►
From YouTube: 2022-09-06 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
B
C
D
C
So
some
time
ago
there
was
a
old
tap
from
johannes
merged.
That
makes
it
clear
that
context.
Propagation
in
messaging
systems
is
to
be
done
via
the
message
itself.
So
the
context
navigates
through
the
to
the
systems
in
the
message.
So
the
attack
was
merged
and
yeah.
He
he
was
on
vacation
and
he
tasked
the
the
messaging
instrumentation
group
to
move
forward
with
his
hotep
and
and
bring
the
context
to
the
spec.
C
So
that's
what
I
did
and
I
opened
an
issue
because
I
opened
a
new
pr
yeah
so,
but
it
was
asked
for
more
discussion
or
or
or
more
context
on
it,
and
I
I
provided
more
information
as
a
comment
on
the
on
the
on
the
issue
plus
there's
the
original
tab,
all
the
discussion
there.
So
I'm
not
sure
if
there
is
anything
that
I
can
provide
more.
C
B
Yeah
yeah,
I
see
yeah
you're
for
the
first
one.
You
actually
provided
like
yeah
a
good
explanation
there,
but
it
was
yesterday,
so
I
suspect
that
some
people
were
out
because
it
was
a
holiday
in
the
states.
So
let's
hope
that
people
can
take
a
look
at
that
now
today,
yeah
the
second
one
similar
situation,
you
see
some
feedback.
Well,
let's
check
it
out.
I
see
that
little
miller
already
approved
that.
So
that's
that's
great.
B
Also.
Amir
bloom
also
here
prove
that
so
yeah
anything
else
riley.
Maybe
you
have
something
yourself
or
just
let's
get
rid
of
that.
B
Perfect,
okay,
thank
you!
So
much
okay.
So
let's
take
a
look
at
that
perfect
all
right!
That's
that's
it
from
that!
One!
Let's
go
to
the
next
one
and
the
next
one
is
an
important
one
that
I
mean
just
pointed
me
out
to
it's
about.
Let
me
share
my
screen.
I
just
have
to
something
here,
but
basically
it's
about
scope
and
achievements
and
whether
they,
whether
attributes
for
a
scope,
library,
do
represent
or
not,
identity,
and
whether
this
is
something
that
is
considered
breaking
change
and
this
in
turn.
B
Because
of
this
this
this
could
be
breaking.
You
know
it
could
be
holding
the
actual
release.
You
know,
so
there
we
are.
So
this
is
the
one
I
don't
see
tigran,
I
don't
see
buckthorn
in
the
call,
but
maybe
we
can
take
a
look.
The
rest
of
us,
as
I
said
before,
this
is
very
important.
B
B
And
then
there
was
some
discussion.
Let
me
just
move
out
here
and
basically
tiran,
giving
his
point
of
view
about
the
consequences
there.
D
D
B
No,
I
think
I
agree
with
that.
I
don't
know
whether
this
is
enough
for
what
tyranny
had
expected,
and
probably
we
can
ask
this.
I
think
what
you
just
said
about
just
having
one.
You
know
a
scope.
Yeah,
this
wouldn't
be
breaking
anything.
You
know
yeah.
I
wonder
what's
the
expectation
here,
if
anybody
here
thinks
that
a
scope
scope
may
actually
be,
you
know,
dude
like
you
are
trying
to
use
different,
multiple
scopes
with
different
attributes,
and
these
scopes
have
the
same
name.
B
D
I
mean,
I
think,
in
the
long
run,
people
do
want
that,
but
I
I
feel
that
where,
where
we
are
today,
these
attributes
will
more
likely
be
used
to
add
single
attributes
here
to
existing
payloads
and
the
the
hold
up
was
all
about
metrics.
We
do
know
that
if
you
have
a
metrics,
payload
and
and
there's
no
threat
of
breaking
the
single
writer
violation
or
no
threat
of
breaking
the
single
writer
rule,
then
it's
safe
to
remove
all
the
attributes.
D
So
in
the
sense
where
there's
only
one
of
those
payloads,
you
can
just
remove
those
scope,
attributes
safely
it's
when
you
have
more
than
one
that
you
could
break
something
downstream.
Yeah.
B
D
Yep,
it's
unfortunate.
We
don't
have
tigran
here
or
bogdan,
especially
who's,
the
one
who
put
up
this
obstacle
and
I
do
feel
we
can
work
this
out,
but
it's
a
legitimate
obstacle.
I
will
say.
B
B
D
D
A
Is
that
really
an
issue
on
the
on
the
protocol
at
all,
because
I
don't
think
it
it
reflects
or
or
is
opinionated
about
whether
this
is
part
of
the
identity
or
not?
