►
From YouTube: Istio Community Meeting 6.14.18
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
All
right,
good
morning,
everyone-
hopefully
you
can
hear
me
all
right,
awesome.
Thank
you
yeah.
So
we've
got
a
full
agenda
for
today,
which
is
awesome
because
I
am
fighting
a
cold.
So
keeps
me
from
having
to
talk
a
lot.
We've
got
some
great
demos
that
are
set
up
for
today
and
if
you
I
put
the
link
to
the
working
doc
in
the
chat.
B
Thanks
April,
so
I'm
me,
I
love
this
tattoo
and
I'm
also
involved
in
the
Filipino
project
and
I
actually
have
a
couple
of
my
teammates
here
on
the
on
the
cold.
Cussin
and
luck.
Mouth
and
Carson
have
been,
has
been
working
on
the
East
EO
integration
and
what
Mel
did
some
work
on
our
kubernetes
integration?
B
So
hopefully
you
should
be
able
to
see
my
screen
now.
So
ballerina
is
it's
a
programming
language
that
has
been
designed
for
micro-services,
and
so
it
has
a
lot
of
native
functionality
for
things
like
type
safety
and
distributed
transactions
and
beady
eyes,
and
so
on.
So
I'll
just
do
a
quick
kind
of
hello
world
service
just
to
give
you
a
feeling
of
what
it
looks
like
and
then
I'll
deploy
it
in
kubernetes
and
then
Katherine
will
take
it
from
here
and
until
they're,
at
least
your
part.
B
So
since
it's
hell
hello,
world
service
and
I'll
just
use
HTTP,
we
have
support
for
the
protocols
like
WebSockets,
cherkassky
and
so
on.
I'll,
do
it
so
and
I'll
do
just
simple
again:
HTTP
service.
That
would
do
a
color
well
thing,
hello,
let
the
resource
will
call
the
resource
hi
and
then,
if
I
want
to
just
send
some
HTTP
response,
create
response,
object,
I'll
set
some
payload
to
say
hello,
world
and
then
I
will
just
use
the
colorant
point
that
got
passed.
B
B
Bind
my
endpoint
my
my
service
to
that
endpoint.
So
that's
sort
of
the
simple
scalar
well
teen,
and
then,
if
I
want
I,
can
customize
it
a
little
further
and
use
annotations
for
that.
So,
for
example,
I
want
to
do
to
say
that
my
service
should
not
be
listening
to
hello
path,
but
it
should
help
some
make
a
different
base
path
or
I
can
customize
my
resource
and
say
that
for
the
resource
again,
if
I
want
again
change
paths,
I
can
change
methods.
B
Example
to
post
I
can
burn
to
something
fancier
like
I
can
use
that
request
to
actually
extract
something.
So
I
can
say
that
if
I
want
to
get
like
a
screen
from
the
request
and
I
can
do
that,
I
can
say:
request
get
text
payload
and
it's
a
it's
Union
type
and
it
returns
it
can
return
the
screen
or
an
error.
I
can
do
a
match
and
then
figure
out
what
I
do
in
each
case
or
I
can
just
use
the
check
you
would
which
would
propagate
the
error.
B
Right
and
if
I
so,
that's
sort
of
the
simple
way
to
do
things
and
the
cool
thing
is
that
so
which
gives
a
full
language
it
gets
compiled
and
run.
What
if
I
want
I
can
use
cementations
and
in
that
case,
I
would
compile
it
not
an
executable
but
zero
to
my
to
community
side
effects,
so
example,
right
now,
if
I
do
LS,
I'm
gonna
have
that
Bell
file,
but
if
I
had
some,
if
I
had
kinetise
functionality.
B
Can
say
that
I
want
to
change
that
I
want
for
that
listener
to
be
to
generate
an
a
note
for
service,
and
they
want
for
that
service
to
add
some
arm.
Kinetic
functionality
as
well
so
I
could
add
a
config
map.
I
could
send
some
ingress
or
do
some
scaling
frames,
etc.
I'll
just
do
a
deployment
and
I
can
out-cook
in
paste.
B
You
will
see
that,
instead
of
just
creating
a
run
executable,
it
will
go
and
use
these
annotations
to
generate
kubernetes
artifacts.
So
with
what
generated
there,
the
service
definition-
and
it
will
generate
me
the
docker
image
and
that
blender
definition
and
then
I
would
be
able
to
just
curious,
then
I,
come
in
I
will
be
able
just
to
do
a
few
control
and
apply
them.
So
now,
if
I
do
three,
you
can
see
that
it's
not
just
generating
me
their
legs,
give.
Oh,
you
generate
the
whole
yellow
structure,
so
I
can
do
you.
B
B
B
See
all
the
examples
of
lots
of
things
that
it
can
do.
It
has
all
their
concurrency
functionality
with
the
workers
at
the
forks
and
joints.
It
knows
datatypes
different
formats
and
so
on
so
forth.
But
since
it
is
an
u0,
we
can
go.
I
think
I'll.
Stop
sharing
on
that
left.
