►
From YouTube: Layer5 Community Meeting (Feb 15th, 2019)
Description
Recording of weekly Meshery community meeting. See https://layer5.io/meshery for more project details.
A
Oh,
very
good,
all
right
we're
five
after
gentlemen
today
today
we
are
recording
so
just
like
many
other
open-source
communities.
We
record
the
meeting
minutes
and
then
post
on
the
YouTube
channel
for
those
that
might
want
to
catch
up
later.
So
look
I
just
want
make
everyone
aware.
So
you
know
keep
your
keep
your
curse
words
to
yourself.
Unless,
unless
you
want
mom
to
hear
those,
that's
fine
and
then
like
I
was
telling
pronounced,
I
mean
I.
Will
we
can
talk
about
how
to
cheat
on
final
exams?
You
know
later,
but
just
like.
A
A
It's
in
the
measure
e
slack
channel,
if
it's
not
handy
for
you,
I
do
want
to
welcome
new
members
today,
although
I
don't
think
we
have
any
that
have
showed
up
just
yet.
No
one
knew
other
than
I.
Think
there's
a
couple
of
folks
that
haven't
met
each
other
before
so
one.
You
do
go
to
record
your
attendance
today
and
then
to
allow
me
to
introduce
yogi
Paula
who's
on
the
call
now
to
pranaam
and
critique
yogi.
A
A
A
I'm
gonna,
you
know
what
I
was
going
to
request
for
topics,
but
I
suspect
that
that
we
don't
we
don't
need
to
I,
think
I
know
a
couple
of
the
topics
give
him
the
on
today's
audience
that
we
need
to.
We
need
to
cover
so,
let's
jump
in
one
I
want
to
bring
the
room
up
to
speed
with
a
couple
of
conversations
have
been
happening
in
the
community
about
the
mystery.
A
A
Benchmark
results
specification,
it's
not
it's
not
even
it
is
a
performance
benchmark
specification,
not
necessarily
results,
as
well
as
capture
the
definition
of
the
environment
and
the
type
of
performance
test
that
was
run.
I
know,
we've
talked
about
this
in
the
past,
with
a
few
of
you
is
everyone?
A
A
A
Agnostic
way
would
describe
the
details
of
the
tests,
that's
being
run
so
the
environment
details
the
type
of
nodes
that
there
are
they
my
memory,
that's
you
know
a
lot
of
those
variables
and
then
describe
the
type
test
to
be
run,
and
so
this
is
just
an
initial
draft
on
what
that
spec
looks
like
the
type
of
tests
to
be
run
and
actually
I've.
Seen
now
that
the
there's
a
couple
of
improvements
need
to
be
done,
we
need
to
categorize
environmental
concerns,
the
types
of
tests
that
are
being
run
and
described
those
and
then.
A
Lastly,
the
test
results
themselves.
I
think
that's
been
confused
a
little
bit
in
the
spec,
that's
there,
but
the
intention
is
for
measure
ease
you
probably
both
ingest
and
read
from
this
specification
as
well
as
when
it's
done
we're
doing
test
that
needs
to
spit
out
its
results
to
spit
it
out
in
this
format.
This
being
the
format
that
we
would
use,
then
to
facilitate
the
apples
to
apple
the
cross.
Comparison,
literary
sharing,
your
screen,
no
I,
guess
so
yeah,
and
here
here,
I'm
sitting
here
talking
over
the
phone.
A
A
Well,
at
some
point
we
will
get
maybe
Google
meat
meats,
not
hangouts,
yeah,
okay,
okay,
hopefully
everyone's
singing
service
mesh
benchmark,
spec,
yeah
yeah.
This
is
a
repo.
That's
out
there,
just
adjacent
to
measure
e,
so
service
measure
x,
mark
spec,
a
very
simple
description.
You
know
kind
of
what
we're
trying
to
accomplish
as
I.
Why,
since.
A
A
You
know,
hopefully
hashey
corp,
that
a
console
connect
folks
will
in
fact
send
their
developer
advocate
to
participate,
but
part
of
what
I
wanted
to
update
today
was
that
I
think
near
as
I
can
tell
both
Mandar
and
Syria
are
earnestly
wanting
to
engage
on
that
spec.
So
that's
good
at
mesh
team
I
think
I've
described
it.
I've
met
with
them
a
couple
of
times.
I
think
we
just
need
to
get
them.
Some
specific
asks.
The
Black
team
is
gung-ho
about
participating
here,
and
so
we
just
again
need
to
get
them.
A
B
Haley
yesterday
we
have
to
check
carefully
and
not
buy
some
others
when
it
comes
to
the
benchmarking
right
so
like.
How
do
we
do
that?
How
do
we
check
that
line
carefully?.
A
Yeah
I
think
as
we
go
to
so
for
some
of
the
numbers
they
will
they're
factual
and
will
speak
to
themselves
or
speak
for
themselves.
That
said,
if
we've
not
accounted
for
the
appropriate
configuration
of
them
of
that
particular
mesh,
then
we
may
end
up
making
a
full
mockery
of
this
project
and
of
the
project
that
we're
testing
so
I.
Think
in
that
respect,
even
though
numbers
are
kind
of
cold
hard
facts,
if
if
the
environment
was
poorly
configured
or
wasn't
tuned
in
the
right
way,
then
we
would
be
presenting
misleading
information.
A
So
we
got
to
make
sure
that
we
do
that
correctly.
