►
From YouTube: Istio Community Meeting 4.18.19
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
Right
now,
there's
nothing
on
the
agenda
for
today.
So
if
y'all
have
anything
you
want
to
talk
about,
please
add
it
to
the
agenda,
the
link
to
which
I
just
posted
in
the
chat-
and,
of
course
you
can
ask
stuff
here
as
well,
but
yeah
I,
think
that
was
dancer
really
but
I
didn't
hear
you
very
well.
Can
I
ask
you.
B
A
B
B
A
A
A
B
To
the
I
think
it's
sometimes
called
be
containers
101
meet
up
last
night
here
in
South
Bay.
There
were
three
speakers:
good
John
from
Palo
Alto
Networks
you
get
to
talk
about
admission
controllers
peer
kubernetes
I
spoke
about
is
true,
sto
and
I,
really
kind
of
gave
the
history
of
how
we
evolved
internally
and
how
we
arrived
at
the
architectures
vide
studio
has,
and
then,
after
that,
Chris
and
Karthik
from
Guerra
talked
about
network
policy
and
SCO
policy
and
how
a
combination
of
layer,
3
and
layer
7
are
very
powerful
and
useful.
B
As
always,
when
I
go
to
one
of
these
I
overestimate
the
amount,
the
level
of
sophistication
people
have
a
biased
view.
It
is
not
so
late
in
the
adoption
curve
that
you
should
assume
everybody
knows
about
is
Geo
and
the
parts
of
the
CEO
and
I
always
think
I.
Don't
you
know
and
I
don't
mind
explaining
and
like
starting
at
the
very
beginning,
but
I
worry,
oh,
if
everybody
knows
what
it
is,
this
will
be
a
waste
of
their
I
realized.
B
From
now
on,
I
should
just
always
have
my
get
to
know
his
do
deck
kind
of
and
start
off
by
saying,
hey,
who
feels
really
really
comfortable
with
what
he's
doing
so
in
the
process
of
giving
our
two
talks
that
were
about
is,
do
we
ended
up
kind
of
explaining
a
bunch
about
the
yeah?
What
is
this
T?
So
that's
my
my
overwhelming
thing
is
that
it's
it's
easy
to
talk
Castillo
all
day
and
assume.
Everybody
knows
everything.
B
For
the
most
part,
most
people
who
show
up
at
some
random,
kubernetes
or
container
meetup
on
don't
understand
what
most
of
the
people
on
this
call
know,
which
is
one
of
these
components,
and
why
do
you
use
them?
And
you
know
just
just
basic
stuff
like
that.
That
said,
you
know
really
interesting.
6070
people
at
least
lots
of
questions.
Yes
topic.
People
are
interested
I.
D
E
A
A
F
E
B
A
Custom
URL
but
I
guess
we
didn't
have
enough
subscribers
at
the
time
yet
so
tell
your
friends
to
subscribe.
That's
like
the
default
YouTube
statement
hit
that
like
button
and
subscribe
smash
that,
like
button
smash,
sorry,
you
know
for
someone
who
doesn't
really
know
how
YouTube
works.
You
sure
know
a
lot
about
what
youtubers
say.
G
Always
when
here
I
was
I
was
gonna,
integrate
some
some
virtual
machines,
possibly
next
week
into
the
history
of
mesh
and
I
know.
We
have
some
documentation
here
in
terms
of
meshing,
expansion
I
was
wondering
if
anybody
had
any
experience
doing
that
at
one
point
one
if
there
are
any
anything's
to
look
out
for
I'm
doing
I'm
doing
it
on
Google
cloud
as
the
instructions
kind
of
have
set
up
here,
but
does
interested.
If
anybody
has
any
experience
working
with
that
they're
good
I
think
database
workloads
are
running
on
I
believe
sent
OS
machines.
G
Know
sounds
good
I'll,
I'll
I'll,
just
report
back
then
on
the
next
call
and
kind
of
give
an
outline
of
what
happened.
G
I'm
not
sure
roaring
I
was
just
trying
to
see
if
I
could
get
it
to
work.
