►
From YouTube: 2023-01-26 meeting
Description
Open cncf-opentelemetry-meeting-3@cncf.io's Personal Meeting Room
A
A
B
A
I
just
thought:
I'd
drop
in
didn't
have
nothing
on
my
agenda
today.
Just
thought
I'd
see.
What's
going
on.
C
Okay,
thanks
for
updating
me,
the
document.
A
C
Yeah,
because
there
was
a
soccer
match
near
my
office
and
it
was,
it
was
gonna,
be
full
of
people,
so
yeah
I
prefer
to
yeah.
Okay,
yes,
thank
you,
yeah,
okay,
so
yeah
I
have
one
topic
for
today.
That
is.
C
Okay,
yeah:
do
you
have
any
topic
Stephen
that
you
want
to
talk
later.
B
A
C
Okay,
so
let's
review
last
meeting
actions
yeah
we
had
the
we
had
removing
the
open,
Telemetry
SDK
object
from
the
code.
I
did
that
I
think
on
Thursday,
maybe
I
don't
know,
and
it's
removed
and
I
also
updated
all
the
sample
code.
C
Maybe
if
you
are
accessing
the
Tracer
provider
for
example,
then
now
you
will
have
to
cast
it,
but
probably
when
you
are
building
it,
you
can
get
the
instance
and
act
with
it
and
all
the
logic
you
should
do
is
probably
done
with
the
Tracer
provider
Builder,
so
I
don't
see
many
difficulties
there,
and
only
the
only
thing
will
be
probably
getting
the
Tracer
and
you
will
have
to
catch
that
to
a
tracer
SDK.
C
B
Yeah
I
updated:
let's
see
what
did
I
update.
C
B
I
think
that
was
because
I
I
don't
think
we
were
sure
last
meeting
like
if
you
were
going
to
remove
it,
remove
it
or
just
update
it,
okay,
yeah
and
so
for
the
instrumentation
I
noted
that
we
needed
to
call
the
urban
Telemetry
SDK
for
setting
anything
and
that
the
instrumentation
needs
to
be
done
before
I
get
yeah
okay
before
set
okay,
so
now
I,
remember
now:
I,
don't
think
I
removed
anything
regarding
open
to
entry,
SDK
and
the
docs.
But
I
did
note
in
the
auto
instrumentation
that.
C
Yeah
that
that's
true
so,
okay,
so
then
we
put
this
one
in
your
dog
I,
don't
know
if
it
appears
in
many
places,
probably.
C
Of
that,
and
just
just
check
that
this
is
done.
C
I
did
so.
C
Yeah
I
did
change
that
because
it
I
mean
it
wouldn't
compile,
if
not
so,
I
I
thought
that
just
changing
it
will
help
in
in
in
understanding.
If
it
was
a
good
change
or
not,
do
you
know
what
I
mean
yeah
if
it
was
if
it
was
gonna
need
lots
of
changes
or
if
we
really
had
to
remove
it
fully
so
I
updated
a
yeah
I
deleted
examples
and
also
added
resources
through
the
through
the
Tracer
provider,
Builder
that
were
initialized
and
not
used,
so
that
that's
done.
D
You
know
I
think
I
can
I
can
refresh
and
look
at
those
Prometheus
issues
again.
C
Okay,
great
yeah,
that's
right,
yeah
check
if
we
can
make
something
compatible
with
the
output,
because
I
we
have
a
very
simple
export
way.
If
we
can
just
fix
that
by
Yeah
by
taking
some
code,
it
will
be
better
that
I
need
a
a
big
dependency,
because
we
are
not
sure
that
many
people
is
using
it
seriously,
so
they
they
yeah.
They
usually
that
saw
this
issue
is
a
Prometheus
user,
but
I
don't
know
if
he
is
using
our
library
or
just
trying
to
get
started
at
all.
C
The
exporters
work
already
done
that.
That's
my
concern
on
on
changing
a
lot
of
code.
Yes,
for
aspect
being
good,
and
now
now
it's
quite
simple.
So
if
we
could
keep
Simple,
it
would
be
great
yeah,
so
yeah.
C
C
B
C
But
at
the
other
side
we
are
not
adding
a
special
functionality,
so
it's
not
a
big
update,
yeah
I,
don't
know,
I
mean
if
we
go
with
going
with
1.4
Maybe
people
who
updates
will
see
the
errors
and
they
we
document
it.
Clearly
I,
don't
know
yeah.
Theoretically,
we
should
go
with
2.0,
but
then
we
should
then
we
might
maybe
also
change
other
things
like
minimum
supported
versions
or
I.
Don't
know
so.
C
We,
for
example,
when
we
added
log
support
we
moved
to
1.3.
We
also
changed
to
1.2
because
we
had
a
source
code,
Breaking
Chains
in
the
TLP
exporter,.
D
1.2
was
you
know,
what
was
the
thing
when
we
made
1.2?
Is
that
just
because
we,
if
we
made
a
ga
or.
C
C
Yeah,
no,
that
was
because
it
was
the
first
version
that
that
covered
the
all
the
all
the
stuff
needed
for
having
a
1.0
version
that
was
yeah
yeah.
C
D
C
Were
some
requirements
to
be
in
1.0,
you
know
pentelementary,
it
was
a
probably
adding
I,
don't
know
which
one
probably
the
otlp
exporter,
or
something
like
that
and
it
was
labeled
with,
because
when
we
thought
that
it
was
more
or
less
finished
and
the
spec
was
also
validated
by
the
committee.
That
was
a
reason.
C
B
Yeah
I
think
that
if
we
were
following
some
of
our
convention
it
should
be
2.0,
but
you
know
if
we
were
doing
that,
we
would
be
at
like
5.0
at
this
point
going
fast
through
the
other
one.
So
yeah
1.4
is
fine.
C
C
Probably
also
they
also
I
have
I
had
been
thinking
about,
but
that
that's
also
a
long
trip,
I
didn't
I
think
await
methods
to
some
of
our
apis
that
are
really
well
suited
for
a
single
weight
like
the
expert
methods
that
you
can
wait
for
them,
and
things
like
that.
That
should
work
really
well,
but
that
will
imply
also
using
a
a
minimum
version
or
OS
version,
and
maybe
that's
a
good
moment
when
you
change
the
the
deployment
I
mean
the
the
version
where
you
can
deploy
could
be.
B
C
Point
to
update
yeah
because
we
couldn't
yeah.
C
Right
now
we
keep
compatibility
with
what
was
released
with
1.0
I
would
say:
we
have
supporting
very
old
versions
of
all
of
all
systems,
so
yeah.
C
C
C
Yeah
I
think,
as
we
have
sort
of
code
changed,
maybe
we
should
have
documentation
updated
before
so
people
can
relate
to
that
yeah,
okay,
cool.
C
Okay,
any
other
thing
topic
change,
hey
you
have
in
mind
any
feedback,
Stephen.
C
A
Don't
I've
got
nothing
to
add
today.
I've
been
I've
been
swamped
with
other
housekeeping
stuff
here
at.
C
A
Yeah
I
know
what
you
mean:
yeah
it's
getting
yeah
it's
pretty
much,
but
oh
well,
Keeps
Us,
employed,
I,
guess
yeah
yeah.
C
That's
something
that
developing
for
Mac
or
iOS
is
something
that
you
always
keep
things
to
do.
An
update
with
yourself.
You
are
never
done
always
changing,
so
it's
always
a
problem
to
have
changes,
but
that
means
that
you're
still
good.