Those
are
good
notes
to
take
as
well
and
agree
if
you've
got
three
fingers
and
you
can
take
these.
The
second
component
to
that
is
with
respect
to
the
landscape,
where
we're
identifying
all
the
the
players
in
the
you
know,
the
offerings
in
the
space
I
think
one
thing
to
lean
into
as
we
go
right
now.
It's
a
listing
of
all
of
the
all
the
players
in
the
landscape
and
that's
informative,
but
it's
not
super
helpful.
A
Try
to
do
right
by
all
projects
involved,
you
know
and
actually
provide
useful
information
is
probably
to
highlight.
You
know,
focus
on
the
areas
that
the
individual
projects
do
well
speak
to
those
you
know,
review
those
and
to
the
the
areas
that
they
don't
do.
Well,
you
know
they
need
to.
If
it's
give
me
valuable
feedback,
it
may
need
up
I'll,
be
you
know,
be
forthright
and
and
I
didn't
your
highlight
those,
but
maybe
now
it's
prominently
is
the
value
that
they're,
providing
that
makes
sense.
I
mean
yeah.
A
B
What
would
it
make
sense
to
actually
delegate
the
confirmation
part
basically
by
not
to
these
individual
lenders,
and
then
we
could
give
them
a
static
configuration,
but
the
memory
consumption
or
at
least
the
memory
CPU
etc.
Then,
maybe
ask
them
to
provide
that
configuration
or
maybe
work
with
them
to
compare
this
up
or
to
bring
up
this
configuration
right.
So.
A
Yeah
exactly
yeah
I
know
you
that's
exactly
right
with
each
of
the
vendors
that
we've
engaged
with
each
of
the
projects.
That
has
been
my
explicit
ask
of
them,
and
it's
not
in
fact
I'll
paste
and
copy
that
or
copy
paste
that
into
here,
which
has
been
your
blessing
on
the
configuration
to
make
sure
that,
because
they
know
best
how
their
mesh
be
deployed
in
configured,
I
think.
B
Bowen
woman
would
be
ideal
for
this
because
he's
been
doing
some
configuration
so
maybe
pulling
him
in
to
this
color
I'll.
Try
to
see
helping
him
to
see
is
available
right
now.
If
so,
you
can
be
pulled
in,
but
it
would
be
ideal
for
him
to
go
and
Eric
is
get
him
on
board.
With
this.
B
A
In
terms
of
the
configuration
of
using
okay,
so
Preity
can't
pronounce,
does
that
any
other
comments,
questions
towards
the
performance,
the
specification.
D
Regarding
go,
but
you
know
he
was
saying
we
can
also
like
pick
up
the
sample
apps,
which
most
of
these
printers
provide,
because
that,
in
some
way
is,
is
the
best-case
scenario,
probably
what
they
have.
So
we
can
try
to
port
those
apps
into
different
measures
and
see
how
that
particular
app
works
well
with
other
measures,
or
does
it
do
best
with
the
provided
mesh
provider.
A
Okay
and
then
young
yo,
you
can
attach
it
the
note
about
the
landscape
of
how
to
be
sensitive
to
folks.
I
did
speak
with
octarine
a
little
bit
earlier
today.
Another
player
who
you
know
earnestly
wants
to
participate
here
as
well.
We
got
a
message
from
OpenShift
from
Red
Hat
earlier
today
about
their
involvement
in
the
landscape
and
so
a
lots
of
good.
A
Lots
of
good
contributions
as
a
matter
fact,
a
different
part
of
Red
Hat
contributed
last
week.
They
put
key
Olli
on
there.
So
one
note
there
was
the
East
EO
technical
oversight
community
call
earlier
today,
they'd
highlighted
asked
the
mesh
ask
the
mesh's
vet,
which
is
a
tool
that
we
have
inside
the
landscape
as
potentially
being
incorporated
into
isseo
itself.
I
know
that
that
was
something
that
we
had
discussed
or
that
have
been
highlighted
as
a
possibility
within
Missouri.
C
E
A
E
You
are
absolutely
right
in
the
sense
that,
like
you
know
about
to
like
last
week,
you
are
actually
using
concrete
Bustillo
types,
but
in
the
last
week,
like
any
change,
which
I
may
that's
actually
more
away
from
concrete
types,
which
means
that
yeah,
the
use
of
the
steel
might
actually
not
be
the
best
or
much
like
us.
Looking
at
this
point
in
time,.
A
B
A
Pavan,
hey,
hey!
You
did
great
good
to
meet
you
quick,
disclaimer,
yeah
nice
to
meet
you
as
well,
quick
disclaimer,
just
because
I
had
to
tell
petit
to
stop
swearing
earlier.
We
are
recording
the
meeting
and
a
Soviet
puppet
hippy
published
publicly
later
I
said
just
that
yeah
I
know
we're
happy
to
have
you
good
to
meet
the
some
of
the
community
contributors.
E
Are
introducing
yourselves
Colin
I
bring
up
the
mystery.
Please
go
over
to
okay,
no
I
just
get
started.
The
hippo
mean
like
in
a
longish,
probably
yeah.
You
probably
would
have
seen
an
email
from
yogi.
We
produced
us
for
thread
so
yeah
great
to
have
you
here,
I
hope,
like
I,
and
actually
doing
this
back.
We
already
started
really
getting
past
us
to
you,
Thank
You
yogi,
for
actually
you
guys
should
see
your
name
in
the
meeting.