I've
got
some
take
over
and
talk
specifically
about
their
easier
integration
that
we
have
what
we
have
right
now
and
what
we
are
working
on
so
Carson.
Can
you
yeah
take
over
the.
E
D
Hope
you
hope
you
can't
see
my
screen
right.
Yes,
okay,
so
as
part
of
the
steer
integration
with
ballerina,
what
we
have
done
was
so
very
nice
programming
language
and
we
develop
services
with
ballerina,
and
so
now
we
need
to
do
deploy
our
service
in
inside
steel
service
mesh.
So
the
deployment
looks
like
this.
D
This
particular
diagram,
so
you
have
a
ballerina,
a
time
service
running
I'm,
just
just
giving
the
time
current
time
and
which
is
four
which
is
running
alongside
my
proxy
and
again
managed
by
a
steel
control
train
and
for
for
the
demonstration
I
use
graph
on
ax
for
observability
purposes.
So,
as
Dmitry
showed
you
earlier,
we
have
native
support
for
kubernetes
through
annotations
right.
D
You
can
just
put
couple
of
annotations
inside
your
ballerina
service
and
it
you
can
generate
kubernetes
deployment
descriptors
so
when
it
comes
to
steal
again
prior
to
that
so
I'm
using
steo
point
seven
release
for
this
particular
demo.
I
know
there
are
some
changes
after
that,
like
in
the
point,
eight
release
I'll
come
back
to
the
those
changes
later
but
based
on
the
point,
seven
release
osteo.
D
D
So
ingress
class
is
steel,
and
so
this
is
the
community
service
definition
which
is
annotated
alongside
the
inbound
listener
of
ballerina,
and
this
is
my
ingress,
and
so
this
is
the
deployment.
So
we
use
this
parameter
to
generate
a
single
file.
That's
a
very
nice
Pacific
thing:
that's
all
this
is
the
ballerina
cord
right.
So
now,
I
have
my.
D
So
it's
generating
the
service
deployment
and
also
creates
the
respective
docker
image,
and
so
once
week
we
created
those
artifact.
We
need
to
do
some
changes
because
by
default
the
artifact
generated
by
Ballerina,
so
we
are
using
fabricate
client
library
to
generate
these
artifacts.
There
are
some
manual
changes
that
we
have
to
do
so
I
have
actually
included
them
in
our
documentation.
There
are
a
couple
of
changes
like
removing
this
host
and
removing
this
ingress
class
stuff,
like
Indian,
X,
stuff
and
and
this
protocol
changing
protocol
TCP
to
name
it
TTP.
D
D
D
D
We
can
see
on
graph
on
ax,
so
I
have
installed
Prometheus
and
graph
on
R
and
ons
and
I
should
be
able
to
see
the
inbound
traffic
to
the
mesh,
as
well
as
the
HTTP
service,
which
is
ballerina
service,
so,
like
it's
quite
pretty
straightforward
changes
that
we
have
to
do,
but
but
we
are
planning
to
do
as
the
next
step.
As
you
know
like
in
the
point
8
release,
we
don't
have
ingress
right
and
we
have
the
gateway
and
virtual
service
configuration
in
steel.
D
So
we
are
planning
to
support
this
natively
through
a
specific
Steel
annotation.
This
is
like
a
rough
sketch,
a
rough
idea
that
we
came
up
with.
So
what
we
are
going
to
do
is
we
are
going
to
include
virtual
service
and
gateway
annotations.
As
part
of
ballerinas
service,
so
this
is
actually
a
ballerinas
service,
accordionist
stupid
from
a
ballerina
service,
annotations
so
similar
to
kubernetes.
E
I
have
a
question
so
so
it
seems
to
me:
I
am
NOT
from
very
familiar
with
ballerina,
although
I
do
like
the
logo,
so
ballerina
has
a
bunch
of
its
own
reasons
you
might
want
to
adopt.
It
looks
like
the
way
it
handles
data
is,
is
compelling.
Parallel
processing
processing
seems
interesting,
but
but
you're
also
here
showing
something
that
is
separate,
which
is
in
your
stores
files.
Essentially,
you
have
the
ability
to
auto
generate
a
bunch
of
the
yam.
All
that
other
people
in
other
languages
are
composing
by
themselves.
D
C
B
E
Yeah,
thank
you
very
much
for
the
demo
I
mean.
Certainly
you
can
see
that
other
people
in
other
languages
might
do
well
to
take
the
same
idea
of
it.
Hey
if
everybody
in
the
world
is
having
to
compose
these
yellow
files.
Have
them
be
an
output
of
the
build
process
instead
of
having
everyone
just
composed
of
my
hand
or
community
and
that's
cool?
That's
that's
a
absolutely
a
Productivity
that.
D
A
D
D
A
B
And
actually
yeah
so
April
if
you
don't
and
and
I
we're
working
on
getting
an
sto
talking
to
the
conference
as
well.
So
it's
July
18th
in
San,
Francisco
and
I
laughs,
like
I'm,
posed
in
the
chat
window,
I'll
post
a
link
to
the
conference
and
also
the
coupon
code
that
you
can
use
to
register
for
free.
