►
From YouTube: 2023-02-01 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
B
A
Okay,
we
can
start
okay,
now
so
I
updated
your
PR
to.
A
Can
you
take
a
look
and
I
can
share
my
screen?
Yeah.
B
A
In
a
draft
state
and
I
was
wondering
if
you
know
you
can,
you
know,
take
that
draft
tag
off
so
that
people
can
review
it.
B
Yeah
I
haven't
had
like
I,
saw
your
published
late
yesterday,
so
I
guess.
If
you
want
to
go
through
it
here,
there's
nothing
else,
but
probably
won't
matter
allocate
the
time
to
go
through.
A
Okay:
let's
do
that.
It's
only
a
couple
minutes.
A
So
yeah
I
just
added
these
two
attributes
to
the
emit
event.
A
The
the
emit
event
API
currently
it
takes
as
per
the
spec.
It
takes
an
event
name
and
the
log
record
we
anyway
wanted
to
avoid
the
term
logs
in
in
the
events
documentation,
so
I
removed
this
it's
an
event,
name
and
event.
Data
and
other
attributes.
B
A
Yeah,
it's
I
know
it's
confusing.
B
Yeah,
you
can
leave
it
as
attributes
and
then
we
just
say
internally:
we
assign
that
argument
to
the
event
data
which
is
sort
of
already
saying.
Well,
you
mentioned
log
record
again,
yeah.
A
That's
all
just
a
small
change,
at
least
sorry.
A
Okay,
yeah,
that's
all
I
think
if
I
I
don't
know
how
we
can
remove
this
draft
level.
B
D
A
Yeah
there's
only
one.
There
are
two
comments,
the
one
from
tigran
we
kind
of
addressed,
but
that
that
will
need
to
be
further
discussed.
D
A
Is
there
a
term
we
want
to
use
for
the
other
attributes,
I
left
it
as
other
attributes
right
now,
yeah.
B
Yeah
I
think
other
attributions,
because
we
don't
know
what
they're
going
to
be
yeah
yeah.
Okay,
let
me
take
the
across
my
God.
B
I
think
I
commented
on
that
in
the
the
description,
and
we
also
talked
about
it
in
a
sec.
It
was
mainly
so
that
the
shape
we
can
say
there's
a
log
event
or
a
span
event,
but
the
only
difference
being
if
we
can't
get
a
map
of
events
in
a
span
event,
it's
going
to
have
to
be
serialized
or
something
that
was
the
main
reason.
Obviously
we
could
use
body.
A
Yeah,
it
will
be
confusing,
although
practically
in
the
short
term,
it
will
be
easier
like
we
don't
need
an
extra
attributes
yeah.
B
A
Yeah
but
I
think
pushback
is
unlikely,
given
that
we
are
going
with
necessary
attributes
across
multiple
signals
anyway,
keeping
it
in
the
signals
attributes
you
know,
enables
you
know
having
all
these
things
at
one
place.
A
Okay,
so
now
that
this
PR
is
open,
will
will
pursue
it.
Yeah.
A
I
mean
there
are
typos
and
grammatical
changes
needed.
We
will
will
do
it
yeah,
but
I
think
we
can
start
getting
comments.
Feedback,
yep.
A
So
the
next
one
I
have
is
I,
think
I.
Think
the
main
thing
at
this
point
is
we
complete
the
logs
and
events
SDK
work
so
that
we
can
start
building
prototypes
Martin.
C
May
I
update
the
size
so
I
have
that
that
PR
to
separate
the
API,
that's
green,
it's
been
approved,
but
it
hasn't
been
merged.
Yet
so,
okay.
D
A
Yeah,
okay,
I
I,
don't
know
what
what
is
the
best
way
to
test
things
with
changes
being
in
multiple
reports
like
multiple
PR's
I
have
to
be
just
build,
separate
packages
and
then
test
that
will
be
a
painful
rather
if,
if
all
of
them
were
in
one
repo
like,
for
example,
we
create
a
branch
from
from
your
PR
branch
and
then
merge
the
logistic
into
that
things
will
all
be
in
one
place
and
then
it
might
be
easier
to
test
locally.
Just
between
us.
C
Yeah
yeah,
but
that's
that's,
that's
a
hard
point.
I
mean
we
can.
We
can
just.
We
can
just
build
the
whole
prototype
in
a
in
a
fork
and
that's
as
at
the
same
time
as
we're
working
on
these
PRS
yeah.
C
So
I
think
the
next
The
Next
Step
would
be
to
to
finish
the
log
SDK
with
the
exporter
or
I.
Think
you
know
there's.
If
this
other
person
is
still
available,
then
yeah.
A
I'll
I'll
ping
him.
He
did
respond
to
me
on
slack,
okay
once
so,
but
I
think
he's
waiting
on
your
PR
being
merged.
D
A
It's
dependent
yeah
so
but
I
believe
you
you'll
be
checking
with
Daniel
today
to
get
that
moving.
Yes,.
B
Technically,
I
have
the
ability
to
push
that
in
right
now
being
an
approver
but
considering
the
JS
meetings
after
this,
let's
wait
till
then.
C
Yeah
yeah
there
was.
