►
From YouTube: 2022-06-14 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
B
B
Okay,
so
c
joe
wouldn't
be
joining
today,
so
he
had
asked
me
to
from
the
meeting.
There
is
no
agenda,
so
let
me
share
my
screen.
First
and.
B
So
yeah
so
you'll
see
my
screen
right.
B
B
So
we
recently
moved
stack
exchange,
redux
instrumentation
from
the
main
repo
to
the
contra
people,
and
we
haven't
had
a
stable
release
for
it.
Yet
we've
only
been
doing
pre-release
versions.
B
So
there's
this
pr
which,
like
on
the
contract
repo
that
came
in
where
so
this
pr
is
essentially
adding
an
option
to
not
have
activities.
You
know
not
have
activity,
events
added
to
the
activity
so
like
yeah,
the
the
reason
they
give
is
activity.
Events
are
also
built
like
individual
spans,
but
regardless
of
that
in
general,
so
we
in
our
instrumentation
we
in
which
we
actually
add
events
to
the
activity.
B
Now
a
better
thing
would
be
to
have
an
option
which,
when
set
to
true,
will
add
activities
instead
of
having
them
on
by
default
and
then
have
another
option
called
disable
them.
So
you
set
this
disable
option
to
true.
So
whenever
you
don't
want
it,
so
so
this
pr
essentially
like
initially,
we
were
just
adding
it,
regardless
of
any
like
because
we
didn't
have
any
options
so
by
default
it
will
get
added
to
the
events.
So
in
this
pr
there
is
he's
introducing
another.
C
B
B
If
that
is
set
to
true
only
then
we
add,
but
we
already
have
an
enrich
option
in
the
options
class.
So
this
enrich
has
all
the
information
that
all
the
data
that
one
would
need
to
be
able
to
do
this,
so
you
are
given
a
command
and
all
of
these
events
are
created
using
that
command
itself.
So,
if
you
look
here
like
send
and
cue,
all
of
this
is
like
all
you
need
is
that
command,
so
so
what
cj
was
proposing
was
that
we
get
rid
of
this
all
together.
B
B
B
B
Yeah
anyway,
I
think
this
doesn't
mean
that
we
cannot
just
do
it
through
enrich
so
asked
me
to
add
this
to
the
meeting
agenda
and
like
does
this
sound
okay
to
people
where
we
can
just
use
the
existing
enrich
option
and
add
these
events
if
we
want,
instead
of
always
adding
them
and
they're,
introducing
an
option
to
disable
that.
C
Logical,
it's
like
the
default
events.
I
mean
I
might
be
speaking
out
online,
but
it
seems
like
it's
similar
to
the
asp.net
core
startup
or
any
other
application
startup
hosting
the
connect.
Here's
all
the
default
stuff.
If
you
don't
like
the
default,
then
you
can
clear
it
add
your
own,
like
even
like
the
logging
provider,
say.
B
Now
she's
saying
we
should
always
have
it
by
default,
like
we
should,
like
the
the
instrumentation
library,
should
always
add
these
events.
C
Yeah,
it
didn't
sorry.
I
haven't
really
looked
at
this
before
enrich
activity
with
events
and
then
you
have
an
enrich
yeah.
E
B
So
yeah
I
mean
like
this
is
essentially
is
capable
of
enriching
with
events
tags
or
whatever
like
so.
If
someone
doesn't,
someone
still
wants
that
information,
but
wants
it
as
part
of
a
tag
because
like
in
this
in
their
this
customers,
like
this
person's
case
there,
the
issue
was
that
events
are
getting
billed
as
spans.
So
maybe
that
information
is
tagged.
They
can
still
do
that
using
language
like
now
like,
we
wouldn't
want
to
introduce
another
option
like
in
which
right.
C
Right
yeah,
we
are
exposing
a
lower
primitive
that
could
just
use
without
you
know,
adding
one
all
right.
F
C
E
C
I
have
a
strong
opinion.
It
does
seem
kind
of
duplicit,
like
the
lower
level,
is
exposed
next
to
something
that
adds
default
or
like
or
canned
events,
if
you
will,
because
they
got
to
be
canonical
right,
the
they're
just
using
strings
to
say
these
are
the
ones
right.
So
if
someone
were
to
discover
what
those
events
were
called,
I'm
again,
I'm
a
completely
outsider
looking
at
it,
I'm
not
that
close
to
the
code
like
how
would
I
know
what
events
they're
going
to
be
adding
when
you
enable
that
flag.
C
I
don't
see
that
even
the
method-
yeah
yeah.
B
So
I
think,
like
yeah,
if
you
look
at
cjos
yeah.
B
So
yeah
yeah
I
mean
the
spec
does
not
probably
does
not
have
he's
going
to
check
if
the
spec
has
an
issue,
but
I
don't
think
it
has
like
they
have
to
standardize
this.
And
if
the
spec
has
something
about
this,
then
we
can
follow
that,
but
otherwise,
like
he
said
it's
going
to
be
done
differently
in
different
implementations.
B
C
B
B
So
so
they
introduced
an
option
which
then
said
to
true
only
then
we
could
be
adding
events,
otherwise
we
won't,
but
then
I
think
cjo's
suggestion
is
that
we
can
get
rid
of
this
all
together
like
like,
let
not
add
any
events
at
all
in
the
library
and
if
the
user
wants
it,
they
can
have
this
logic
built
in
their
enrich
option
in
the
existing
enrich
option.
