►
From YouTube: 2022-08-24 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
A
A
Or
no.
A
A
A
If
you're
watching
the
recording,
I
guess
all
right.
So,
let's
move
on
to
the
first
first
js
topic
of
the
day
here
in
the
pr
to
add.
You
know
the
there's
a
minor
feature
here
to
add
to
pg
this
person
upgraded
the
node
engines
from
8
to
14.
A
Their
reasoning
is
that
the
sdk
doesn't
support
versions
below
14
anyways,
so
they're
completely
correct
about
that,
but
I
don't
think
it
should
be
included
in
a
like
a
fixed
pr.
A
A
So
the
obvious
answer
is
that
we
should
just
support
node
version
14
and
higher,
but
we
could
also
support
older
node
versions
just
for
the
instrumentation,
because
the
api
supports
those
older
versions.
A
B
A
We
have
also
upgraded
a
bunch
of
our
other
dependencies
already
a
bunch
of
our
build
tooling
and
eslint
and
lerna
and
stuff
like
that
which
some
of
those
don't
support,
node
old,
node
versions.
That's
why
we
were
holding
off
on
upgrading
them
for
a
long
time
now,
they've
been
upgraded,
so
that
would
make
it
even
more
difficult.
A
B
Probably
just
a
general
quest:
is
there
anything
specific
to
node
14,
that's
being
used
like?
Could
we
say
from
a
dev
dependency,
we
need
14,
but
from
a
peer
dependency
we
will
let
it
support
older
versions.
B
Unless
we
define
it
and
saying
well,
we
will
set
the
peer
dependency
at
whatever
base
level,
but
it's
completely
untested
and
any
any
bugs
that
may
exist.
I'm
sorry
they're
just
telling
us.
We
should
then
bump
up
our
minimum
peer
dependency
for
that
package.
A
B
A
B
This
is
defining
that
this
package
needs
this
node.
I
I
don't
have
this
coming
from
the
browser
space:
how
to
define
the
engine
so.
A
A
Like
build
the
packages
and
then
install
them
as
dependencies
on
some
test
furnace
that
uses
older
versions
of
node,
some
sort
of
you
know
integration
test
type
of
deal.
I
mean
it
would
be
a
lot
of
work.
It's
not
something
that
that
I
think
I
would
have
time
to
set
up.
A
I
think,
since
the
sdk
doesn't
support
them
anyways
and
they
are
deprecated
node
versions.
I
think
that
I
agree
with
amir.
I
would
say:
let's
just
drop
those
versions.
I
don't.
I
think
we
should
do
it
as
a
a
pr
on
its
own,
though,
or
dropping
support
for
the
old
node
versions
not
having
it
done
as
a
part
of
some
unrelated
pr.
A
I'm
only
going
to
just
briefly
mention
this,
because
pa
will
come
next
week
to
talk
about
this,
but
for
those
that
don't
remember,
they
had
been
working
on
a
package
which
adds
hotel
tracing
to
the
fastly
compute
at
the
edge
service,
which
is
a
restricted
edge
computing
service.
A
This
is
not
really
an
instrumentation
as
much
as
it
is
like
a
fully
featured
sdk.
Almost
they
do
use
some.
A
A
They
are
using
this
in
production,
so
I'm
fairly
confident
in
it,
but
it
will
require
obviously
some
review
before
before
being
merged,
and
I
would
expect
that
pa
or
one
of
his
co-workers
become
the
component
owner
of
this
package,
but
I'm
sure
he
will
he.
I've
already
talked
to
him
and
he
said
he's
gonna
come
join
next
week
to
talk
about
what
they
did
so
that
we
can
effectively
review
this.
A
Does
anyone
have
questions
for
me
on
this
before
we
move
on
keeping
in
mind
the
pa?
Will
be
here
next
week
to
answer
questions.
B
A
A
A
A
Nope,
okay,
again
same
thing
as
every
week
we
are
still
working
towards
metrics
general
availability.
There
are
two
pr's
here
that
I
wanted
to
draw
attention
to.
One
is
the
per
metric
reader
aggregation
pr
that
was
open
last
week.
A
This
is
a
fairly
large
body
of
work
and
definitely
needs
reviews
before
it
can
be
merged.
There
is
some
discussion
in
here
about
the
interface
for
exporters
and
he's
not
here.
