►
From YouTube: 2022-11-22 meeting
Description
Instrumentation: Messaging
A
B
B
B
Yeah
I
haven't
been
to
the
group
in
a
bit,
so
I
was
popping
in
to
just
see
how
things
were
going,
but.
A
A
A
Good
I
mean
mostly
we've
been
working
on
on
some
semantic
conventions
for
the
different
types
of
events,
and
you
know,
ram
ram
has
been
doing
a
good
job
of
beating
that
effort.
Great
I
think
at
this
point
like
we're,
trying
to
decide
like
how
to
put
those
schemas
into
the
existing
specs
like
where
they
would
fit,
because
the
you
know
like
the.
A
Conventions
are
per
event
type,
so
there's
no
such
thing
right
now
in
the
spikes,
so
right.
B
I
I
definitely
I
proposed
reorganizing,
some
of
that
work,
a
while
back
yeah
and
then
went
on
vacation
instead
of
doing
it,
but
I
definitely
think
we
could
reorganize
how
we,
how
we
present
the
semantic
conventions
at
minimum
I
think
having
them
the
way
they're
kind
of
buried
in
the
different
signals.
I
think
it
would
be
better
if
it
was
at
least
yeah
organized
by
like
domain
or
something
like
that.
B
The
only
thing
that's
annoying
there
is
like
that
stuff
is
auto-generated
machine
generated
out
of
a
yaml
file,
so
actually
that's
that's
kind
of
where
I
got
stopped.
Actually,
it's
like
moving
it
all
around
a
little
bit.
Annoying
yeah
you
have
to
you
have
to
go
mess
with
whatever
that
script
is
to
do
it.
A
Yeah,
it
seems
like
last
week
they
talked
about
also
a
common
like
sharing
attributes
like
across
signals
and
and
like
it
was
brought
up
that
you
know
that
there
is
already
a
duplication
right
now
like
for
certain
things
like
there's
a
for
example,
like
HTTP
semantic
events
are
duplicated
for
traces,
tracing
and
metrics.
B
A
B
Many
different
ways
do
we
need
to
tell
you
the
duration
of
this
request.
Yeah,
but
I
mean
I,
think
some
of
that
duplication
is
okay
in
the
sense
of
like
if
you're
going
to
admit
this
is
a
metric.
Here's
how
you
should
do
it,
but
I
think
we
haven't
written
down
like
to
do
you
need
to
do
all
of
this
at
the
same
time,
or
is
it
more
just
like
guidance
for
how
you
should
do
it?
If
you
are
going
to
do
it,
I.
A
Do
you
do
you
attack
the
this
month
in
convention,
seg,
yeah.
B
That
group
we're
just
kind
of
I'm
gonna,
propose
next
week,
trying
to
just
plan
out
how
we're
actually
going
to
tackle
all
these
different
semantic
conventions
just
based
on
how
HTTP
and
these
other
groups
have
gone.
You
know
I
have
a
concern.
It's
going
to
take
us
10
years
to
get
through
through
all
of
these,
so
I'm
I'm,
trying
to
put
together
kind
of
like
like
a
schedule
for
like
how
many
of
them
I
think
we
could.
We
could
tackle
per
quarter
and
get
feedback.
So
that's
one
thing
happening
there.
B
The
other
thing
is
Josh,
serif
and
others
are
trying
to
defining
like
stability
and
compatibility
and
stuff
for
metrics
is
a
little
bit
trickier
than
for
tracing
and
so
they're
kind
of
focused
on
on
that
the
the
same
way
we
had
it's
actually
the
other
side
of
like
us
needing
ephemeral
resources.
A
B
Want
to
do
they
need
they
have
like
the
same
issue
of
like
what
are
like.
What
are
the
subset
of
resources?
You
should
use
as
identifiers
when
you're
generating
these
metrics,
because
it
can't
be
all
of
the
resources,
because,
if
you
change,
if,
if
you,
if
you
add
or
remove
one
of
these,
it's
going
to
break
break
metrics
in
a
way
that
doesn't
really
break
tracing,
so
that's
that's
kind
of
where
that
group
is
focused
right
now.
It's
just
right.
A
Right
so
those
like
those
attributes
like
I
was
like
the
identifying
attributes
for
metrics
are
kind
of
like
the
things
that
you
know
like
those
that
you
Group
by.
