►
From YouTube: Policies and Telemetry WG Meeting: 2020-05-06
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).
B
Sure
so
I
think
probably
everyone
on
here
may
know
some
of
the
context
behind
this,
but
we're
changing
how
we
do
the
demo
installed
east
geo.
So
we
don't
want
to
offer.
You
know,
unbounded
customization
on
all
these
add-ons
that
we
don't
control.
We
want
to
just
provide
a
plain
VM
will
very
simple
and
though
customization
offered
to
get
started
and
make
it
very
clear,
it's
for
demos
and
then
improve
the
experience
for
integrating
with
you
know,
production-grade
diego.
B
B
Sometimes
it
doesn't
really
make
sense
for
a
demo
to
have
two
options
that
do
the
same
thing
or
are
these
effectively
the
same
thing,
but
it's
also
hard
to,
like
you
know,
pick
a
favorite.
You
know
you
know
issues
involved
with
that.
So
basically,
what
this
is
is
we
need
to
figure
out
what
we're
gonna
do
around
that
there's
a
few
options.
B
The
obvious
ones
are
pick
either
just
if
accan
or
just
jäger,
but
we
could
also
do
things
like
provide
both
or
provide
like
one
as
the
default
and
then
one
and
like
some
extras,
slash
the
zip
kannur
extra
/igor
folder
that
they
could
optionally
use
or
you
know
something
else,
I
haven't
thought
of
so
I,
don't
really
I,
don't
particularly
have
an
opinion
on
neither
of
these
I
don't
really
care
too
much.
I,
don't
know
enough
about
racing
to
make
it
inform
decisions.
So
that's
that's
why
I'm
here
mostly.
C
A
B
B
B
B
Yeah,
so
I
don't
actually
think
that
overhead
of
maintaining
is
it
in
deployment
is
an
issue
it's
more
about
the
user
choice
right,
because
you
know
these
are.
This
is
my
four
demos
only
and
if
the
first
thing
we
present
is,
do
you
want
to
use
the
printer
Jaeger?
You
know
adds
some
extra
overhead
and,
in
terms
of
you
know,
the
we
had
a
stir
was
complicated
enough
right
right.
D
So
I
think
I
think
that
that
we
are
gonna
have
to
choose.
One
thing
has
to
favorite,
because
we
don't
want
people
to
start
with
this
question
when
they
haven't
even
thought
about
facing.
Yet,
however,
if
someone
has
thought
about
racing
and
they
and
they
want
to
test
it
out
with
like
zip
came
or
a
girl
or
whatever
the
opposite
of
the
default
is
then
it
should
just
be
in
the
talk,
so
I
mean
I
I,
like
the
way
I'm
thinking
about
this.
Is
that
if
the?
D
A
The
point
that
was
being
raised
on
the
comment
thread
there
on
the
side
was
significant
and
eager
aren't
the
entire
universe
of
tracing
backends,
and
there
is
a
ton
of
places
that
supports
a
Zipkin
API.
So
I
do
want
to
like
us
to
think
about.
If
we
start
saying,
oh
well,
we'll
put
other
options
in
another
folder.
D
Be
so
yeah
I
think
that's
a
bit,
that's
a
great
question,
but
so
yeah.
If
there
is
another
open
source
option
which
is
kind
of
not
vendor
specific,
really,
then
we
should.
We
do
need
to
think
about
it,
and
part
of
it
has
to
be
in
response
to
user
and
kind
of
community
demand
right.
Just
because
an
option
exists
doesn't
mean
we
need
to
add
it
like
these
two
are
the
prominent
ones.
D
B
Create
may
disagree
yeah
well,
there's
also
a
we
don't
necessary
need
to
put
everything
in
the
in
the
demo
install
right.
We
could
have
a
doc
about
integrating
with
sky
walking
without
having.
Yet,
despite
of
the
demo,
that's
I,
don't
I,
don't
think
we'd
be
up
against
more
documentation
of
how
to
integrate
with
things.
A
B
A
E
E
E
But
yeah
I
mean
that's
why
this
problem
is
its
preference.
You
know
when
you,
when
you
provide
a
demo
version,
users
will
then
probably
go
on
and-
and
you
know,
if
they've
got
no
strong
opinion
on
the
tracing
system,
they're
more
likely
to
then
just
go
with
that
same
system
when
they
deploy
into
production.
So
I
mean,
from
my
point
of
view,
I
think
there
is
more
production,
so
deploying
than
scene
meshes
of
Jaeger.
E
A
F
G
E
G
Don't
I'm
sort
of
unbiased
a
little
bit
even
though
key
ollie
integrates
with
Jaeger
I
mean
maybe
try
converting
a
test
and
see
how
easy
it
is,
and
so
only
jaegers
Jaeger
is
there,
and
but
if
you
can
just
use
the
same
API
that
you're
already
using
for
your
test,
that's
easy,
then
maybe
it's
not
too
big
a
deal.
I
mean
the
fact.
Now,
from
the
bias
side,
I
mean
since
well,
since
kiala
is
pretty
much
goes
side
by
side
with
this,
do
it
would
make
sense,
I
think
to
have
Jaeger.
D
A
D
D
So
I
think
if
we
hear
from
sto
community
that
this
is
something
that
people
are
using
then
I
think
then
then
we'll
certainly
have
a
discussion.
But
as
of
now,
I
would
say
that
everything
else
can
just
stay
in
in
Docs
and
we
are
simply
trying.
We
simply
want
to
be
nice
to
go
to
the
Zipkin
community
and
the
fact
that
we
are
using
Zipkin
api
you're.
Maintaining.
