
this
URL
here!
If
we
were
to
go
to
that
URL.
All
that
is
returned
is
a
very
simple
JSON
key
value
pair
and
the
key
is
called
total
runs,
I
guess,
I
guess:
I
lied,
it's
not
500!
C
It's
four
hundred
and
fifty
five
total
performance
test
tests
that
people
have
run
and
what
we'd
like
to
do
is
let
people
know
how
many
hamburgers
we've
sold
you
kind
of
like
the
McDonald's
sign.
If
you
guys
are
familiar
with
it,
McDonald's
says
you
know
billions
and
billions
of
hamburgers
sold
and
we'd
like
to
let
people
know
how
many
performance
tests.
C
C
How
good
okay,
I
man
I'm
always
excited
Nino
was
on
the
call
earlier
you
get
a
hymen.
Did
you
have
it?
Were
you
gonna
say
something
or
mm-hmm?
Oh
I
thought
you
might
have.
You
were
about
I
thought
you
were
gonna,
say
something
and
so
I
know.
Oh,
oh
good,
good,
good!
Okay,
all
right!
Yeah!
Thanks
up
very
good
to
see
you
guys
see
you
next
week.