►
From YouTube: Policies and Telemetry WG - 2020-11-04
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
C
B
D
Question,
maybe
you
stayed
up
late
and
maybe
just
wake
up
to
it.
I
don't
know.
D
E
Yes,
like
you
need
to
stop
presenting,
then
I
I
could
prison
or.
E
So
like
in
1.8,
we
introduced
the
seo
awesome
extension
dashboard,
which
will
have
a
collection
of
well
some
related
matrix.
E
We
gather
them
to
into
the
dashboard
in
the
charts,
which
is
basically
the
dashboard,
looks
like
this.
E
In
the
first
section,
we
have
the
was
on
active
one
created,
and
then
we
have
the
vassar
module
remote
load,
which
has
cash
entry
and
cash
visits
which
will
track.
The
cash
has
misses
an
active
case,
and
the
third
chart
is
the
remote
fetch
which
is
corresponding
to
the
remote
fetch
failures
and
successes,
and
the
last
section
we
present
some
of
the
proxy
memory
and
cpu
usage,
so
those
charts
are
corresponding
to
each
of
the
existing
ones.
On
matrix,
we
have
right
now
yeah.
It
basically
looks
like
this.
E
This
is
a
status
for
1.8
and
in
the
future
we
might
need
to
upgrade
it
to
the
eco
data
point
dashboard,
which
I
think
john
maybe
have
some
similar
work,
which
is
going
on
like
in
this.
In
this
issue,.
E
D
D
Daniel
or
any
or
peter,
or
anyone
that
has
done
a
lot
of
playing
around
with
and
development
work
with
wasm
extensions
does
this
look
like
something
that
would
be
useful
and
helpful
for
you.
C
A
D
Okay,
great
thanks
karen,
I
guess
that's
a
good
transition.
Daniel
did
you
want
us
to
look
at.
Did
you
want
to
present
or
or
walk
through
the
simplified
extensions
api
here
as
the
first
pass?
Or
did
you
just
want
people
to
take
some
time
this
week.
A
I
can
walk
you
through
real,
quick
yeah.
It's
just.
I
just
want
people
to
review
it
exactly.
Let
me
see:
where
is
the
present
button?
A
window?
No
entire
screen.
A
A
Yeah,
maybe
you
do
it.
I
think
google
meet
is
acting
up
in
firefox.
Again
I
have
to
use
chrome
when
I
present.
D
Oh
okay,
yeah,
just
let
me
know
when
you'd
like
me
to
scroll
and
all
that
good
stuff.
A
Right
sure
yeah,
so
I
just
tried
to
based
on
our
discussion
last
time.
I
just
tried
to
keep
the
stock
to
the
parts
that
I
think
are
crucial
to
get
right.
First,
so
the
objective
I'm
describing
a
bit
what
it's
about.
So
it's
it's
really
about
the
the
user
facing
api
and
the
design
for
packaging
format.
So
what
I'm
not
touching
touching
upon
is
the
distribution
of
filters,
because
there
has
been
quite
some
work
by
other
people
looking
into
that.
A
First
of
all,
but
also,
I
think
it's
not
necessarily
so
for
me,
it's
an
implementation
detail
that
doesn't
really
matter
as
much
to
how
the
api
should
be
designed,
in
my
opinion,
so
I'm
right
now,
I'm
actually
not
mentioning
it
at
all.
I
kept
it
in
there
up
there
just
in
case
it
comes
up,
and
I
have
some
requirement
that
actually
touches
upon
it,
but
I
didn't
I
didn't
find
anything
to
be
honest.
So
if
you
look
at
the
requirements
part,
it's
really
the
interesting
part.
A
I
tried
to
write
down
what
I
think
are
the
requirements
for
an
api,
so
how
users
will
actually
use
extensions,
what
they
see
the
part
that
they
have
to
touch.
So,
as
I
said
last
time,
my
opinion
is
that
it
should
be
its
own
crd.
That
should
be
a
first
class
api
instead
of
envoy
filter
because
onward
filter
is
this
low
level
thing
where
you
can
actually
dig
into
envoy
config
and
change
low-level
things
that
are
on
that
are
generated
by
scd?
A
So
you
have
to
be
an
expert
to
actually
get
that
right
and
you
can
break
things
and
I
think
extensions
should
be
available
to
normal
users
as
well,
and
people
who
just
get
started
with
istio,
because
it
can
actually
show
them
the
possibilities
that
this
gives
them.
You
know,
there's
there's
so
many
things
you
can
actually
do
by
by
using
wasn't
filters
that
I
don't
think
we
should
hide
it
in
this
power
user
api.
A
A
A
Building
these
images
is
a
solved
problem.
People
know
how
to
handle
them.