B
Yeah
yeah
yeah,
exactly
the
the
metrics,
are,
are
more
brittle
when
it
comes
to
the
dimensions
that
you're,
including
in
them,
and
so
we
have
to
strictly
Define
what
those
are
going
to
be,
which
isn't
the
thing
we've
done
yet
because
you
did
tracing
it,
doesn't
matter
as
much.
You.
B
So
so
that's
kind
of
where
that
group's
at
that
I
think
once
they
they
get
once
once
they've
they've
gotten
the
the
rules
written
down
for
how
how
we
stabilize
metrics,
then
then
they'll
be
able
to
move
forwards
on
to
like
some
of
the
domains,
but
it
might
be.
It
definitely
might
be
helpful
if
there's
things
you
know,
this
group
needs
for
someone
to
go.
B
There
I'm
happy
to
pass
pass
information
along
as
well,
but
where
I
feel
like
this
and
the
the
messaging
group
we're
we're
like
the
first
groups
to
be
like
trying
to
really
stabilize
these
like
big
chunks
of
of
instrumentation,
so
I
think
that
group
would
love
feedback
from
us
if
there's
stuff
missing,
where
we're
like,
we
can't.
We
can't
finish
because,
because
we're
missing
some
important
piece-
or
we
don't
understand
where
we're
supposed
to
put
this
in
the
spec
or
like
how
this
interacts,
they
would
definitely
love
that
feedback.
Yeah.
A
Yeah,
it
looks
like
nav
already
reached
out
to
Josh
Josh,
okay,
perfect
and
I
might
I
might
start
attending
the
Samantha
convention.
Seg
yeah.
B
Yeah,
but
it's
it's
I'm.
It's
promising,
though
it's
I'm
happy
to
see
the
core
group
like
shift
to
paying
a
lot
more
attention
to
this
stuff.
I
think
part
of
the
slowness
in
the
past
was
just
everyone
was
busy,
so
people
outside
of
these
groups
weren't
really
like
paying
attention
to
it.
B
B
We
have-
maybe
not
everything
done,
but
all
the
all
the
major
stuff
is
done
so
between
this
configuration
like
yaml
file
based
configuration
for
the
the
clients
and
like
op
amp,
like
the
control
plane
for
being
able
to
push
out
configuration
if
we
get
all
of
that
done,
plus
plus
stabilizing
instrumentation.
B
For
for
everything
then,
like
end
of
2023,
open
Telemetry,
like
the
core
project
is
like
would
be
I
would
say
like
pretty
much
pretty
much
done
at
that
point.
I
mean
it'll
never
be
done,
but
that
that's
like
we
will
have
delivered
like
a
stable,
completely
coherent,
easy
to
use
Telemetry
system,
and
then
we
can
start
applying
for
graduation
at
the
cncf.
B
Yeah
and
I
don't
know
if,
if
has
alolita
or
anyone
from
Apple
been
attending,
do
you
know.
B
No
I
think
that
that
will
happen,
but
now
that
Alita
works
at
Apple
she's
promised
to
get
the
like
the
Swift
team
involved
a
bit
more.
So
we
should
be
able
to
do
something
coherent
on
iOS
that
that
has
Apple
actually
actually
like
paying
attention
to
what
we're
doing,
which
I
think
would
be
really
useful.
A
Yeah
yeah
for
sure
cool,
yeah,
there's
been
there's
been
much
much
progress
on
mobile.
B
Yeah
yeah
I
figured
we
would
get
through,
I
I
haven't
been
pushing
on
it
because
I
figured
like
it
would
be
helpful
to
get
farther
on
browser
before
we.
We
talk
to
the
mobile
people,
because
they're,
probably
gonna
they're,
less
involved.
They're
gonna
want
more
of
a
plan
from
us,
but
iOS
has
worried
me
just
because
so
much
of
iOS
development
is
done
with
proprietary
Apple
Frameworks.
B
B
Exactly
either
them
directly
baking
in
open,
Telemetry
or
just
making
sure
that
we're
all
on
the
same
page
about
what?
What
it
means
to
to
to
Loop
that
stuff
in
yeah.
So
but
yeah.
A
B
B
All
right
well
have
a
good
holiday
break
yeah.