►
From YouTube: Meshery Community Meeting - March 8th, 2019
Description
Today's Meshery community meeting walk's through @pkumrsv's Terraform scripts to deploy Meshery SaaS.
See https://layer5.io/meshery for more project details and other recorded meetings.
A
Okay,
very
good,
all
right,
we're
three.
After
welcome
to
another
measure,
e
community
meeting
pay,
no
attention
to
the
grandmotherly
knickknacks
I
have
in
the
background
here,
I'm
traveling
today,
so
this
week
is
scale
out
in
Pasadena
California.
So
fortunately,
for
me
and
I
get
to
visit
some
some
family,
that's
it's
nice!
A
It
looks
like
we've
got
per
Nagas
with
us,
Pavan
and
gireesh
be
sure
to
record
your
attendance
in
the
in
the
meeting
minutes
and
as
always,
we
are,
we
are
recording
and
so
be
on.
Your
best.
Behavior
looks
like
one
of
our
other
regular
is
a
vanilla
message
just
a
little
bit
ago.
That
he's
also
at
the
same
conference
with
me
scale.
So
hopefully
we'll
get
to
meet
up
in
person
and
don't
be
good
nice
with
that,
let's
jump
over
into
the
meeting
minutes
sounds
like
Harish
we
may
have.
A
You
may
have
an
update
in
terms
of
some
of
the
changes
to
measure
e
sounds
like
this
very
good
I.
Think
this
time,
in
terms
of
a
general
community,
update
a
couple
of
things
for
us
to
go
over
and
then
maybe
we
can
do
a
round
a
round
table
in
terms
of
I
know:
Pronovias
made
some
some
movements
open,
oven
and
so
a
view
Carrizo
the
general.
So
the
general
update
is
oh
a
couple
of
the
Aspen
mesh
folks
reached
out
to
join.
A
So
Zach
and
Niraj
Zach
being
heading
up
marketing
and
Aspen
mission
in
Rodge
I
think
leading
up
the
probably
a
few
different
things,
but
one
of
them
being
so.
He
was
interested
in
the
innovation
effort
that
we
given
and
that
Vanille
had
had
led
that
they're
looking
to
join,
to
discuss
that
and
make
sure
that
we
had
a
good
integration.
A
Nick
Oh,
whose
last
name
we
need
to
go
look
up.
Nick
B
was
very
excited
by
when
he
learned
in
the
project,
and
so
I
half
expected
him
to
see,
see
him
here
today,
so
I'll
make
sure
we
invite
them,
and
then
we
ended
up
speaking
with
the
ISTE.
Oh
I'm,
C
and
I
are
the
contributors
of
this
TOC
and
I,
and
you
can
see
have
a
long
chat
that
we
had
had
in
the
general
channel.
A
Other
than
that
I
think
that
that
that's
probably
the
only
general
update
some
gireesh
do
you
have
in
a
general
update
with
respect
to
the
octarine
service.
No,
we
don't.
Okay,
so
they're,
tentatively
or
I
think
I
think
they
committed
to
contributing
an
adapter,
but
just
haven't
made
material
progress
just
yet
exactly.
B
So
early
I'm,
actually
typing
some
updates
from
the
team
in
general
I
can
also
I
was
working
with
critique,
mainly
in
helping
with
their
set
up
awful
communities,
enrollment
plus
getting
Missouri
to
work
with
it.
We
have.
We
did
have
some
challenges
with
setting
up
the
kubernetes
environment,
but
ma
cherie
just
ran
like
no,
especially.
There
are
new
updates
with
the
dr.
Campos,
like
invited
a
lot
easier.
So,
as
of
yesterday,
pratik
has
a
functional
and
run
that,
like
you
know,
where
he's
able
to
test
like
you
know
all
the
pieces,
this
was
like.
B
You
know
the
update
I
got
from
pratik
as
of
24
hours.
Back
Pavan
and
I
actually
had
several
charts
during
this
week,
and
we
met
yesterday
evening,
like
you,
know,
discuss
some
of
the
infrastructure
stuff,
and
you
know
I
think
often
already
mentioned
that,
like
he
completed
like
you,
know
most
of
the
terraforming
stuff,
so
I've
kind
of
recorded
that
as
well.
Apart
from
that
one
other
thing
is
Reynolds
PR,
it's
not
yet
emerged.