People
understand
the
references,
so
the
urls
that
reference
container
images
is
it's
a
well-known
thing,
because
everything
is
based
on
containers,
especially
if
you
get
to
the
point
where
you're
using
istio,
then
you
know
how
to
use
a
container,
so
I
think
it's
it's
a
natural
fit
really.
A
A
I
I
have
a
paragraph
here
about
solos
packaging
format,
which
I
think
needs
some
work,
and
I
already
spoke
to
solo
about
it
and
they
didn't
really
understand
what
my
problem
with
it
was.
So
I'm
still
I'm
I'm
keeping
that
conversation
going.
A
I'm
hoping
I
can
get
them
to
understand
that
a
more
standards,
compliant
format
would
help
everyone,
and
I
think,
ideally
we
should.
We
should
agree
with
them
on
a
format
so
that
we
we
don't
have
multiple
standards
in
in
the
ecosystem
and
yeah.
That's
about
it
actually
just
have
a
look
check
out
the
requirements
and
if
you
have
any
ideas
or
comments,
please
leave
them
there
and
then
we
can
go
from
there.
C
So
I'm
wondering
why
container
format
would
affect
the
extension
api
like
what
the
api
looks
like?
Is
it
like
yeah?
I
don't.
I
don't
see
the
connection
there
to
me.
C
A
Right,
it's
a
url
right,
I
mean
you
could
say
the
format
has
to
be,
but
so
the
format
basically
defines
how
publishing
of
an
extension
works
right.
So
if
you
say
the
format
is
a
tarball
and
in
the
api
you
have
a
url
and
it
has
to
point
to
an
http
server.
A
Then
somehow
you
have
a
connection
there
between
the
api
and
and
the
format.
Although
it's
just
a
url
right,
we
could.
We
could
say
it's
we're
not
touching
upon
that.
It's
just
url,
but
there's
other
things.
So,
for
example,
I
wanted
something.
That's
the
the
fourth
requirement
for
a
container
format.
The
ability
to
set
vendor-defined
defaults
phase
is
a
bad
word,
that's
actually
from
another,
so
this
is
filter
class.
So
what
I?
What
I
would
like
to
see
is
that
wait?
Where
is
the
dock?
A
Developers
of
an
extension
can
actually
pre-fill
some
of
those
fields,
so
you
don't
have
to
as
a
user.
So,
for
example,
filter
class
is,
is
basically
predefined
for
for
an
extension
right.
So
if
you're
writing
something
that
authorizes
against
an
authorization
server,
you
know
the
filter
class
is
going
to
be
authorization.
A
At
the
same
time,
I
would
want
a
user
to
be
able
to
overwrite
that,
though,
if
they
have
the
requirement
in
their
environment
to
execute
it
before
that
or
after
that.
Okay,
so
that's
the
kind
of
the
points
where
I
see
interconnections.
It's
not
super
interdependent.
I
agree,
but
there's
small
points
where
I
think
they
have
to
match.
C
Okay,
okay,
so
basically
it
is
possible
to
control
from
my
wrap
some
default
settings
you
need
which
okay,
I
see
and
there's
your
question
I
have,
or
maybe
suggestion
I
have
is
that
recently,
like
we
add
a
support
for
external
rc
customization
with
our
early
start
division
policy,
so
that
might
be
helpful
to
for
this
design.
I
mean
like
if
that
api
goes
through
a
lot
of
so
and
eventually
it
settles
down
on.
C
I
think
reference
the
mesh
config
for
the
extension
for
external
rc
configuration
filter
configuration.
So
I
think,
like
like
this
new
simplified
extension.
Api
might
follow
the
same
path.
So
I
would
suggest
that
to
look
at
that
design,
I'm
not
sure
whether
you
aware
you
are
aware
of
that
or
not.
A
A
C
It's
over
to
you,
then
yeah.
So
it's
it's
going
to
be
over
quick.
So
this
this
is
our
rfc
pc
to
remove
angular
filter.
Remove
I
refer
to
the
basic
based
telemetry
configuration
and
move
them
to
is
to
dc.
Basically
programmatically
implement
all
those
temperature
filters.
We
have
the
the
goal
being
that,
since
we
are
going
to
introduce
telemetry
api
and
that
that
means
that
we
cannot
use
our
filter
to
configure,
say
permission,
stature,
filter,
standard
filter
anymore.
C
We
need
to
implement
them
in
in
2d,
so
this
rfc
is
like
a
step
like
it's
a
prep
work
for
the
temperature
api
implementation,
and
also
we
have
several
other
filters
which
does
not
need
to
be
included.
It
doesn't
doesn't,
doesn't
have
to
be
configured.
We
are
in
our
filter
as
well.
We
can
move
them
to
study
in
like,
in
the
same
time
like
hp,
meta
exchange
and
tcp
meta
exchange.
