►
From YouTube: 2022-09-15 meeting
Description
Instrumentation: Messaging
A
A
A
Okay,
so
yes
about
the
splitting
I
think
that's
a
good
idea,
because
what
they
can
do,
then
in
the
pr
for
the
attributes.
We
can
use
what
glutamine
I
started
doing
that
we
actually
put
the
attributes
already
in
the
yaml
file
and
then
generate
the
table
from
from
it,
and
then
the
old
type
actually
has
the
table
already
from
the
source
and
then
moving
that
to
the
spec
will
be
easier
than
yeah.
A
B
A
B
A
Bruno
had
some
comments
on
the
on
the
depth
on
the
original
one.
Now
about
that,
the
diagrams
they
don't
have
what
it
was
egg
on
the
producer.
I
think
he
mentioned
something
like
that
that
the
diagrams
they
don't
show
up,
there's
no
examples
for
AG
on
the
producer
side
and
he
left
some
suggestions
on
the
diagram.
How
that
would
look
like
so
it's
buried
in
one
of
the
comments
it's
not
solved
yet
so
it
should
be
easy
to
find
I.
Think
it's
one
of
the
last
ones.
B
A
B
Maybe
and
maybe
try
to
fix
right
away.
What
is
what
is
in
those
comments.
A
Yeah
I
think
that
one
he's
already
did
he
suggestion
and
I
think
I
if
I
remember,
I,
think
it
made
sense,
so
should
be
good
just
to
to
the
example
and
put
it
there.
I
think.
A
Yeah,
so,
okay,
so
now
we
have
zoom
back.
So
this
was
the
first
topic
about
the
having
the
owners
and
and
the
Specialists,
and
things
like
that.
Sometimes
is
it
so
I'll
just
put
to
you.
B
A
On
the
on
the
selector,
so
for
the
Eagle
tap
them
we'll
try
to
split
it.
A
Okay,
genuine
somehow
help
with
any
of
this.
How
would
we
maybe
split
the
word,
maybe
on
this,
you
want
to
try
get
to
get
get
you.
B
B
C
C
A
Okay,
yeah
with
the
attributes
just
reach
out
to
me,
if
you
need,
if
you
need
help
with
anything
like
I,
can
also
use
some
Cycles
to
help.
You
understand.
A
Okay,
perfect
right,
so
this
sounds
good,
because
the
next
thing
that
I
I
was
thinking
is
about
like
how
we
move
forward
because
I
was
thinking
like.
Should
we
focus
on
on
energy
on
the
old
tab,
get
it
get,
it
emerged
or
something,
but
I
think
with
we
discussed
now
that
I
think
it
makes
sense
right
splitting
it
and
it
will
be
easy
to
review.
It
will
be
easy
to
get
approval,
I
think
in
smaller
things.
A
So
I
think
this
should
it
looks,
looks
looks
good
as
our
Focus
for
me
focus
on
I
think
this
is
small
things
that
we
can
do.
C
This
is
up
for
grabs,
so
anyone
who
has
capacity
can
start
working
on
it
and
this
way
we
can
basically
work
on
multiple
streams.
So
some
people,
the
the
research
and
figure
things
out,
then
we
have
a
consensus
and
once
we
have
a
consensus,
some
people
can
just
go
and
grab
this
work.
Thank
you.
A
C
I
think
we
can
maintain
one
issue
where
we
just
list
things
and
mark
them
with
some
words
or
we
can
create
multiple
issues
and
link
to
them.
C
A
C
B
Think
yeah
I
mean
it's
a
future
possibility
here
we're
just
supposed
to
think.
We
also
have
our
project
board,
because
if
you
still
put
it
on
the
board
and
then
people
could
see
like
three,
what
I
mean
they
can?
We
can
adapt
our
board
to
our
needs
and
then
what
could
be
a
column?
What's
ready
to
be
picked
up
and
people
could
trust,
pick
up
things
and
work
on
those
yeah?
But
when
people
start
to
write
like
one
issue
with
like
check
marks
is
fine.
A
C
C
Yeah,
so
if
you
guys
don't
mind,
maybe
I
can
take
a
stab
at
create
like
it's
like
I'm,
dating
the
project
board,
to
include
this
details
and
we
can
review
it
next
time
and
see
if,
if
we
have
a
consensus
on
on
the
list
of
topics,
we
want
to
make
conclusions.
