►
From YouTube: 09.09.2020 Community Meeting
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
B
C
B
No,
so
it's
basically
the
fire,
it's
the
particulates
from
the
fire
fires
in
california,
but
the
it's
like
the
smoke
and
everything
is
high,
high
up
and
high
enough
up
in
the
atmosphere
that
it
doesn't
smell
smoky
outside.
But
it's
created
this
weird
filter.
So
it's
actually
it's
super
dark
outside.
You
know
it's
10
at
10
a.m,
but
it's
super
dark
and
it's
like
orange
yeah.
B
B
So,
oh
joe,
I
see
you
typing
away
in
here.
I
was
going
to
say
like
if
we
can
kick
it
off
with
a
couple
of
updates
from
the
solo
crew
on
we've
got
some
progress
happening
within
blue
on
the
open
source
side.
So
I
wanted
to
have.
We
have
some
of
the
glue
team
here.
They
can
give
a
little
update
on
what's
in
the
open
source,
betas
right
now
and
then
service
mesh
hub.
B
A
Yeah
for
sure
I
can
speak
to
that
zero.
Seven,
two
release
we're
now
up
to
zero,
seven,
four
following
a
couple
of
more
fixes,
but
if
you
do
download
the
latest
service
mesh
hub
you'll
find
that
we
now
have
support
for
open
service
mesh,
the
new
service
mesh
project
led
by
microsoft,
and
you
know
that
just
kind
of
continues
the
story
of
service
mesh
hub,
providing
a
way
to
configure
any
service
mesh
that
you
bring
to
it
next
up,
we're
targeting
mesh
support.
A
But
again,
if
you
install
open
service
or
sorry
service,
mesh
shop,
zero,
seven,
two
or
later
you
will
find
that
you
are
now
able
to
configure,
discover
et
cetera
your
local
installation
of
open
service
mesh.
So
yeah,
please
give
that
a
spin.
Let
us
know
what
you
think
and
feel
free
to
file
any
issues
or
reach
out
on
slack
with
any
questions
or
additional
feature
requests.
A
Absolutely
yeah
on
the
open
service
mesh
front
that
project
is
is
fairly
young,
so
the
support
that
we
have
isn't
like
super
robust
in
terms
of
the
like
api
that
is
provided
by
service
meshup,
like
the
features
that
are
available
on
osm,
are
a
subset
for
now.
It's
largely
based,
since
their
api
is
largely
based
on
the
smi,
spec
and
api
there.
A
But
you
know,
as
features
are
added
on
that
front,
we
will
do
our
best
to
keep
up
with
the
state
of
the
art
for
osm
beyond
that
on
service
mesh
hub.
Our,
like
main
focus,
has
been
adding
app
mesh
support
to
our
like
existing
service
master
controller,
to
allow
app
mesh
to
app
mesh
communication
and
configuration
through
our
api.
A
We
get
some
technical
details
if
you
think
now's
the
time.
Betty
scott's
here
he's
been
kind
of
leading
our
effort
or
we
could
move
along
to
the
next
topic.
B
So
we
can
add
a
little
color
there,
and
then
we
can
see
I'm
just
scrolling
actually
through
the
service
mesh
hub
channel
in
the
community
to
see,
if
there's
any
kind
of
maybe
a
question
that
has
come
up
recently
that
we
just
want
to.
You
know
share
out
here
in
this
call,
but
scott,
if
you
want
to,
if
you
want
to
talk
a
little
bit
about
the
what's
coming
up,
that'd,
be
great,
give
a
little
more
detail.
D
So,
let's
see
for
at
mesh
we've
gone
with
a
few
constraints
that
were
not
there
before.
Basically,
we
are
we're
building
on
top
of
the
atmesh
controller
for
eks,
which
is
basically
the
recommended
way
of
using
app
mesh
with
kubernetes
today,
and
so
there's
just
a
bit
of
a
number
of
extension
points
for
service
mesh
hub
that
we're
currently
implementing
basically
to
support,
as
as
many
of
the
features
that
we
can
on
at
mesh,
which
means
you'll
be
able
to
use
the
same
api.
D
The
same
semantics
that
you're
already
familiar
with
with
the
service
mesh
shop
to
control
app
mesh.
It's
particularly
nice
because
app
mesh
has
a
bit
of
a
complex
workflow.