That's
already
so
you
know
so
I
think
you'll
have
fun
here.
E
All
right,
okay,
so
I
can
you
can
see
on
screen
I
try
to
share
a
browser
window.
Yes,
the
Firefox,
like
you
know,
okay
cool,
so
I
know
there
are
several
new
people
on
the
couple,
so
maybe
I
think
it
would
be
wise
to
actually
start
the
demo
over
like
it.
On
last
week
was
the
first
week
like
I'm
gonna,
actually
try
the
demo
make.
You
know,
functional
Mestre
with
all
the
confidence
to
talk
about
that,
like
I
mean
I'm,
not
gonna,
talk
about
that
is
a
pretty
today.
E
We
have
some
other
topics
to
cover,
so
I'm,
just
gonna
touch
very
briefly
on
some
recent
changes.
You
know
probably
layer
file,
IO
is
actually
our
main
site
and
inside
that
like
no,
there
is
a
mystery
section.
Once
you
go
to
the
site,
you
will
be
able
to
see
it
now
in
the
mystery
section
like
now,
we
have
a
snapshot
and
we
have
the
architecture
diagram
now
in
the
architecture
diagram
again,
like
you
know,
this
is
I
would
say
like
massive
today
like
no.
E
It's
not
yet
hosted,
so
you
know
we
will
be
hosting
that
eventually,
and
you
know
I
think
yogi,
it's
already
like,
given
your
name,
but
you
know
for
helping
us
with
define
that.
So
you
know
hopefully
like
I'm,
very
some
like
numbers,
how
many
SAS
a
deep
fly
somewhere
on
the
club
and
the
Mystery
Science.
Has
a
persistence?
Confident-Looking
sense
it.
You
know
it
has
a
capability
to
persist.
Some
result,
data
that's
sent
to
it
plus
the
mystery
sauce
is
actually
amazing,
which
helps
with
single
sign-on.
E
At
the
moment
later,
we
have
single
sign-on
with
a
few
providers,
so
I
can
I
have,
but
for
this
demo,
I'll
be
showing
Twitter,
a
tester
could
get
up,
but
the
rest
like
I
have
interested
because
you
know
we
have
to
create
a
token,
a
key
and
a
secret
time,
so
I'm
at
the
moment,
like
I'm,
just
leaving
it
off
with
Twitter
and
now
the
other
major
component
is
a
mystery
app
or
the
mesh
we
app
is
actually
sort
of
with
you
can
think
of
it,
like
you
know,
I
see
as
the
controller
for
memory
in
this
context,
because
it
has
a
uie.
E
It
has
the
app,
but
you
know
which
takes
care
of
the
user
sessions.
It
has
capable
of
collecting
the
results,
sending
it
to
the
mesh
resize,
confident
and
also
it
is
a
piece
which
will
be
working
with
dimensionally
adapters,
which
essentially
will
be
the
pieces
that
will
be
talking
to
the
appropriate
precious
like
so
for
examples
for
sto
will
have
one
mesh
readapt,
which
we
already
have
today.
Eventually
you
have
one
mishra
dr.
Foley
31
for
sofa
man,
ashore
and
so
on
and
so
forth.
E
E
So,
as
so
I've
created
some
changes
to
this
document
where
I
pray,
you
don't
kind
of
list,
the
confidence
in
my
Sri
and
cry
makes
what
they
do,
like
you
know,
pretty
briefly,
I'm
not
trying
to
pull
it
on
extensive
details,
while
I
try
to
appeal
an
overview
offer.
You
know
what
the
compliments
does.
So,
hopefully
they
got.
One
Sally
merges
that
PR
this
page.
Without
a
little
more
information-
oh
you
know,
which
might
that
is
some
more
context.
Oh
jump
me
back
to
the
demo,
so
you
can
see
the
URL
might
restart
again.
E
Like
I
mean
this
is
supposed
to
be
hosted,
but
now
it's
just
running
on
local
and
again
we
have
a.
We
have
put
together
a
dr.
Campos
file,
so
with
the
one
command
like
you
can
actually
get
the
entire
infrastructure
off,
just
very
handy,
that
is
in
the
mesh
we
github
repo
now
so
this
is
the
single
sign-on
page,
like
I'm,
sorry
I,
sign
in
with
Twitter,
because
I've
already
signed
on
in
this
browser.
E
So
it
will
ask
me
to
authorize
the
app
all
right
after
authorization
I'll
be
taken
back
to
my
Sri,
and
you
can
see
here,
I'm
logged
in
on
the
right
top
now,
in
other
browsers,
you
probably
would
see
a
and
I
with
our
I've
been
in
Firefox.
You
know,
I
have
a
security
setting
which
kind
of
prevents
it
from
loading
external
images,
so
which
is
reason
why
that's
not
shown
here.
There
are
few
important
deals,
context
name
to
support
in
these
contexts.
E
Here
we
have
to
provide
the
config
file
or
if
we
have,
if
the
instance
of
measure
is
running
alongside
on
the
same
cluster
of
kubernetes,
then
you
can
actually
opt
in
to
choose
the
in
cluster
contact
now.
The
other
thing
that's
needed
is
the
mission
after
location,
so
added
I'm
sure
we
can
work
with
on
a
Napper,
but
you
can
always
come
back
and
like
it
also
and
change
it,
which
means
that
you
don't
have
to
restart
battery
or
do
like
I
mean
everything
is
like
dynamic.