A
So
I
don't
know
why
those
are
even
there.
I
think
that
pre-exists,
even
when
I
started
contributing
to
open
telemetry,
I
had
problems,
so
I
don't
actually
use
the
redis
instrumentation
because
it
didn't
capture
enough
useful
stuff.
So
I
just
wrap
the
stack
exchange
client
with
my
own
stuff,
because
the
information
that
was
really
pertinent
to
me
was
like.
Was
it
a
cash
hit
or
a
miss?
B
Okay,
so
just
like,
like
keeping
the
need
of
events
aside
like
if
we
can
just
have
the
user,
do
it
like
you
know,
we
give
the
flexibility
to
let
the
user
do
whatever
to
the
activity
like
add
events
and
tags
with
the
enrich
method,
so
would
wouldn't
that
be
better
instead
of
like
adding
an
option
which
then
does
something
conditionally,
and
then
we
also
have
to
document
what
events
are?
Is
this
adding?
B
Because
this
is
achievable
through
the
existing
option?.
A
B
A
G
B
D
D
Paulo
that
wrote
the
original
instrumentation
for
redis
kind
of
reaching
out
to
him
and
whether
he
would
have
any
insight.
A
Okay,
yeah
has
anyone
checked
any
of
the
other
languages.
B
Okay,
I
think
I'll
do
that
I'll,
like
I
think
you
can
reach
out
to
paulo
as
well
as
check
the
other
languages
instrumentation
for
latest.
D
B
B
Cool,
so
did
we
have
any
other
things
to
discuss.
C
Doesn't
sound
like
it,
although
it
sounds
like
the
conversation,
isn't
three
camps,
it's
like
one.
Is
this
the
right
method
to
do
to
add
to
utilize
the
options,
object
and
configure
this
second
thing
I
heard
was
who's
doing
it
that
way,
it's
the
standard,
this
really
the
spec,
doesn't
say
anything
about
this,
and
the
third
thing
is:
why
is
it
even
there
so
by
next
by
this
time?
C
I
guess
it's
in
the
evening.
The
next
time
this
group
meets
we'll,
have
those
three
answers.
Perhaps.
B
Thanks
thanks
for
summarizing,
that's
actually
a
very
good
summary.
I
think
we
can
start
the
other
order
like
we
can
change
the
order
like.
Why
is
it
even
needed
and
then,
if
it
is
needed,
then
we'll
have
to
look
at
the
other
two
things
like:
how
are
the
other
languages
doing
it
and
then
do
we
need
to
expose
a
new
option
for
not
doing
it
or
just
let
them
use
the
existing
enrich
option.
B
Cool,
I
guess
we
can
end
the
meeting
then.
A
A
D
Yeah,
that's
a
good
call
out
there.
There
are
filtering,
is
actually
kind
of
an
interesting
use
case.
There's
the
specification
actually
has
a.
I
recall
a
sentence
in
there
somewhere
that
in
the
span
process
or
part
of
the
specification
that
suggests
like
oh
yeah,
like
filtering,
is
one
use
case
that
is
a
span
processor
is
well
suited
to,
and
I
actually
don't
know
that
a
processor
is
quite
the
right
tool,
at
least
how
they're
currently
designed
to
do
filtering.
So
then
it
kind
of
raises
the
question
of
like
what
is
the.
D
D
But
to
my
knowledge,
I
don't
think
a
processor
can
necessarily
in
the
middle
of
the
pipeline
say
like
nope.
Not
this
span
anymore.
A
A
D
D
D
The
example
has
some
problems,
and
then
it
doesn't
invoke
certain
things.
Like
he's
trying
to
wrap,
I
think
he's
trying
to
create
a
processor
that
wraps
the
the
simple
activity
export
processor,
but
in
that
wrapping
it
causes
the
initialization
of
the
processor
to
not
be
to
not
work
correctly.
So,
like
I
think,
studying
the
parent
provider
makes
certain
that
things
like
resource
attributes
and
so
on,
get
properly
applied
at
export,
but
through
this
setup
that
does
not
happen.
D
Which
is
then
kind
of
led
me
down.
This
train
of
thought
like
processor,
is
not
actually
how
you
want
to
do
this
anyways,
and
then
I
you
know
we're
talking
about
instrumentation.
I
don't
know
whether
instrumentation
is
necessarily
the
best
solution,
either.
B
B
Wait
so
I
think
I'll
just
write
this
for
now
that
maybe
in
the
next
meetings
next
few
meetings,
we
should
on.
B
B
D
I
can
help
help
with
some
of
that
I
can.
If
we're
curious
about
that
question,
I
can
ask
my
colleague
who
works
on
the
java
project.
They
can
talk
and
talk
to
him
about
what
java
has
and
doesn't
have.
B
Yeah
yeah,
I
think
that
would
be.
That
would
be
good
to
know
because
I
think
filter
is
still
like.
Somehow
there
are
workarounds
using
processor
to
filter
out
stuff
in
like
we
can
even
set
those
recorded
flags
to
fall
like
in
such
a
way
that
it
won't
get
exported.
But,
like
yeah
like
how
michael
said
for
enrich
and
like
in
asp.net
and
http,
we
do
get
hold
of
the
payload
and
the
request
body
and
whatnot.
So
I
don't
know
how.
E
B
G
B
E
D
No
yeah,
this
is
a
good
topic.
I.