►
From YouTube: 2022-09-06 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
Hello,
yo
yo,
I
I
might
be
having
like
a
stroke
or
something
I
just
smelled
digiorno,
which
is
not
it's
not
cooking
in
my
house.
There's
no
there's
there's
coffee
brewing,
but
that's
not
what
that
smells
like.
A
Know
great
question:
hey
burning
feathers,
burning,
feathers,.
A
Reuben,
if
you're
watching
this
in
you
know
20
45,
you
know.
C
Yeah,
let's
go
and
get
started
I'll
go
ahead
and
share.
My
screen
do
a
specific
recap
and
then
jump
into
hotel,
ruby
issues.
C
All
right,
the
agenda
was
pretty
short
today.
The
first
item
up
was
integrating
the
message:
context:
propagation
otep
into
the
spec.
I
think
we
talked
about
this
a
little
bit.
I
think
they
were
differentiating
between
multiple
types
of
context.
B
C
I
think
I
forget
all
the
details.
I
know
riley
commented
here
that
there
there
are
some
fundamental
issues
and
the
tc
might
not
have
capacity
for
everything,
but
this
is
the
pr
that.
D
C
Was
the
one
that
had
the
most
discussion
and
it's
potentially
the
most
relevant
to
this
sig
and
this
is
there-
are
now
scope
attributes,
so
for
a
when
you
get
a
tracer
provider,
you
provide
the
library's
name
and
version,
but
there's
also
potentially
other
scope.
Attributes
now
and
bogdan
is
pointing
out
that
this
is
kind
of
a
breaking
change.
C
If
you
want
to
use
those
attributes
as
kind
of
identity
of
the
thing
generating
the
telemetry,
it
used
to
be
that
all
the
telemetry
for
a
given
instrumentation
library
and
version
would
group
together,
but
now
you
can
kind
of
get
a
tracer
provider
for
that
same
instrumentation
library
inversion,
but
have
varying
scope
attributes.
So
if
you
were
to
try
to
kind
of
normalize
all
that
to
say
where
this
telemetry
is
coming
from,
it
will
actually
be
coming
from
multiple
places.
C
I
think
that's
the
crux
of
this
issue,
so
I
think
people
are
debating
to
what
degree.
This
is
a
breaking
change
and
I
think
people
are
also
debating
whether
all
these
attributes
should
be
like
something
that
you
use
for
identity.
C
D
I
couldn't
think
of
a
use
case
that
isn't
multiple
scopes
for
the
same
instrumentation
library,
but
it's
to
have
what
is
essentially
resource
attributes,
but
at
the
instrumentation
library
level,
so
like
all
spans
produced
by
this
instrumentation
library,
should
have
this
attribute
so
that
you
don't
have
to
repeat
that
thing
on
every
span
that
that
that
instrumentation
library
produces
that's
not
the
only
use
case,
but
that's
another
one
that
isn't
about
having
multiple
scopes.
D
But
I
don't
know
how
to
resolve
identity
from
from
at
least
my
back
ends
perspective.
We
don't
care,
it's
the
name,
inversion
and
anything
else
that
comes
along.
C
Yeah,
I'm
I'm
not
sure
how
much
my
back
end
cares
either
about
any
of
this,
like
I'm,
not
sure
how
much
any
back
end
has.
All
of
the
I
don't
know
has
a
complete
like
hotel
modeling,
where
there's
separate
resource
attributes
and
those
are
somehow
different
than
your
spam
attributes,
which
are
somehow
different
than
your
scope.
Attributes
I
think
there
are.
C
I'm
guessing
some
back
ends
might
might
have
some
issues,
so
at
the
time
this
was
brought
up.
The
most
current
comment
was
tigran's,
so
there
has
been
some
some
further
discussion
here,
but
yeah.
I
think
he
was
asking
about
it
a
couple
of
questions
saying
within
the
same
process,
if
you
use
an
hotel
api
to
simultaneously
obtain
a
tracer
provider
using
the
same
name
version
but
different
attributes,
what
should
happen?
Do
you
get
the
same
different?
C
Is
it
an
error
or
is
it
undefined?
I
guess
over
time
if
your
process
changes
and
your
tracer
provider
uses
different
attributes
than
what
you
used
to
obtain
previously,
what
should
happen.
B
D
C
So
this
is,
this
is
a
very
active
issue,
like
things
are
coming
in
as
as
we're
talking
about
it,
but
I
think
definitely
something
to
follow,
because
I
think
it
will
have
some
implications,
at
least
for
you
know
the
hotel,
ruby
implementation,
just
like
the
feasibility
of
trying
to
keep
track
of
all
of
these
tracer
providers
and
media
providers
and
what
actually,
what
you
know
should
we
turn
a
new
instance
a
distinct
instance
and
what
you
should
not
start
of
curiosity
have
we
have
we
implemented
scope
attributes
at
all
yet,
or
is
this
something
still
kind
of
future.
B
No,
I
was
just
going
to
comment
that
yeah
we
haven't
implemented
this
yet
and
we've
yeah
we've
just
kind
of
punted
this
for
a
bit.
C
So
yeah,
I
think
the
key
takeaway
is
to
maybe
watch
and
participate
in
this
issue.
C
C
C
C
B
There's
a
link
in
the
issue
like
the
pr
description,
there's
a
link
to
the
other
one.
D
In
the
use
case
of
a
specific
environment
variable
that
was
boolean,
they
spun
out
the
platonic
ideal
of
the
bike
shed
for
this
boolean
value,
so
that
we
could
decide
the
color
and
then
come
back
to
the
specific.
C
D
D
C
Was
it
for
the
spec
sig
recap,
questions
comments,
concerns.
B
Was
there
any
discussion
about
booleans
actually
being
ternary
in
the
sense
of
true
false
or
null.
C
Not
during
this
session,
I'm
not
sure
I
am
not
up
to
date
on
this
full
issue
to
understand
if
that
has
been
there,
but
I
think
I
think
they
are
trying
to
make
them
binary
and
I
think
only
one
of
them
makes
sense.
It's
like
explicitly
true
or
one
variation
of
true.
I.
D
C
D
But
then
the
phrasing
of
your
setting
is
if
the
default
is
enabled
your
setting
is
now
you
have
to
name
your
environment
variable
disabled
because
the
default
is
false,
and
so
now
you
have
a
double
negative.
Those
break
my
brain,
the
disabled,
true
like
but
okay
is
it
on
disabled
is
false.
So
it's
enabled
okay,
but.
C
A
I
believe
that
releases
from
contrib
are
not
working
right
now,
because
the
sidekick
sidekick
build
is
failing
on
main,
I'm
actually
going
to
try
to
take
a
look
at
that
right
after
this
meeting.
So
maybe
that's
do
we
need
to
release
instrumentation
all
is
maybe
putting
the
cart
before
the
horse.
Do
we
have
the
ability
to
release
currently.
C
Okay
got
it
so
we
we
need
to
fix,
fix
the
bill
first
and
what
was
failing,
sidekick.
B
B
Yeah,
I
think
it's
like
the
sidekick
spec
face
with
ruby
3.1.
It's
actually
something
I
think
eric
likes
before.
C
B
B
B
Yeah,
I
still
have
an
issue
with
that
risky
yeah.
Sorry,
I
don't
know,
but
I
still
have
an
issue
with
the
moss
being
unreliable
on
bb
3.1
and
I
wasn't
sure
if
there
was
anyone
who
was
maybe
familiar
with
interesting
ones,
issues.
That's
it's
kind
of
intimate.