E
E
So
at
this
point,
if
you
anchor
things
currently
and
hit
enter,
what
measure
does
is
in
the
background
it
actually
will
pass
on
the
configuration
to
the
measure
adapter
and
if
measure
adapter
is
able
the
cluster
an
established
chef
direction,
then
it
will
able
to
take
us
to
the
next
phase,
always
with
Sean
header,
on
this
page.
So
you
can
see
like
now
at
this
point
that
coronal
header,
so
which
means
that
you
know
the
connection
is
established.
E
Basically,
the
test
run
and
once
the
tests
you
know
runs
and
the
results
of
back
to
the
browser
it'll.
Actually
you
know
kind
of
charted
out
in
this
graph
and
at
the
same
time,
we're
gonna
like
invention
and
architecture.
Diagrams
results
gets
into
the
machine
SAS
which
persisted
in
time,
oh
and
again,
for
my
local
test,
I'm
using
a
local
diamond
Institute.
So
so
that's
what
the
you
know
the
load
test
part
of
actually
is
as
of
today.
Now
this
is
a
part
which
is
kind
of
the
interesting
part.
E
You
know
where
this
section,
like
you
know,
lets
us
play
with
the
with
the
mesh
or
coordinate
is
like
not
at
this
point
in
time
and
like
last
week,
I
have
like
you
know
more
rules
because
they
can
open
the
last
week
and
all
the
rules
are
very
specific
to
an
application.
But
this
tweak,
one
of
the
main
changes,
was
actually
to
move
away
from
specific,
are
specific
rules
which
are
tied
too
tight
down
to
the
app.
So
the
rules
right
now,
like
I,
mean
like
on
the
install,
is
still
was
actually
one.
E
What
you've
been
allotted
he's
not
here,
but
I'm
I'm,
not
sure
if
it'll
work,
but
the
conflict
which
he
has
put
in
the
rebel
or
repose
actually
put
one
so
I'm
kind
of
left
it
there,
the
other
two
ones,
a
little
circuit,
breaker
conflict,
one
connection,
all
that's
a
generic
one,
so
you
know
it'll
it'll
change
the
circuit
breaker
sitting
across
the
cluster.
So
you
know
that's
a
good
candidate
to
actually
have
now
and-
and
the
next
thing
which
was
done
or
accomplished
in
the
last
few
days,
is
actually
this
custom
option.
E
When
you
click
on
discuss
tamarkan,
it
will
actually
show
this
open.
This
section
now
I
know
it's
black
white
video
when
you
click
on
it
like
you'll,
actually
see
a
editor.
Oh,
this
is
actually
a
Yama
literature.
Now
one
thing
to
confess
that
I
have
linking
turned
on,
but
for
some
reason
like
no,
it
doesn't
yet
work,
but
the
libraries
up
there,
the
methods
are
there,
so
I'm
still
like
figuring
it
out,
but
but
that's
not
going
to
stop
us
from
this
demo
now.
E
The
good
thing
with
this
is
that
we
can
actually
take
any
kubernetes
Yammer
and
then
you
know
apply
it
on
the
cluster
or
if
it's
a
delete
operation
you
can
actually
delete
it
or
if
you
want
to
make
an
update,
an
existing
configuration,
you
can
actually
do
the
same.
So
the
back
end
will
actually
kind
of,
like
you
know,
intelligently
understand
that,
like
knowledge,
a
particular
thing
that
already
exists
and
appropriately
it'll
actually
updated.
Now
this
was
actually
one
of
the
major
changes
in
the
last
week.
E
E
This
is
actually
made
possible.
So
what
do
you
lose?
It
will
dynamically,
you
know,
construct
the
types
and
it
will
actually
pass
on
the
conflict
to
the
cluster
and
work
with
the
cluster
pre
round.
So
it
was
a.
It
was
an
interesting
project
to
the
last
week,
and
you
know
we
haven't
worked
enough
now
before
I.
Do
this
like
now
just
go
back
here
and
you
know:
first
apply
the
configuration.
E
One
thing
I
thought
is
that
I
actually
applied
one
of
the
companies
of
light
on
it
succeeds.
It
will
show
a
success
message
and
it
will
close
the
window
now
if
we
really
apply
the
same
conflict
like
I
said,
like
you
know
this
time
it's
actually
not
going
to
create,
but
rather
update.
Now,
if
we
want
to
believe
it,
but
there
we
can
actually
put
a
lead
up
well,
I
think
hangouts
does
not
allow
me
to
share
multiple
applications
just
going
to
check
if
accrual.
If
it
will
then
yeah
I,
don't
think
it
does.
E
I
also
wanted
to
actually
show
the
terminal
again
I
connect
to
the
cluster,
and
you
don't
show
that,
like
every
time,
I
change
this
happening
from
the
UI.
You
know
the
changes
are
really
applied
on
this
steel
cluster,
but
yeah
like
oh,
maybe
I
can
actually
show
it
in
a
minute
or
two
I
can
after
I
finish
this
demo
so
and
then
now
coming
to
the
other
part
that
which
is
a
custom
part
and
I'm
gonna,
remove
the
delete
off
because
I'm
going
to
create
one.
E
So
I
have
a
bunch
of
configurations
like,
for
example,
so
the
first
one
is
that
you
still
get
weight
so
and
then
I'm
just
calling
it
a
gateway
one
and
you
can
see
the
analytic
or
come
to
come
to
life.
