►
From YouTube: 2022-08-10 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
For
me,
it's
been
very
hot,
I'm
in
austin
it's
been
basically
100,
plus
for
like
two
months
straight
up.
No,
no,
no
rain.
A
I
was
here
last
summer,
like
you
know,
on
and
off,
and
it
it
had
real
weather
like
it
was
real
hot
and
then
suddenly
it
was
thunderstorms.
And
then
you
know
it
flipped
like
twice
a
day
or
so
it
was
quite
interesting.
But
now
it
was
just
a
hundred
hundred
four
under
three
like
day
after
day.
B
B
Sorry,
I
think
my
mic
was
off,
but
I
was
just
asking
how
you,
like
austin
other
than
that.
A
I,
like
I
love
it.
I
think
it's
great
I'm.
I
was
moving
from
the
bay
area
right,
so
I
was.
I
was
in
bay
area
for
20
years
and,
like
you
know,
it's
pretty
nice
in
california.
You
know
so
there's
nothing
particularly
there's
nothing
especially
wrong
with
it
there,
but
it
you
know.
With
all
this
covert
thing.
You
know
people
started
moving
away.
Suddenly
it
it
never
crossed
my
mind.
You
know
that
that
there
was
a
life
outside
of
the
bay
area.
A
You
know,
given
what
I
do
and
like
you
know
like,
and
I
just
that's
just
where
I
ended
up
coming
from
germany
and
blah
blah
blah
right,
but
the
kobe
thing
just
remixed
everything
and
you
know
it
just
suddenly
felt
like
might
as
well
live
somewhere
else
and
after
20
years.
You
know,
you
know
why
not
mix
it
up
a
little
bit
and
we
also
have
an
office
out
here
in
austin.
This
game
is
an
acquisition,
so
I
had
been
around
like
a
couple
of
times.
A
A
Good
david
is
good
again
taken
to
take
infrastructure.
Now,
yes,
bunch
of
techies,
it's
you
know
and
everything
around
it.
So
I
like
it
very
cool,
it's
a
little
bit
smaller.
So
you
know
everything
is
either
10
or
20
minutes
away.
So
you
could
pretty
much
bounce
anywhere
at
any
time,
so
austin
it's
comparable
to
terrible
traffic
terrible
traffic
all
the
time
but
like
coming.
B
C
Yeah
I
was
going
to
bring
up
the
the
pr
that
I've
got
there
about
clarifying
nested
attribute
support,
I'm
getting
lots
of
pushback
on
that.
C
C
So
I
was
trying
to
clarify
it
and
say:
hey.
We
have
this
single
thing
called
attributes,
which
is
whether
it's
logs
metrics
or
spans
it's
attributes,
rather
than
forcing
everyone
to
effectively
identify
log
attributes
to
something
separate,
but
yeah.
It's
getting
lots
of
pushback,
really
driven
from
no
spans,
can't
no
won't,
never
go
away.
Evil,
evil,
creature.
A
C
No,
it's
more
a
case
of
like
the
the
protocol
supports
it.
The
back
ends
could
could
could
receive
it
and
therefore
have
to
deal
with
it.
Someone
can
go
and
create
their
own
open,
telemetry
sdk
and
actually
send
their
attributes
in
spans,
because
the
protocol
supports
it
from
the
rum
perspective,
which
is
where
I'm
coming
from
with
this.
As
part
of
the
attribute
definition,
we
are
going
to
support
or
actually
define
it,
that
the
data
is
will
be
nested,
attributes
and
long
term.
C
C
But
I
don't
like
our
chances
of
getting
that
the
based
on
the
comments
that
I'm
getting
as
well
as
the
other
linked
issue
376,
which
is
on
that
pr,
which
has
been
open
for
a
very
long
time,
which
also
talks
about
general
domestic
attribute
support.
So
yeah,
really
just
looking
for
some.
A
C
Yeah,
I
I
don't
think
it's
a
it's,
not
a
simple
problem.