B
So
these
are
kind
of
the
general
updates,
like
you
know,
based
on
what
I've
heard
from
the
other
than
that
I'd,
like
you
know,
I
was
able
to
make
some
progress
with
measure
in
terms
of
displaying
and
comparing
whistles
John,
hopefully
demonstrate
today
or
present,
like
you
know
the
next
few
minutes.
So
these
are
like
general
updates
from
my
side,
like
you
know,
reflecting
some
of
the
information
I've
got
from
others,
but
yeah.
B
Yeah
so
so
yeah,
that's
pretty
much
looking
like
that
page
from
my
site,
unless,
like
you
guys
had
any
other
updates.
C
B
Very
informative
he
actually
was
was
was
way
good
like
a
night
in
terms
of
like
you
know
what
he
was
doing.
So
you
know
a
chat
with
him
trying
to
understand
what
he
was
talking
about
in
the
other
community
meeting
was
actually
very
helpful
for
sure
that
was
an
amazing
conversation
other
than
that
other
than
that.
Actually
I
do
have
one
thing,
which
is
the
sto
one
point,
one
performance
document
or
it's
the
performance
sizing
tuning
guidance
document
looking
at
which
we
can
probably
bring
up.
B
You
know
during
this
call
and
I
will
have
a
discussion.
I'm
pretty
sure,
like
you
know,
it'll
be
in
line
with
what,
for
now
and
fatigue
like
you
know,
might
be
interested
in
looking
at
there
is
predict
so
awesome,
so
there
is
ta
one
point.
One
performance
document
like
very
interesting
to
both
pratik
and
print
also,
hopefully
like
we
can
actually
bring
it
up
and
share
it
with
them
from
there.
The
other
thing
is,
the
other
thing
is
actually.
A
A
No
yeah:
let's
no
no
go
ahead.
Boutique.
D
B
So
I
was
just
you
know,
so
the
last
point
was
that
we
we
have
a
recent
performance
of
day
document
from
this
do
team,
which
might
be
of
interest
to
you
guys
so
we'll
we'll
probably
never
bring
it
up
and
like
I
know,
the
document
is
a
bit
long,
but
it's
it's
something
like
a
know.
We
can
briefly
have
a
touch
on
and.
C
B
A
Well,
yeah
I
guess
I've
got
a
few
different
Doc's
up,
but
so
one
of
those
Doc's
is
is
a
well
alive.
I
know,
rightly
live
online
training
that
we're
getting
ready
to
do
for
estia
anyway,
back
to
to
Greece's
point.
Here's
the
other
document
on
that
comes
out
of
the
SEO
performance
and
scalability
working
group,
IBM
and
Google
it's
by
and
large
those
two
groups
that
perform
much
of
the
performance,
testing
and
scale
testing
for
is
deal.
A
A
Totally
in
there
and
in
some
respect,
I
think
you
know
if
you
took
that
at
face
value,
you
might
say:
well,
you
know
how
valuable
is
that,
like
you
know
the
is
it
worth
repeating
the
same
test?
That's
like
well,
the
most
prominent
reason
to
do
so.
There's
a
number
of
reasons.
I
think
that
we've
discussed
in
the
past
of
what
why
it
makes
sense,
it's
a
kind
of
quote
unquote
worth
to
redo
some
of
this
work.
E
A
I
think
if
you
guys
have
looked
at
this
there's
way,
there's
all
kinds
of
permutations
and
combinations
of
different
tests
and
environments
that
you
can.
You
know,
do
performance
measurements
within
that.
This
is
a
good
indicator
for
us
in
terms
of
where
to
spend
the
time
in
measuring
what
tests
to
run,
but
then
taking
these
tests
and
running
them
against
the
other
meshes.
A
C
A
A
But
but
the
other
requests
that
I
have
guys
is
that
we
revisit
our
design
doc
for
measure
II,
in
which
what
we're,
what
we've
begun
to
articulate
is
hey.
We
think
that
we
should
measure
latency
and
CPU
overhead
and
memory
consumption
and
extension
like
there's
various
things
that
we
were
calling
out
initially,
as
we
reflect
on
this
document,
we
should
probably
go
back
and
update
our
approach.
