►
From YouTube: 2022-09-12 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
A
B
Yeah
thanks
thanks
for
joining,
let
me
check
if
tom
is
going
to
join
and
we
just
bring
in.
A
Okay,
just
to
say
that
I
have
a
cold
tonight,
so
I
hope
you
can
hear
my
voice.
Okay,
so.
A
B
Okay,
I
think,
let's
start
we
can
start
with
appear.
I
don't.
I
haven't
recently
done
the
tracking
for
the
specs
changes
from
last
week.
Probably
we
can
discuss
it
next
week
in
case
in
case
somebody
want
to
really
discuss
one
of
these
specs.
We
should
use
over
here
this.
We
add
it
here.
Otherwise,
I
think
it's
discussing
next
week.
B
I'm
just
going
from
top
to
bottom.
Hopefully,
that's
fine.
Okay
at
the
ci
test
for
with
grpc
latest
release,
hey
thanks
astronaut.
I
think
this
looks
good
in
general
to
me.
I
just
had
a
few
comments
here.
As
of
now,
we
are
adding
a
new
ci
test
for
the
latest
grpc.
B
B
A
B
Okay,
that
that's
for
bazel,
right,
yeah,.
B
Yeah
I
mean
I
was
thinking
like
if
going
ahead,
I
mean
our
mainstream
ci
pipeline
should
be
always
on
the
latest
grpc,
so
that
from
that
perspective
I
was
taking
the
first
point.
B
So
if
we,
if
at
least
for
a
non-gcc
4.8
for
cmake,
if
we
can
make
that
as
a
default
grpc
version,
so
that
will
allow
us
to
upgrade
to
the
latest
version
for
the
cmake
bazel,
I
think
probably
we
can
start
working
start
looking
into
it
separately.
I
think
there
are
multiple
issues
I
could
see
as
of
now
invasive.
B
I
don't
think
it
would
be
possible
to
do
it
coming
pretty
soon.
We
may
have
to
work
with
grpc
team,
totally
get
it
in
case.
There
is
an
issue
there
I
mean,
because
I
I
remember
it,
there
was
a
there
was
a
segmentation
part
on
the
c-line
compiler
or
something
while
working
with
this
version
with
the
grpc,
which
will
with
the
new
version
of
you,
not
the
latest,
but
the
new
version
of
grp,
which
means
that
that
may
be
something
to
do
with
either
grpc
or
bazel
or
somewhere.
B
B
Which
should
be
fine
just.
A
B
B
It
should
be
okay,
as
long
as
the
mainstream
is
c
plus
mainstream
is
the
latest
version
of
grpc.
A
B
Yeah,
I
think,
as
as
we
discussed,
I
think
this
this
one.
Basically,
I
think,
let's
track
it
separately,
I
think
it
won't
be
something
which
we
can
do
it.
A
B
B
B
A
B
B
B
Yeah
I
mean
I'm
not
putting
anyone
as
of
now
on
this.
Probably
let's
see
who's
your
gets
trying
to
start.
Looking
into
that,
I
was
able
to
reproduce
this
issue
locally.
So
probably
I
could
spend
some
time
I
didn't
get.
I
was
thinking
to
spend
some
time
last
week,
but
somehow
didn't
have
time
to
do
it,
but
let
me
see
if
I
can
get
some
time
to
test
bazel
with
the
latest
grpc.
B
A
B
B
Yeah
I
have
gone
through,
the
change
looks
good
to
me.
Basically,
I
think
we
are
using
just
to
summarize
we
are
using
this
open,
telemetry
api
singleton
macro
for
all
these
static,
local
static,
singletons
methods,
and
then
we
have
this
test
to
test
this
scenario
and
yeah.
I
think
then,
although
this
is
nicely
documented
here,.
A
So
I
simplified,
for
example,
to
only
show
how
to
do
it
with
gcc
and
and
you
know,
yeah
the.
B
A
Push
requests
which
is
still
working
for
us
for
windows
as
all
the
code
there
for
the
other
windows
case,
but
unfortunately
it
still
needs
some
work.
So.
A
B
B
Yeah
on
on
linux
and
mac-
right,
yes,
okay
and
we
don't
have
so
this.
This
is
nothing
to
do
with
windows,
any
compiler
on
windows.
B
B
B
B
A
B
B
B
B
So
it's
a
small
comment.
I
think
it
can
be
done
as
a
separate
pr,
because
I'm
already
working
on
that,
we
have
to
write
the
example
for
grpc
exporter
for
matrix.