Now.
That's
like
you
know
that
it's
a
value
camel
and
you
can
see
like
you
know,
it's
a
it's
a
gateway
configuration
I've,
given
it
a
name
booking
to
keep
it.
One
can
be
anything,
and
here
is
a
sample,
functional
gateway.
E
So
now,
when
I
apply
this,
but
you
know
you
can
see
it
successfully
applied
this
on
the
cluster
and
the
same
way
like
if
I
actually
again
reapply
this,
it
will
it'll
update
the
existing
conflict
and
the
same
way
if
I
actually
do
a
delete
operation.
The
same
exact
flow
works
because
the
same
way
like
you
know,
based
on
the
given
llamo,
it
will
construct
a
binary
point
and
then
it
will
work
with
the
cluster
same
way.
We
can
actually
do
a
deployment
now.
E
This
is
not
an
aesthetic
one
like
you
can
see
that
this
is
a
very
general
Navis
deployment
configuration
which
essentially
just
applies
a
for
tile
container.
No,
like
oh
you're,
like
last
week,
like
you
know,
we
were
not
able
to
actually
do
this
because
we
were
had
using
concrete
types
now
that,
like
you
know,
this
is
actually
not
using
concrete
types.
E
It
will
be
able
to
work
with
the
cluster
I
could
delete
something
which
was
not
there
our
message
so
now
you
can
see
that
that
you
know
it
successfully
applied
it
and
now
I'm
going
to
try
to
delete
it
and
the
preparation
works.
The
same
way
for
virtual
services.
Same
way
for,
like
you,
know,
I
eat
any
corn,
eighties
ballad
kubernetes
types
are
going
to
work
again
by
violet
coordinators
types.
It
is
very
specific
cluster
because
one
cluster
can
have
more
custom
CRD
supply
while
another
one
might
not.
So
anything.
E
That's
valid
on
your
cluster
will
actually
work.
That's
kind
of
the
advantage
of
using
a
dynamic
line
so
now
before
I
actually
finish
this
so
I
thought
well.
Actually
we
could
be
show
this
I'm
going
to
apply
this
circuit
breaker
configuration
again
on
the
cluster
and
I'm,
going
to
just
switch
screens.
I
clicking
on
show
the
other
window
in
just
a
second.
E
Yes,
now
awesome,
so
this
is
the
plus
er,
which
you
know
the
mystery
is
steal.
Confident
is
talking
to
so
what
I've
done
is
actually
apply
a
destination
rule
you
know
by
by
applying
that
from
the
user
interface.
So
if
I
could
go
here
and
an
electronic
check
or
destination
rules,
you
can
see
that
right
now
there
was
a
new
destination
rule,
that's
applied,
and
you
know
for
those
who
are
interested.
You
can
actually
see
that
it
is
a
circuit
breaker
configuration
assuming
like
all
it
says.
E
Is
that,
like
you
know,
a
maximum,
you
know
request
can
only
give
you
one.
Maximum
request
for
connection
is
one
a
total
of
a
maximum
one.
Pcd
connection
is
what's
above,
etc.
So
if
these
conditions
are
not
met,
then
it'll
automatically
trip
the
circuit.
That's
what
this
configuration
is
informing
is
here
too
I'm,
actually
going
back
to
the
UI,
of
course,
like
you
cannot
see
that
for
a
second
and
I'm
going
to
do
a
delete
operation
from
there
so
now
I've
deleted
it.
E
So
if
I
actually
know
try
to
list
the
destination
pose,
you
can
see,
the
destination
goals
is
gone.
So
so
pretty
much
like
we
have
this
whole
piece
functional.
It
was
merged
to
the
master
yesterday.
So
you
know
that
pretty
much
includes
the
contrary,
steel
demo.
Now
the
last
piece
is
that
last
pieces,
so
the
same
exact
cluster
I
have
a
product
page
in
for
a
product
page
looking
for
a
bee
fly.
So
every
time
you
reload
this
page
you'll
see
a
different
version
of
the
review
start
up
here.
One
version
has
a
black
star.
E
One
word:
who
has
a
red
star?
The
other
one
has
something
start
so
I'm
gonna
actually
be
using
this
URL
for
load
testing
and
I'm
gonna
actually
set
a
concurrency
of
three
I'm
not
going
to
limit
the
queries
per
second,
and
the
duration
is
one
so
you
can
see
like
and
how
the
test
was
wrong.
So
the
test
is
passed
to
a
poor
tile
container
and
that
foreign
container
has
initiated
a
lot.
This
real
request
to
be
clustered
on
this
URL.
So
once
the
succeeds
yeah
the
beers
to.
D
E
Yeah
yeah,
okay,
excellent
priority,
so
I,
essentially
like
I,
mean
so
I
have
a
you,
can
see
the
countdown
timer
in
action
now
once
since
it's
like
at
10
seconds
I'll,
let
this
run
and
then
I'll
show
you
the
remaining.
You
can
see
I'm
asking
about
this.
Url
I've
set
the
concurrency
request
to
three
duration
to
one
minute
and
the
phrase
per
second
I'm,
not
too
limiting
it.
E
So
once
the
test
successfully
around,
you
can
see
that
right
now
it
actually
chartered
it
out
and
or
the
lower
desk,
so
I
can
I
have
an
app
which
is
running
on
sto
on
the
same
exact
environment
like
I,
show
it
in
the
terminal
know.