A
Does
this
Translate
what
we
discussed
I
guess
all
right.
A
A
A
I
can
try
to
ask
a
Christian
because
yeah
we
work,
we
work
in
the
same
company,
so
I
could
could
ask
him
if
he
could
be
more
close,
because
I
think
he
was
also
one
of
the
authors
of
the
current
messaging
specification
he's
quite
knowledgeable
about
tracing
in
general
and
all
these
things
so
I
could
ask
him,
will
not
do
the
same
team,
but
I
could
ask
him
if
he
could
be
have
more
eyes
on
it.
A
No
I
will
reach
out
to
him
we'll
see
what
happens
it
might
be
busy,
but.
A
Maybe
I
can
actually
I'll
see
he's
quite
approachable,
so
I'll
I'll
make
him
maybe
next
week
he
can
try
show
me.
Thank
you.
Let's
see
if
my
persuasion
skills
are
good.
A
Okay,
I,
don't
know
if
you
want
to
discuss
something
else,
otherwise,
I
think
we
went
through
all
the
topics.
We
can
discuss
the
time
that
we
have
to
discuss
whatever
you
want.
I,
don't
have
any
other
topics,
I
think.
C
I
have
a
small
question
about
the
the
pr
drugs
they
have
for
that
tribute,
so
I
didn't
want
to
switch
or
so
to
Market
ready
for
review,
because
I'm,
not
sure
I,
wasn't
sure
about
the
procedure,
and
so
perhaps
the
plan
here
is
we
wait
for
like
we'll
build
this
list
of
people
who
will
review
right
and
once
we
have
at
least
some
idea
about
maybe
Bruno
and
whoever
else
wants
to
join
I
can
then
convert
it
to
ready
to
review
and
also
tag
people
there.
A
C
I
think
there
are
a
couple
of
points
on
whether
we
want
to
specify
anything
else
about
the
links,
but
I
intentionally
want
to
stay
away
from
it
right.
C
A
Exactly
okay,
okay,
let's
I'll,
take
a
look
tomorrow
and
then
I
think
if
we
have
our
at
least
our
approved,
that
we
we
are
being
close
together,
working
so
I
think
if
we
have
our
approval
at
least
then
we
can
Mark
is
that,
as
as
we're
ready
for
a
review,
I
think.
C
C
It
the
rename
attributory
names
they
are
breaking
well,
I
mean
it's
an
experimental
spec
and
if
we
are
ready
to
spec
our
destination
name.
A
C
The
break
then,
at
some
point
technical
committee
members
will
get
worried
right
every
time
you
get
through
some
pushback
on
this
Breaking
part,
and
even
though
it's
an
experimental
stack,
even
though
it's
okay,
it's
just
could
be
harder
to
do
it
and
also
for
users.
I
think
backhands,
adopt
versions
as
soon
as
they
can
and
if
we
introduce
breaking
changes
every
time
it
will
just
piss
people
off.
B
I
really
think
we
should
bundle
breaking
changes
as
much
as
we
can
actually
because
the
I
mean
the
spec
is
experimental
or
the
so
many
conventions
are
experimental,
but
there
are
still
libraries
out
there
implementing
them
and
using
them
and
basically
when
we
of
course
they're
experimental.
But
if
we
do
breaking
changes,
then,
basically,
we
more
or
less
a
break,
open,
Telemetry
compatibility
for
those
libraries.
B
So
we
should
try
to
minimize
that
as
far
as
we
can
and
bundle
the
braking
changes
as
much
as
possible
for
sure
and
also
thought
middle
said
just
also
for
kind
of
our
how
to
say
credibility,
because
people
might
get
their
pissed
off
if
we
just
put
up
breaking
powers
all
the
time.
But
then
the
PRS.
C
Yeah,
it's
a
trade-off
right,
I
mean
there
are
just
a
handful
of
more
attributes
there
and
all
the
things
are
related
to
destination
right.
So
if
we're
going
to
rename
our
destination
attributes-
and
maybe
I
think
on
top
of
what
there
isn't
spr
is
just
the
two
or
three
more.
A
Could
we
the
part
that
you
have
opened
our
goodly,
only
introduce
the
new
attributes
but
still
keep
the
old
ones?