A
The
third
reason
why
it
makes
sense
to
repeat
some
of
these
tests
is
because
pretty
you're
running
we're
gonna,
take
this
to
a
variety
of
different
hardware,
environments
or
infrastructure
drivers,
and
so
you,
you
guys,
have
one
and
then
we
collectively
have
access
to
the
Texas
Advanced
Computing,
Center,
yeah
that
and
then
we're
gonna
get
access
to
the
CNCs
infrastructure
lab,
and
so
anyway,
that's
the
third
reason
to
rerun.
Some
of
these
is
because
an
interesting
insight
that
I
just
recall
off
the
top
of
my
head
about
the
variety
of
insights.
A
This
document
that
you
know
that
this
team
has
come
to
is
that
specific
to
IBM
kubernetes
service
specific
to
IKS?
They
IKS
not
like
a
Tobias
and
Azure
and
GCP.
They
have
a
load
balancing
service.
They
have
an
Ald,
I,
think
I,
I
think
iBM
is
even
maybe
borrowing
some
of
the
same
terminology
from
AWS
in
that
regard.
They
call
it
an
al,
be
an
application
load,
balancer
Whitacre,
which
isn't
really
borrowing.
A
D
A
Let
me
see
if
we
can
break
it
down,
so,
if
you
guys
can
so
that
this
is
results
of
I
think
two
different
tests.
So
if
we
look
at
this
first
figure,
this
first
collection
of
figures-
this
is
the
amount
of
CPU
consumed
cluster
CPU
distribution
with
no
SEO.
For
this,
are
you
sharing
your
screen?
Oh
man,
Wow
this
entire
time,
I've
not
been
sharing
yeah
man
I
was
waving.
My
mouse
frantically
at
various
areas
of
my
screen.
I
can't
believe
you
guys
him.
A
A
A
You
know
we're
being
pulled
out
and
put
into
a
longer
farm
word
doc
that
just
sort
of
writes
up
what
some
of
the
tests
were
about.
You
know,
so
it's
an
excellent
document
right
what
the
tests
were
about
and
what
some
of
the
results
and
what
the
insights
gleaned
were
and
of
this
visit,
one
of
the
specific
tests
or
environments
that
I
was
talking
about
was
like
the
amount
of
CPU
consumed
when
you
used
two
different
types
of
ingress
gateways.
A
D
A
D
A
So
no
mesh,
the
app
was
taking
up
10%
the
application
load
balancer
that
they
were
using
was
taking
up
1%
okay
over
here
they
put
that
same
app
same
nodes
in
the
same
environment
and
how
they
put
it
on
is
to11
and
so
sto
is
taking
up.
You
know
in
this
environment,
based
on
all
the
other
variables
and
etc
that
it
was
taking
up
20%
average
CPU
utilization
across
that
ok
a
bit
so
and
what
you
can
see
is
like.
A
Of
this
second
set,
this
is
really
what
I
was
trying
to
say
earlier.
That
I
was
just
like
hey
yeah
there's
a
number
of
interesting
insights
that
came
out.
I
said
nginx
may
be
interested
in
this
insight,
which
is
here's
the
sample
app
running
on
e
steel
on
one
when
it's
using
nginx
as
the
application
Allah
balancer
and
here
is
the
same
sample
app
running
on.
You
see
a
1:1
in
this
case.
A
My
understanding
is
that
the
it's
written
a
bit
different
than
what
I
understood,
but
because
it
looks
like
the
difference
being
that
that
in
peer-to-peer
security
was
enforced
here.
But
but
I
think
I
think
the
difference
that
they
were
trying
to
highlight
was
that
nginx
was
the
sto
gateway.
The
ingress,
whereas
over
here
envoy,
was
the
technology
used
as
the
ingress
and
that
the
difference
in
how
much
the
1%
CPU
that
nginx
takes
up
versus
the
higher
percentage.
I,
don't
know,
I,
actually
I'm
a
little
bit
concerned
about
yeah.
D
A
D
D
A
A
Can
do
that
while
we
are
talking
about
a
couple
of
other
other
things,
so
so
that
was
one
I
think
that's
very
informative.
We
should
definitely
digest
it
holistically.
We
should
update
our
design
Docs
based
on
it.
We
should
try
to
make
sure
that
our
ministry
itself
facilitates
for
those
tests.
I
don't
know
if
we're
gonna
be
able
to,
you
know,
perform
all
the
same
ones,
but
I
think
it
would
be
very
helpful
to
the
East
EO
community.
