►
From YouTube: 2022-11-02 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
A
D
C
Tiring
I
was
I
was
much
more
tired
after
kubecon
than
I
expected
I
I.
Don't
know
why
I
mean
I,
always
I
always
travel
to
kubecon
and
then
I
come
home
and
I'm
super
tired,
but
I
think
I
usually
blame
it
on
the
Travel.
C
B
C
C
Yeah,
it
was
great
I
mean
it's
it's
convenient
and
inconvenient,
so
it
was
really
convenient
that
I
didn't
have
to
travel,
but
it
was
inconvenient
because
normally
when
I
travel
I
get
a
hotel,
that's
close
to
the
conference,
but
since
it
was
in
my
hometown,
I
stayed
at
home,
so
I
had
to
drive
every
day
and
park
and
all
of
that
I
kind
of
wish
that
I
had
gotten
a
hotel
anyways.
But
next
time
it's
in
Detroit.
C
I
expect
that
this
meeting
will
be
likely
quite
a
bit
shorter
than
normal,
since
last
week
was
kubecon
and
I'm
still
honestly
catching
up
on
things.
So,
if
anyone
has
anything,
they
want
to
bring
up
feel
free
to
add
it
to
the
agenda.
Otherwise,
we'll
just
go
through
as
normal.
C
C
I
suppose
there
is
a
PR
open
to
release
everything,
but
actually,
after
discussions
with
with
the
maintainers,
we've
decided
to
merge
the
API
metrics
package
into
the
main
API
package,
the
primary
reason
not
to
do
this
originally
had
to
do
with
bundle
size,
but
xinjiang
was
able
to
show
that
tree
shaking
can
be
done
effectively
as
long
as
we're
careful
about
how
we
structure
the
package.
C
So
this
PR
that
he's
open
does
make
the
required
changes
to
enable
effective
tree
shaking
in
the
API
and
merges
the
metrics
API
into
the
main
API.
So
I
encourage
everyone
to
look
at
it.
It
is
mostly
just
merging
the
packages
together.
So
there's
a
lot
of
files
changed,
but
most
of
the
the
changes
are
just
moves
and
copies.
C
There's
very
little
done
in
the
way
of
actual
changes,
so
it
should
be
relatively
straightforward
to
review
and
when
that
merges,
I
will
plan
on
releasing
the
new
version
of
every
package,
including
the
API,
which
does
include
a
handful
of
important
bug,
fixes
so
I
I,
definitely
think.
Certainly
by
the
end
of
this
week,
I
will
try
to
merge
the
pr
which
merges
the
API
metrics
into
the
API
today,
if
possible,
and
merge
the
release
before
the
end
of
the
week.
C
So
I
encourage
anyone
that
has
strong
opinions
on
this
to
look
into
this
PR
before
the
end
of
the
day.
I
probably
would
have
already
merged
it,
except
that
I
wanted
to
give
people
a
chance
to
see
it
before
I
merged
it.
Anyone
have
immediate
questions
or
concerns
on
that.
C
Okay
also
related
to
the
metrics
GA.
There
is
a
documentation.
Pr
I
am
going
to
probably
merge
this
PR,
also
at
the
same
time
as
the
release
and
then
once
that's
merged.
We
will
move
it
over
to
the
website
repo
I'm,
going
to
merge
it
in
this
readbook
so
that
there
is
some
live
documentation.
While
we
work
on
the
website
and
also
all
right,
more
accurate
contribution
stats,
because
if
we
just
closed
it,
it
would
be
I,
don't
know
some
people
care
about
things
like
that.
C
That's
all
I
had
on
the
metrics
GA.
Does
anyone
have
questions
on
that
or
or
comments.
C
Okay,
the
esmpr
is
updated
now
I
know
this
has
been
an
ongoing
topic
for
a
long
time.
There
is
now
a
PR
that
is
up
to
date,
is
not
a
draft
and
introduces
esm
support.
C
I
know
this
is
something
a
lot
of
people
want
I,
but
hasn't
really
been
a
strong
priority
for
a
while,
but
I
would
encourage
anyone
who
cares
about
esm
support
to
look
into
that
PR
oud.
There
are
still
some
limitations
like
you,
can
only
have
a
single
loader
per
process
and
some
things
like
that,
but
something
is
better
than
nothing
in
this
case
it
does
use
Import
in
the
middle,
which
I
know
is
used
by
a
few
different
vendors
in
production,
so
it
should
be
fairly
stable.
C
Okay,
I
saw
that
mat
Ware
did
join
the
call,
so
he
can
speak
to
this
if
he
wants
to,
but
I
notice
that
he
assigned
himself
to
work
on
the
exponential
histogram
support
Matt.
You
want
to
give
us
a
quick
update
on
where
that's
at,
or
are
you
not
at
the
point
yet.
F
I
picked
it
up
at
the
end
of
last
week
and
have
been
working
on
the
the
mapping
functions
right
now,
so
it's
it's
still
pretty
early
on,
but
generally
I
think
things
are
going
well.