So
I
can't
ask
him,
but
I
believe
he
actually
removed
that
part
from
this
pr
and
at
created
a
specification,
follow-up
issue.
A
The
specification
has
it
as
a
part
of
the
metric
reader
section,
but
it
doesn't
explicitly
say
that
the
metric
reader
makes
that
decision,
and
you
can
see
here,
the
sdk
implementations
are
a
little
bit
split
so
java
and
go
in
particular.
A
Leave
it
completely
up
to
the
metric
reader.
So
we
based
a
lot
of
our
implementation
on
the
java,
so
we
are
also
currently
falling
in
this
category,
but
when
gennik
has
opened
a
issue
here
on
the
spec
to
get
some
clarification
on
that,
but
that
part's
been
removed
from
this
pr.
So
I
encourage
you
to
review
the
pr
so
that
we
can
get
it
merged.
This
is
like
the
last
big
feature
required
for
ga,
so
this
is
an
important
one.
A
A
So
looking
forward
to
that
as
well,
and
then
there's
a
pr
here
from
mark
mark,
do
you
wanna
talk
about
this
pr,
real,
quick.
B
Yeah,
so
basically
I
just
removed
some
of
the
exports
from
the
sdk
matrix
package
should
be
fairly
straightforward.
B
Most
of
it
hasn't
been
used
in
the
other
packages
anyway,
so
the
diff
is
rather
small,
and
most
of
them
are
very
obviously
internal,
internal
classes
and
functions,
and
things
like
that,
so
yeah
would
be
happy
if
you
would
take
a
look
and
review
that,
because
that
would
probably
make
our
lives
easier
when
we
want
to
change
something
in
the
future.
A
Yeah
so
obviously
removing
this
export
star.
It
makes
it
a
little
bit
more
explicit
what
we're
trying
to
export
yeah.
That
looks
like
it's
just
formatting,
okay,
so
this
should
be
a
relatively
straightforward
pr
to
to
review
and
get
merged.
A
That's
also
very
important
for
ga,
though,
because
anything
that
we
anything
we
publish
needs
to
be
maintained
in
perpetuity.
So
if
we
accidentally
export
something,
we
don't
want
exported,
that's
a
problem.
So
it's
a
it's
a
relatively
straightforward
pr,
but
I
encourage
people
to
review
that
one
as
thoroughly
as
they
can.
C
Yeah
and
not
not
much
of
an
update,
but
it's
I
fixed
the
tests
and
rebased
on
on
latest.
I
think
there's
still
some
conflict,
I
see
new
one,
but
but
I
just
there
are
a
few
comments
from
you.
Daniel
is
from
yesterday
that
I
responded
to
and
I'm
I
think,
santosh
who's
working
on
the
corresponding
api.
C
A
A
A
A
Sandbox,
I
assume
that
nev.
This
must
be
you
asking
if
I
have
gotten
an
answer
about
the
cla
issue:
correct
yeah,
so
I
opened
a
service
deck
desk
ticket
for
this
and
have
not
gotten
a
response.
Yet
I've
also
reached
out
to
I
reached
out
to
the
gc
on
slack
and
they
did
not
have
an
answer.
They
essentially
didn't
know.
A
The
their
good
feeling
was
that
it
would
be
probably
okay
to
just
leave
it
in
the
js
repo,
because
there's
no,
no
ip
added,
that's
not
covered
by
a
cla.
So
it's
not
a
problem
but
we're
working
on.
You
know
what
the
cncf
wants
us
to
do
here
in
terms
of
what
the
sandbox
should
do
about
it.
I
don't
have
any
answers
for
you
at
the
moment.
A
I
I
think
this
is
not
an
issue
that
they
run
into
very
often,
and
I
think
nobody
really
knows
what
to
do.
I
think
the
the
most
correct
solution
would
be
for
us
to
re-author
that
commit
in
our
repo,
but
that's
obviously
pretty
impactful
as
there's.
A
A
A
C
A
That's
not
a
very
satisfying
outcome
or
not
a
very
satisfying
update.
Have
you
had
any
further
thoughts
or
ideas
on
what
to
do
here
on
how
to
handle
this?
I
know
you're
using
just
git
merge,
which
holds
the
full
history,
and
I
don't
know
if
you
can
filter
just
that
commit
or
anything
along
those
lines.
I'm
not
that
familiar
with
the
git.
