►
From YouTube: 2022-10-18 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
B
B
I
did
see
like
Dan,
but
I
I,
don't
know
whether
you
ever
had
a
chance
to
reach
yourself,
because
I
can
see
your
name
in
the
last
few
meetings.
Probably
I
did
not
notice.
B
B
B
Well
because
I
think
we
can
just
cover
the
agent
items,
we
have
any
smaller
yeah
one
of
the
museum
in
discussions
that
might
take
okay,
yeah
yeah
released
one,
not
4.0
beta
2
version
yesterday,
and
this
has
the
min
max
feature
of
histograms
and
yeah.
We've
also
released
the
other
like
the
instrumentation
packages,
rc98,
so
yep,
and
let's
look
at
this
request.
A
B
Okay,
we
did
this
because
this
like
briefly
last
week,
but
then
we
gave
everyone
time
to
think
through
so
and
I
would
suggest
to
like
pick
something
today,
because
we
have
a
bunch
of
other
instrumentation
libraries
which
would
require
a
similar.
A
B
No
I
I
wouldn't
have
like
very
strong
opinion,
but
my
only
preference
is
we
need
to
close
it.
You
won't
be
able
to
release
other
things
together.
Yeah
makes
sense.
B
B
So,
with
this
case,
I
think
we
would
explicitly
have
it
on
the
name
and
also
the
type
would
be
we
would.
You
can
see
the
arguments
type
yeah
either
way.
I
don't
have
like
a
very
strong
opinion,
and
one
thing
is
that
it's
not
our
goal
to
have
the
excite
same
name
for
each
instrumentation,
Library,
I.
Think,
for
example,
if
you
have
SQL
there
would
be
like
something
else
like
enrich's
own
command
start
or
enrich
from
SQL
command
like
it
would
be
different,
so
I,
don't
think
that
we
need
to
aim
for
that
thing.
A
B
A
B
Last
three:
at
the
bottom,
there
enrich
with
base
TTP
requests.
Http
response
with
exception.
It'll
seem
pretty
good,
yeah
I,
think
that
has
two
Watts
now.
So,
if
you
want
to
write
it
down,
okay,
yeah.
B
B
Okay,
yeah
like
so
who
ordered
the
pr.
Is
it
so?
You
can
like
put
this
into
the
summary
and
say
that
we
discussed,
and
this
is
what
we
agreed
and
we
can
proceed
sorry
for
the
folks
who
did
not
get
a
chance
to
work,
but
it
would
just
save
us
time
because
it's
already,
therefore
yeah
I'll
I,
can
put
a
commenting
yeah.
So
the
only
thing
is
like
we
need
to
follow
the
consistent
thing
with
other
one
like
for
secret
to
be
enriched
with
SQL
command,
enrich
with
secret
connection.
B
B
Okay,
yeah
so
I
mean
it's
still
like
possible
that
we
can
break
it
because,
anyway,
this
instrumentation
libraries
will
remain
like
not
stable,
so
we'll
still
have
opportunity,
but
we'll
try
to
stick
with
this
name
when
we
actually
do
this
table
so
yeah,
it's
not
like
end.
There
is
probably
one
more
opportunity
when
you
will
be
the
public.
Api
maybe
decide
to
do
stable,
so,
okay,
so
like
we
should.
You
will
update
the
pr
and
the
other
instrumentation
libraries
matching
the
same
thing
like
that.
B
Yes,
okay
yeah,
after
that
yeah
there
might
be
few
in
the
country
report.
I
do
not
know
whether
all
of
them
offer
enrich,
but
just
take
a
look
like
if
there
are
like
few
in
the
contract
report.
Let's
try
to
make
them
consistent
with
them
sure.
B
Okay,
yeah
the
next
one,
I
added
it
to
the
agenda,
so
this
is
something
which
was
open
for
quite
some
time
and
it
was
still
in
time.
Then
it
got
active
again.
So
I
took
a
look
at
it.
It
looks
reasonable.
We
need
to
First
merge
it.
Then
it's
currently
a
standalone
thing
nobody's
using
it
by
passing:
no,
no
SDK,
no
exporter,
no
instrumentation
libraries
consuming
it.
B
So
we
need
to
like
start
modifying
the
instrumentation
libraries
to
depend
on
this
instead
of
their
own
hard-coded
version
of
semantic
conventions
and
eventually
release
this
as
a
standalone
package
for
people
to
consume.
So
all
I'm
asking
is
just
take
a
look
and
if
okay,
let's
merge
it
in
So.
Currently
we
do
linking
between
the
contribute
before
all
of
these
libraries
right
yeah,
but
this
this
way
like
it's
all
generated
from
the
specification
itself
yeah.
So
if
you
want,
we
can
take
the
linking
approach.
Okay,
yeah!
B
We
cannot
make
the
core
components
depend
on
it
directly
linking
it
should
be
fine,
but
we
cannot
ask.
Has
the
SDK
to
take
a
dependency
on
this
package,
because
this
package
will
be
experimental.
Sdk
wants
to
post
table
so
we'll
have
to
do
like
some
extra
thing
to
let
me
get
now
we
have.
We
are
still
stable,
even
though
we
depend
on
something
so
might
be
easier
to
packages.
B
Yeah
yeah,
so
just
asking
everyone
to
just
take
a
quick
look.
I
think
there
are
like
few
open
question
like
what
should
be
the
namespace.
