►
From YouTube: 2022-08-02 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
A
A
A
A
A
Yeah,
it
makes
sense,
though,
I
mean
that
time
is
kind
of
annoying
bow
for
you
to
ask
a
little
bit
for
me
with
the
impact
representation
gone
through
the
APAC
meetings,
yeah.
B
Yeah
just
like
that,
just
like
that
yeah
we
already
canceled
on
when
honor
told
us
on
Thursday
we
already
canceled
that
meeting
and
the
Consolidated
it
into
the
John
said
he
would
start
coming
to
the
Thursday
morning.
Meeting.
A
B
And
we
also
canceled
the
the
jvm
metrics
well,
half
of
the
jvm
metrics
meeting
I
was
supposed
to
Ping
Tommy
plaque
to
see
if
he
wanted
to
keep
the
the
APAC
every
other.
B
Yeah,
we
were
well,
we
merged
it
into
the
Thursday
meeting.
We
just
don't
have
enough
topic.
We
don't
need
as
many
meetings
anymore
these
days,
we're
just
not
as
robustly
active
as
we
once
were,
that
could
change
in
maybe
after
summer,
but
we'll
see
we're
all
we're
also
losing
people
whether
it
was
Nikita
you
John,
more
or
less
foreign.
B
B
A
B
Yeah
I
had
recommended
to
jack
on
Thursday
to
like
trying
to
take
the
new
projects
like
say,
like
a
profiling
and
really
try
to
advertise
that
for
somebody
new,
because
again
it's
a
lot
easier
to
get.
You
know
some
vendor
to
yeah
Pony
up
somebody
to
work
on
a
thing,
as
opposed
to
just
maintaining
the
project,
doing
code
code
reviews.
B
Yeah
another
one
of
the
projects
upcoming
that
I
I
think
will
be
able
to
get
like.
A
B
John's
time
is
will
potentially
profiling,
but
also
the
static
instrumentation
stuff.
Do.
B
Yeah
yeah
I'm
very
I,
I'm,
very
keen
on
that,
partly
because
I'm
we
get
a
lot
of
complaints
about
startup
overhead
in.
B
B
A
I,
don't
know
what
his
feelings
on
that
are,
but
his
time
is
very
close
to
zero,
which
is
the
same
for
Rana,
unfortunately,
so
I'm
pretty
sure
they
would
want
the
thing
to
be
continued
somehow
by
somebody
and
might
even
have
some
ideas,
for
you
know
things
things
that
should
be
done
next
or
what's
what
should
be
fixed
or
improved
in
the
existing
static
instrumenter?
A
So
it
might
be
worth
it
to
ask
them
for
some
opinion
other
than
that,
but
I
don't
think
that
they
will
actively
contribute
to
any
anything
in
the
near
future.
A
So
it's
it's
really
great
that
if
Jean
would
be
able
to
do
something.
B
Yeah
he
just
went
on
paternity
leave,
but
when
he's
back
in
a
couple
weeks
because
he's
splitting
his
paternity
leave,
we'll
probably
bring
that
up
as
a
topic.
A
B
A
couple
months
ago
joined
from
he's
in
France,
oh
okay.
Unfortunately,
no
well,
unfortunately,
unfortunately
I
don't
know
the
the
APAC
time
zone
is
really
tricky.
I
can
see
why
I
remember
when
the
project
when
you
started,
joined
the
project,
and
you
expressed
that,
like
projects,
don't
tend
to
ever
prioritize
or
have
any
kind
of
APAC.
A
B
A
A
B
Yeah
but
I
did
research
once
there's
even
one
that
is
like
a
15
minute,
timer
I,
don't
think
anybody
lives
there
that
I've
seen
but
or
no
it's
doing
some
time
zone
math
and.
B
Yeah
that
the
config
like
through
a
big
old
wrinkle
at.