I
think
it's
only
about
the
the
api,
whether
you
you
can
have
the
same
instrumentation,
scope,
name
and
different
attributes
or
not
right,
because
I
think
on
the
on
the
buyer.
It
will
always
look
the
same,
so
there
shouldn't
be
any
breaking
changes
based
on
that
decision
here,.
D
My
understanding
of
the
failure
mode
we're
talking
about
here
is
that
a
new
producer
produces
these
attributes
and
then
it
gets
passed
through
some
sort
of
processor
that
has
no
idea.
Those
attributes
were
there,
so
it
loses
those
attributes
at
that
point,
it's
possible
that
you
know
multiple
senders
could
have
sent
exactly
the
same
metric.
That
was
not
what
they
intended,
so
it's
breaking
by
by
ignoring
something
new
in
the
sense.
D
What
I
realized
from
this
issue
is
that
we
can
never
add
anything
identifying
the
protocol
right
now
or
it's
always
going
to
be
breaking
by
this
definition,
and
I
think
we
need
to
find
a
path
forward.
I
feel,
like
we've,
consistently
rejected
the
idea
of
putting
the
version
identifier
into
the
actual
protocol
payload,
but
maybe
that's
another
way.
We
could
do
this
here.
I'm
not
sure
why
we
always
seem
to
object
to
that
approach.
A
B
E
B
E
Okay,
well,
yeah.
I
would
appreciate
having
clarity
on
that
for
sure,
because
I
know
go
has
added
this
already
so
it'd
be
nice
to
not
release
something
that
we
have
to
take
out
eventually,.
B
Right
right,
that's
what
like
once
we
know
this
and
we
know
it
will
not
impact
the
api
or
the
sdks.
Even
you
know,
once
we
know
we
can
work
on
fixed
after
the
release.
You
know
otherwise
yeah
yeah,
yeah,
totally
so
yeah.
I
would
be
happier
to
either
you
know
to
instead
just
delay
the
release
a
little
bit
more.
B
Okay,
that's
that
one
yeah
we
don't
have
button
here.
Let's
hope
we
can
follow
up.
Thank
you,
mcd
for
the
for
the
alternative
proposal,
and
I
just
I
haven't
seen
that
tigger
had
answer
to
your
comment
already
so
time
to
read
that
one:
okay,
let's
let's
move
then
to
the
next
one
bruno
definition
of
boolean
environment
variables
on
the
sdk.
Please.
F
Hello,
so
we
have
a
couple
of
approvals
there
and
some
minor
comments,
but
the
the
last
comment
that
we
have
here-
voices
concerns
about
warning.
Casey
insanity
case
insensitive
false
because
there
was
made
a
suggestion.
If
we
were,
if
we
had
any
configuration
that
was
not
explicitly
false,
we
would
warn
that
an
invalid
configuration
was
being
added
and
that
and
that's
it.
So
I
suggested
that
we
could
remove
crazy
sensitive
falls
from
the
warnings.
F
Also
there
was
concern
about
all
boolean
environment
variables
should
be
named
and
defined
such
that
false
is
the
expected
safe
default,
behavior
yeah,
so
mark.
F
Mark,
I
don't
think
he's
here:
yeah
express
that
in
case
give
gave
an
example
that
that
would
not
fit
this,
but
I
I
argue
that
should
is
not
mandatory
and
would
still
encompass
you.
It
would
still
include
the
case
that
he's
mentioning,
so
I
don't
think
this
is
a
showstopper
as
well.
Let's
see
how
it
how
he
reacts
to
this.
B
So,
for
me,
it's
fine
the
way
it
is
now.
Probably
it's
a
little
bit
friendlier
to
when
you
get
false
to
not
generate
a
warning,
but
it's
a
minor
thing
relatively:
it
doesn't
change
the
semantics.
B
F
I'm
fine,
I
think
it's
a
minor
detail,
I'm
not
crazy
about
the
warning
in
first
place,
but
I
I
I'm
not
I'm
not
gonna
argue
against
that.
I
think
it's.
It
makes
sense.
Yeah.
B
F
B
Going
once
going
twice:
okay,
thank
you!
So
much
everybody
yeah!
Let's
go
back
to
this
cause,
especially
you
know
the
point
regarding
scopes
and
attributes.
I
think
that's
the
one
that
could
be
blocking
us,
but
please
take
a
look
at
that
one
and
yeah
and,
of
course,
jaws
and
bruno's
peers
stay
safe,
ciao.