►
From YouTube: 2022-09-14 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
A
A
D
Yeah
I
had
a
quick
question,
so
one
thing
looking
at
the
browser
instrumentation,
so
we
have
like
user
interaction
document
load
Etc.
D
A
A
There
is
an
effort
going
on
to
standardize.
We
collected
the
scheme
of
the
different
events
sent
by
when
used
by
the
different
vendors,
and
we
are
trying
to
identify
now
a
common
structure
that
we
could
use
as
a
standard
in
the
open
television.
A
So
as
part
of
that
yeah,
we
intend
to
move
some
of
the
objects
that
are
currently
represented
as
spans
as
a
zero
duration
spans.
We
intend
to
move
them,
as
you
know,
log
events
and
which
type
of
events
you
said,
can
you
say
them
once
again.
A
So
user
interaction
I
think
yes,
but
document
load.
Can
you
explain
why,
like
as
in
the
document
load,
does
have
a
duration?
A
So
we
are
thinking
of
keeping
or
emitting
both
a
span
and
an
event.
The
idea
is
that
you
know
the
the
moment.
You
load
a
page.
You
you
want
to
emit
an
impression
event
that
hey
your
page
has
loaded,
but
you
don't
yet
have
any
of
the
timing
information
yet
and
we
we
let
the
page
load
completely
and
after
that
we
will
generate
a
span
with
with
the
duration,
a
field
set.
A
D
Awesome
yeah
that
does
make
sense
how
it's
kind
of,
because
those
signals
we're
just
kind
of
curious
as
where
working
on
some
instrumentation,
that
is,
you,
know,
heavily
moving
towards
using
logs
for
that
and
so
appreciate
the
context.
Sir.
Thank
you
yeah.
A
D
First
time
yeah,
so
we're
prototyping
an
agent
based
on
hotel
right
now.
Okay,.
A
Because
if
I
was
asking,
I
was
going
to
ask:
if
you
already
have
a
product,
then
you
know
you
might
want
to
include
your
current
schema
in
in
in
in
the
documentation
where
all
other
vendors
have
put
theirs.
That
way
we
can
combine.
We
can
look
at
you
know.
What's
common.
D
Yeah
I'll
pass
that,
along
to
the
to
the
team,
that's
working
on
that
we
do
have
some
kind
of
like
conventions,
but
because
it
is
so
early
on
kind
of
our
plan
is
to
to
go
along
with
whatever
this
sig
ends
up
kind
of
standardizing
on
shoulder
flexible.
On
that
end,
okay,.
A
I
have
a
question
for
those
who
joined
the
JS
Sig
I
I,
don't
I'm,
not
so
familiar
with
the
you
know
how
things
work
there.
I
have
a
question,
so
we
are
introducing
some
new
attributes
in
the
browser
detector
and
we
want
to
drop
some
of
the
attributors
that
are
already
there
and
it
turns
out
that
the
open
Telemetry
resource
package
is
already
stable
and-
and
so
we
cannot,
you
know,
remove
anything
and
the
suggestion
provided
in
the
pr
was
to
introduce
a
new
detector
and
Mark
the
current
one
as
deprecated.
A
But
we
don't
want
to
let
go
of
the
name.
Browser
detector
sounds
more
relevant
than
the
suggested
names,
so
in
general
there
has
to
be
a
way
to
phase
out
functionality,
or
you
know,
make
some
breaking
changes
even
in
stable
code,
especially
you
know.
If
we
are
able
to
plan
ahead,
does
anyone
know
if
there
is
any
precedence
to
doing
such
a
thing.
C
Yeah,
it's
normally
a
case
of
brand
new
version.
You
know
a
brand
new
instance
or
you
go
through
this
phase
of
having
both
and
then
you
have
to
make
a
major
version
change
to
say
that
you're
dropped
in
the
previous
one.
The
alternative
is,
you
have
a
config
that
would
have
to
default
to
the
way
it
currently
operates,
that
you
can
flip.
B
A
Okay,
yeah:
we
can
discuss
and
confirm
in
the
in
the
next
the
JS
signal.
A
And
there
is
one
more
FYI,
there
is
a.
A
For
for
the
ephemeral
resource
attributes,
I
I
see
there
might
be
a
potential
Solution,
that's
being
discussed
in
a
different
context.
E
Sure,
yeah
yeah
I
haven't
had
time
to
work
on
it
over
the
past
couple
weeks.
It's
definitely
high
up
on
my
list:
I
I'm
trying
to
balance
being
able
to
do
open,
Telemetry,
stuff
and
and
other
other
work
things,
but
yeah
I'm,
hoping
to
get
that
PR
into
a
reviewable
state
in
the
next
week
or
so.
If
anybody
like
wants
to
help
or
yeah
like
if
you
want
to
help
at
all
I'm
I'm
open
to
like
chatting
about
transferring
that
knowledge
and
getting
getting
other
folks
to
to
contribute
to
so.