This
is
the
product
page
app.
It
has
three
versions
of
the
reviews
running,
so
one
version
returns
back
no
stars,
the
other
one
like
an
over
tons,
a
red
star
gather
one
returns,
a
black
star,
usually
like
know
round
robins.
So
this
is
what
I
was
actually
using
for
s-salam.
E
So
that's
pretty
much
like
not
the
entire
Rush
Rhees
functionality.
As
of
today.
Of
course,
we
know
we'll
be
adding
more
and
growing
my
sri,
improving
the
UI
excedrin,
but
then,
as
of
today,
that's
this
is
my
sri.
I
try
to
give
the
entire
demo
because
pollen
is
actually
new,
so
I
thought
when
I
took.
We
are
hopeful
for
him,
but
overall
again
in
the
last
week
the
P
stuff
was
accomplished
was
actually.
This
is
custom.
Llamo
part
which
gives
the
person
who's
using.
E
My
sri,
complete
control
over
the
appropriate
is
cluster,
so
whether
they
have
controlling
their
names,
wait
I
mean
so
it
all
depends
on
the
axis.
The
you
know
the
the
Q
config
file,
you
know
ass,
pretty
much
well
or
the
service
account
associated
Hakeem,
concrete
file
that
you
have
to
watch.
So
if
a
person
has
I've
been
perfect,
more
so
in
commodities,
llamo
file
will
actually
work
with
this
configuration.
So
this
way
you
know
we're
not
limited
to
just
providing
some
default
options,
but
rather
we
are
giving
them
the
option
to
actually
apply
any
changes.
E
E
So,
if
you
guys
have
any
questions
like
I'll
feel
free
to
reach
out
to
us-
and
you
know
all
the
code
is
pretty
much
like
an
open
source.
It's
all
I
ask
you
know
I,
ask
how
actually
yeah
absolutely
so
guys.
You
have
any
questions.
You
want
me
to
go
over
something
specifically.
You
want
me
to
show
the
source
for
run
over
something
you
know
feel
free
to
ask
us.
E
So
one
other
mention
is
that
I
mean
all
the
instructions
who
run
this
run.
All
the
pieces
is
actually
in
a
dr.
Campos
file,
so
you
don't
feel
free
to
run
it
up.
I've
even
shared
the
my
Twitter
and
get
her
butt
API
keys
in
the
mesh
we
channel
last
week,
so
people
who
just
want
to
give
it
a
try
there
is
this
dr.,
Campos
command
and
with
everything
all
the
pieces
in
there,
you
should
be
able
to
try
that
you
just
scroll
up
a
bit,
but
you
know
this
is
I.
E
D
So,
as
we
can
see,
there
is
a
peak
of
requests
in
the
beginning
and
then
low
counts
towards
the
end
like
it
might
be
better
to
have
a
look
at
what
kind
of
CPU
usage
is
also
there
across
threads
or
the
work.
What
is
the
memory
configuration
and
stuff
like
that?
Because
I
mean?
Why
is
the
difference
of
these
gums
stuff
that
that
might
be
helpful
to
have
a
insight
into
what
with
the
measures,
because
this
is
40
steal,
a
different
submission
mesh
might
have
a
different
kind
of
the
jar.
D
E
Absolutely
so
you
you
are
right,
so
that's
essentially,
like
you
know,
one
of
the
things
like
know
that
we
have
planned
for
future
where,
right
now,
this
cluster
has
a
Prometheus
and
a
crow.
For
instance,
I
think
I
actually
showed
the
graph
on
aunty.
Last
week
we
have
all
the
metrics.
So
you
know
from
the
mesh
from
the
container
is
a
CPU
memory,
like
not
all
number
of
her
coming
in.
You
know
the
response
time
on
the
surface
I
like
all.
We
have
all
these
metrics
in
Prometheus.
E
That's
actually
like
you
know,
and
we
have
a
profile.
On
instance:
that's
starting
to
grow
for
me
is
I'm,
charting
it
out
now
or
future.
One
of
the
tasks
are
one
of
the
plans
we
have
are
actually
do
it
all
see
if
we
can
grab
the
the
chart
from
Prasanna
and
embedded
in
this
user
interface,
so
that
now
this
Lotus
results
are
very,
very
specific
client
side
view
of
the
results.
E
It
does
not
have
any
set
aside
information,
but
once
we
are
once
we
have
the
ability
to
actually
really
the
graph
on
our
charts
and
embed
it
probably
on
this
UI,
then
you
know
the
end
user
of
measure.
You
will
actually
have
the
capability
to
up.
You
know
see
both
like
and
compare
it
in
the
sense
they
can
actually
see
the
client
side
results
and
compare
it
with
looking
all
the
other
charts.
So
that
way,
you
know
they'll
actually
have
a
very
clear
understanding
of
like
what
was
going
on
the
time.
E
What
went
wrong
if
something
really
went
wrong,
like
you
know,
why
was
there
a
spike
like
you
know,
and
why
you
had
to
request
your
decline?
What
it
says
like
I?
Why
did
the
CPU
here
increase
our
memories?
Our
network
I
mean
yeah,
it's
actually
yeah
we'll
be
without
plan
for
it,
but
you
know
unfortunately,
maybe
have
given
the
toss
to
Union.
You
know,
you
probably
don't
know
yogi
look
like
he
hardly
getting
work
done
so
so
tough
to
have
him
here.
