►
From YouTube: 2023-03-01 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
A
C
All
right,
I'm,
not
sure
if
everybody
is
here
but
like
yes,
we
can
get
started
so
as
always
feel
free
to
add.
Your
topics
to
the
agenda
below
I
have
already
put
a
few
of
those
myself
on
there.
So
the
first
one
is
to
please
review
this
PR
right
here.
It
should
be
rather
simple
one,
it
just
renames
the
logger
options
to
Diego
options
and
we
haven't
exported
that
type
yet
probably
by
accident.
So
it
shouldn't
be
a
breaking
change,
so
I
already
approved
it.
C
If
any
of
you
have
some
time
and
would
give
it
a
review,
I
would
greatly
appreciate
it
just
to
make
sure
I
didn't
miss
anything
there
and
that
we
can
make
sure
that
it's
actually
not
the
breaking
change
right.
Other
than
that,
the
next
point
is
contribci
is
failing.
C
Amia
I
think
you
have
tried
to
to
increase
the
memory,
but
it
didn't
really
help.
We
had
I
thought
a
similar
issue
on
the
Quarry
ball
recently,
but
it
seems
to
me
like
that
is
not
the
case.
It
seems
to
me
that
the
the
problem
in
in
country
is
a
different.
One.
I
tried
removing
the
the
coverage
collection
entirely
and
it's
still
running
out
of
memory.
So
if
anyone
has
any
ideas
or
yeah
any
yeah,
any
ideas,
what
what
that
could
be?
We
would
greatly
appreciate
someone.
A
B
C
Yeah,
that
would
be
great
I'm
a
bit
puzzled
by
all
these.
All
these
similar
failures
that
we
had
have
had
yes
in
in
the
past
few
days,
I'm
not
sure
exactly
where
they
come
from
I
think
it
may
be
some
dependency
that
updated
in
the
that
that
we
hadn't
pinned
yet
and
that's
causing
the
problems,
but
I'm
I'm
not
really
sure.
What's.
D
Good
yeah
I
I've
had
this
issue
with
with
other
projects
as
well.
So
it's
not
just
open
Telemetry.
Looking
at
the
1416
PR,
it's
only
increasing
the
memory
to
six
gig
I
found
locally.
When
I
was
testing
the
the
windows
tests
for
instrumentation
HTTP,
six
wasn't
enough:
I
had
to
bump
it
to
eight
but
yeah
your
finding
of
stopping
the
source
mappers.
That's
that's
working
for
me
locally!
Now
so
I've
just
closed
my
PR,
so
it
might
be
a
case
of
you
know
bumping
it
might
actually
make
it
work.
E
Do
we
know
whether
there's
even
Aid,
available
on
the
GitHub
Runners
I.
E
D
Yeah
for
the
yeah
in
GitHub,
sorry.
B
There
was
an
issue
in
the
past
where,
like
I,
think
webbeck
is
using
a
coverage,
a
istanbula
loader,
and
it
is
using
another
package
which
is
called
converter
Source
map
and
this
package
changed
from
using
like
third
party
dependency
to
using
the
built-in
buffer
from
a
node
and
I
think
there
was
sometimes
where
it
had
a
bug
that,
like
it,
didn't
clean
the
memory
until
they
eventually
had
the
chance
to
run
so,
if
you're
having
a
sink
Loop
really
long
and
you
consume
more
and
more
memory,
nothing
cleaned
it.
B
So
it
was
just
building
up
more
and
more
and
more.
Maybe
if
we
change
the
node
version
in
the
CI,
it's
it's
14
I
think
it's
good
time
to
to
try
maybe
16
or
18..
C
Yeah
it
that
is
maybe
a
good
idea.
I
was
also
looking
at
the
note
version
there
and
it
was
it
yeah.
It's
it's
at
14,
so
yeah
may
help.
D
C
And
it
didn't
help
I
I
didn't
remove
it
completely.
So
basically
what's
happening
in
the
in
the
tests
that
are
failing.
Is
there
their
browser
tests
and
there
is
an
a
kind
of
a
loader
in
the
webpack,
config,
I,
think
or
the
karma
config
or
something
and
I
think
that
one
may
still
run
by
default.
So
it's
just
the
NYC
thing
that
I
removed
and
tried
out
so
that
there
may
still
be
some
some
Istanbul
stuff
in
the
tests.
D
Yeah
and
I
guess
for
reference.