C
They
don't
need
any
configuration
so
far,
so
it's
pretty
safe
to
move
them
to
s2d,
and
the
benefit
of
this
is
that
we
don't
like
obviously,
no
ammo
filter
installation
is
a
it's
a
great
plus
point,
because
everything
stories
are
friction
and
and
it's
like
not
it's
a
low-level
api
that
is
not
user-friendly
and
also
like.
I
said
it:
it's
going
to
help
the
temperature
api
implementation
yeah,
so
I'm
I'm
just
trying
to
gather
some
feedback.
C
Whether
is
there
any
any
objection
to
the
for
this
move
or
if
there's
any
concerns.
Basically,
the
only
concern
I
see
is
that
we
need
to
maintain
backward
accountability
for
the
installation
options
if
you
screw
down.
This
is
currently
telemetry
installation
options.
C
The
only
part
that
is
not
modeled
by
the
10
gpa
are
the
bus
will
be
wasn't
enablement,
and
that
could
I
mean
the
proposed
in
this
stock
is
to
make
them
any
variable
since
since,
like
the
currently
russell
environment
is
mostly
for
demo
purpose
or
experiment
or
like
develop
development
purpose.
C
I
don't
see
much
report
that,
or
I
don't
see
the
user
starting
to
use
awesome,
telemetry,
v2
filters
in
production,
so
I
think
in
environmental
is
appropriate
for
this
or
for
these
options
so
yeah
that
is
only
catch.
I
see
in
the
installation
options
other
than
that.
I
think
we
can
achieve
parity
with
potential
api
yeah.
C
C
Yeah
that
is
basically
about
this
proposal.
I
mean
please
take
a
look
and
let
me
know
if
you
have
any
concerns,
if
you
don't,
if
that
looks
good,
I
think
we
can
start
implementation
and
so
that
with
telemetry
api
landed
lens.
Oh,
we
can
integrate
with
that.
D
C
F
Was
wondering
this
this,
I
think,
makes
total
sense
if
we're
doing
the
telemetry
api,
which
I
hope
I
hope
happens
and
I
hope
happens
in
one
night,
but
what?
What
does
it
make
sense
to
ship
this
without
the
telemetry
api
like?
If
we,
if
we're
not
going
to
ship
the
telemetry
api
in
1.9,
would
we
still
ship
this
in
1.9,
because.
C
We'd
have
migrations.
If
we
don't,
if
we
don't
ship
telemetry
api,
then
we
don't
need
to.
We
need
to
like
keep
promises
instead
of
configured
with
angular
filter,
but
for
other
parts
like
metadata
exchange
or
tcp
meta
exchange.
There's
no
reason
to
use
our
feeder
anymore.
I
mean
this
is
like
a
intermediate
state:
okay,
yeah.
F
F
C
A
switch
from
ammo,
fitter
and
telemetry
there
will
be
no
inter
interstate
yeah,
cool.
D
Yeah,
I
think
this
looks
reasonable
to
me.
I
want
to
read
through
some
of
the
comments
just
to
catch
up
on
all
the
discussion,
but
I
think
this
looks
reasonable.
Okay,
cool
thanks.
A
Peter,
could
you
move
the
document
into
the
istio
drive
because
I
don't
have
access.
I
think.
C
D
C
So
since
we
want
to
have
some
talk
about
webassembly
extension,
you
wanna
see
it.
Currently,
I
have
two
docs
to
add
in
my
mind.
First
is
a
blog
about
the
awesome
extension
ecosystem
ripple
and
in
that
doc
we
will
basically
advocate
for
that
report
and
also
a
guide
about
how
to
develop,
testing
and
maintain
a
supply
plus
awesome
extension
and
busy
to
show
the
best
in
practice
that
we
learned
from
development.
C
We
learned
from
telemetry
to
filter
and
yeah.
That
is
the
first
stock
and
the
second
is
the
basic
off
muscle
filter.
So
I
have
chat
with
security
team
and
they
it
seems
to
be
fine
too
listen.
This
seems
to
feel
I
feel,
okay
to
put
this
in
authentication
section
in
the
security
category
and
the,
but
they
have
some
concern
about
the
the
user
perception
about
this
filter.
So
they'd
like
to
a
data
like
to
be
very
clear
in
that
doc
that
this
basic
oscillator
is
going.
C
It's
not
recommended
for
production
usage
because
for
sure,
basically
off
is
not
secure
enough.
Also,
no,
it
is
not
like
the
first
class
authorization,
api
authentication
api
and
we
don't
intend
to
promote
it
to
the
first
class
authentication
api,
so
so
yeah,
so
that
is
that
is
going
to.
We
need
to
be
very
clear
about
that
in
the
in
the
dock
and
also
they
they
like
to
make
sure
that
the
block
also
put
advertise
on
like
this.
