►
From YouTube: 2023-01-11 meeting
Description
Open cncf-opentelemetry-meeting-3@cncf.io's Personal Meeting Room
A
B
C
D
D
A
E
A
Have
a
topic
or
which
is
everyone's
favorite
thing,
to
do,
naming
things
proposing
a
new
path
in
ottl
context,
which
can
act
like
a
a
temporary
place
to
save
information
while
you're
doing
Transformations
so
that
you
don't
have
to
like
muck
with
attributes
or
whatever
along
the
way
it'll
be
super
useful
for
complex
Transformations,
especially
with
logs,
and
this
particular
field
will
only
live
for
the
duration
of
the
transform
context.
So
once
you're
done
messing
with
a
particular
set
of
data,
it
just
disappears.
A
It
doesn't
get
automatically
added
to
attributes
or
anything
like
that.
It's
truly
just
a
temporary
place
to
save
some
data
before
you
go
to
the
next
transformation
where
we've
been
noodling
on
names
right
now.
The
the
name
is
TMP
temps,
something
really
small
and
explicitly
temporary,
but
it
doesn't
really
do
a
good.
D
A
F
G
It's
about
making
the
collector
resource
attributes
available
to
the
extensions
I
also
have
a
use
case
where
I
needed
to
access
this,
a
specifically
service
instance
ID
inside
the
processor
component,
because
when
I
have
the
collector
scale
horizontally,
it
was
inside
the
process
component.
It
was
necessary
to
add
this.
Add
this
unique
ID
so
that
the
when
data
reaches
the
back
end.
G
E
H
Yeah
I
mean
I
I,
agree.
I
also,
don't
have
any
opposition
at
this
point,
since
it's
kind
of
identifying
the
The
Collector
the
same
way
that
some
of
the
other
Telemetry
settings
would
be
helpful
for
specific
instance
of
the
collateral
make
son
said:
I
might
live
there,
but
I'd
have
to
think
a
little
bit
more
about
it.
G
J
I
I
made
a
PR
to
basically
add
it,
but
that
num
workers
field
has
could
probably
use
better
wording.
So
I
was
hoping
to
get
eyes
on
this
just
to
see
if
anyone
has
recommendations,
especially
if
you've
used
the
processor
on
like
how
we
should
recommend
this
to
end
users
or
possible,
you
know
more
accurate
verbiage,
so
yeah
I,
just
hoping
they
get
a
some
extra
set
of
eyes
on
this.
K
E
K
No,
no
I'm
not
concerned
that
it's
it's
not
doing
the
right
thing.
I'm
just
concerned
that
we
don't
have
good
guidance
for
users
on
how
to
determine
what
the
appropriate
value
for
a
number
of
workers
to
use
is
I.
Think
the
the
initial
guidance
that
Brian
gave
may
have
been
copied
from
somewhere
else.
I,
don't
know
if
that
was
from
the
cute
retryer.
K
If
that
was
from
comments
in
in
this
code
here,
but
it
talked
about
physical
processors
and
setting
it
to
the
number
of
physical
processes
which
I
stood
out
as
kind
of
weird
to
me
since
I
think
a
lot
of
people
run
this
collector
and
virtualized
context
or
container
context
where
you
might
be
talking
about
millicorders
instead
of
number
of
processors,
anyways,
so
deciding
how
many
workers
should
you
have
working
on
this
queue
and
how
to
figure
that
out
is.
Is
the
question
here:
okay,.
K
E
Just
just
one
comment
about
DNS
name
and
namespace
prefixes
I
believe
we
better
follow
the
schema
that
you
already
have
for
for
The
Collector
and
like
for
the
scope,
for
example,
we
have
a
schema
field
and
it's
a
bit
different.
It's
not
like
not
java-like,
but
starting
with
I
believe
go
bro
like
what
what's
being
used
for
go
packages.
So
if
you
take
a
look,
there
it'll
be
nice,
but
we
can
probably
take
it
offline
because
I
don't
have
any
any
other
suggestions
here.
If
anyone
else
have
this,
the
share.
M
That's
been
flagged
some
time
ago
by
by
Ryan
actually
on
by
request,
but
actually
from
the
community.
He
opened
an
issue
asking
if
we
could
make
this
table
and
then
another
one
which
is
more
recent,
that
to
cover
a
maintain
component
called
the
HTTP
forwarder
extension
and
I
wanted
to
understand
what
it
would
look
like
to
mark
them
as
table
in
some
some
feedback
from
Liberty
here
that
this
might
just
be
used
by
single
FX,
which
specifically,
do
we
have
have
you
folks
seen
any
use
of
those
extensions
in
the
world.
M
E
E
C
There
is
an
open
issue,
I
just
shared
specifically
about
the
stable
stability
and
what
that
means
for
breaking
changes.
I.
My
personal
opinion
is
that
we
should
not
Mark
any
country
components
as
stable
until
we
have
made
a
decision
on
that
issue.
I
don't
know
if
everyone
else
agrees,
but
it
feels
ambiguous
to
call
something
stable.
If
we
don't
know
exactly
what
that
means
in
terms
of
breaking
changes,
for
example,
yeah.
E
E
A
Don't
remember
where
we
talked
about
this,
but
I
think
one
time
when
we
were
talking
about
component
stability,
we
were
also
talking
about.
Can
something
be
marked
stable
at
the
semantic
conventions
that
that
component
interacts
with
aren't
stable,
I?
Don't
know
if
that's
applicable
for
these
components,
but
it's
something
to
think
about
that.
