►
From YouTube: 2022-08-15 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
B
You
all
good
yeah.
How
are
you
doing
okay,
yeah,
since
five
minutes.
A
A
A
Yeah
so
for
release
candidate.
I
think
we
are
in
good
shape.
I
mean
just
let
me
know
if
you
think
otherwise
I
mean
I
see
there
are
lots
there.
There
are
a
few
memory,
leaks
and
race
condition,
issues
which
you
have
already
fixed.
B
B
B
B
A
A
B
B
B
And
we
confirm
this
version
number
will
be
1.0.1.6.0.
A
I
think
we
can
start
focusing
on
the
ga
for
this
going
forward.
A
So
I'll
spend
some
time
to
create
the
issues
for
the
ga
I
mean
I
I
plan
to
go
through
the
matrix
specs.
I
know
there
have
been
some
changes
in
the
matrix
specs
in
recent
times
and
probably
need
to
create
the
issues
for
that
but
feel
free
to
any
any
gap.
You
feel
I
mean
feel
free
to
add
it
on
the
ga
milestone
already
the
game
and
zone.
Is
there
so
probably
just
start,
adding
it
to
the
milestone
right
now.
A
A
A
Based
on
the
feedback
which
we
have
got
from
him,
it's
not
very
it's
it's
a
bit
complicated
and
even
I
think
specs
is
not
yet
stable.
For
that
I
mean
I
don't
think
specs
is
completely.
It's
has
identified
all
the
all
the
things
which
we
need
to
be
there
for
implementation,
so,
based
on
that
feedback
feedback,
I
think
we
can
discuss
if
we
want
to
keep
it
in
ga
or
not.
B
A
Hey
anything
else
we
want
to
add
before
we
discuss
the
other
points,
anything
you
feel
it
should
be
there.
Otherwise,
you
can
start
discussing.
A
A
We
have
all
patrick's
obsolete,
so
we'll
create
an
issue
for
this
and
we
will
target
it
for
the
for
the
ga.
So
I
think
it's
good
to
do
a
proper
cleanup
before
ga
and
I
think
we
can
create
once
once
the
release
candidate
is
done.
We
can
create.
We
can
update
our
documentation
to
to
clearly
mention
that
this
is
dedicated
and
it
would
be
cleaned
up
while
we
go
for
the
g8
for
matrix.
A
A
B
A
We
can
do
some
macro
updates.
Also,
if
you
haven't,
if
you
want,
we
can
do
the
update.
For
the
I
mean
we
can
create
a
macro
for
the
depletion
and
probably
update
the
the
current
implementation,
the
old
implementation.
With
that
not
sure,
if
you
should,
we
should
do
it,
but
at
least
at
least
to
start
with.
We
can
at
least
document
that
yeah
yeah
and
probably
probably
in
the
slack
channel,
we're
gonna.
We
can
announce
it,
which
should
be
enough
to
start
with.
B
B
A
A
A
B
A
And
I
think
discussing
continuing
discussion
with
this
etw
exporter.
We
got
some
issue
of
use
after
free.
That
was
more
of
internal
to
microsoft,
but
I
think
this
is
something
as
the
as
that
code
is
in
public
repo.
I
think
probably
we
can
discuss
that.
So
I
think
tom.
You
was
also
a
part
of
that
made
discussion
right,
yeah.
A
A
A
A
A
It's
if
we
go
to
trace
so
we
have
a
tracer
provider.
A
A
B
B
B
A
A
A
So
now
tracer
will,
what
tracer
will
do
tracer
will
have
with
con
yeah,
so
tracer
tracer
maintains
yeah,
so
yeah
yeah
this
one.
So
instead
of
maintaining
a
reference
to
tracer
provider,
tracer
will
have
a
shared
context,
shared
pointer
to
the
tracer
context,
the
same
way
as
we
are
doing
it
in
the
open,
telemetry,
sdk
and
now.