F
Yeah
I
think
my
only
concerns
at
this
point
is
the
JavaScript
number
type
that
that,
in
the
end,
you
know
some
modifications
may
need
to
be
able
or
may
need
to
be
made
to
to
accommodate
it.
But
right
now
like
it's,
it
hasn't
been
a
blocker
and
I
haven't
found
anything
that
actually
has
to
change.
It's
just
like
it
has
been
a
thing
on
my
mind:
I
started
working
through
it.
C
I'm
curious:
what
what
about
the
number
type?
That's
a
problem.
Is
it
the
lack
of
integer
support
or
the
lack
of
exact
decimal
support.
F
Support
more
than
anything
like
as
I've
been
writing
things
I've
been
drawing
inspiration
from
some
other
implementations
and
as
I'm
writing
test
coverage
I'm
trying
to
have
like
the
same
test
coverage,
but
I
definitely
noticed
that
these
other
implementations
are
testing
much
larger
integers
larger
than
Max,
safe
integer.
F
So,
like
yeah
and
I
know,
the
bucketing
is
kind
of
set
up
for,
like
the
full
I
believe
that
the
bucketing
is
set
up
to.
You
know
accommodate
the
full
space
of
numbers
that
you
can
work
with,
so
maybe
that
will
have
to
be
truncated
just
a
little
bit
but
for
JavaScript
but
I'm,
not
I'm,
not
positive,
I'm,
still
kind
of
I'm.
Not
to
that
point.
Yet,
where
I've
run
in
into
any
issues,
it's
just
like
I'm
I'm
waiting
for
that
point
or
fearing
that
point.
C
A
C
Update
I
think
most
people
already
know
this,
but
for
those
that
don't
I
actually
reverted
the
pr
that
used
the
anchored
clock
to
calculate
span
times.
C
C
I
am
open
to
suggestions
for
more
long-term
solutions
for
this,
but
it's
clear
based
on
just
the
number
of
issues
that
have
been
opened
and
and
people
that
have
reached
out
in
slack
and
things
like
this-
that
the
negative
durations
definitely
affect
people
significantly
worse
than
the
shifted
span,
timings
that
ever
did
so.
That
will
go
out
with
the
next
release
and
a
more
long-term
solution
is
welcome.
If
anybody
has
the
time
or
inclination
to
join
that
discussion,
there
is
a
discussion
issue.
C
Here
so
I
will
move
that
up
for
anyone.
That's
interested.
C
Okay,
I
added
this
next
item
here,
just
to
ask:
if
there
is
any
update
on
the
logging,
SDK
or
exporter
I
know
there
was
some
work
ongoing,
but
I
don't
know
what
the
current
status
of
that
is.
D
C
D
I'm
here
so
I'm
the
the
update
this
I'm
still
working
on
it.
It's
been
slow
progress,
progress,
unfortunately,
I
think
what
I,
what
I
I
wonder
if
what
I
should
do
is,
if
is
just
open
a
draft
PR
at
this
point
like
with
at
least
like
the
high
levels
structure,
so
I
can
get
some
initial
feedback
and,
while
I'm
still
working
through
like
tests
and
stuff
like
that,
so
that's
that's
I'm,
working
on
the
SDK
itself
and
I
someone
else,
an
engineer
from
I
think
app.
D
Dynamics
Cisco
is
gonna
help
with
the
exporter
and
I.
Don't
know
what
the
what
the
update
on
that
is.
But
I'll
check
with
Santosh
is
the
point
person
on
that.
So
I'll
check
with
him.
C
Okay,
yeah
I
mean
a
draft.
Pr
would
be
very
reasonable.
I
also
think
there's
there's
no
requirement
for
it
to
be
like
a
full,
complete
implementation
in
order
to
get
something
into
the
repo
that
can
be
iterated
on.
C
Can
even
mark
it
as
private
so
that
it
won't
be
released.
Yeah
there's,
no
reason
that
that
the
first
PR
has
to
be
like
a
full
complete
implementation.
I,
don't
think
anybody
expects
that.
D
Okay,
yeah
also
I'll,
do
my
best
to
open
a
PR
by
by
next
week,
so
that
we
have
something
to
look
at
before
the
holidays.
Yeah.
B
A
No
I,
just
just
a
bit
like
it,
didn't
help
that
tigarden
was
looking
to
change
the
API
for
the
event
stuff
as
well.
D
A
C
Okay,
I
guess
I
missed
that
one.
But
if
it
got
closed
then
I
will
Let
It
Go
yeah.
A
C
Okay,
I
mean
I,
assume
that
that's
in
good
hands
with
the
log
Sig
I
do
Wonder
a
little
bit.
Why
up
here
like
that,
would
have
been
opened
in
the
specs,
since
the
Otep
should
have
hashed
out
all
those
issues?
That
seems
like
something
kind
of
overriding
the
Otep
that
had
a
lot
of
work
put
into
it.
C
C
Thanks
for
that,
update,
though
I'm
not
sure
who
added.