So
it's
not
breaking
right
now
that
removes
them.
A
Yeah
because,
for
example,
at
least
we
can
get
emerged
now
and
then,
like
once,
we
have
all
of
the
let's
say,
ranking
changes
in
in
a
way
that
they're
not
breaking
they're
like
still
compatible
and
then
because
they
did
this
with
the
Proto,
for
example,
with
the
stimulation
scope.
They
did
this,
so
they
first
renamed
it
they
still.
They
left
the
old
one
there
and
then
in
the
future
of
it
I
removed
it.
A
A
C
I
think
it
helps
and
it
gives
people
an
opportunity
to
gradually
update
to
either
new
and
old
version.
The
question
remains
for
the
the
instrumentation
owners.
I
guess:
I
can
try
playing
with
attribute
requirement
level,
so
the
old
attributes
become
recommended
or
optional,
we'll
see.
If,
if
you
can
do
it
and
then
the
new
ones
get
the
proper
levels
and
then
it
means
that
either
of
versions
would
work.
C
C
So
there
is
a
list
in
the
pr
with
breaking
changes.
I
think
there
is
a
handful
of
attributes
there,
but
okay.
So
then
the
plan
is
we're
trying
to
keep
the
old
attributes.
We
add
new
ones
and
we'll
see
how
it
goes,
because
there
is
a
versioning
system
right
and
we're
also
putting
things
in
parallel,
this
version,
but
we'll
try.
It
will
be
the
least
breaking.
A
Yeah,
because
when
I
think
it
was
nice
with
the
approach
that
they
did
with
the
instrumentation
scope,
it
was
that
there
was
a
period
for
for
people
to
still
to
still
use
the
the
old
and
and
have
a
have
some
time
for
them
to
to
change
their
code.
A
But
now
now
in
the
in
the
late
distribution
19,
they
removed
it
completely
yeah
and
they
have,
for
example,
they
had
this
for
the
instrumentation
Library.
They
moved
into
a
different
index
in
the
Proto,
but
they
have
a
huge.
They
had
a
huge.
The
like
the
disclaimers
say
like
don't
use
this
move
your
move
your
code
to
using
this
deal.
So
something
like
this
like
we
can
still
keep
the
Old
Fields.
The
old
attributes
until
like
I,
don't
know
something
like
update.
B
B
I
mean
for
the
attribute,
so
many
there's
also
this
whole
schema
transformation,
stuff
I
think
that's
just
mentioned,
and
then
but
but
I'm
not
sure
how
much
in
how
far
we
can
use
that
for
for
extra
experimental
semantic
conventions.
A
B
So
we
couldn't
trust
them
in
that
that
at
least
would
I
think
mitigate
the
impact
of
any
kind
of
breaking
changes
to
attribute
names
and.
B
C
Let's
try
to
default
building
the
pr
I
have
I
will
try
to
incorporate
all
the
destination
changes
we
discussed
so
far.
I
will
keep
it
in
draft.
It
will
give
us
an
opportunity
to
review
those
new
attributes.
We
have
better
and
well
talk
again
after
we
have
it,
but
the
general
direction
yeah.
Let's
not
try
not
to
break
people
unless
we
have
to
and
bundle
things
up.
A
Good
idea
internet
because
it's
still
in
draft
so
we
can
take
a
look
and
raise
our
concerns
and
I
guess:
have
the
user
head
of
of
I
don't
like
breaking
changes
and
try
to
spot
them
and
see
it's
something.
C
Oh
on
this
note,
since
we're-
oh
sorry,
I
I
wouldn't
want
you
to
leave
this
early,
not
four
minutes
back
on
this
note.
The
span
structure
is
a
break,
can
change
for
sure,
and
there
is
no
schema
transformation
for
this.
C
So
once
we
get
to
the
span
structure,
basically
anything
that
worked
before
might
stop
working,
because
everyone
would
need
to
change
the
relationships
into
expense
right
and
we
can
keep
it
quiet
and
say:
okay,
it's
an
experimental
stock.
You
should
expect
it.
There
is
no
precedent.
There
is
no
procedure.