There
was
one
comment
from
about
the
name
of
the
the
class.
C
A
C
But
that's
that
API
is
like
that's
not
not
valid.
In
my
opinion
anymore,
so
I
mean
that
spec
is
not
really
really
a
representative
of
of
what
we're
working
on
now.
Okay,
I
mean
we
can.
We
could
decide
to
call
it
even
blogger,
but
you
know
like
we're
separating
this.
This
API
completely
from
blog
blog
signal
or
like
from
the
log
SDK
API.
So
I
guess.
The
question
is
like
what
does
event
loggers?
C
Don't
make
sense
as
a
name
I
Know,
That
Jag
has
a
prototype
in
Java
and
he
calls
it
event.
Emitter.
C
C
Yeah,
even
there
is
an
interface
it's
built
into
node,
oh
okay,
okay
I
mean
I,
don't
I,
don't
if
it's.
C
If
it's
like
it's!
This
comes
from
this
different
package.
I
mean
I,
don't
see
a
real
I
wouldn't
spend
too
much
time
worrying
about
it.
I
guess
but
but
I
can
I
can
ask
what
other
others
think
in
in
the
js6.
Today,
okay,.
A
Okay,
yeah
I
mean
any
any
name
is
fine,
yeah,
okay,
yeah
actually.
B
B
Yeah
the
fact
that,
like
regenic,
has
got
there
it's
like
event,
emitter
from
node
events.
That's
how
node
effectively
sends
a
dependent.
You
can
actually
extend
the
event
emitter
as
well,
and
no
yeah
so
yeah
that
there's
potential
someone
might
want
to
effectively
be
emitting
node
events
and
also
trying
to
emit
your
our
events
and,
if
they're,
not
careful,
it's
it
could
be
a
problem.
B
I
I
don't
have
a
recommendation
for
a
name
at
this
point.
It
would
be
nice
to
keep
parallel
with
with
Java
but
yeah,
but
yeah
I,
don't
think
event.
Logo
makes
sense
either
I
think
I'm.
Still
a
fan
of
having
eye
on
the
front
for
interfaces,
so
I
seem
to
be
in
the
minority
nowadays.
C
A
Martin
other
than
the
logs
SDK,
you
will
also
need
to
build
event
SDK,
but
the
event
SDK
I
believe
can
wait
right.
You
can.
You
can
always
Plug-In
or
rather
ship
the
events
via
log
SDK.
Is
that
correct
or
no
like.
A
We
prevent
the
events
from
being.
Let's
say
there
is
your
your:
you
have
instrumented
things
with
both
logs
API
and
events
API,
but
then
at
the
application
Level
somebody
only
configured
a
log
SDK
a
log
pipeline.
Then
it
will
take
the
events
as
well.
C
Well,
no
I
think
I
think
the
goal
is
to
have
to
have
a
complete,
separate,
separate
the
events
SDK
that
you
would
have
to
configure
in
your
application,
yeah,
so
yeah,
you
will
configure
it
by
also
configuring.
The
log
SDK.
A
Yeah,
but
that
means
that
the
log
SDK
and
the
event
SDK
spec
need
to
explicitly
you
know
avoid
looking
at
the
other.
Other
type
like
the
log
SDK
should
should
leave
the
events
untouched.
That's
that's
right,
yes
right,
so
that
on
change
needs
to
go
into
log
SDK.
C
A
Not
just
in
terms
of
code
I'm
talking
about
the
log
SDK
should
explicitly
look
for
the
event.name
parameter,
because
that's
what
we
use
to
distinguish
logs
from
events
and
then
exclude
them.
A
A
We
don't
want
just
the
presence
of
a
log
SDK
shouldn't
so
by
default,
when
SDK
is
not
configured,
we
we
use
the
no
op
processors
right
yeah.
So
unless
an
SDK,
an
events
SDK
is
configured
is
setup
for
events.
It
should
continue
to
use
an
OP
processors,
so
just
the
presence
of
a
log
SDK
should
not
enable
processing
of
events.
B
You
would
need
to
also
set
up
the
event
SDK,
which
is
you've
set
up
to
use
logs
like
we
could
have
a
an
event
SDK
that
emits
to
the
console
just
like,
like
there's
a
trace
span,
console
thing.
B
B
It
doesn't
need
to
know
anything
about
events.
It's
just
gonna,
be.
C
So
so
the
instrumentation
and
the
instrumentation
will
will
talk
to
the
events
API,
which
will
be
linked
to
events
SDK
and
the
events.
Sdk
is:
has
an
implementation,
that's
linked
to
some
instance
of
the
log
SDK,
and
it
can
be
you
know.
So
if
you
didn't
configure
the
events
SDK
at
all,
then
like
no
events
would
be
flowing
to
the
logs
SDK
yeah.
A
I'm
concerned
about
the
other
way
where
there
is
no
events,
SDK
setup,
the
instrumentation
is
emitting
events
as
log
records
with
log
records,
as
the
underlying
mechanism
underlying.