If
we
atleast
repeated
a
few
of
those
same
ones,
yeah.
A
A
I'm
gonna
try
to
update
you
know
or
feel
free
if
anyone
else
wants
to
update
the
notes,
as
we
go
update
the
notes
to
reflect
that
that
another
topic
reach
that
you'd
brought
up
is
that
if
we
take
a
moment
to
go,
revise
the
benchmark
specification
that
once
we
have
a
first
draft,
we
can
share
that
with
other
service
master
project
projects
and
and
they
can
begin
to
bear
weight
on
that
spec.
A
Not
capturing
but
for
describing
the
environment
and
the
type
of
test
that
a
given
test
runs
like
a
given
benchmark,
ran
I.
Think
you
need
to
store
the
results
of
the
test
separately,
but
if
the
goal
here
of
this
type
of
the
specification
was
to
say,
hey
this,
these
are
the
environmental
x'
for
this
test.
The
you
know
like
this
many
nodes
of
this
size
and
this
type
of
an
environment.
You
know
of
this
build
of
of
what
mesh
that
was
of
what
build
it
was.
You
know
all
of
those
details
is
the
first
part.
A
A
What
tests
did
you
actually
perform?
Was
it
you
know
a
thousand
requests?
A
second
force
of
you
know
four
five
minutes
or
what
was
the
type
of
test.
So
if
we
reflect
on
this
and
ask
ourselves
earnestly,
if
trying
to
create
a
standard
benchmark,
a
performance
benchmark
is
beneficial
to
the
community
or
not.
A
What
I
want
to
ask
you
guys
that
question?
So,
let's
start
thinking
about
that
before
you
answer,
I'll
say
that
there
there
are
standards,
bodies
that
work
on
and
put
forth
specifications
like
this
gireesh
when
we
talked
with
Surya
a
few
weeks
ago.
You
recall
the
standards
body
that
he
had
noted
I
did
we
put.
B
A
A
Try
the
IBM's
gonna
continue
to
run
their
SEO
performance
tests,
they're
all
into
sto
as
the
mesh
that
they
want
that
they're
behind
you
know.
Obviously
they
helped
start.
Seo
and
they're
gonna
keep
running
these
performance
tests
and
publishing
them
and
they're
gonna
use
their
own
tooling
they're,
not
gonna,
they
might
use
measure
I
mean
they
were
excited
that
we're
doing
the
project,
but
that
they're
clearly
invested
into
their
their
proprietary
closed
sourced,
blooper,
tooling,
that
if
we
have
a
common
format,
we
can
interchange
between
them
and
have
some
common
language.
A
The
same
thing
for
Google
they're,
using
a
bunch
of
scripts
around
for
tayo
to
perform
all
of
the
test
that
they're
performing
and
that,
if
we're
using
a
common
format
that
between
these
three
projects,
we
might
be
able
to
interchange
them
and
have
a
common
language
that
we
would
then
be
able
to
use
this
common
form.
You
know
that
we
would
then
take
this
common
format
to
buoyant.
You
know
the
link
of
D
folks,
the
Hoshi
Corp,
the
console
connect
folks,
AWS,
the
app
mesh
folks.
A
D
D
A
Critique
it
see
it's
good
good
feedback
and
maybe
and
I
would
regurgitate
what
you
said
and
say.
Look
this
specification
is
it's
probably
only
compelling
to
the
extent
that
measure
e
is
successful
and
well
adopted?
It
would
be
subsequent
to
that
subsequent
to
the
success
and
prominent
use
of
misery.
That
people
would
then
be
interested
in
or
the
other
project
would
be
interested
in
providing
some
sort
of
using
some
sort
of
common
language.
But
you
know
to
be
able
to
compare
and
exchange
tests.
D
A
There
right
I
mean
yeah
anyway,
so
this
being
kind
of
a
secondary
effort
but
and
one
that
we
really
haven't,
discussed
much
because
it
is
such
a
secondary
thing.
The
reason
I
bring
it
up
now
is
because
of
the
conversations
that
I've
had
with
those
other
project
teams.
They
have.
Some
of
them
have
said
we're
not
sure
if
we
can
spare
the
time
the
engineering
time
per
se
to
go
right
and
adapter
for
measure
E,
but
but
we
do
want
to
review
before
you
publish
any
test
results.