A
But,
on
the
other
hand,
it's
good
that
we
got
this
very
straight
and
we
have
some
kind
of
approach.
That
is,
that
makes
sense
overall,
even
when
you
combine
the
SDK
and
the
instrumentation
Ripper
yeah.
B
A
Yeah,
it
probably
needs
a
little
bit
more
of
work
like,
for
example,
some
of
the
matchers
that
we
expose
are
like
implements
method
or
something
that
we
shouldn't
really
exposed
as
a
tight
matter
and
to
be
released.
So
there
might
be
a
couple
of
things
that
we
should
remove
from
there,
but
it's
also
a
little
really
good.
B
B
B
A
A
Okay,
yeah
it's
and
it's
by
the
way,
it's
very
similar
to
messaging
operation,
where
we
also
don't
set
it
when
it's
send.
A
A
Oh
okay,
I
understand,
so
you
can
still
set
initial
okay
yeah,
so
I
got
confused
by
it.
You're.
B
And
yeah
this
one
is
yes,
thank
you
for.
Oh,
let
me
make
sure
I
did
what
you
were
thinking
for
the
backward
compatibility.
B
Yeah
yeah
and
then
so
this
one
I
had
to
add
a
default,
but
then
I'll
remove
the
default
once
we.
B
A
Yeah
so
it's
kind
of
a
connected
topic,
but
I
mentioned
last
time:
I
started
looking
at
the
HTTP,
all
the
HTTP,
symmetic
conversion
changes
and
there's.
A
A
Just
just
for
the
net
attributes,
but
when
you
look
at
the
like
some
HTTP
simcon,
it
obviously
includes
some
of
the
net
attributes,
but
it
has
some
different
conditions.
It
says
a
little
bit
different
things
about
how
they
should
be
filled
in,
for
example,
like
the
net
so
host
address
or
not
appear.
A
Yeah,
so
when
you
look
at
the
comments,
okay,
you
scroll
a
little
bit
down
yeah.
So
the
for
example.
The
number
three
which
is
like
the
host
name
should
be
determined
by
the
primary
server
name
or
the
request
Target
or
a
post
and
like
none
of
these
can
be
really
described
by
the
net
attributes.
Scatter
I
mean,
and
it
I
started
thinking
that
it
might
make
sense
to
sort
of
merge.
A
The
HTTP
gather
within
that
color,
and
just
so,
and
only
and
extract
these
net
attributes
in
the
HTTP
attribute
extractor,
because
the
HTTP
attribute
extractors
has
knowledge
about
the
headers,
the
URL,
especially
in
case
of
the
clients,
because
the
servers
don't
have
the
full
URL
and
the
other
individual
server
name
is
another
thing.
But
so
I
came
up
with
the
idea
that
maybe
it
would
make
sense
to
make
the
http.
A
A
Which
again
doesn't
really
complete
with
the
changes
in
this
PR,
but
well
we're
going
to
have
breaking
changes
anyway,
which
also
connects
to
another
topic,
which
is
the
incubator
module
which
so
what
I'm
saying
that
we
should
probably
do
the
work
connected
with
moving
to
the
new
incubator,
module
really.
B
B
But
certainly
we
could
have
an
incubator
for
I,
mean
and
I
think
that
oh
right
right
and
so
the
thought
was
we
could
do
that
as
sort
of
the
first.
The
1.0
for,
like
DB
semantic
convention,
would
get
to
own
that
package,
name
in
the
incubator
module
so
that
it
could
become
stable,
1.0.
A
Yeah
but
in
case
of
I,
mean
I
did
the
same
kind
of
are
incubator
module.
However,
you
name,
it
is
still
Alpha
right,
so
we
are
allowed
breaking
changes,
but
are
only
fine
with
for
anyone
from
the
HTTP
convention
is
breaking
for
everybody.
