►
From YouTube: 2022-08-01 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
B
B
C
I
think
we
can
quickly
start
with
the
agenda.
I
wanted
to
share
about
the
I
want
to
talk
about
the
microsoft
about
the
sorry,
not
microsoft,
but
matrix
status.
C
Okay:
okay,
we
don't
want
to
talk
about
the
header,
only
singleton,
right
or
just
in
case
mark
to
feel
that
something
we
want
to
discuss
here.
C
C
C
So
for
this
pr,
multiple
async
callbacks,
this
one,
I
think,
is
on
review.
I
saw
asylum
and
some
review
comments.
Add
configuration
options.
This
I
already
reviewed.
It
looks
good
and
add
in
memory
matrix
exporter.
This
is
in
review.
I'm
still
not
very
convinced
whether
we
should
have
a
dependency
on
otlp
for
this.
C
So
probably,
if
required,
we
can
redo
it.
I
can
take
it
up
once
this
async
callback
is
done.
We
we
don't.
I
mean
if
it's
feasible
and
easy
to
do
it
without
having
a
dependency
on
otlp
yeah,
so
the
rest
default
metric
view.
I
haven't
seen
that
I
think
I
thought
you
have
done
some.
B
C
C
C
C
Yeah
example,
I
mean,
if
it
is
not
done,
I
think
we
should
be
okay
to
move
it
post
rc,
that's
something
something
which
we
can
add
it
even
even
before
we
reach
to
the
ga,
that's
not
something
which
is
at
least,
that
example
should
not
be
affecting
the
core
matrix
aggregation
feature.
So
we
even
if
we
add
it
afterwards,
that
should
not
have
a
major
impact,
but
I'm
not
sure
how
much
how
much
effort
is
required
for
this.
I
think
you
can.
C
Even
you
will
be
knowing
better
than
what
we
know,
but
I
know
about
this,
so
I
mean
we
can
take
a
call
probably
next
week
if
this
is
something
which
is
a
major
work
and
we
cannot
really
achieve
it
before
it
release
candidate.
We
can
move
it
post
release
candidate
and
we
can
do
it
before
ga
and
the
rest
should
be
a
minor
work.
I
think
probably
we
should
be
good
to
take
it
before
doing
it
before
working
on
a
release,
candidate
and
yeah,
so
that
that's
it
that's
a
quick
overview.
C
Let's
try
to
target
it
by
the
date
of
15th
august.
I
mean
if
we
can
do
that.
That
would
be
good
enough
for
us
yeah.
C
Unless
somebody
has
any
concerns,
we
do
need
to
do
some
benchmarking
on
the
current
implementation.
Probably
I
thought
that
we
can
take
it
up
once
we
have
all
the
feature:
implementation
done
as
part
of
release
candidate
before
doing
a
ga.
We
should
have
a
proper
benchmarking
for
the
aggregation
I
mean
for
each
of
those
aggregations
where
we
exactly
are
in
that,
but
as
of
now,
that's
not
part
of
release
candidate,
we
may
have
to
do
some
optimizations
if
we
feel
that
well,
probably
it's
not.
C
C
C
C
A
Yes,
so
what
I
did
is
some
investigation
to
try
to
understand.
What's
going
on
there,
I
was
able
to
find
something
that
works
with
gcc
by
doing
exports
with
with
attributes
to
declare
explicitly
that
things
should
be
exported
and
as
far
as
eyes
understand
it
works
for
gcc.
A
B
A
Also
worked
with
windows
so
for
windows.
I
used
something
different
to
okay,
respond
with
cmos.
C
C
It
is
not
working
too
much
for
what
he
has
said,
but
yeah
probably
need
to
test
need
to
test.
Once
with
this,
I
did
test
your
changes
on
linux,
as
I
mentioned
here,
and
I
think
it
looks,
looks
good
for
at
least
linux
I
mean
okay,
I'm
not
sure
if
c
bank
should
be
the
same
api.
As
for
linux,
I
mean,
if
cba
on,
if
you're
using
c
sorry
favoring
c
link
on
mac,
it
should
be
using
elf
right
from
it.
C
A
C
Yes,
that's
what
I
was
even
thinking.
We
can
add
it
as
part
of
our
ci,
somehow
somehow
we
can
test
it.
Yes,
that
would
be
good
to
catch
any
errors.
Yes,.
A
A
I
found
this
select
any
based
on
some
research
looking
at
on
the
web,
but
I
have
first
of
all,
I
don't
have
enough
even
a
windows
platform
to
put
tests-
and
I
don't
know.
B
A
C
C
C
Yeah,
that's
what
I
was
thinking,
even
if
you
can
scope
it
scope,
the
current
change
to
linux,
at
least
with
the
shared
library,
with
visibility
as
hidden.
If
that
works,
I
think
we
can
actually
have
should
be
good
to
at
least
fix
fix
the
issue
for
set
of
users
who
are
using
linux.
A
So
yes,
I
can.
I
can
scope
that
with
for
linux
and
mac
os
to
start
with,
and
when
later,
when
we
have
the
the
proper
export
for
this
decay
and
on
windows,
we
can
see
how
to
make
that
work
on
windows.