It's
been
around
for
a
long
time
to
me.
Looking
at
the
history
and
looking
at
the
comments,
it's
it
seems
to
be
driven
by
the
fact
that
jager
and
prometheus
don't
allow
storing
of
nested
attributes
and
that's
fine.
I
don't
know
I'm
not
talking
about
nested
attributes
being
being
on.
The
protocol
means
the
back
ends
have
to
store
it.
That
way.
You
know
they
can
store
it.
C
However,
they
like,
and
there
is
some
wording
already
in
the
atari
definition
to
say
if
it's
not
a
primitive
or
whatever,
that
should
be
passed
as
a
json
string,
but
I
you
know,
I
really
think
it's
at
the
export
level
or
the
persistence
level
should
be
where
it
gets
converted
into
whatever
the
back
end
needs.
C
Rather
than
limit
the
entire
sdk
api
definitions
to
say,
nested
attributes
or
even.
C
Yeah,
because
otherwise
it
means
every
sdk
has
to
go
and
define
at
the
sdk
level
that
log
attributes
are
different
from
standard
attributes
which
okay,
if
you've
got
hierarchies,
you
can
can
extend
it,
but
it
then
these
attributes
aren't
attributes
which
are
just
going
to
get
confusing.
A
A
C
A
So
there's
two
things
that
I
have
come
on.
Sorry,
I
have
a
very
unhappy
little
critter
here
that
yeah
wants
to
go
back
to
mom,
but
like
mom's,
not
here
so
so,
there's
two
things
that
I
am
trying
to
track
like
just
from
the
perspective
of
trying
to
be
a
good
spec,
approver,
there's
the
events.
Pr.
I
think
I
think
satoshi
you're
kind
of
involved
in
that
right.
If
I'm
mapping
githubs
to
names
properly
yeah-
and
it
looks
like
this-
is
you
know
you?
A
I
I
think
we've
been
through
the
process
and
I
was
kind
of
ready
to
approve
it,
and
then
I
didn't
because
I
forgot
it
and
now
it
is
more
commenced
here
so,
but
I
kind
of
feel
they
need
to
sort
of
do
something
and-
and
I
think
by
and
large
we
had
sort
of
agreed.
You
know,
while
widely
that
this
is
the
right.
You
know
that
that
this
that
this
pr
should
actually
you
know,
be
approved
right.
So
with
these
additional.
A
B
A
Up,
I
missed
the
window,
so
let
me
know
you
know,
feel
free
to
bring
me
directly
on
the
slack
or
something
okay
and
I'm
I'm
also
going
to.
Obviously
you
know
track
the
incoming
changes.
I
think
my
github
filters
are
set
up
well
enough.
Now.
D
C
Yeah-
and
we
talked
about
the
javascript
sig,
which
was
the
meeting
just
before
this
and
daniel
was
looking
to
get
more
more
eyes
on
martin's
implementation
of
that
too.
So
my
only
comment
that
I
had
on
there
is
rather
than
us,
defining
the
shape
of
the
event
to
include
data.
Do
we
want
to
include
it
in
that
pr,
as
well,
rather
than
having
to
spin
up
another
one,
but
I'm
open
it's
fine.
We
don't
have
to
define
it
now.
D
Well,
on
on
one
hand,
I
feel
it's
it's
better
for
that
data
attribute
to
be
present
from
day
one
yeah,
but
it
also
requires
the
nested
attribute
support.
So.
D
A
A
There's
a
bunch
of
open
comments
there
also
still,
but
I
think
again
the
folks
are
not
here,
so
I
guess
we'll
just
hold
off
on
that
one,
but
it
seems
overall
reasonable.
I
think
we
basically
said
we
want
lock
records
to
be
mutable,
because
you
know
otherwise
why
I
have
processors,
so
it
seems
like
there
might
be
a
technicality
or
two
left.
Just
looking
at
the
comments,
I
don't
know
if
anybody
here
has
any
insight
into
that.