A
We
sure
do
want
to
review
how
you
configured
our
mesh.
What
environment
you
ran
it
in
if
you
rent
specific
types
of
tests
that
were
particularly
that
our
mesh
is
particularly
poorly
positioned
to
handle.
Well,
you
know,
basically,
they
want
to
make
sure
that
they're
not
going
to
be
misrepresented
that
if
we
were
able
to
exchange
this
this
specification
within
this
document
that
says
well,
this
is
the
type
of
environment.
A
This
is
the
configuration
of
your
mesh,
and
these
are
the
types
of
tests
that
yeah
that
this
spec
might
help
in
that
regard
as
well,
and
so
so
they
had
been
much
more
willing
to
sign
up
to
review
the
spec.
Then
they
had
necessarily
to
go
and
write
code
to
contribute
an
adapter,
okay,
okay,
cool,
so
I'm
glad
we
talked
to
talk
about
that
I
mean,
let
me
be
quiet
and
let
us
reflect
a
bit
more.
A
D
D
So
I
was
thinking
of
running
like
really
high
load
test
like
10,000
requests
over
30
minutes
or
something
of
that
sort
and
trying
to
look
at
what
the
distributions
are
and
if
we
are
ever
being
bottlenecked
on
any
of
the
hardware
resources.
Or
should
we
just
look
at
Network
interactions
and
if,
like
probably
give
some
feedback
on
what
other
kind
of
graphs
should
be
prioritised
to
bring
it
on
the
battery
UI
of
that
sort?
That's
what
my
thought
process
was.
A
A
One
of
the
things
that
in
part
to
what
pratik
it
initially
said,
was
yeah
I
found
myself
leveraging
some
you
know,
basically,
when
we
want
to
measure
things
other
than
latency
and
throughput
things
other
than
Oh.
What
for
I/o
measures
there
isn't
the
need
to
retrieve
those
metrics
from
Prometheus,
you
know
from
from
yeah
and
then
probably
the
need
to
critique
a
lot
of
times
when
people
are
monitoring
a
given
performance
of
a
given
application
a
lot
of
times.
They
end
up
with
multiple
charts
on
the
same
dashboard
and
a.
B
Yeah
so
once
we
have
some
well
again
like
I
know,
there
are
multiple
points
like
I
mean,
so
so
we
definitely
are
progressing
in
that
direction.
I
think
we
did
bring
this
out
like
you
know
where
I
think
like
now,
I
can
even
bring
up
the
Kofun
or,
like
you
know,
from
a
sto
install.
Is
that
yeah?
Definitely
like
not
those
charts
are
very
valuable.
You
know,
and
there
are
a
lot
of
charts
like
it.
Definitely
like
you
know.
B
We
could
actually,
like
you
know,
have
a
dashboard
experience
here
which
definitely
something
like
an
I
definite
my
mind
but
yeah.
Well,
we
actually
went
sure
in
that
direction.
We
I
just
need
it
to
like
it.
I'll
get
some
there's
stuff
done,
which
is
the
reason
why,
like
we
haven't
gone
there,
but
but
yeah
absolutely
that'll,
be
very
happy
for
us.
D
If
we
did,
we
had
discussed
this
I
think
a
couple
of
weeks
back
about
this.
That
is
why
I
brought
up
like
I'll,
try
to
look
into
the
load
tests
and
try
to
give
feedback
that
what
kind
of
charts
are
more
important
than
like
as
I
was
mentioning
that
the
computer
utilization
will
not
being
the
bottleneck.
So
we
can.
D
Since
we
are
a
bit
short
on
the
engineering
side,
we
can
prioritize
what
kind
of
charts
are
more
valuable
and
others
rather
than
putting
all
like
all
the
definitely
I
will
I
see
a
benefit
in
putting
all
the
graph
on
our
charts
so
on
to
the
dashboard,
but
we
can
probably
prioritize
some
over
here.
That's.
B
E
In
my
kids,
like
I,
think
I
completed
this
terraformed
for
deploying
of
V
PC
ec2
instance
and
API
gateway
and
everything,
but
still
I
need
to
finalize
I
need
some
more
time,
hardly,
but
not
what
words
to
spend
on
these,
to
complete
it,
easy
to
instance,
to
the
API
gateway.