Should
it
be
dot
semantic
convention
store,
trace
or
the
class
name
could
include
that
and
namespace
can
be
shorter,
but
those
are
collectively
smaller
things
which
we
can
always
iterate
on
later,
as
well.
B
Okay,
actually
I
had
one
more
thing,
but
I
did
not
have
time
to
prepare
the
pr,
so
I
want
to
invade
vishwesh
to
join
us
and
up
over
and
I
want
to
invite
pre-order.
That's
the
approval
and
the
contribute.
I
already
spoke
to
them
individually
and
already
discussed
with
other
maintenance,
but
I
just
didn't
have
time
to
measure
PR,
so
I
will
be
sending
a
PR
shortly
after
the
meeting
yeah.
That's
all
the
agenda
from
me.
So
there
are
any
other
things.
A
A
Counters,
what
was
anything
more
on
your
mind
as
far
as
validation
that
you'd
like
to
hopefully
see
before
we
go
stable.
B
Yeah,
the
one
thing
which
I
want
to
say
is
like:
if
someone
is
in
neurotic
like
you,
are
consuming
up
down
counter
via
otlp,
it
would
be
good
to
get
a
confirmation
that
they
are
okay
with
the
temporality
thing
we
are
doing,
because
we
are
keeping
up
down
counters
as
cumulative,
even
though
the
aggregation
temporal
T6
Delta.
That's
following
the
spec
assist,
but
just
one
like
someone
to
confirm
that
they
are
indeed
seeing
what
they
expect.
A
Yeah
I
know
if
so,
yeah
I,
don't
I,
don't
have
any
net
users
as
examples
of
course,
since
we
don't
have
a
release
yet.
But
the
I
did
note
that
we
have
many
users
that
are
using
the
Java
runtime
instrumentation,
which
does
use
up
down
counters
and
they
are
coming
across
with
cumulative
temporality
and
I
could
actually
probably
share
that
with
you.
Just
it
was
useful
to
to
show
you
what
it
looks
like.
A
B
Yeah,
please
go
ahead.
We
have
like
some
samples
running
already
to
show
if
you
want
to
run
some.
A
A
Let
me
think
about
it,
maybe
maybe
I'll,
maybe
I'll,
show
you
next
time
around
yeah.
B
That
would
be
fine,
so
the
main
reason
why
I
asked
what
was
like
within
Microsoft.
We
heard
it
back
in
which
only
accepts
deltas
and
we
had
some
trouble
with
up
down
so,
but
that's
mostly
because
of
our
own
incorrect
understanding,
so
I'm
still
testing
it
out.
So
if
I
see
anything
like
wrong
and
if
I
see
the
need
to
have
a
pure
Delta
for
up
down
I'll
raise
it,
but
so
far
it
looks
like
we
will
just
export
it
as
great
and
it
should
just
focus.
B
A
Yeah
I
think.
The
one
thing
that
you
you
pointed
out
is
that
it
sounded
like
the
Microsoft
system
did
not
have
a
last
value,
aggregation,
yeah
and
so
you're
using
average,
which
will
come
up
with
a
different
result.
You
know,
but.
B
Oh
interesting,
okay,
yeah,
it's
just
a
lack
of
understanding
like
with
me
and
because,
like
you're
trying
to
export
it,
but
we're
not
sure
like
why
is
it
still
accumulating?
Even
though
it
says
we
asked
the
SDK
to
use
Delta,
so
we
had
to
like
dig
some
history
around
it,
but
I'm
99
sure
like
we
wouldn't
have
any
issues
with
the
way
it's
currently
setting.
A
Yeah
I
think
the
the
history
there
is
well
there's
temporality
and
then
there's
the
thing
that
we're
calling
at
the
spec
level,
temporality
preference,
which
admittedly
I
think
all
of
this
stuff
is
confusing.
But
the
temporality
preference
of
Delta
means
basically
Delta
for
everything,
but
up
down
counter
instruments.
A
B
The
good
thing
is
I,
don't
really
think
our
end.
Users
would
really
need
to
know
that
they
should
they
usually
use
some
launcher
or
destroying
the
individual
vendors
they'll
set
it
up
the
right
way
as
they
expect.
So
it's
usually
not
required
for
induces
to
know
what
all
these
things
means.
It's
no
statement
for
exporter
writers
and
the
people
who
ship
the
distros
so
yeah
I
can
be
with
that.
It's
it's
not
that
bad.
B
A
The
moment
any
longer
have
our
own
distros.
We,
we
are
asking
people,
you
know
with
the
TLP
exporter,
to
configure
the
temporality
preference
too
be
Delta
okay,
so
it
is
kind
of
a
it's
in
the
face
of
end
users.
In
that
respect,.
B
Okay,
okay,
yeah,
so
I
still
plan
to
like
go
with
the
stable
release
in
end
of
November
or
mid-number.
Whenever
we
feel
comfortable
I
think
the
there
are
like
few
changes
from
Plants
about
the
configuration.
Those
are
the
only
things
which
we
need
some
validation.
B
Apart
from
that,
the
only
key
thing
means
we
added
One
Stop
downtown
I'll
be
testing
Max
from
the
Microsoft
perspective.
There
is
no
Prometheus
for
I
mean
there
is
no
place
in
permit
just
to
put
in
Air
Maxes.
We
don't
really
get
any
validation
from
Primitives
right
yeah.
So
like
build
start,
a
API
review,
like
maybe
two
weeks
from
now,
just
to
see
all
the
things
which
have
public
are
intended
to
be
public
by
1.4
time.