B
Merge
I
I
could
probably
modify
the
script
to
do
it,
but
apart
from
that,
I'm
actually
still
an
admin,
because
I
got
the
admin
privileges,
elevated
to
add
the
the
bot
token.
A
Yeah,
I
don't
think
you
can
do
that.
I
I
know
that
specifically
the
cncf
has
said
that
the
cla
check
has
to
be
enabled
on
every
repo
required,
so
you
we
would
definitely
want
to
get
some
sort
of
exemption
from
them
on
that.
B
B
C
A
B
B
A
A
Okay,
so
yeah
that
that's
where
we're
at
now,
I
will
keep
you
updated
as
I
can.
I
am
going
to
be
on
vacation
for
the
rest
of
this
week,
but
I
do
have
slack
on
my
phone,
so
I
will,
if
somebody
pings
me
about
this,
I
will
be
able
to
reply
to
it
and
I
do
get
my
email
so
I'll
see
if
there's
any
updates
to
the
service
desk
ticket.
B
A
Okay,
I
looked
this
morning
before
the
meeting
and
there
actually
are
no
new
issues,
but
what
I
thought
we
would
do
is
go
through.
There
are
four
here
that
have
information
requested
and
I
thought
we
would
go
through
each
of
them
and
maybe
close
them
since
we're
not
obviously
getting
information
from
the
original
authors.
A
So
it
looks
like
does
this
failure
occur
only
when
exported
yeager,
or
does
it
also
fail
in
something
to
zipkin?
I
guess
I
could
probably
test
this
myself.
Instead
of.
A
I
think
this
was
a
cores
issue,
they're
exporting
to
the
zipkin
receiver
on
their
jaeger
server.
So
I
can
probably
test
this
to
see
if
see
if
this
is
still
a
problem
or
if
it's
a
problem
on
just
yeager
or
and
not
not
zipkin,
there
is
a
work
around
here
as
well.
A
I
guess
I
will
assign
this
to
myself
and
try
to
reproduce
this.
I'm
not
going
to
close
this
one.
Quite
yet.
A
C
A
A
A
A
C
I
didn't
have
one
word
on
I'm
on
the
other
issue
like
yeah.
I
would
be
interested
if
you
do
look
into
it,
but
just
some
background.
I
think
that
the
open
tracing,
shim
and
open
telemetry
are
kind
of
incompatible
in
many
situations.
Just
because
open
tracing
javascript
did
not
have
in-process
context
propagation.
C
So,
like
everything
was
kind
of,
it's
contexts
were
explicitly
passed
around
so
anytime.
You
were
like
relying
on
there
being
a
kind
of
a
current
context
to
pull
like
a
parent
span
out
of
like
things
break.
So
if
you,
if
you
are
trying
to
like
interrupt
between
open
tracing
and
open
telemetry
in
the
same
process,
I
think
you're
not
going
to
have
very
much
success.
A
Yeah,
I
think,
open
tracing.
It
didn't
even
have
a
context.
You're
passing
around
right.
You
just
passed
around
the
spam
yeah.
B
A
I
have
worked
with
users
that
use
the
shim
and
in
every
case,
that
I've
seen
so
far
they've
just
done
manual
context
propagation,
instead
of
using
the
the
context
manager,
which
is
obviously
unfortunate,
but
it's
it's
kind
of
a
half
step
right.
If
you're,
if
you're
not
wanting
to
completely
switch
away
from
open
tracing,
your
options
are
limited,
no
matter
what.
C
A
Yeah,
I
know
that
the
the
attach
and
detach
methods
that
were
proposed
in
the
api
were
proposed
as
a
potential
solution
to
this,
but
I'm
not
even
sure
if
that
will
actually
help
because
of
like
the
the
fundamentally
nested
nature
of
the
context.
Even.
C
A
A
B
A
Of
yeah,
so
valentine
had
the
same
suggestion
that
it
was
a
load
ordering
problem
that
should
be
fixed
with
the
new
proxy
tracers
since
the
user's
not
responding
here,
I'm
gonna
close
this
issue,
I'm
pretty
sure
it
is
actually
fixed,
though.
A
And
that
was
all
of
the
information
requested
ones.
I
will
make
sure
to
look
into
this
one.
That
was
everything
that
I
had
on
the
agenda
today.
Anyone
have
anything
before
we
before
we
go.
A
Okay,
well
thanks
everybody
for
your
time,
and
I
will
speak
to
you
next
week:
cheers.