In
order
to
achieve
some
of
the
simple
use
cases
like
traffic
shifts,
and
we
will
yeah,
I
mean
if
anyone's
interested
in
the
internals
of
service
mesh
hub
and
how
things
are
actually
working
internally.
We
can
get
into
that,
but
from
a
high
level.
D
Essentially,
what
it
means
is
we'll
be
able
to
integrate
into
you'll,
be
able
to
install
service
mesh
hub
register,
a
cluster
that's
running
on
eks,
and
we
will
be
able
to
detect
that
app
mesh
has
been
integrated
into
that
cluster
and
then
assume
control
of
it
as
well
as
monitoring
it
and
so
on.
C
D
D
Are
going
to
be
deferring
that
behavior
to
the
app
mesh
controller?
So
what
you'll
do
is
you
have
an
eks
cluster?
That's
running,
you
install
the
app
mesh
controller
there
as
a
part
of
the
workflow
for
setting
up
the
cluster.
The
atmesh
controller
will
handle
interacting
with
the
api,
which
means
all
I
am
just
can
be
handled
in
the
standard
fashion.
C
D
C
D
C
Okay,
I
guess
part
of
the
reason
I
was
asking
that
is,
I
mean
we'll
talk
about
this
with
a
limited
trust
and
some
of
the
identity
issues
that
come
up.
So
I
was
wondering
what
you
might
have
decided
there
and
it
sounds
like
you
were
able
to
get
around
it
for
the
api
access.
But
I
wonder
when
we
come
around
to
identity
between
app
mesh
and
istio
or
whatever
we
might
still,
we
might
get
right
back
to
that
same
problem.
You
know
yeah.
D
D
Will
provide
you
know,
native
ways
of
integrating
with
the
im
on
various
platforms
and
like
we've
discussed
in
the
past.
Limited
trust
will
allow
us
to
map
trust
across
domains
through
service.
Bishop.
B
Cool
thanks:
if
there's
no
additional
questions,
you
can
have
someone
from
the
glue
team
talk
about
some
of
the
enhancements
we've
been
making
on
the
glue
open
source
site
and
if
there's
any,
if
there's
any
kind
of
like
issue,
you
know
outstanding
issues
or
requests
that
we
want
to
discuss
on
this
call.
E
Sure
yeah,
so
I
can
talk
to
some
of
the
changes
coming
in
glue,
so
the
it's
hard
blue,
it's
harder
to
summarize,
because
we
have
it's
a
lot
more
smaller
different
requests
or
feature
work,
touching
a
lot
of
different
parts
of
glue,
but
it's
in
terms
of
like
long
big
story
not
as
much
for
one
to
five,
but
some
of
the
things
you
can
expect
in
1.5
includes
like
exposing
the
grpc
to
json
transcoding,
filter
kind
of
raw
from
envoy.
E
We
did
some
improvements
to
the
route
table
delegation
model,
so
you
can
have
more
powerful
selectors
and
you
can
use
more
different
kinds
of
route
table
matchers
in
your
delegation
models
we're
currently
in
the
process
of
reworking
and
should
have,
by
the
time
of
the
release,
improvements
to
the
rate
limit
api
that
allows
you
to
handle
wildcarding.
You
know
more
generically
support
for
you
know:
aws
sts,
credentials,
simple
censorship,
token
service-
and
this
is
just
kind
of
giving
an
idea
like
those
are
the
kinds
of
things
we've
been
working
on.
E
But
in
terms
of
you
know,
big
story
picture,
I
think
glue
federation
and
multi-clusters.
It's
kind
of
the
bigger
push
there.
I
could
speak
to
that
for
a
bit.
Joe
would
also
be
a
good
person
to
talk
about
that.
If
we
had
follow-up
questions
happy
to
go
into
depth
on
any
of
those,
if
you
want
to
talk
about
them
further,
if
you
have
any
questions
or
things
you
want
to
raise,
that
now
would
be
a
great
time
to
do.
C
E
Yeah,
let
me
try
to
dig
that
up.
We
have
some
some
stuff
written
down.
Let
me
give
you
a
second
and
I
should
drop
it
in
there.
C
B
Well,
actually,
when
we
have
some
anything
on
the,
I
only
listed
service,
my
shopping
glue
in
the
projects
here,
but
I
think,
if
there's
any
other
anyone
else
from
the
solo
side,
if
we
want,
if
there's
something
to
share
on
some
of
our
open
source
and
list
everything
so.