C
We
cannot
do
specific
major
version
for
messaging,
but
it
will
have
to
be
a
new
major
version
for
the
whole
open,
Telemetry
specification
schema
version,
and
this
is
the
some
difficult
procedure
so
I
think
the
the
one
opportunity
to
consider
is
we
just
use
the
different
namespace
like
messaging
with
you
or
I.
Don't
know
we
can
be
creative
and
basically
say
it's
a
new
specification.
C
C
Or
a
revolution
overhauling
I,
don't
have
any
strong
opinion
here,
but
it's
something
to
consider.
A
Yeah,
okay,
I
see
yeah.
That's
that's
a
good,
very
good
point
because
yeah,
especially
if
we're
gonna
use
links
which
I
guess
not
a
lot
of
back-end
support,
which
is
still
like
I
guess
we
still
is.
A
This
is
one
of
the
big
things
that
we
still
need
to
discuss
or
get
approvals
on,
because
if
we
allow
or
if
we
depend
on
the
links
to
be
added
after
the
Spain
creation
thing,
which
is
not
possible
today,
right
but
anyway,
so
it
will
break
because
yeah,
if,
if
I'm,
using
a
system
that
doesn't
support
this,
then
my
Trace
will
be
someone
broke,
broken
but
yeah,
but
I,
I,
I,
guess
that's
a
good
idea.
A
A
You
can
still
use
the
other
thing
if,
if
can
still
support
your
thing,
if
you
don't
one
two
yeah
move
to
the
new
I
guess
it
will
be
a
sub
spec
new
version
or
something
like
that.
Great
conversion.
C
I
think
the
current
spec
it's
just
basically
can
evolve
for
non-batch
scenarios
right.
Maybe
we
can
do
this.
We
will
consider
this
pack
as
a
power
message
scenario.
Sorry
per
single
message,
single.
C
Deprecate
the
password
batching
and
basically
introduce
the
second
part
or
second
document
or
a
second
name
space.
However,
we
will
do
it
for
it.
It
covers
batching
really
well,
and
then
people
can
I
the
simple
messaging
like
people
who
do
messaging
kind
of
simple.
We
can
just
go
and
use
the
simple
one.
Yeah.
A
Also,
a
good
idea,
I
think
that's
one
thing
that
I
thought
is
that
maybe
we
think
that
this
is
a
big
problem,
because
the
like
that
this
is
a
huge
breaking
change,
but
I
think
what
Christine
mentioned
in
one
of
the
comments
in
the
old
tab.
A
He
said
like
he
seem
to
be
doing
like
an
overhaul
like
we're,
basically
redoing
the
whole
thing
right,
and
he
mentioned
like
instead
of
calling
the
Otep
as
stabilizing
call
it
like
overhaul
and
then
take
parts
of
the
overhaul
and
put
it
on
the
document
to
stabilize
it
so
like
if
I
interpret
what
he
said
correctly,
it
is
will
be
like
an
incremental
stabilization
thing
and
do
a
thing
that
is
experimental.
A
That
should
be
fine,
but
I.
I
can
see
both
sides
in
if
people
are
already
rely
on
this
today
that
it
will
be
breaking
it's
not
great,
but.
A
But
maybe
like
we
think,
I'm
sorry,
but
maybe
we're
thinking
like
this
is
a
big
problem
and
maybe
we,
if
we
discuss
with
somebody
else
or
with
other
community
members,
they
can
tell
us
no
like
I,
guess
you
can
go
ahead
with
like
just
redoing
everything.
It's.
It
is
experimental
anyway.
So
maybe
we're
thinking,
that's
a
problem.
Maybe
it's
not
a
problem.
So
maybe.
C
I
think
it's
just
the
phasing
approach,
so
the
first
phase
is
get
what
we
have
but
covers
single
message:
scenarios
reasonably
well
after
the
stability
and
or
to
the
place
we
want
it
to
be
and
and
then,
however,
the
the
second
part
comes,
which
is
still
where
she
won.
We
need
both,
it
will
come
so
the
first
part
is
evolution.
We
try
to
just
just
with
attribute
her
name,
so
we
can
stay
within
the
realm
of
the
current
specs.
A
C
A
All
right,
okay,
all
right,
then
I-
guess
we're
through
I-
will
update
the
document
with
more
things
afterwards,
with
the
meeting
notes
so,
but
I
think
we
should
be
good
right.
Awesome.