E
That
one
is
applicable
for
receivers
for
script,
script
and
receivers,
because
that's
true
and
when
you
I
believe
we
have
a
document
for
scrape
and
receiver.
Specifically,
we
don't
add
that
requirement
there,
because
there
is
still
work
on
the
semantic
conventions.
Right
like
this
new
working
group
but
yeah.
That's
that's
probably
needs
to
be
one
of
the
criteria
or
marketing
company
that's
stable,
but
for
for
companies
like
like
the
one
that
Antoine
is
talking
about
APM,
for
example,
it's
just
sending
data
I.
Believe
it's
okay,
it's
good
enough,
but
Eric.
B
If
it,
it
was
the
original
question
about
whether
people
just
had
anecdotal
issues
with
the
exporter
or
it's
more
about
the
semantics
that
the
protect,
because
if
it
helps
we
used
it,
the
sapm
exporter
for
a
year
or
so
at
scale,
I,
don't
know
big
whatever
scale.
That's
the
big
and
it
worked.
Fine
could
have
been
cheaper
good
so,
but
that
helps
anecdotally
and.
E
E
So
we
can
proceed
with
sap
and
I
believe,
but
I'm
not
sure
about
the
HTTP
exchange,
HTTP
forwarder
extension,
so
that
one,
which
is
which
was
added
for
our
Forex
for
Splunk
needs,
and
we
are
not
sure
that
we
will
use
it
going
forward
because
we
need
to
like
that
need
will
go
away
at
some
point.
So
we
are
looking
for
some
evidence
that
this
particular
extension
is
being
used
by
someone
else
as
well
to
like
to
decide
yeah.
H
Great
so
just
to
just
I,
don't
know
regarding
you
mentioned,
you
know.
If
it's
a
vendor
component,
they
should
be
able
to
Market
stable
if
they
wanted
to
I.
Think
the
the
issue
that
had
originally
created
the
issue
that
Pablo
linked
to
was
a
vendor
component
right.
It
was
like
the
log
xio
exporter
who
had
changed,
configuration
and
broken
backwards,
compatibility
or
something
like
that
and
I
guess.
H
The
question
is
what
what
does
that
mean
for
The
Collector
contribute,
for
example,
where
like
we're
the
ones
publishing
these
components,
even
if
it
is
a
vendor
component
right
like
if
we
mark
it
as
if
we
accept
that
it's
marked
as
stable
and
there's
a
backwards
breaking
change?
Well,
if
we've
published
the
artifact,
doesn't
that
still
fall
on
us
for
breaking
end
users
deployments
and.
H
C
I
yeah,
basically,
what
you
said
is
my
concern:
I
was
taken
with
components:
our
Market,
stable
I
think
there
are
only
two
as
far
as
I
can
see
the
signal,
the
fixed
receiver
and
the
carbon
fever,
which
were
both
marked
in
a
PR
by
Alex,
so
I
think,
probably
like
that
happened
before
this
discussion
on
Civility,
and
maybe
one
can
argue
that
there's
not
really
any
component
that
we
have
decided
to
Mark
a
stable
after
thinking
about
this
breaking
change.
Issue.
J
Is
my
microphone
working
now?
Yes,
yes,
oh
nice,
wonderful,
okay,
yeah!
So
just
circling
back.
Nothing
should
be
stable.
If,
if
the
underlying
components
are
not
say
what
that
means,
if
the
component
is
using
TLS,
config
or
HTTP,
client
or
HTTP
server,
or
anything
like
that,
that
should
not.
That
component
should
not
be
marked
as
stable
right
so
because
those
configs
can
still
be
changed
and
if
those
are
not
stable
and
I
think
they're.
Not
then
nothing
that
depends
on
that
should
be
stable.
J
Now,
if
those
the
vendor
components
do
not
depend
on
anything
like
that,
if
they're,
which
is
very
unlikely
but
it's
possible,
then
then
I
would
say
it's
up
to
the
to
the
vendors
discussion,
whether
it's
his
team
or
not,
but
then
to
answer
Alex
question
I
think
it
was
someone
from
the
vendor
has
to
be
the
code
owner
and
has
to
be
responsible
for
fixing
Nations.
If
they
are
not,
then
the
complaint
will
then
eventually
be
marked
as
deprecated
and
then
unmobtained
or
I
think
only
unobtained.
M
It
doesn't
make
sense.
Thank
you
for
that
clarification.
I
can
definitely
comment
on
the
bugs
and
and
explain
that
so
I'll
have
that,
and
it
will
probably
also
give
us
a
timeline.
So
now
we
can
have
some
concrete
expectations
and
we
know
where
to
go
what
to
do
about
this.
Thank
you.
Everybody
really
appreciate
it.
M
It
looks
like
you
didn't
know
about
the
process.
I'm
sorry
I'm
still
standing
around
that
this
goes
through
elimination
by
a
maintainer.
Don't
have
to
take
action
here.
I
just
want
to
let
you
know
that
I'm
interested.
If
you,
if
there's
an
objection
or
is
there
specific
requirements,
please
let
me
know,
as
far
as
I
can
tell
I've
been
somewhat
active,
so
I
have
some
history.
J
I
think
the
only
requirement
is
to
be
a
member
of
open,
Telemetry
and
scqr.
If
you're
a
member
of
open
language
organization,
then-
and
if,
if
the
only
other
thing
that
you
need
is
a
is
a
maintainer
to
to
refer,
you,
then
from
what
I
from
what
I'm
concerned,
you
are
a
triage
already,
so
we
just
need
to
formalize
that
you
have
my
my
nomination.