B
Yeah
this
is
community,
so
I
think
we
should
keep
it
open
source
because
then
it's
something
that
folks
can
contribute
into
as
well,
but
we
do
have
a
dev
portal
for
folks
who
are
using
glue
and
or
seo.
We
do
have
a
dev
portal,
but
that
is
that's.
That
is
kind
of
that
is
in
the
enterprise
version.
So
we
could
discuss
if
people
have
interest
or
questions
here,
but
I
think
it's
you
know
it's
not
the
same
as
like.
E
Sure
I
will
say,
I've
noticed
an
uptick
in
some
of
the
community
contributions,
at
least
in
glue,
and
one
of
the
other
ones
I
was
really
excited
about.
Is
the
community
took
it
upon
themselves
to
help
glue,
build
on
arm
64
architectures,
which
has
been
great
for
running
it
and
like
raspberry,
pi's
and
stuff
like
that?
So
just
one
more
thing
we
can
look
forward
to
at
1.5
and
we
appreciate
all
the
contributions
that
we're
getting
fantastic.
B
Oh,
that's
a
good
point.
We
should
actually
ask
ask
ask
them
to
join
the
next
call
to
just
talk
about
it.
That'd
be
fun.
B
Okay,
then,
let's
see
here,
there's
there's
a
question
here
on
the
discuss:
the
limited
trust
implementation
options.
So
I
don't
know
who
added?
That
john
was
that
you.
C
B
C
Yeah
daniel's
on
he's
been
doing
a
lot
of
the
work,
he's
actually
gotten
some
prototypes
up
and
working,
but
we're
not
really
in
a
position
to
show
it
today.
C
The
other
one
mihai
has
been
on
holiday
for
a
little
bit
so
kind
of
slowing
us
down
a
little
bit
but
anyway.
I
think
that
the
main
point
that
we
wanted
to
get
to
there's
a
couple
macro
points.
One
is
that
there's
a
couple
different
ways
you
can
do
this.
C
Obviously
you
know
how
whether
whether
you
do
something
like
even
christian
pointed
to
in
the
issue
with
this
deal,
where
you
have
trust,
bundle
all
the
way
from
work
or
workload
or
if
you
do
something
with
gateways
as
daniel
put
together
or
even
something
that,
within
your
current
documentation
with
respect
to
limited
trust,
and
it's
mostly
about
how
you
want
to
manage
their
certificates
and
how
they're
rooted
and
stuff
like
that.
C
So
I
don't
know
if
there's
strong
opinions
here
in
the
group,
I
think
my
opinion
is
that
we
probably
want
to
have
some
ability
to
support
multiple
modes,
even
not
necessarily
out
of
the
gate,
but
over
time
and
and
the
other.
C
So
that's
one
aspect
of:
do
we
want
to
go
to
a
specific
sort
of
model
of
how
we
do
this
or
not,
if
there's
any
strong
opinions
and
then
the
second
is
that,
with
respect
to
this,
we're
going
to
need
some
way
to
manage
the
credentials
that
need
since
they're
not
commonly
rooted,
there's
going
to
be
some
api
implications
of
how
that
credential
information
gets
around
and
the
exact
credential
information
will
depend
a
little
bit
on
the
model
so
being
a
little
fuzzy
here
on
the
exact
details,
mostly
it's
just
to
collect
input
as
to
whether
there's
strong
opinions,
one
way
or
the
other
for
either
of
these
topics
before
we
just
get
a
design
together,
probably
in
time
for
the
next
community
meeting.
C
So
I
I
don't
know
betty
if
you
want
to
just
share
the
issue
I
did
between
daniel
mihai
and
I
we
kind
of
at
least
enumerate
some
of
the
implementation
models
that
are
currently
listed.
Christian
christian
kind
of
roughly
pointed
to
istio
that's
sort
of
the
trust
bundle
model.
Then
mihai
had
an
original
architecture,
daniel
added
something
a
little
more
specific
on
using
gateways,
and
then
I
just
referenced
the
currently
documented
the
model.
F
So
I
don't
think
in
general,
there's
gonna
be
a
lot
of
strong
opinions,
but
I
would
the
one
that
I
would
have.
Is
that
to
your
point
about
supporting
multiple
different
types,
the
the
apis
that
we
build
are
generic
enough
or
or
not
generic,
but
like
they
account
for
the
fact
that
there
will
be
multiple
different
identity
models
in
in
the
different
meshes
that
we
use.
So
just
as.