Essentially,
they're
they're,
they're
database
work
ones
that
don't
have
any
toiletries
so
that
that
would
be
kind
of
a
start,
but
maybe
to
just
I'm
guessing
you
also
get
the
mutual
TLS,
but
I
I,
don't
think
so
because
nothing's
running
on
the
on
on
the
database
nodes,
so
yeah
I,
I'm
I'm,
just
kind
of
seeing
if
I
can
get
it
to
work
and
then
see
what
work
services
I
can
get
running
from
there.
So
I.
B
The
vulnerabilities
are
centered
on
the
fact
that
envoy
did
not
normalize,
HTTP
URI
paths
and
did
not
fully
validate
HTTP
one
one
header
values
and
then
so.
The
vulnerabilities
impact
of
a
stele
features
that
relied
on
I'm
going
to
enforce
authorization
routing
early.
There
is
a
in
the
1.1.2
announcement,
I'll
post
I'll
post.
The
link
here,
which
has
in
turn
links
to
the.
H
B
That
is,
that
is
quite
worrying.
I
think
that
we
should
bring
it
up
to
the
TOC
there.
There
is
everyone's
probably
aware
that
we're
we've
declared
a
code
moth
on
sto
and
encode
moths,
the
term
we
use
internally
at
Google
that
basically
says
the
project
itself,
the
infrastructure
around
the
project,
the
testing
and
release
of
the
project.
It
needs
some
work,
and
so
our
emphasis
in
1.2
is
much
more
project
and
quality
related
than
then
feature
related.
B
B
F
F
And
I
and
I
know:
I
was
just
looking
at
an
issue.
I
think
that
came
in
this
morning
that
somebody
had
mentioned
that
it
was
fixed
in
1.1
or
maybe
they
thought
it
I
forget
they
thought
it
was
fixed
in
one
and
we're
wondering
the
same
thing,
but
I
did
go
and
look
and
it
did
get
moved
to
the
other
branch,
so
they
weren't
seeing
the
same
problem
in
both.
H
H
D
So
it
was
two
distinct
clusters
with
a
single
control:
yeah
single
control
plane.
So
all
traffic
went
into
one
cluster
and
then
it
would
distribute
to
the
second
cluster
I
think
so
we
may.
We
were
calling
cross
cluster
when
you
have
multiple
control
planes
and
you
can
enter
any
control,
plane
and
so
I
think
technically
I
think
we
actually
had
that
set
up.
But
that
wasn't
the
purpose
of
the
talk
and
stuff
that
we
were
doing.
H
C
D
The
use
case
here
is
like
okay:
I
want
to
have
failover,
but
realistically
it's
like
hey
man,
if
you're
coming
in
through
the
uswest
server
I,
don't
want
your
traffic
to
get
routed
across
to
the
East
Coast
for
no
reason
and
I
want
you
to
get
the
the
local
pods
here
and
I
didn't
see
a
way
to
make
that
happen.
I
have
been
assured
that
it's
possible
for
that
at
about
how
to
do
it.
The
other
people
you
might
want
to
talk
to
are
the
guys
from
stack
point
cloud.
D
They
were
acquired,
but
they,
their
entire
focus,
has
been
around
doing
setups
for
multi
cluster
multi-cloud.
Remember
what
they're
called
now,
because
they
were
acquired
and
they're
part
of
like
NetApp
or
something
no,
but
they
actually
built
a
bunch
of
stuff
around
cross
cluster
using
Sto
the
guys
to
talk
to
yeah.
They
still
have
their
select
channel.
D
D
B
B
Just
like
I
mean
with
with
kubernetes
ra
he's
done
updates
over
time,
as
kubernetes
has
changed.
I
know
he's
working
with
with
another
Googler
someone
who's,
I'm
on
the
SEO
team,
but
is
very,
very
new.
Still
has
that
outsider
perspective,
which
is
very
important
and
which
kelsey
always
is
able
to
maintain
no
I,
think
it's
great
I
think
that
existence
of
kubernetes
the
hard
way
actually
hold
kubernetes
as
a
project.