Yes,
the
good
news
is
that
this
select
anything
seems
to
work.
So
very
hopefully
I
guess.
A
Okay,
just
some
some
heads
up,
I
will
I
will
be
away
for
two
weeks,
so
I
will
probably
work
back
work
on
that
when
I
come
back.
A
Oh
I'm,
starting
in
two
days
from
now,
and
I
will
be
back
so
I
will
basically
miss
two
meetings.
C
C
C
In
our
project
also,
so
if
you
want
to
be
a
member,
I
think
you
can
so
these
are
the
basic
requirements
for
the
member
and,
I
think,
mostly,
should
be
met.
There
should
not
be
any
problem
in
meeting
it.
You
have
to
raise
one
pr
on
this
repo
on
the
community,
repo,
okay,
and
that
should
mention
it.
It
says
everything
says
here
that
you
have
to
mention
whatever
contributions
you
have
done
in
that
pr
and
then
that
pr
should
be
sponsored
by
any
two
approvers,
so
probably
any
anyone,
any
two
of
us
can
approve.
C
C
A
So
it's
on
community.
I
will
take
a
look.
C
Yes,
like
request
new
membership,
so
just
like
list
of
contributions
to
open
telemetry
project,
and
you
can
mention
in
any
two
sponsors
either
me
or
tom,
as
could
be
one
of
the
sponsors.
One
of
us
could
be
the
sponsors
because
the
two
sponsors
should
be
from
different
companies,
so
I
think
we
can
keep
either
me
or
tom,
as
one
of
them
and
assange
should
be
good
as
the
another
other
one.
A
Okay
sounds
good,
but
by
the
way
asanois-
and
I
don't
know
you
yeah.
B
A
Hey
son,
do
you
work
for
a
company
also,
or
is
it
just.
B
A
B
B
C
Oh,
is
that
with
open,
telemetry
membership.
C
B
C
A
So
should
I
mention
lolita
and
azan,
then
yeah.
B
A
C
B
A
C
C
C
C
If
that's
possible
right
now
what's
happening,
is
that
meter
contacts
maintain
the
set
set
of
all
the
meters
which
are
created
and
the
meter
is
using
meter
context
to
get
all
the
configuration
information
and
in
tracer
in
tracer
at
least
for
for
trace
ap
sdk.
We
all
these
tracers,
which
are
created,
are
maintained
in
tracer
provider,
and
I
think
I
remember
there
were
some
issues
because
of
that
I
moved
moved
to
the
meters
from
meter
provided
to
meter
context.
C
C
C
B
C
B
We
also
have
a
race
in
one
of
benchmark
tests:
okay
and
attribute
hashmap
benchmark
test.
C
C
C
C
C
Where
should
I
go
in
matrix
and.
B
Okay,
okay,
so
you
see
this
alarm
dies
accessing
this
hash
map
line
23.
It's
just.
A
C
C
B
Is
it
supposed
to
be
thread
safe
supposed
to
be
okay,
I'll
check
that
so.
C
C
B
B
B
A
I've
seen
something
like
that
when
I
created
my
own
repo
of
open
telemetry,
it's
basically.
What
I
got
is
an
audit
saying:
well,
you
have
some
code
somewhere
doing
http
and
not
https.
Therefore,
it's
an
audit
violation
of
some
sort.
A
B
A
C
B
C
C
C
C
C
This
is-
and
this
is
ongoing,
I
think
still
still
the
test
unit
tests
are
has
to
be
changed.
So
the
matrix,
positive
exporter
is
modified
by
adding
resources,
but
the
unit
test
has
not
been
changed.
So
I'll.
Just
comment
on
that
to
change
the
unit
tests
also
and
yeah
this
one,
I
think,
you'll
take
care
of
it.
C
C
I
don't
see
anything
decode
values
for
hotel
resource
attributes.
This
is
some
changes
in
the
way
for
the
package
baggage.
The
way
the
attributes
are
getting
the
format
of
the
baggage
attributes
using
the
confirmations.
I
think
that's
probably
that
requires
some
changes
in
this
text,
so
we
may
have
to
modify
our
code
for
that
yeah.
I
think
the
part
is
not
merged,
so
this
should
be
good.
You
know
this
is
the
pr
and
see
that
is
not
yet
it's
approved,
so
I
think
hopefully
it
could
be
merged.
C
There
has
been
some
discussion
going
on
just
just
to
update,
probably
that
in
in
the
specification,
any
change
in
the
specification.
If
that
needs
a
change
in
the
six,
I
think
there
would
be
auto
generation
of
the
issue
for
all
the
all
the
language
sets
for
any
specs
changes
so
that
somehow
they
are
planning
to
do
auto
generation.
It's
just
on
discussion,
I'm
not
sure
if
that
will
get
materialized,
but
if
that
happens,
that
this
kind
of
manual
creation
of
issues
will
not
be
required.
C
C
C
Yeah
and
see
you
mark
probably
after
the
vacation
have
a
good
vacations.
A
B
A
So
I
will
do
a
paperwork
to
to
join
and
otherwise
I
will
be
back
in
two
weeks.