My
test,
I,
don't
use
webpack
I
use
rollup
for
but
I
was
using
Karma
TS,
which
does
include
your
support
indirectly
from
that.
So.
A
C
All
right,
I
guess
we
were
just
have
to
to
look
at
all
these
things
afterwards
and
see
which
which
solution
helps
but
yeah
I
guess
we
have
a
few
points
to
go
off
on
now.
Thank
you.
Everybody
right.
If
there
is
a
no
more
discussion
on
that,
then
I
think
we
can
move
on
to
the
next
point.
With
the
exponential
histogram
PR,
then
I
think
you
added
this.
Do
you
want
to
say
a
few
words
about
it?
I.
E
Don't
have
anything
new
to
say:
it's
been
sitting
for
a
long
time
and
I
was
waiting
for
someone
else
to
review
it
since
it's
a
fairly
impactful
PR
touches
a
lot
of
places,
but
at
this
point
I
can't
wait
forever.
So
if
it's
not
reviewed
by
anyone
else
by
the
end
of
the
day,
I'm
just
gonna
merge
it.
So
we
can
move
on.
C
Yeah
sounds
good:
I
haven't
had
time
to
in-depth
review
it,
but
I
think
it.
It
looks
good
and
there's
also
part
three
that
still
is
coming
up
so
I
think
it.
It
makes
sense
to
merge
it
yeah
all
right,
then
the
next
one
is
the
logs
SDK.
That's
ready
for
review.
E
The
user
that
opened
to
this
PR
I
don't
know
how
it's
pronounced
f-u-y-a-I
or
something
like
that
mentioned
it
early
this
morning
or
early
this
morning.
My
time
I
suppose
just
asking
for
more
reviews.
I,
don't
know
if
Martin
has
any
more
information
or
if
the
pr
author
is
on
the
call
or
not.
A
I,
don't
have
any
more
information,
I
already
reviewed
it
once
now
that
it
has
the
missing
tests
that
I
can
I'll
do
another
pass.
C
Correct.
Thank
you.
C
Yeah
s
then
already
said:
please,
please,
review
this
PR
and
then
moving
on
is
another
topic
right
here
related
to
the
auto
instrumentation
PR.
That's
opening
country
right
now,
which
adds
this
agent-like
experience
to
the
auto
instrumentation
node
package.
There's
this
this
Otep
right
here
then
I
think
you
added
this.
E
Yep
so
I
just
wanted
to
mention
this
configuration
proposal
was
opened,
I
think
on
Friday
may
have
been
Thursday,
but
certainly
after
the
Sig
meeting
last
week,
so
we
haven't
had
a
chance
to
talk
about
it.
Yet
there's
nothing
specific
that
I
really
want
to
bring
up
about
this
PR
exact
or
about
the
show
tip
exactly,
except
that
the
auto
instrumentation
PR
has
some
configuration
concerns
in
it.
I
know
that
the
author
was
at
least
considering
adding
some
sort
of
environment
variable
for
a
configuration
and
I.
E
Think
I,
don't
know
if
Sam
is
on
the
call
or
not
looks
like
no
I
was
hoping
to
be
able
to
talk
about
it
on
the
call.
But
it
looks
like
the
pr
author's
not
here
so
we
can
discuss
it
offline,
but
I
just
want
to
I
think
it
affects
this
PR.
So
I
wanted
to
make
sure
it
was
brought
up
right.
C
Thank
you,
yeah
I.
We
had
to
talk
about
this
earlier
and
we
will
probably
also
have
to
make
some
changes
to
the
to
the
core
review
repo
and
the
node
SDK
package.
Once
this
Firebase
configuration
comes
along
to
account
for
all
the
different
ways
you
can
get
the
configuration
Dennis
currently
we're
passing
this
environment
variable
thing
around
which
may
be
hard
to
maintain
in
the
future
as
more
and
more
ways
of
configuring
stuff
come
comes
along.
E
Yeah
I
think
we'll
likely
have
to
introduce
a
configuration
component
which
hides
the
details
about
how
exactly
it
was
configured
in
the
current
Otep
state.
If
you
have
a
configuration
file,
all
environment
variables
will
be
ignored,
so
we
don't
want
to
have
every
place
in
the
code.
Have
to
check.
Is
there
a
file?
Is
there
not
I?
Think
I
a
unified
configuration
component
would
make
more
sense
right.
C
All
right,
so,
if
there's
no
more
comments
on
this
I
guess
we
can
move
on
to
the
next
one.