A
I
mean
all
events
come
as
log
records
right,
so
so,
when
you,
okay,
I
I
I'm,
not
100
sure
about
this,
but
I
have
a
feeling
that
the
log
processors,
you
know
we'll
be
looking
at
all
log
records
and
therefore
it
will
be
looking
at
events
as
well.
Yep.
B
If
someone's,
even
the
logs
SDK
a
log
record,
it
should
just
process
it.
It
shouldn't
weed
out
events,
because
if
it
did
that,
when
we're
using
the
event
SDK
to
send
it
through
the
logs,
it
would
then
drop
them
right
like
if
someone
wants
to
go
and
explicitly
use
the
logging
API
to
create
a
log
record.
That
looks
like
an
event.
That's
perfectly
legal
to
do
not
recommended,
but
perfectly
legal
to
do,
because
that's
affecting
all
the
event
SDK
is
going
to
be
doing.
C
Yeah
and
I
suppose,
if
you
wanted
to
you,
could
have
you
could
have
a
completely
separate
blog
SDK
instance
just
for
the
events
yeah,
if
you
wanted.
C
Way,
but
if
your
back
end
can
handle
like
distinguishing
logs
from
events,
then
you
can
just
hit
one.
B
A
I
I
didn't
fall,
I
wasn't
Following
last
one
minute,
I
was
trying
to
take
out
an
example.
So
let's
say
you
will
have
a
logger
provider
right.
Your
provider
is
a
logger
provider
and
then
you
set
up
a
lock
processor.
A
B
So
yeah
that
that
would
have
to
configure
both
where
what
I
thought
you
were
saying
is:
if
someone
decides
to
use
that
logger
instance
directly
to
create
something
that
looks
like
an
event.
No.
D
D
A
Okay,
now
I
I
I
I
miss
the
part
that
you,
you
cannot
create
events
using
a
logger
provider,
correct
yeah,
okay,
foreign.
A
Martin,
we
will
then
need
the
events
SDK
as
well.
Are
you
planning
to
start
on
that
too?
On
the
side.
C
Yeah
yeah
once
so
once
the
is
it
has
yes,
yes,
yes,
it
has
dependency
on
the
log
SDK.
So
but
it
can
it.
You
know:
I'm
I,
don't
I,
guess
I,
don't
anticipate
that
to
be
in
too
much
work
if
it's
just
the
first
Vlog
SDK,
but
so
like
the
main.
The
main
thing
to
work
on,
in
my
opinion,
is
it's
like
to
help
help
with
the
log
SDK
and
unlock
exporters.
A
Okay,
actually,
when,
when
you
work
on
the
so
the
log
SDK
PR
should
come
along
with
some
exporters
in
in
the
same
PR.
That
way,
you
know
it
will
be
easier
to
test.
Otherwise,
how
will
you
test
the
log
SDK
standalone.
A
I
think
his
name
is
Hawaii
I
hope
I'm
pronouncing
his
name
correct,
but
we
should
ask
him
to
include
the
logs
some
exporter
HTTP
exporter,
in
into
the
in
the
same
PR.
C
A
A
So
yeah
I
think
if,
if
we
can
make
progress
on
the
current
open
PRS
that
that
will
be
great
this
week,
the
your
PR
and
the
logistic
APR.
Could
you
could
you
ask
my
Daniel
also
to
look
into
that
PR,
the
log
SDK,
so
that
we
can
make
some
progress
I.
A
Yeah,
but
if
we
can
get
some
early
feedback,
it'll
be
good.
A
C
A
I
I
hope
there'll
be
some
decision
on
it
in
the
jset
today
so
that
it
doesn't
remain
pending.
A
All
right
so
the
next
one
on
my
list
was
I,
I,
I,
think
Ted
about
a
week
ago.
He
said
he
was
going
to
check
with
the
TC
in
you
know,
Wednesday
meeting
last
week,
but
I
haven't
heard
further
from
him.
Have
any
of
you
heard
from
him
on
having
a
TC
member
join
our
sick,
no.
A
Martin
I.
A
C
I
will
think
that
yeah,
okay.
B
I
I
I
guess
the
sandbox
I
I
think
you
like
yesterday,
which
is
probably
probably
too
late,
that
I
I
have
confirmed
that
the
API,
compile
and
test
is
working.
B
The
pr
still
needs
one
minor
update
once
you've
successfully
finished
testing
in
that
at
the
moment,
the
test
flag
is
the
default
and
I
don't
have
a
switch
to
unset
the
test
flag.
So
I
need
to
change
that
back
to
the
so
that
it'll
be
full
so
that
when,
when
I
activate
the
task,
it
won't
be
in
test
mode
that
just
changes
the
location
of
that
dot.
Auto
merge,
folder,
that's
really
wow.
B
B
Time
well,
the
change
is
there
at
the
moment
so
that
all
you
have
to
do
is
say
no
PR
or
you
can
just
run
it.
Okay,
so
yeah.
So
once
you're
happy
with
that,
I'll
then
do
one
one
more
final,
push
to
I
think
revert
those
flags.
Now,
okay,
I'll
highlight
those
in
the
I'll.
Add
comments
on
them
in
the
piano.
You
know
which
ones
I'm
talking
about.
A
All
right,
we
can
end
here
thanks
everyone.