Once
we
I
mean
everybody
mean
probably
on
this
pink
suit
right
now,
but
one
once
we
switch
over
to
the
new
conventions
that
I
use
I'm,
pretty
sure
that
there
will
be
some
at
least
some
changes
that
are
not
quite
compatible.
B
Oh
yeah,
so
it
wasn't
break
it.
Wasn't
the
time
of
the
concern
wasn't
about
breaking
by
changing
the
package
name
now,
but
changing
the
package
name
later.
B
Basically,
how
do
we
go
from
RC
to
stable?
Then?
B
B
B
Wouldn't
require
was,
it
was
a
smooth
transition
from
RC
to
stable,
I
can't
remember
what
the
complication
was
at.
One
point
were.
A
I,
remember
that
with
the
metrics
API,
there
used
to
be
a
separated,
I'm,
not
sure
about
the
SDK,
but
the
API
used
to
be
used
in
a
separate
module
with
the
same
package
name.
So
it
the
only
thing
that
broke
was
the
like
reference
and
greater
laws
from
files
to
the
to
the
old
module.
B
And
then
there
was
the
the
Java
9
module
issue
would
then
be
sort
of
solved
by
that
approach
by
the
semcon
packages.
I.
B
A
It
it
makes
sense,
but
there
are
still
some
classes
that
are
in
the
comment
package
like
I,
think
the
peer
attributes,
extractor,
is
for
some
reason
and
instrumental
package
instead
of
the
net
package
that
yeah,
we
just
have
to
move
it
somewhere
else.
A
Wasn't
thinking
about
I?
Wasn't
thinking
really
about
the
release,
but
my
idea
about
this
was
the
like:
replicate
the
whole
sentence,
module
and
basically
create
a
copy
of
it
called
incubator.
You
know
with
the
change
package,
but
yeah
I
agree
that
it's
a
lot
more
work
and
probably
the
the
thing
that
Lori
came
up
with
his
it's
enough.
We
don't
really
have
to
do
that.
So.
A
Yeah,
but
anyway,
this,
what
bloody
purpose
should
be
were
probably
much
easier
to
implement
and.
A
There's
this
parameters,
and
now
by
default
it's
also
instrumentation
API.
We
like
that
I
just
basically
did
the
same
thing
that
the
SDK
does
everywhere,
which
is
requiring
on
now
and
package
info.
A
A
I
also
have
the
ingredients
here
open,
but
it
fails
with
a
thing
that
I
completely
do
not
understand
and
I'll
probably
wait
for
the
Lord
and
he
comes
back.
A
Yeah,
it's
very
interesting
because
it
fails
on
spring
season
and
if,
if
you
scroll
up,
perhaps
a
little
bit,
it
says
that
it
can't.
Finally,
it
cannot
see
the
defined
class
method
in
the
class
loader
because
of
some
module
visibility,
restrictions
which
is
really
weird
because
it
all
passed
before
and
just
started
failing
when
I
updated.
You
know
just
Gradle
radar,
yeah,
greater.
A
Just
like
Griffin
New,
Deals
or
whatever
it
we
don't
even
modify
them
in
the
in
the
agent.
So,
even
if
you
were
to
run
this
test
without
the
legion,
it
would
probably
still
fire
you.
Oh
perhaps,
unless
the
agent
does
something
really
weird.
B
B
A
I
have
no
idea,
I
wouldn't
be
I
even
checked.
If
maybe
there
was
some
kind
of
a
bug
and
we
weren't
really
running
Java
17
before,
but
now
we
actually
were
so
I
had
no
idea
what
has
changed,
but
it's
it's
one
of
those
funnier
errors
that
you
just
don't
understand.
B
B
Yeah,
let
us
know
if
you
do
any
what
you
end
up
doing
next,
if
you
end
up
back
in
the
Java
world,
so
we
can,
we
can
harass
you
with
more
Gradle
questions
like
not
fair,
to
ask
you
cradle
questions
once
you've
like
turned
that
part
of
your
brain
off.