Matt
I
think
you
added
this
right
here.
The
process,
detector
improvements.
F
Yeah
I
I
have
this
PR
open
there
yeah.
Basically,
we
already
have
a
process.
Detector
I
was
just
flushing
it
out
and
collecting
some
additional
attributes
that
were
that
were
already
expect
and
yeah
I
think
slight
change.
Behavior,
like
this
kind
of
used
to
be
like
an
All
or
Nothing
detector,
in
that
it
would
kind
of
enforce
that
it
found
all
process,
attributes
that
I
thought
it
would
otherwise
would
collect
none
which
was
kind
of
weird,
so
I
kind
of
made
it
a
best
efforts.
F
Detector
and
kind
of
the
other
thing
that
I
did
is
that
this
was
shared
between
browser
and
node
and
we
kind
of
short
circuit
if
it
was
in
browser,
so
I
actually
moved
the
move.
This
so
there's
a
node,
specific
implementation
and
then
just
a
no-op
detector
for
browser.
F
That's
the
summary
of
the
changes,
but
my
actual
question
is
that
nav
had
a
question
when,
when
reviewing
this
around
collecting
specifically
process
are
the
one
mainly
about
like
might
be
possibly
collect,
something
sensitive
and
I.
Think
I
think
there
is
a
risk
of
that.
But
I
was
just
acknowledging
that.
A
F
Already
collecting
process
argument
one
with
this
detector,
as
is
today,
so
we
could
decide
to
add
an
option
to
omit
collecting
some
of
these
attributes.
F
F
Some
of
this
stuff
up
just
trying
to
make
it
so
that
you
know
detectors
from
JavaScript,
collect
more
or
less
the
same
thing
as
you
would
get
from
go,
or
something
like
like
that,
and
nobody
else
seems
to
have
an
option
to
to
disable
collecting
portions
of
of
the
command
so
I'm
just
going
to
bring
this
to
folks
attention
if
you
want
to
put
eyes
on
it,
if
you
have
opinions
about
how
we
should
handle
collecting
this
information.
D
Yeah
the
reason
I
had
the
comment
was
I.
Don't
think
we
have
this
today
and
please
correct
me
if
I'm
wrong,
but
we
probably
need
some
form
of
generic
scrubbing
utility
that
we
could
then
just
apply,
and
it
probably
should
be
done
at
the
spec
level
so
that
everyone
does
it
the
same
but
yeah
so
I
just
added
the
comment:
I
didn't
block
it
because
I
don't
think
it's
a
blocking
issue,
because
it
is
what
it
is,
but
we
should
definitely
have
something
to
scrub.
D
I
think
pii
in
general
isn't
addressed
when
it
needs
to
be
yeah.
E
I
mean
it's
not
addressed
in
the
product
as
a
whole.
In
many
ways,
I
apologize,
I
thought
that
I
had
responded
to
this,
but
I
guess
I
didn't
my
reading
of
the
spec
is
like
for
us
following
the
spec
is
not
a
which
is
not
an
option
right.
We
have
to
do
it,
and
this
is
what
the
spec
recommends
we
do
there
is.
You
can
always
question.
Is
the
spec
correct,
but
that's
a
question
for
the
spec,
not
a
question
for
us.
E
As
far
as
this
PR
goes,
I
think
I
suspect
these
questions
were
already
brought
up
during
this
semantic
convention
process
or
I
at
least
hope
they
were
so
I
think
they
probably
have
been
at
least
considered.
We
may
want
to
look
into
the
the
spec
history
just
to
like
you
know,
conversation
history,
just
to
see.
E
Validate
whether
that
assumption
is
correct,
but
as
far
as
this
PR
goes,
I
think
that
we
can
accept
this
PR
with
the
understanding
that,
yes,
there
is
potentially
some
leak
on
the
command
line,
but
I
think
in
general,
I'm
I'm,
certainly
not
a
Windows
expert,
but
in
Linux
at
least
the
process.
Command
line
is
not
typically
considered
to
be
a
secure.
So
if
you
have
secrets
in
there,
they're
already
potentially
leaked
in
other
ways,
I'm
not
saying
that.
E
That's
not
you
know
it
doesn't
invalidate
the
concern,
but
it
is
generally
not
considered
to
be
somewhere
where
you
should,
where
you're
likely
to
find
secrets.
D
And
it
may
not
be
Secrets,
which
is
why
I've
got
pii
there.
It
may
just
be
a
user
ID
or
something
yeah,
yeah,
I,
I
suspect.