So
what
it
means
is
that
racer
pro
tracer
will
not
be
dependent
on
tracer,
provided
now
I
mean
afterwards
after
tracer
tracer
is
created,
it
just
used.
It
has
to
use
tracer
context.
B
B
A
A
I
think
that's
all
we
have
header
only
api
for
singleton
is
something
which
we
are
waiting
for
mark
added
memory
exporter
is,
it
has
multiple
ci
failures,
probably
ideally,
we
should
not
have
dependency
to
otlp,
but
probably
as
of
now.
I
think
this
is
how
it
is
implemented
and
we
can
decide
afterwards
if
we
want
to
have
a
separate
implementation
without
opnp,
or
we
want
to
fix
this
pr
and
upgrade
to
grpc.
I
think,
let's
probably
start
looking
into
that
now,
once.
A
For
the
issues
otlp
http
live
is
not
built,
this
is
for
windows
and
it
is
not
being
built
in
windows.
I
know
I
know
I
was
checking
the
ci
in
the
ci.
Also
it
is
not
being
built,
I
think
so,
probably
I'll
look
into
once
again,
but
because
we
don't
install
kern
in
our
cio.
So
if
the
kernel
is
not
being
installed,
then
it
will
not
be
built,
but
I
think
looks
like
the
way
he
is
testing
this
guy
here
already
installed,
kern
and
still
otlp
is
not
being
built.
A
A
B
B
A
Your
metric
provider-
yes,
if
we
see
this
so
battery
reader
is
created
by
application
owner
and
then
they
have
to
pass
it
to
the
metric
to
the
meter
provider
and
meter
provider
is
going
to
call
this
metric
reader
based
on
will
will
call
dr
will
internally.
Then
it
assumes
that
this
metric
reader
is
a
is
a
periodic
pull
reader,
so
it
is
which
is
going
to
call
which
is
going
to
do
which
the
matrix
data
will
do
going
to
do
collect
internally
in
its
logic,
but
that
may
not
be
true
for
a
full
exporter.
A
A
So
this
needs
some
changes.
Probably
we
have
to
provide,
add,
pull
metric
reader
and
add
push
metric
reader,
two
different
apis
for
different
types
of
readers,
and
then
they
should
behave
according
to
the
requirement.
A
A
It
has
added
all
the
information
along
with
the
problem,
so
I
think
this
should
be
sufficient
and
just
go
through
this.
I
think
it
explains
everything.
B
A
A
A
Yeah
this,
we
already
discussed,
add
matrix
exemplar
in
example.
There
is
a
boy
this
is
already
there
in
the
ve,
so
I
build
this.
This
is
also
in
the
deal
set
is
provided
in
the
new
program
we
have
to
see.
There
is
a
check
fault.
I
think
this
probably
is
good
to
have
and
wanted
somebody
even
I
think
this
also
we
can
somebody
want
to
really
take
it
further.
A
A
Oh
nothing
important
in
support
upgrade
to
grpc.
Here
we
have
to
see
that
vr,
which
is
created
by
mark
this
one
feels
yeah.
I
think
this
is
a
sound
breathing
sound
whenever
you
have
time
nothing
about
nothing
very
urgent.
Whenever
you
get
time,
you
can
look
into
this
jagger
remote
sample
this
also
tom.
Whenever
you
have
time,
I
mean
again
nothing
very
urgent,
which
you
need
to
fix
more
implementation
from
sdk
header
points
yeah.
This
is
something
mark
already
has
with
him.
A
This
also,
I
think
we
have
to
reconcile
the
names
for
the
targets
for
bazel
and
c
magnet,
which
are
different
as
of
now
moving
last
year.
I
think
this
we
have
already
discussed
I've
just
created
this
to
track.
A
A
A
Yeah,
restaurant
old
ones,
I
think,
probably
nothing
important
which
we
want
to
discuss
immediately
now,
just
let
me
know
if
something
we
want
to
talk
in
the
issues.