You
know
it's
like.
Unfortunately,
you
know
we
have
to
do
with
it.
E
Actually,
you
know
that
was
that
definitely
was
indeed
like
you
know,
one
of
the
main
goals
of
memory
or
I
mean
it
writes,
predecessor
like
in
which
we
used
to
call
these
do
playground.
That's
essentially
what
it
was
doing.
What,
with
a
different
product
like
you
know,
it
used
to
be
called
I
mean
so
they've
had
integration
with
a
optics
recently,
it's
just
in
the
last
two
weeks.
I
can
really
rip
that
out
and
we
are
planning
on
using
graph
honor
because
prefer
Nozick
an
open
source
project.
So
so
the
plan
is
like
that.
E
If
a
kubernetes
cluster
has
profile
instance,
then
make
sure
you'll
be
able
to
talk
to
the
piranha
instance
and
hopefully,
like
you
know,
get
the
charts
from
it
and
then,
like
my
short
in
the
metro,
UI
is
one
about
man.
Now,
if
that
doesn't
work,
I
mean
there's
a
possibility
that,
like
might
have
difficulties
like
you
know,
in
which
case
we'll
probably
be
directly
talking
to
Prometheus,
getting
the
data
and
then
charting
it
out
with
some
kind
of
a
charting
like
really
eat
through
yourself.
So
you
know
army
yeah.
E
While
we
are
here,
I'm,
just
going
to
like
I,
mean
I
since
Robin,
exactly
Newton's,
call
and
I
hope,
I
can
Oprah
teak
and
he
can
turn
out,
but
you
know
would
have
seen
the
info
sto
github,
sorry,
the
layer,
five
github,
so
you
know
layer,
five,
exactly
the
site.
My
sri
SEO
is
the
SEO
doctor
which
essentially,
like
you
know,
I
tried
the
demo
during
this
call.
I'll
make
sure
the
link
there
D
is
actually
the
next
adapter.
We
would
probably
be
working
on
that's
for
interfacing
with
liquor,
so
you
can
see
that.
E
But
you
know
it's
actually
that
the
plan
is
to
actually
have
adapters
for
every
mesh
so
that
you
know
make
sure
you'll
be
able
to
dynamically
talk
with
any
like.
You
know,
which
or
is
available
on
the
cluster,
and
my
Cherie
is
the
heart
of
the
entire
application
and
that's
pretty
much
it
so
this
mesh
benchmark
Spector's
work.
We
actually
present
a
bit
of
you
back
so
so
yeah.
This
is
Dora
github.
You
don't
so
feel
free
to
actually
like
to
check
out
the
projects.
A
It's
so
confusing
because
to
share
you've
got
to
share
the
the
hangouts
video
call
itself,
which
is
kind
of
weird,
so
very
good,
so
yeah
so
pumped
to
have
Pavan.
Here
we
did
get
some
contributions
from
the
new
of
VMware.
Not
only
was
he
had
to
call
last
week,
but
he
did
some
contributions
this
week
it
was
great
to
have
Josie
from
IBM
last
week
as
well.
A
So
that's
all
very
encouraging.
As
we
go
to
well,
I
was
going
to
wrap
the
call
I
think
that
the
most
appropriate
thing
to
do
is
well
as
one
to
let
Pavan
digest
get
familiar,
but
also
identify
a
couple
of
explicit
small,
hopefully
small
items
where
he
may
be
able
to.
If
interested
you
know,
take
a
crack
at
and
I
and
Pavan.
Don't
don't
let
me
you
know,
put
words
in
your
mouth
or
or
but
it
sounded
like
hey.
Maybe
part
of
the
provisioning
of
the
mesh
might
be
up
your
alley
and
I.
E
Provisioning
of
SEO,
what
we
have
is
actually
a
good
config
file,
which
was
which
was
created
by
Fennell,
but
now
now
this
is
the
first
part
in
where
we
try
to
use
that
Yahoo
from
the
you
from
the
mesh.
We
interface
I'm,
pretty
sure
it's
not
gonna
work,
but
I
haven't
tried
it
personally,
so
you
know
I'll
have
to
try
it,
but
your
body,
if
it
works,
then
I
think
maybe
you
know
Christy.
E
Oh,
we
can
call
that
it's
functional,
I
haven't
yet
tested
it
like
I'll,
probably
cry
it
on
today,
but
you
know
so
once
once
we
have
that,
like
you
say
that
stillness
is
uz+
losing
one
of
these
people
on
this
place.
Actually,
if
you
want
to
give
an
actual
and
I
have
another
another
interesting
one
which
is
actually
long
and
of
actually
SAS
along
with
the
diamond
hook
tables,
so
I
can
also
wear
on
the
cloud
nice,
but.
A
So
Pavan,
do
you
I'm
not
sure
how
much
extra
time
you
have
or
if
that's
of
interest,
but
the
the
SAS
portion
of
this
project
will
use
a
couple
of
AWS
services
and
they
need
to
be
terraformed
up
is
terraform.
Something
you're
familiar
with.
Is
the
provisioning
of
native.
C
F
A
And
then,
with
respect
to
to
Pavan
&
pratik,
so
pratik,
we
had
a
bit
of
a
healthy
and
very
lengthy
discussion
last
week
about
part
of
the
goals
of
the
project
and
very
end
and
how
that
aligns
and
overlaps,
with
the
focus
that
you
guys
have
both
on
existing
projects
that
you've
got.