If
you
look
back
at
history,
it's
probably
punted,
because
you
could
do
also
do
this
at
the
processor
or
Expo
level,
so
yeah
yeah.
E
Well,
that's
the
Spec's
favorite
thing
is
to
just
punt
in
for
later
yeah,
exactly
I
will
make
sure
to
review
this
PR
I
I,
don't
see
any
reason
for
us
to
block
it
in
JS.
If,
if
there
are
larger
questions,
I
think
they're
better
addressed
at
the
spec
level.
A
F
A
C
All
right,
thank
you
and
if
there's
no
more
comments,
then
I
think
we
can
move
on
to
the
sandbox
update
by
Neff.
D
Me
again
remove
a
lot
of
talking
today,
so
the
sandbox
is
I.
Think
that
to
the
main
JS
encounter
repos
from
a
couple
of
days
ago
after
I
questioned
a
change
to
fix
the
4.6
typescript
thanks
Daniel
for
pushing
that
one
in,
but
in
the
last
two
weeks,
Sammy
actually
added
the
resource
test
to
say,
I
would
also
test
it
with
version
190,
because
the
async
stuff
and
for
some
reason
it
well.
There
was
two
two
things:
the
the
first
one
was
my
script.
D
Didn't
take
this
into
account,
so
I
actually
do
the
script,
renamed
it
to
open
Telemetry,
sandbox
resources
underscore
190,
which
of
course
failed
because
that
doesn't
exist
once
I
fix
that.
But
the
test
is
now
failing
to
do
the
import,
even
though
everything
looks
fine.
So
if
anyone
has
any
idea
why
I'm
checking
the
TS
configs
and
between
the
two
they
look
the
same,
the
sandbox
is
using
a
letter
version
of
typescript,
so
I
tried
bumping
to
a
later
version
of
typescript
on
the
JS
repo.
D
C
All
right,
yeah
I'm,
looking
at
thinking
it
through
I'm,
also
not
really
sure
why
why
it
wouldn't
work.
D
D
D
Like
I
even
tried
bypassing
Rush
and
just
do
an
npm
install
directly
on
that
package
and
that
didn't
do
it
either.
So
it's
not
like
Russia's
cross-linking
something
incorrectly.
D
D
Yeah,
it
was
committed
two
weeks
ago,
so
the
my
previous
runs
in
the
the
previous
merge
was
prior
to
two
weeks
ago.
I
have
actually
just
caused
the
action
to
run
to
effective,
merge
the
auto,
merge
Branch
as
of
yesterday
or
actually,
as
of
like
seconds
ago,
I'll.
Try
pushing
that
in
and
then
running
the
test
again.
Maybe
there's
something
else
that
fixes
it,
but
I'm
not
optimistic,
but
that
will
be
it.
C
All
right,
yeah
I
will
I
will
have
a
look
at
that
and
see
if,
if
maybe
I
can
come
up
with
something
but
yeah
yeah.
D
C
Interesting
all
right,
thank
you
for
the
update.
I
will
definitely
have
a
look
and
yeah.
If
there
are
no
more
comments
on
this
or
suggestions,
then
I
guess
we
can
move
on
to
the
15
minute
back
triage,
let's
see
so
the
first
one
right
here,
I
have
looked
at
that
one
before
seems
to
be
one
person
having
problems
with
the
otrb
grpc
export-up,
but
I
have
asked
them
to
try
something
out
there.
C
It
looks
like
the
that's
just
a
config
problem
right
here,
but
in
general
I
think
they
are
having
a
problem
here
where
the
the
expertise
is
basically
just
dumping.
C
This
long
log
line
there
and
we
may
have
to
improve
vlogging
on
on
the
export
out
there
at
some
point,
just
to
make
sure
that
yeah
people
actually
can
see
what
what
the
problem
is
because,
right
now
it's
it's
very
non
non-descriptive
there,
but
yeah,
let's
see
once
they
they
come
back
on
that
I
think
this
one
we
already
had
yeah
there's
a
bunch
of
discussion
on
here,
looks
like
the
they
are
trying
to
use
the
grpc
exporter
with
with
roll
up
which
I
think
shouldn't
it's
not
working
right
now.
A
E
I
could
be
wrong
about
that
yeah.
It's
hard
to
keep
all
the
exporters
straight
in
my
head,
yeah.
C
I'm,
pretty
sure
that
that
we
don't
support
grpc
right
now
in
the
browser
it
might
be
that
they're
using
roll
up
and
running
it.