So
once
that
is
done,
like
think,
I
need
to
work
on
like
creation
of
docker
containers.
E
In
the
ec2
instance,
only
I
need
to
go
deeper
into
that
like
how
to
create
I'm,
not
sure
like
that
we
can
created
using
data
for
more,
like
I
need
to
find
it
out
the
race,
easy
difference
and
once
the
ec2
instance
comes
up
like,
maybe
we
can
do
darker
kind
of
artwork
containers.
We
I
think
I
presently
have
two
containers
running
in
I
see
two
instance
right,
probably
we
can
make
it
as
out
of
it
like
use
the
telephone
problem.
I'm
not
sure
I
have
to
take
it.
That.
A
Is
ensuring
the
dependency
of
the
well
first
of
all
I
think,
certainly
ec2
AWS
Amazon
Linux
images
come
with
docker
runtime
already
there?
Yes,
so
it's
probably
it's
probably
a
no
op
in
this
case,
but
but
to
the
extent
that
it
wasn't
you
you,
you
could
use
things
like
cloud
in
it
to
assist
with
the
initial
provisioning
of
certain
resources
as
well.
So
so,
in
this
case,
we
pretty
much
yeah.
B
So
Souter
phone
causes
user
data
concept
like
an
old
when
spitting
up
the
ec2
instances.
So
you
just
need
to
give
it
the
user
data,
which
is
you
know
it
can
be
any
shell
script,
so
yeah
I
mean
literally,
you
can
just
give
it
a
command
and
it'll
just
run
it
like.
You
know
when
the
ec2
instance
comes
up
so
pretty
straightforward
to.
E
B
A
F
A
Perfect,
you
know,
you
know,
even
if
you
didn't
have
the
other
project
focus,
maybe
almost
ideal
that
you
guys
are
just
doubling
down
on
the
same
environment,
providing
each
other
kind
of
partner
programming.
If
you
will
can't
you
know
the
back
and
forth
feedback
good.
E
Yeah,
can
you
have
it?
Yes,
just
let
me
go
through
that
each
and
every
I
mean
like
it
won't
be.
It
won't
take
much
time,
but
like
listen,
Cody
and
I
have
created
a
V
PC,
which
is
like
cedar
block
of
10.10
or
0.06.
You
know,
like
I,
have
created
two
subnets.
One
is
for
the
public
subnets,
where
our
ec2
instance
is
gone.
E
Satan
sitting
to
this
public
subnet
and
as
we
have
a
database
like
Souls
like
properly
the
database
D
beat
like
DB,
is
going
to
visit
in
this
poet's
subnet
10
dr.,
not
$20
0:24,
and
we
have
a
secure.
I,
have
created
a
security
group
and
like
given
access
for
the
poor
24
stayed
abroad
and,
as
goodness,
we
have
a
9
8,
7
6
port.
Given
access
for
from
the
outside
world
to
connect
it,
and
we
have
a
top
tip.
I
have
created
a
row,
table
4,
4,
4
row
being
just
came,
2nd
I'm,
yeah,
I,.
E
E
I'm
not
sure
for
that.
I
can't
find
like
that,
like
Internet
gateways
for
connecting
to
the
the
ec2
instance
for
like
to
connect
to
the
ec2
instance,
a
part
of
the
like,
we
don't
write
for
Internet
gateway
over
here
and
and
as
well
as
like.
We
need
to
get
Internet
to
the
ec2
instance
like.
If
we
didn't
have
bet
we
need
to
do
like
you
have
to
do
that,
and
I
have
created
an
API
gateway
and
they
take
it.
Where
has
the
resources
and
HTTP
methods
and
interact
integration?
E
A
In
Pavan,
you
know
grease
correct
me
if
I'm
wrong,
but
it
Pavan
wanting
to
do
is
that
that
am.
I
omit
the
ami
number
that
you're
using
is
to
go,
take
and
confirm
that
that
image
already
has
docker
pre-installed,
because
it
probably
is
it
probably
already.
Is
there
and
is
not
not
something
you
have
to
separately
address
right.
B
E
I
haven't
V,
PC,
miss
finishes
being
created
and
like
Cedar
Falls,
like
ten
dotted
note
16
like
subnets,
there
are
two
subnet
for
here
like
one
is
yes
t,
2a
and
2b
and
I
have
routing
pair
of
tables
like
by
default.