This
is
mostly
about
websomebody
and
the
customization
with
the
web
assembly.
C
So
if
you
have
any
customize
like
the
current
basic
of
the
filter,
is
very
simple:
you
just
do
a
very
simple
match
based
on
path.
It
doesn't
even
provide
red
x
based
match
and
then
check
user
password
from
a
username
password
from
the
authorization
header,
and
so
so
it
it.
It
is
highly
possible
that
if
user
wants
a
real
usage
of
with
this
filter,
they
need
to
to
have
a
customization.
C
So
so
we
really
need
to
provide
steps
or
pointers
in
that
doc
about
how
to
customize
the
basic
pass
filter
yeah.
So
so
that
is
to
prerequisite
to
put
the
stock
into
the
security
category
I
used
to
at
io.
I
think
I
think
that
makes
sense
to
me
yeah,
so
I'm
going
to
create
a
draft
doc
and
share
with
secure
team
and
check
and
make
sure
that
they
they
are
okay
with
it.
C
So
that
is
that
is
basically
that
are
busy
the
two
dogs
that
country
in
my
to-do
list.
So
I
don't
know
whether
you
you
guys
have
any
saw
some,
what
we
can
add
more
or
any
sauce
on
the
the
the
six.
These
two
ducks
that
we
plan.
C
C
D
I
think
it
sounds
good,
I'm
interested
in
what
others
think.
I
think
this
will
be
very
useful.
C
I
think
I
think
the
blog
post
will
help
people
to
I'll
just
say
like
learn
the
best
practice
I
mean.
Currently
I
see
the
issue
these
studios.
C
Sorry
on
github
is
like
most
of
them
are
because
people
are
not
cannot
find
the
canonical
way
to
develop
a
simple
extension,
so
lots
of
issue
from
there.
So
I
think
this
blog
post
will
help
like
to
ease
user
pen
on
that
a
lot
so.
C
C
Okay,
cool,
then,
if
there's
no
feedback,
I
would
just
proceed
with
this
plan,
and
this
doesn't
have
to
happen
by
when
it
deadline,
I
mean,
while
somebody
nature
could
happen
asynchronously.
I
can
like
this
can
finish
after
one
eight.
First
of
the
first
release
of
one
day,
I
think
that
is
fine
unless,
as
people
think
this
need
to
happen
with
one
day.
D
D
G
I
I
was
just
this
is
just
a
curiosity,
so
your
rfc
for
all
those
ideas
that
are
getting
bounced
around
in
the
standard
metrics
rfc
that
you
put
out.
I
didn't
know
if
any
of
that
was
going
to
be
considered
for
1.9,
or
is
that
looking
further
out.
D
I
think,
when
I
had
envisioned
that
that
was
looking
further
out,
okay,
I
think
it
priority
wise.
I
want
to
get
a
final
agreement
on
a
telemetry
api
and
get
that
in
before
sort
of
taking
that
next
step.
Unless
there's
something
you
know,
we
could
take
it,
each
of
those
things
was
meant
to
be
sort
of
separable.
So
if
there's
some
aspect
that
you
think
we
should
prioritize
and
push
on
for
one
nine,
I
guess
now
would
be
the
time
to
to
sort
of
have
that
conversation.
D
G
Know
well,
I
was
just
going
to
sit
and
I
was
really
just
asking
from
the
perspective
of
like
my
own
personal
road
mapping,
for
you
know
when
that
stuff
might
may
materialize,
I'm
not
saying
it
has
to
be
in
one
nine.
Maybe
if
anything
starts
to
raise
in
priority
from
my
perspective,
then
I
I
would
call
that
out,
but
not
really
right.
G
Now
I
I
mean
I
certainly
got
excited
at
the
idea
of
some
of
the
stuff
in
there
you
know
for
reducing
potentially
duplication
for
reducing
just
the
number
of
fields
and
and
the
you
know
the
size
requirements
for
prometheus
and
so
forth.
I
mean,
I
think,
there's
a
lot
of
good
stuff
in
there.
I
just
wanted
to
make
sure
that
you
know
people
didn't
push
it
out
potentially
too
far,
because
I
think
there's
a
lot
of
good
stuff.
There.
D
Yeah
thanks
yeah
there's
some
stuff
in
there
I'm
energized
about
as
well.
I
I
just
worry
that,
with
the
nature
of
the
this
time
of
year,
we
try
and
push
through
this
cycle
that
might
be
shortened
just
by
the
number
of
absences
of
people
who
actually
be
doing
the
work
and
and
and
all
the
other
big
changes
going
on
that
to
me
it
was
sort
of
like
yeah,
let's,
let's
look
at
this
in
the
abstract
and
sort
of
make
it
more
tangible,
maybe
in
the
110
time
frame.