E
They
can
find
it.
So
that's
probably
a
configuration
issue
of
some
kind,
not
a
roll-up
expert,
but
I
suspect
it
should
be
possible.
I
know
that
it's
that
some
people
have
managed
to
make
it
work
with
webpacks,
so
I
don't
see
any
reason
why
Roll-Ups
should
be
different.
C
C
Yeah,
maybe
we
we
have
to
add
some
documentation
that
it's
not
supported
onto
the
the
exporters,
but
I
think
this
looks
mostly
like
a
feature
request.
I
will
leave
the
labels,
as
is
for
now,
and
once
they
come
back,
have
a
come
back
with
more
information.
I
will
update
the
labels
on
on
the
back
here.
C
All
right,
the
next
one
is
already
assigned,
and
the
one
of
the
debt
has
a
PR
open.
C
C
C
E
We've
gone
through
a
few
different
iterations
of
this
one
I
believe
the
most
recent
state
was
that
we're
just
waiting
for
somebody
to
have
time
to
to
do
this.
I
think
Simon
might
do
this
next
week.
I
might
have
time
to
do
it
before
the
end
of
the
week,
but
we'll
see.
C
All
right,
in
that
case,
I,
think
I
will
leave
that
here
at
up
for
grabs,
I.
E
C
Yeah
leave
it
up
for
grabs
yeah.
So
if
anyone
comes
along
and
has
some
time
yeah,
they
can
feel
free
to
take
it
right.
I
think
that
should
be
it
for
Decor,
repo
and
Fork
on
trip.
B
C
I
think
that's
a
good
idea.
She
has
been
she's
very,
very
knowledgeable
in
these
database
instrumentation
packages.
So
I
will
just
ask
her.
B
C
All
right,
I'm
I'm,
not
sure
what
to
make
of
it
right
now,
because
I'm
not
too
familiar
with
the
postgres
instrumentation,
but
I
think
we
can
apply
the
labor
once
others
has
had
some
time
to
look
into
it.
Maybe.
C
All
right
moving
on
to
the
next
one
does
not
instrument
AWS
SDK
3.
D
B
The
issue
is
not
the
version
three,
it's
something
in
its
code.
I
know
that
the
someone
in
the
past
I
have
Tim
and
it
turned
out.
It
was
like
internal
function
called,
don't
remember
exactly
what,
but
something
like
set
promise.
It
changes
like
the
promise,
implementation
and
it
over
hides
the
patch
and
it
causes
the
instrumentation
not
to
walk.
So
it
might
be
the
issue,
it
might
be
something
else
you
can
assign
it
to
me.
I
will
take
a
look
at
it.
C
All
right,
thank
you
for
taking
a
look,
I
appreciate
it.
E
C
E
They're
also
using
multiple
Telemetry
packages,
so
it's
possible
that
they're
that
they're,
conflicting.
C
Right
has
some
Sentry
packages
there
yeah.
E
There's
there's
Century
stuff:
there's
this
code
like
a
carpenter,
Telemetry
I,
don't
know
what
that
is
right,
it's
possible
that
they
just
have
some
conflict.
C
Next
one
something
with
the
yeah.
D
E
Fastify
and
they're
loading
open
Telemetry
in
an
adapter,
so
fastpot
is
loading
open
Telemetry,
which
by
definition
means
that
they're
loading,
Best
Buy
first.
C
All
right
and
then
this
one
is
already
assigned
and
has
information
requested.
Let's
just
take
a
look
if.
C
I
I
think
we've
we've
talked
about
this
enough
in
the
past
meetings.
I
guess
we
can.
We
can
keep
this
one
here.
This
one
is
also
already
assigned.
C
That
disorder
information
requested
so
I
think
we
can
also
leave
it
at
that,
and
that
should
be
it
all
right.
Thank.
E
C
Right,
I
think
that
is
because
this
you
need
these
open,
SSL
Legacy
provider
note
options
because
with
node
18
they
they
don't
allow
some
open,
SSL
provider
anymore,
because
they're
insecure
and
have
been
deprecated
for
a
while
I
will
I
will
send
you
the
flag
on
on
Slack.
B
C
All
right,
thank
you
all
right.
If
there's
no
more
topics,
then
I
guess
we
can.
We
can
end
the
meeting
here
and
I
will
give
everyone
25
minutes
of
your
time
back.
Thank
you.
Yay.