A
F
C
Yeah,
that
makes
sense,
so
should
we
try
to
propose
some
api
thoughts
around
supporting
those
multiple
options.
C
Because
I
you
know,
I
know
api
changes
can
always
really
a
challenging
area,
so
I
wanted
to
just
make
sure
that
that's
a
good
good
good
thing
to
do.
F
C
C
So
daniel
do
you,
do
you
want
to
add
anything?
Are
there
any
specific
questions
you
have
for
the
group
here
that
or
any
clarity
that
you
think
might
be
helpful
as
we
work
through
those
api
and
implementation
details.
G
Yeah
so,
basically
today
I
wrote
a
comment
here
in
the
issue
describing
a
potential
implementation,
so
here
I
have
a
drawing
and
it's
basically
just
referring
to
easter.
G
Basically,
this
this
draft
is
actually
pretty
simple,
just
having
easter,
secure
gateway
with
the
tls
type
mutual
enabled
and
from
the
other
side,
workloads
will
be
configured
with
the
destination
rules
of
type
mutual
as
well,
and
they
will
they'll
basically
have
a
reference
to
their
certificates
and
to
the
other
cluster
to
other
sale
certificate
authorities
that
it
should
trust
so
it
this
model.
I
think
it's
it's
something
might
be
called
the
cross
signing
or
something
like
that.
So,
but
I'm
not
sure
yeah.
G
G
But
it's
it's
still
just
a
sketch.
C
Okay,
so
I'm
based
on
your
input,
question
we'll
we'll
try
to
just
come
out
with
something
a
little
more
concrete
next
week
with
respect
to
api
and
maybe
a
sketch
of
implementation
for
at
least
one
of
the
methods,
and
we
can
iterate
on
that.
I
guess
we'll
enter
it
as
you
put
it
posted
as
a
pr
to
the
design,
doc
folder.
Is
that
what
people
suggest
or
just
add
it
to
the
issue?
What's?
Is
there
a
preferred
approach.
A
C
B
Cool,
let
me
double
check
here.
If
we
have
other.
B
B
If
so,
then
we
will
see
you
next
time
and
yeah.
We'll
probably
have
a
few
more
things
to
talk
about
as
we
get
closer
to
the
end
of
september
and
we
start
getting
ready
to
ship
out
get
make
one
five
also
ship
it
versus
having
all
the
bits
being
beta.
B
C
That's
good
a
question
on
the
webassembly
is
that:
are
you
treating
that
as
part
of
the
service
mesh
hub
project
itself,
or
is
that
sort
of
in
that
context,
of
a
different
project.
B
So
john
one
thing
we
are
doing
with
this
bi-weekly
meeting
is
as
instead
of
creating
separate
forms
for
each
project.
What
we've
done
is
just
bring
all
the
discussions
of
all
the
other
projects
into
this
as
a
standing
community
meeting
and
we've
been
doing
work
with
web
assembly.
For
you
know,
web
assembly
for
envoy
and
kind
of
you
know
the
things
that
are
built
with
envoy
like
istio
and
glue,
and
so
it's
kind
of
in
parallel,
but
you
know
they
can
be
used
together,
so
we're
just
bringing
that.
C
H
I
I
know
harvey
mentioned
a
good
point
in
this
slack
channel
yesterday,
where
I
was
basically
having
some
trouble
getting
the
right
api
to
work,
and
he
mentioned
that
you
know
some
label
shouldn't
be
there
and
I
found
it
to
be
useful.
Although
we
had
some
alternate
strategies,
I
found
out
that
there
is
no
documentation
on
this.
So
do
you
think
it's
useful
to
add
a
developer
documentation
who
might
use
your
api?
F
H
F
H
F
C
I
I
want
sorry
I'm
late
today,
but
I
will
also
add
that
if
you
have
any
dogs
like
because
you're
coming
from
outside
a
lot
of
the
stuff
that
we
are
document-
maybe
you
know
you
know
we
kind
of
like
know
that
so
maybe
we
missed
it
up
and
didn't
document
it
well,
it'd
be
great
if
you
will
be
able
to
request
and
it
talks
like
I
mean
if
you
feel
that
something
is
missing,
that
we
didn't
cover
that
at
all,
or
something
like
that,
mainly
because
you
have
a
fresh
start,
then
we
we
already
opinion.
I
C
B
Awesome
cool,
so
any
other
questions.
Any
other
last
comments.