D
F
B
Right
that
sounds
good
I
will
want
to
let
that
and
then,
if
you
guys,
didn't
take
a
look
at
the
actual
PR.
That's
implementing
this
once
IDEO
said
be
be
great.
I
did
have
one
quick
question
about
that.
Now
would
be
kind
of
a
good
time
to
change
anything
that
we
want
to
change
about.
The
add-ons
like
Jaeger
is
called
East,
EO
tracing
I.
Think
do
we
want
to
rename
that
to
Jaeger
to
be
more
clear
that
it's
not
like
owned
by
East
e?
Oh.
B
A
D
A
D
A
A
A
A
D
So
I
mean
it's
oh
yeah.
What
I
was
talking
about
was
is
related,
but
it
is
very
difficult
to
debug
like
mapping,
sort
of
sorts
of
issues
and
which
is
and
and
I
even
even
the
code
that
I
know
really
well.
There
aren't
enough.
There
aren't
enough
trace
points
and
things
like
that,
so
I
think
in
one
seven
we
need
to
do
work
on
debug
ability
and
like
adding
right
sorts
of
error
messages
and
anything,
and
things
like
that
there
are.
D
D
So,
like
I
think
generally
also,
there
is
the
way
we
are
tracing.
Sorry.
The
way
we
are
producing,
trace,
logs
and
debug
logs
is
is
not
not
very
consistent
right
like
sometimes
they.
Sometimes
these
logs
were
added
as
part
of
development,
so
they
are
extremely
noisy,
but
then,
once
initial
development
was
done
and
everything
was
working
except
they
were
excessively
too
moved.
D
And
now,
if
you,
if
you
run
into
a
problem-
and
you
turn
on
debugging-
you
don't
get
like
high
quality
logs
to
actually
to
much
of
anything
and
I
mean
I've
had
to
like
just
attach
a
debugger
and
kind
of
see
what
is
going
on,
which
is
yeah
that
so
that's
that's
not
where
we
want
to
be
so.
We
have
extensions
debugging
as
a
as
kind
of
a
bigger
theme
and
part
of
it
is
tracing
and
logging.
D
A
D
A
So
do
we
think
that
the
people
who
have
already
written
these
adapters
in
our
troubleshooting
and
maintaining
them
today
are
they
doing
know
if
any
of
them
are
already
starting
this
transition
and
if
they
feel
they
have
the
tooling
to
do
that
Mike?
My
concern
is,
you
know,
wasn't
that
long
ago,
that
who
said
hey
everyone,
stop
writing
mixer
adapters
right
these
G
RPC
adapters
and
now
we're
just
sort
of
leaving
that
out
to
you,
know
people
out
to
dry
a
little
bit
on
on
that
transition.
Yes,.
C
D
C
So
the
problem
is,
we
cannot
guarantee
that
vasami
seen
practitioner
at
least
eight
one
seven,
but
we
are
planning
to
deprecate
mixer
in
y7,
so
that
means
they
will
have
a
period
that
they
don't
really
have
a
way
to
do.
Extension
right
present
support
other
than
like
using
external,
see
and
access
long,
accused
native
and
what
extension
mechanism.
So
that's
why
I
thank
me
make
sense
to
provide
this
Sheen
as
a
for
like
shorter
period
and
also
as
people
to
start
to
think
about
transition
to
a
zone.
D
So,
even
perhaps
even
though
sandbox
will
do
or
no
no
it,
it
has
to
be
true.
Awesome
support,
yeah,
okay,
I
think
I.
Think
that's
that's
a
that's
a
good
point
and
especially
if
we
can
attach
it
to
X
dot
zealot.
Like
you
said,
then
we
don't
need
core.
Like
tech
debt,
it's
it's
kind
of
separate
standalone
piece
of
code
and
we
can
remove
it
once
you're
done.
Yep.
C
D
A
A
So
then
I
guess
we
did
want
to
spend
a
few
more
minutes
as
I
know
some
people
weren't
here
to
say
last
time
on
generating
ideas
for
one
seven.
We
have
about
I,
don't
know
eight
days
before
we
need
to
present
this
to
the
TRC.
So
there
are
ideas,
things
that
have
come
up
that
we
think
should
be
on
this
map
that
aren't
there
now
it
would
be
the
time
so
I'll
open
the
floor
to
one
seven.
A
C
D
D
A
So
what
about
this,
an
ERISA
who
had
to
leave
this
was
hit,
something
you
really
was
adamant
about.
Are
there
other
gaps?
I
know
we
had
they're,
not
the
sort
of
out
of
mesh
contacts.
Are
there
other
things
related
to
lost
behavior
Amman?
We
talked
about
the
geography
adapter,
so
that
is
there
something
else,
we're
forgetting.
C
D
A
D
A
Well
with
the
work
it
can
be
done
over
the
summer,
yeah,
okay,
well,
my
says
other
contributions
here,
I
think
there's
not
much
more
to
say
at
the
moment.
I
will
try
and
get
this
in
the
shape
with
priorities
and
I
will
send
a
notification
on
the
slight
general
and
maybe
discuss
about
the
priorities
just
so
we
can
have
a
group
sort
of
offline
review
before
presenting
to
the
TSE
next
Friday
yep
make
sense
and
then
I
think
the
next
item
on
the
agenda
was
doing.
A
A
D
A
Ok,
well,
then
so
decide
I,
don't
think
we
all
need
to
stay
online
to
do
testing,
but
I
think
we
should
all
take
some
time
today
and
tomorrow
and
Friday
to
look
at
the
p.
0
is
4
1
6
and
look
at
the
new
Docs
that
we've
added
and
do
testing,
particularly
for
features
that
maybe
didn't
implement.
But
you
know
who
are
doing
it
so
I
think
having.