E
I
added
the
instrumentation
guidelines
to
contrib
and
I
started
a
documented
documenting
things
that,
like
it's
very
easy
for
people
to
make
mistakes,
because
they
don't
know
all
the
history
and
all
the
small
details
like
leaking
types
from
the
instrumented
package
and
there's
some
other
issues
and
yeah
so
I
just
want
to
let
you
know
that
I
added
it
and
asked
for
a
feedback.
If
you
think
it's
good,
if
you
think
it's
should
be
something
else
or
or
not
not
to
do
it.
E
C
E
A
C
I
think
that's
the
most.
It's
not
necessarily
complicated
but
difficult
to
understand
and
easiest
to
do
wrong
thing
is
the
the
dependency
handling.
E
B
D
C
It
looks
like
that
was
the
last
item
on
the
agenda
here.
I
will
go
through
the
triage
process
here,
but
I
think
there's,
probably
not
much
spec,
violation
start
and
finish
time
stamps
missing
and
must
include
hotels,
start
and
end
time
stamp.
Fortified,
two
Keys
missing.
You
do
have
a
duration.
C
Where
is
he
logging?
The
oh
console?
Span
exporter
I
see
yes,
this
is
not
a
spec
violation
in
my
opinion,
because
the
spec
is
not
about
the
internal
representation,
but
rather
about
the
data
model
itself,
although
maybe
the
console
exporter
may
need
to
be
updated.
B
C
But
broke
their
mental
ID
that
this
one
we
already
know
about
I
will
put
a
priority
label
on
this,
though.
C
Http
crashes
super
agent,
that's
one.
We
also
already
know
about.
This
has
been
open
for
a
long
time
and
information
was
requested
a
long
time
ago.
I'm
going
to
close
this
because
it's
not
reproduced
reproducible
actually
yeah.
C
I
think
this
has
actually
been
fixed.
It
just
I
released
contrib.
This
morning's,
this
user
probably
doesn't
have
the
new
version.
Yet.
E
I
know
about
graphql
that
it's
linking
the
the
types
the
instrumentation
exports.
E
B
E
Yeah,
if
you
remember
originally,
we
released
a
wssdk
instrumentation
with
version
025,
and
then
we
released
it
again
with
version
0.1
and
it
confused
the
sneak
I
think
oh
okay
and
then
I
bumped
the
version
to
something
higher
than
25,
so
it
will
be
resolved.
But
this
release
was
a
like
missing
change,
log
or
release
notes.
C
Probably
because
the
rate
limiting
issue
yeah
I'll
assign
this
one
to
myself,
I
think
this
issue
should
now
be
shouldn't
happen
anymore,
but
there
probably
were
a
bunch
missed
in
the
past.
Yeah.
C
Well,
it's
partially
resolved
so
the
it
should
no
longer
it's
still
possible
for
it
to
fail.
If
we
release
too
many
packages,
but
the
issue
where
all
of
the
releases
would
be
created,
but
then
nothing
would
actually
be
released
as
resolved
and
GitHub
did
raise
our
rate
limit,
but
I
don't
know
how
they
didn't.
Tell
me
like
what
our
specific
rate
limit
is.
So
it's
possibly
resolved
I'm,
not
sure.
Okay,
we'll
keep
an.
E
C
C
E
This
yeah
I
opened
a
bunch
of
facials
for
them
on
goodbye.
Instrumentation
I
got
the
chance
to
play
with
it
a
bit
and
found
a
few
problems.
So
mongodb
is
calling
the
original
function,
but
it
doesn't
wrap
it
in
a
tie,
catch
or
safe,
execute
in
the
middle.
So
if
an
exception
is
a
phone,
the
span
is
not
ended,
so
it
should
be
easy
to
to
fix
you.
B
Can
up.
E
D
E
C
E
E
Insult
many,
but
the
instrumentation
like
the
hooks
into
the
insert
column,
so
got
it:
okay,
yeah,
so
we're
getting
the
first
yeah,
so
I
I,
guess
people
like
foreign.
They
didn't
want
to
see
an
array
because
it
doesn't
make
sense.
So
they
took
the
first
document.
But
the
now
insult
many
doesn't
make
any
sense,
because
it's
only
one
document.
C
Got
it
okay,
that
makes
sense,
so
it
should
probably
be
an
array
in
both
cases.
Then.
C
E
Don't
know
how
to
fix
it,
because
I
think
it's
a
bit
weird
to
have
an
array
for
insult
one,
because
it's
one
document,
the
DB
statement
should
reflect
that
it's
a
one
document,
but
the
maybe
it
can
be
fixed,
I,
don't
know,
I
didn't
look
into
it.
E
C
Looks
like
nobody
had
it
anymore
topics
here
and
a
little
bit
of
a
short
meeting,
but
with
kubecon
last
week,
I'm,
not
too
surprised
anyone
have
anything
they
want
to
bring
up
before
we
call
it
a
day.
C
All
right
well,
thank
you.
Everybody
for
your
time
enjoy
the
rest
of
your
week
and
I'll
talk
to
you
next
week.