So
I
can
include
it
in
that,
but
we
just
just
wait
for
over
interplay
before
you
can
budget
you
already
have
to
afterwards.
Nothing
should
be
good.
B
Yes,
I'm
just
waiting
for
this
owen
just
had
some
comment
which
whether
we
can
remove
save
the
linking
of
this
library
here,
I'm
just
waiting
for
the
the
earlier
pr
which
we
discussed
to
merge
and
then
I'll
incorporate
those
changes
here.
So
it's
actually
only.
B
B
B
B
So
this
was
discussion
today,
maintainers
hunting
and
I
think
the
documentation
team
is
going
to
come
up
with
some
more
guidelines
before
before
we
finalize
anything
because
people
have
raised
the
concern
that
in
case
we
are
going
to
add
our
user
documentation
in
open,
telemetry
dot
io.
How
are
we
going
to
maintain
the
versions
for
that?
A
B
This
is
already
we
are
discussed.
This
is
here.
Is
there?
Is
there
this
also
pr?
Is
there
basically
the
reason
why
it
pays
to
connect
one
that
it
is
using
the
older
version
of
collector,
which
does
not
have
matrix
photovoltaics
in
place?
So
just
as
part
of
my
pr,
I've
updated
the
readme
to
use
the
latest
version
of
collector,
and
then
it
is
able
to
connect
to
the
current
database.
B
B
B
B
B
B
That's
more
of
a
high
level
based
on
the
very
high
level
estimate
I
may
need
another
one
month.
A
Do
we
need
a
another
incremental
release
this
month
before
the
stable
release.
B
We
can
do
that
yeah,
let's,
let's
do
that,
because
anyway,
we'll
have
substantial
changes
coming
from
both
shared
library,
support
for
linux
and
mac
and
also
yeah.
The
changes
required
for
the
one
which
omit
is
working
on
the
issues
related
to
the
contacts
among
the
lock
locked
here
or
the
lp
exporter
and
the
trace
exporter.
B
B
B
B
B
B
Yeah,
this
is
something
good
first
issue
with
some.
We
just
need
to
see
the
logic.
What
we
have
got
from
how
to
really
do
the
aggregation
for
this
anyway,
you
can
it's
a
separate
component
as
any
new
histogram
aggregation
can
be
done
separately,
without
touching
the
actual
logic
for
delta
and
cumulatives.
B
B
B
B
So
this
remove
circular
differences
is
or
this
is
on
me.
I
just
need
some
time
to
really
think
about
the
design,
how
to
do
it
and
then
start
working
on
that
this
is
p1
here.
Definitely
cleanup
of
all
days
also
clean
up,
but
I
think
we
know
what
to
do
in
this
cleanup,
so
it
should
not
be.
It
would
be
a
big
big
code
change,
but
I
think
should
not
be.
B
B
Recording
my
interface
would
be
metrics,
but
just
a
very
initial
thought,
which
came
to
me
that,
instead
of
moving
the
data
across
from
from
the
storage
aggregations
to
all
the
way
to
exporter,
if
you
can
use
a
recordable
interface
but
yeah
it's
on
priority,
p2
api
documentation
is
priority.
Given
you
have
to
take
it
getting
shutter
documentation
again
same
as
that,
this
is
would
be
fixed
with
the
pr
which
will
be
raised.
B
B
But
just
let
me
know
if
you
guys
feel
this
is
something
that's
pretty
good,
definitely
for
testing
purpose,
or
something
should
be
definitely
there
that
we
can
make
a
priority.
I
just
felt
that
this
is
nothing
to
do
with
the
actual
the
actual
pipeline
of
aggregation
and
export
something
separate,
so
you
can
do
it
even
afterwards.
A
B
A
B
I
mean
there's
nothing
which
is
stopping
us
to
do
it
sooner.
It's
just
that
we
just
wanted
to
give
some
time
two
users,
so
we
already
put
this
as
a
documentation.
I
think,
at
the
time
of
doing
a
release
candidate.
We
have
already
put
this
documentation.
B
B
So
this
was
added
thinking
that
we
are
going
to
do
it
very
near
to
declaring
statement.
B
B
B
B
You
confirm
time
k
for
wednesday.
Nine
pm
pt
should
have
done
it,
and
let
me
put
this
on
slack
channel.
We
have
one
week
from
today
to
wait.
If
there
is
any
objection,
I
don't
think
there
would
be
any
objection,
but
let's
let's
wait
for
one
week
and
then
probably
next
monday
we
should.
We
will
raise
the
pr
to
change
the
timings.