There
is
a
one
row
table
but
I
like
it's
a
good
practice
to
create
another
photo
built
on
a
hand,
routes
to
that,
and
so
I
read
the
same
thing:
Internet
gateway
I
have
an
Internet
gateways
created
over
here
like
which
is
just
necessary,
like
name
and.
E
If
you
go
to
the
AP,
it
may
have
been
a
sleeping
it
and
resources
and,
like
the
resource,
has
been
created.
Our
life
we
go
to
the
any
like
we
have
I'm
gonna
ask
for
the
timing.
I,
just
given
the
URI
as
google.com
and
like
we
haven't
staged
this
over
here
and
like
production
stage,
has
been
created.
I.
E
Mean
like
we
have
me
and
give
it
a
discussion
that,
like
this
one
need
to
be
connected
to
the
me
like
there
is
some
both
need
to
be
done
on
invoke
URLs,
to
make
it
bring
a
certification
for
me
AWS
and
make
it
run
with
our
own
specific
Kathy,
URL
and
so
I
need
to
work
on
that.
But,
like
I'm
not
sure
like
how
much
time
it's
going
to
be
take
I
can't
promise
that
I
can
complete
it
by
next
week
but
lecture
different
ways
and
come
back
on
this.
A
A
E
A
E
A
A
So
you
guys
can
attend
and
participate
if
you,
if
you
want
to
warning
I,
will
if
you
attend
I,
will
call
you
out
and
recognize
you
for
the
work
you've
done
so
so
I
encourage
you
to
attend.
I
will
also
probably
send
you
each
of
you,
some
slides
for
review
both
that
you
know
slides
that
describe
the
project
kind
of
overall
and
I
will
likely
end
up
putting
your
names
in
there
as
well.
So
so,
please
do
review
to
make
sure
I'm
representing
representing
you
in
your
works.
B
Cool
yeah
I
have
one
actually
I,
don't
know
how
to
just
help
along
with
some
more
details.
Like
you
know,
the
user
data
that
needs
to
be
executed
like
I
want
the
ec2
mission
is
coming
up.
I
think
that's
pretty
much
like
my
action
item
for
him,
yeah,
of
course,
he'll.
One
minute
I
can
probably
show
my
Shree
okay.
B
B
Okay,
so
you
guys
can't
see
the
screen.
This
is
the
mystery.
So
some
of
the
changes
like
I've
been
in
the
UI
like
a
where
I've
just
increased
the
width
of
the
container
to
better
utilize.
The
space
in
the
performance
configure
play
sections.
They
have
not
been
much
changes.
The
only
change
is
actually
this
new
section.
It's
called
results
now,
one
of
the
main
pieces
of
measure
is
that,
like
you
know
so,
every
time
we
run
a
load
test,
it
is
also
persisting.
B
The
results
using
mesh
reefs
asks,
which
are
the
one
that
is
actually
on
AWS.
Results
are
stored
in
dynamo,
but
for
testing
like
and
I,
have
a
local
dynamo
and
so
I'm,
actually
using
that
now
till
they're,
looking
at
me
only
been
pushing
results
to
the
remote
endpoint
like
not
to
persist
it.
So
one
of
the
other
interesting
things
we
wanted
to
do
with
measure.
It
was
actually
to
fetch
the
results
and
present
it
to
the
user
selectmen
so
that
they
can
consume
that
in
some
form.
B
So
this
new
section
was
actually
created
to
address
that,
so
this
is
kind
of
an
initial
version.
So
we
have
a
table
which
actually
fetches
the
data
from
dynamo
I've
just
highlighted
some
of
the
attributes,
like
you
know
up
front
and
again,
like
you
know,
all
this
is
actually
fetching
data
from
the
remote
my
Sri
SAS
endpoint.
So
the
results
are
paged.
There
are
some
challenges
when
we
you're
using
dynamo,
it's
not
easy
to
quickly
get
the
total
number
of
results
or
make
it
searchable.
B
B
The
other
functionalities
is
that
now
this
results,
these
results
are
not
fun
if
we
just
want
to
see
pages
after
pages
of
course,
like
I
searching
as
a
functionality
that
we'll
get
out
of
probably
the
next
week
now.
The
other
interesting
thing
is
actually
the
charting.
So
if
you
expand
the
role,
it'll
actually
chart
out
the
chart
for
that
particular
row
of
data.