You
know
either
potentially
with
hallway
or
just
you're.
At
you
know,
some
of
your
exist
focus.
A
Paul
the
further
along
that
measure
II
as
a
utility
gets,
you
know,
as
a
performance
benchmark
tool
gets
maybe
the
the
easier
it
is
for
you
guys
to
accomplish
at
least
one
aspect
of
how
I
think
that
you're
able
to
contribute-
and
that
is
to
actually
perform
the
tests
themselves.
Kind
of
you
know,
go
through
in
an
environment
and
run
the
run,
various
flavors
of
the
tests
and
and
look
at
the
consistency
by
which
the
results
are
returned.
You
know
just
like
for
the
same
tests
in
the
same
environment
run
five
times
over.
D
That's
yeah.
That's
definitely
a
good
idea
like
once
things
are
in
place
like
we
have
steel
and
probably
at
least
one
more
adapter
in
place
for
mesh.
We
can
run
in
different
apps,
look
at
through
codes
memory,
consumption,
network,
I/o,
stuff
and
run
them
various
amounts
of
time,
and
then
probably
looking
at
the
results
we
can
contribute.
A
This
first
question,
which
is
more
or
less,
is
their
confidence
in
memory
as
a
performance
benchmark
tool
in
in
an
in
a
vacuum.
In
a
lab
environment,
where
you
get
that
you
know,
you've
got
the
same
notes:
the
same
resources,
hopefully
not
much
else
going
on.
You
know
that
it's
a
controlled
environment
if
they're
taking
mesh,
ready
and
running
the
same
style
performance
tests
against
interpret
each
point.
You
know
right
now
only
against
the
book
info.
That's
sample
app
yep.
They
can
begin
to
answer
that
first
question:
it's
like!
Does
the
tool
produce
consistent
results?
E
Stupid
what
what
actually
just
before
I'm
sorry,
get
just
a
quick
question
in
place?
Are
you
guys,
like
a
Windows
users
or
our
Mac
users
or
Linux
users,
Linux.
E
Ok,
so
I
can
actually
send
the
instructions
or
probably
using
something
like
have
you
heard
of
micro.
D
E
C
E
All
it
takes
and
you
should
have
a
functional
like
a
noise,
do
posterity
minutes
and
the
other
thing
is
like
you
know,
then
you
can
actually
use
a
dr.
Campos
file
electronically
to
your
local
cluster
and
probably
like
that.
That
will
probably
be
the
fastest
way
for
you
to
actually
pick
your
hands
dirty
or
if
you
have
some
admissions
to
spare,
then
we
can
actually
I
can
help
you
with
setting
up
operators
cluster.
E
There
is
a
tool
called
queue
idiom
which
can
actually
help
you
get
up
super
fast
and
go
from
there,
but
yeah
essentially
like
I,
mean
to
answer
all
these
questions.
Yes,
I
can
definitely
like
you
know,
have
a
conversation
with
you
guys
and
let
you
know
probably
like
I,
don't
get.
It
is
going
very
quickly.
Yeah.
D
Sure
we
have
machines
so
like
we
might
have
to
just
configure
them
with
something
like
a
lot
of
the
machines
are
being
Co
used
by
users
who
require
to
look
into
the
logistics,
and
then
we
can
get
started
soon
and
yeah,
as
we
mentioned,
the
getting
our
hands
dirty
and
looking
at
the
current
stats
that
we
are
getting
and
looking
what
kind
of
results
we're
getting
or
the
consistent
doesn't
make
sense
is
what
kind
of
improvements
we
can
do
just
from
day,
zero
yeah.
We
can
start
working
towards
that
cat
awesome.
G
A
Wrote
down
maybe
two
questions
that
you
might
be
able
to
answer.
You
know
back
by
the
next
time
we
meet
and
that's
you
meet
you.
Potentially
these
two
questions
do
do
we,
you
know
yeah,
okay
and
then
for
Pavano,
where
we're
right
at
the
top
of
the
our
I
think
I
think
we
probably
do
need
to
describe
to
you
a
little
bit
of
the
mess
resets
itself
a
little
bit
of
credential.
You
know
some
credentials
to
AWS.
A
And
so
with
that,
maybe
we
will
describe
that
after
we
end
the
community
meeting,
so
I
was
gonna,
wrap
up
here,
yogi
pratik
phenomena,
anything
else
that
we
wanted
to
chat
about
this
week.
H
B
Forgot
things
of
that
task,
so
I'm
glad
and
actually
reminded
me
yesterday
a
couple
of
questions.
One
was
related
to
initially.
My
plan
moves.
Okay,
if
we
are
going
to
pull
the
data
from
Prometheus
that
Prometheus
already
has
an
integration
with
woody
call
refiner,
so
so
absolutely
I'm.
Thinking
of
going
to
these
two
approaches,
which
is
one
directly
to
graph
honor,
how
do
we
integrate
it
directly,
super
fauna
and
then
the
second
one
is:
how
do
we
actually
I
pick?
B
E
Disability
yeah
absolutely
sounds
good
yeah
Griffin
already
has
a
charge
like
and
also
the
easiest
thing
I
want
her
to
see
was
the
chart
to
grab
fauna.
If
that
is
hard,
yeah,
because
I
know
it
might
have
some
challenges,
then
we
could
probably
let
you
know
rightly
pipe
from
Prometheus
of
how
awesome
some
pretty.