A
C
So
that's
kind
of
what
I
released
about
two
days
ago,
which
I
called
this
tio
and
it's
a
direct
implementation
of
the
visceral
add-on
for
SPO,
so
very
targeted
for
this
deal
and
I'm
going
to
share
my
screen
here.
I'll
walk
through
a
demo
I
think
my
medium
post
is
posted
in
the
working
doc,
so
I'm
actually
going
to
follow,
along
with
it
and
probably
copy/paste
some
commands
from
there.
But
I
have
two
demos
to
show
in
case
one
of
them
fails
so
I'll
share
my
screen
we'll
get
started.
C
C
C
C
Do
it
locally
with
doctor
and
changing
the
values
Curtis
do
so
generates
a
pretty
cool
match
and
I'll
show
you
if
you've
blown
the
example
Packard
folder
in
the
project,
you
just
need
to
do
a
doctor
compose
up
build,
and
so
once
is
what
this
is
going
to
do
is
stand
up
all
of
the
this
do
components
and
then
also
stand
up
a
mock
mock
service.
There
kind
of
feeds
Prometheus
data
that
you
can
visualize.
So
what
you
need
to
see
here
is
the
dista
web
is
compiling
uses
by
pack.
C
So
it's
a
front-end
framework
and
once
that's
ready,
we
should
be
able
to
open
up
this
yellow,
and
here
it
comes,
and
so
this
is
all
mock
data
for
my
example-
and
this
is
for
you
to
play
around
with
if
I
edit,
a
custom
label
onto
my
SEO
request,
count
metric
so
that
you
could
do
it,
you
could
display
by
cluster.
So
each
of
these
clusters
are
actually
clickable.
You
can
zoom
into
them
and
view
them.
So,
let's
view
the
mesh
in
this
clot
in
the
u.s.
West
to
cluster.
So
this
is
the
generated
mesh.
C
It
kind
of
simulates
real
world
is
theƶ
mesh
insert
an
example
in
certain
ways.
So
I
have
an
issue:
ingress
controller,
that's
running
and
there's
traffic
coming
from
the
air.
So
we
can
click
on
that.
So
you
can
see
how
much
traffic
is
coming
in,
how
much
is
going
out
and
the
load
actually
that
I
put
on
it
is
quite
large.
It
looks
like
four
hundred
thousand
requests
and
that
the
load
actually
doesn't
really
affect
the
flow
or
the
data.
C
C
Yeah,
so
for
the
demo,
the
regally,
nope
I.
Guess
we're
aggravated
look
up
there
so
thought
there
doesn't
work
in
that
down
what
you're
getting
me.
There
is
some
alerting
that
you
can
do
as
well,
and
they
don't
have
this
in
the
demo,
but
you
guys
you'll
be
able
to
see
little
pop-ups
that'll.
Allow
you
to
put
a
debug
a
little
bit
where
your
network
is
going
bad.
So
if
you
see
a
red
line
between
two
services,
you
know
something
bad
is
happening.
You
can
quickly
identify
those
services
and
hopefully
fix
them.
C
So
this
is
a
reskin
UI
from
the
visceral
that
I
kind
of
match
is
Kiyo's
color
palette.
Nothing
really
fancy
here,
the
eye
reskin
did
you
can
also
search
for
so
there's
an
ingress
gateway.
If
you
need
to
search
for
certain
components,
so
a
very,
very
nice
UI,
there's
some
cheap
fun
you
can
have
with
the
display.
So
you
can
drag
him,
though,
then
you
can
add
some
physics
and
so
there's
some
fun
stuff
that
I
don't
know
if
it
really
hasn't
used.
But
so
that's
the
docker
compose
demo.
C
C
C
We
actually
have
to
run
kupono
port
forward
for
both
of
the
api
and
the
web,
and
we
can
also
shouldn't
actually
note
that
you
should
actually
I'm
going
to
generate
some
load
as
well
on
the
so
I'm
logged
into
the
making
service
and
I
have
here.
I
just
have
a
simple
curl
to
call
the
product
page
pretty
much
over
and
over.
So
this
is
good.
C
C
There
we
go
okay,
so
the
API
server
is
now
up
and
running
a
9091,
and
so
we
can
go
over
and
check
to
see
that
we're
actually
getting
data.
So
it's
a
good
check
to
make
sure
that
everything's
working
well
then
so
you
felt
well
close
9091
graph
and
we
got
a
lot
of
data
lifts
it
here.
So
that's
good!
That
means
we're
getting
data
from
Prometheus
and
that
this
is
going
to
be
fed
to
the
UI
for
rendering.
So,
let's
expose.
C
C
So
this
is
more
of
just
a
an
example
to
get
your
demo
up
and
running.
Let's
see
that
at
first
and
I
think
that's
pretty
much
it.
It's
really
fun.
We're
actually
going
to
be
using
it
at
the
weather
company
because
we
have
quite
a
large
mesh
and
it's
gonna
really
help
us
identify
where
issues
are
arising
between
services.
So
it's
a
really
quick
view
to
start
with
to
determine
where
you're
having
issues.