So
the
same
way
like
you
can
actually
have
multiple
chart
like
you:
can
you
can
visually
compare
them
this
way
and
I
mean
so
on
and
so
forth?
B
So
you
can
actually
do
this,
like
you
know,
for
every
row
in
the
table,
but
again,
looking
at
all,
since
the
results
are
paged,
it's
not
easy
to
actually
compare
them
across
pages,
so
what
I've
done
is
I've
actually
have
made.
This
selection
persist
across
pages,
so
so
right
now
it
says
one
row
selector
long.
B
They
are
on
page
one,
so
I'm
just
going
to
go
to
a
different
page,
select
a
different
row
and
you
can
see
the
number
of
rows
got,
updated
and
so
on
and
so
forth,
and
then
the
right
table
can
also
have
this.
Compare
selectively.
You
like
Anna,
where
it'll
it'll
compare
all
the
selected
charts.
So
here
this
is
a
complex
chart
like
you
know
where
it
actually
is
comparing
four
data
points,
so
they
are
vertically
distinguished.
So
it's
like
one
two,
three
four
essentially
and
you
can
actually
hover
over
them
and
interact
with
them.
B
This
is
charges
yeah
same
charges.
No,
the
way
this
works
is
that
if
you
click
on
one
of
course
like
you
can
expand
and
see
the
chart.
But
if
you
just
click
on
one
and
like
it'll
hit,
compare
selected,
you
will
see
the
single
chart
now
you,
if
you
click
on
two
and
hit
this
you'll,
see
two
charts
overlaid
over
one
another,
but
then
from
three
onwards
it
actually
switches.
Now
again,
this
is
not
something
totally
new.
B
This
is
I'm
piggybacking
on
work
for
IO
does
so
more
than
three
and
more
charts
like
and
I
will
actually
use
this.
So
you
can
now
see
three
three
vertical
points
and
looking
on
so
all
the
other
values
are
actually
you
know
graphed
around
that,
so
the
same
way,
if
you
actually
use
for
you,
will
see
four
columns
and
so
on
and
so
forth.
So
you
know
so
now,
since
this
is
across
pages,
you
can
actually
go
crazy
and,
like
I'll,
compare
how
many
ever
you
want,
so
so
that
six
yeah.
B
So
this
was
one
of
the
like.
You
know
the
major
work
like
not
that
it
got
done
this
week.
Of
course,
like
you
know,
so
the
back
end
has
a
support.
My
Cherie
has
a
support
and
the
mushriq'
I
has
the
support
so,
but
they
try.
These
changes
have
not
yet
been
merged,
master
they're
still
in
the
in
the
branch
you
know,
I'm
just
working
on
some
other
fine-tuning
stuff,
once
those
are
fine-tuned,
I'll
be
merging
into
master
now.
B
One
other
thing
is
this:
D
selection,
so
here
you
so
now
it
shows
that
six
fours
are
selected,
but
but
we
know
that
it's
across
pages,
if
we
just
want
to
deselect
the
pay,
the
selection
within
this
page,
we,
you
can
actually
use
the
button
here
at
the
top,
which
will
deselect
the
things
on
that
one
particular
page.
What
if
we
want
to
deselect
across
all
pages,
you
can
use
the
one
at
the
top.
So
if
you
click
on
that,
you
can
see
that
it
be
selected
all
the
items
across
all
the
pages.
B
A
B
B
I
can
openly
embed
it
on
a
separate
page,
but
that's
something
even
if
I
reload
like
and
it
doesn't
make
any
difference.
It's
still
stuck
at
like
you
know
that
particular
range.
So
that's
going
to
be
an
interesting
thing
with
how
we
get
the
charge
from
graph
Anna
or
you
know
worst
case.
You
know
just
go
the
Prometheus
route
but
know
where
we
go
talk
to
Prometheus
fetch
the
data
and
strain
it
to
some
charge
in
the
UI.
That's
pretty
much
like
know.
What
do
you
have
to
show.
A
B
A
A
B
A
Nice
all
right
guys,
I'll,
go
I'll,
tell
you
about
how
scale
the
scale
conference
was
next
week.
I'm.
Actually
literally
next
week,
I'm
presenting
out
at
open
source,
Leadership
Summit,
probably
I'll,
be
traveling
that
week
as
well,
but
will
be.