E
I
need
to
make
sure
that
I
think
the
tests
are
failing
for
some
of
the
node
stuff,
so
I
need
to
verify
that,
like
I
haven't
broken
any
functionality
for
the
node
Proto
exporter
by
the
changes
I've
made
so
yeah
basically
need
to
test.
I
can
make
a
list
on
the
pr
for
I.
Think
I
have
a
list
up
there,
but
the
tests
are
failing,
so
I
need
to
fix
all
of
that
up
before
it
can
go
in.
B
Yeah
so
in
terms
of
like
the
current
support,
like
it's
not
already
working
with
protobuf
like
we
can,
can
not
release
three
spacing
in
protoba
correctly.
B
Okay,
so
this
one
is
just
bringing
touching
also
to
the
browser.
C
Yes,
I
feel
your
pain
with
it,
with
balancing
with
the
two
Kirby,
probably
on
that
the
sandbox
I
end
up
closing
the
previous
PR
for
merging
everything
in
because
the
JS
API
repo
is
now
merged
back
into
the
main
JS
repo.
C
So
I
have
a
PR
App
that
Martin's
already
signed
off
for
making
it
the
the
contrib
repo
and
the
jspo
mergy
into
the
sandbox
I
want
to
get
Daniel
Smith
up
on
that.
Once
that
happens,
I'll
run
my
script
again
to
merge
all
the
history
into
the
new
Sandbox
and
once
that's
done,
I
don't
have
to
find
the
same
time
as
Hershey
was
talking
about
to
massage
that.
C
C
I
was
hoping,
Ram
was
going
to
join
today,
but
it
looks
like
you
might
not
make
it
so
or
we've
continued
our
discussion.
We
had
yesterday
on
the
rum
events.
A
Yeah,
do
you
want
to
talk
about
it?
Nev
I
can
share
the
screen,
but
I
think
if
you
want
to
talk
through
it,
we
can
yeah.
C
I,
don't
really
have
any
more
I
didn't
think
with
ram
after
the
meeting
yesterday,
so
I
don't
really
have
any
more
context
but
yeah.
If
you
want
to
bring
it
up
because
people
in
this
video
that
weren't
in
that
one,
okay.
B
C
So,
while
Santosh
gets
that
ready
effectively,
Ram
is
managed
to
find
some
time
now,
he's
back
from
holidays
to
start
solidifying,
all
this
back
together.
C
As
part
of
that
he's
created
a
spreadsheet
which
he's
linked
off
in
the
doc
to
try
and
identify
what
the
common
attributes
are
going
to
be
yeah,
which
is
the
spreadsheet
here
so
along
the
bottom.
We've
got
I
think
the
set
of
events
or
states
in
the
case
of
the
span
ones
that
we
want
to
try
and
Tackle.
C
First,
where
the
end
goal
is
that
First
Column
a
is
what
we
want
to
establish
and
say
well
for
in
this
particular
case,
the
page
view
event:
what
are
the
fields
that
otel
is
going
to
mandate
and
as
part
of
a
collecting
that
we've
got
all
the
companies
across
in
each
other
column
to
say?
Well,
this
is
what
you
call
it
and
they're
aligned.
C
So
in
the
case
of
row,
six
Cisco
call
it
page
referral
where
Microsoft
calls
it
ref,
URI
I,
think
Ram's
missed
a
few
others,
so
he
effectively
harvested
all
this
information
from
the
word
doc.
So
the
word
doc
is
still
the
the
initial
place
of
establishing
where
everything
is.
C
So,
if
you
put
your
name
down
in
in
the
list,
RAM
will
probably
reach
out
to
you
on
Slack
to
try
and
get
some
information,
especially
for
the
columns
that
are
currently
missing
in
this
initial
table.
C
If
there's
any
particular
events,
you
would
like
to
have
promoted
I
guess
these
are
the
ones
we
came
up
with
in
our
Tuesday
meeting
the
reason
for
the
page
view
and
the
page
view
span
is
while
the
pageview
event
will
go
off,
the
span
will
be
a
grouping
mechanism
to
group,
a
bunch
of
other
stuff
together
very,
very
closely
related
Some
People
Want
it
some
people
won't.
C
This
does
not
include
or
hasn't
defined
at
this
point,
whether
we'd
we
have
this
in
the
event
data,
what
the
shape
of
the
event
will
look
like
or
whether
we'll
just
reuse
existing
attributes,
like
the
HTTP
attributes,
I
think,
is
like
the
referral
and
the
URL
we're
not
yet
at
that
point
we're
just
saying:
well,
what
are
the
common
set
of
data
that
we
want
to
want
to
send
over,
and
how
do
we
want
to
send
it.
C
It's
good
to
be
funny
starting
to
get
to
that
point.
I
I
had
planned
on
getting
there
like
six
months
ago,
but
I
just
haven't
been
able
to
balance
internal
on
Hotel
work
to
the
extent
that
I
would
have
liked.
A
Okay,
we
can
end
here:
okay,.