►
From YouTube: 2022-07-28 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
B
C
B
C
B
A
A
I'm
trying
I'm
wondering.
C
A
A
No
one
has
topics
for
the
day.
Do
we
want
to
go
around
and
add
anything
I
want
to
talk
about,
or
do
we
want
to
just.
A
C
A
Not
here
reference
architecture,
working
example,
so.
A
So
there's
a
I
put
the
link
in
the
docs,
but
there's
a
demo
application
that
is
currently
in
flight.
That
is
at
least
a
reference
on
the
so
for
v1
it's
going
to
be
kind
of
a
more
of
an
instrumentation
reference
versus
a
collector
deployment
reference,
but
there's
a
goal
to
like.
I
think,
once
we
get
the
once
we
get
support
for
all
of
the
sdks
in
there,
then
we
can
kind
of
go
back
and
we
can
add
in
variations
that
focus
more
on
like
different,
collector,
setups
and
stuff.
A
But
this
is
it's
targeted
september.
A
A
Here's
a
working
sample
there
could
be
a
complete
subsection
about
the
demo
like
there
could
be
it's.
You
know
like
right
now.
If
you
go
to
the
site.
B
For
the
grpc
website
we
kind
of
kept
things
separate
per
language
and
then
linked
back
as
appropriate,
so
you'll
see
if
you
go
under
languages,
see
tutorials
or
guides
or
examples.
I
don't
know,
choose.
B
In
some
cases
for
some
languages,
so
you
see
under
learn
more
there's,
there's
an
external
link
to
under
learn
more
so
we
tried
to
get
a
consistent
setup
so
that
anybody
who's
using
a
particular
language.
Will
it's
a
one,
stop
shop
type
of
thing
yeah,
including
linking
to
examples.
A
A
B
A
B
But
independent
of
that,
if
you
go
back
to
the
website,
you'll
see
how
you
know:
we've
listed
all
the
languages
there
without
distinguishing
necessarily
which
ones
are
derived
from
the
core.
A
A
So
what
would
look
like
in
here
in
the
demo
is
there'll,
be
which
one
of
these
is
c
plus
plus.
B
A
Like
the
rust
the
service-
that's
written
in
rust
in
the
demo
repo
like
this,
so
you
could
have
some
I
mean
you
could
see
something
like.
If
you
go
to
rust,
there
would
be
a
a
link
out
to
the
demo
app,
but.
A
Here's
documentation
about
the
demo
from
the
if
you're
like,
using
this
as
an
end
user,
so
you're
someone
that's
coming
into
this
trying
to
like
understand,
open
telemetry
at
a
high
level,
or
you
want
to
know
about
all
the
different.
You
know
it's
like
here's,
the
different
features
and
here's
all
the
different
metrics
and
attributes,
and
things
like
that
that
the
demo
exposes
and
then
have
the
documentation
for
the
demo
be
more
focused
on.
A
People
developing
the
demo,
because,
right
now
it's
like
you,
know,
here's
this
big
list
of,
like
all
the
main
of
all
the
attributes
that
each
service
emits
you
know
like
this
is
useful
if
you're
trying
to
use
it
if
you're
trying
to
like
go
into
jager
or
if
you're,
trying
to
go
into
whatever
you
know,
honeycomb
or
pic
pic.
You
know
pick
whatever
thing:
you're
using
you're
trying
to
set
up
a
query.
Then
it's
like
okay,
here's
like
here's
the
stuff
I
can
query
on
so
that
was
kind
of
the.
C
Well,
the
demo
correct
me
if
I'm
wrong,
but
I
understood
that
the
demo
would
show
instrumenting
your
app
or
whatever
it
is
so
that
you
can
export
or
open
telemetry
data
in
an
open,
telemetry
format.
Is
it
also
showing
how
you
might
consume
that
data?
So
it
goes
a
lot
beyond
collecting
it?
C
I
guess
what
I'm
wondering
is
if
in
information
architecture-wise,
if
it
fits
in
under
collector
or
if
it's
almost
its
own
thing,
which
is
a
guidebook,
you
know
something
that's
sort
of
meta
and
apart.
A
A
You
know
it
comes
with
like
prometheus
and
jager
right
and
grafana,
so
that,
if
you
have,
you
know
that
you
should
be
able
to
check
this
out
and
run
this
stuff
and
get
kind
of
a
end-to-end
instrument
system
and
a
way
to
visualize
the
data
that
open
telemetry
is
producing,
so
it
it's
very
cross-cutting
because
it's
so
it's
showing
how
to
instrument
and
showing
sort
of
like
best
practices
for
instrumenting
or
what
we
think
like.
What
does
it
look
like
to
have
an
observable?
A
You
know
if
you're,
if
you're,
actually
instrumenting
a
real
production
system
like
what
are
the
attributes
that
make
sense
to
add
in
yourself.
You
know
like
what
what
is
open
to
them
to
provide,
and
then
what
are
you
as
the
programmer
as
the
code
purse
code
code
type
person
like?
What
do
you
need
to
add
to
get
value
out
of
open
telemetry
for
this
particular
type
of
application?
A
But
it
also
shows
stuff
about
the
collector
right.
It's
like.
Okay:
here's
how
you
like
here's,
how
you
would
deploy
the
collector
to
collect
this
stuff
or,
if
you're,
using
kubernetes
like
here's,
how
you
deploy
the
collector
to
also
get
metrics
about
your
kubernetes
cluster
and
add
labels.
You
know
add
attributes
to
the
telemetry
coming
from
the
process
itself,
so
it
is
kind
of
its
own
thing
because
it
is
so
cross-cutting.
A
A
A
A
So
philip
has
been
doing
some
research,
some
user
research-
about
how
people.
D
D
Is
like
I
what's
the
status
and
I'm
like
okay?
Well,
what
do
you
mean
they're
like
well,
I
went
to
the
status
page
and
I
got
nothing
and
what
they
actually
meant
was
like.
Well,
I
am
a
go
developer
and
I'm
trying
to
use
metrics
in
the
sdk,
and
I
don't
know
if,
like
it's
something
that
I
should
be
able
to
use
or
like,
if
I
can,
is
it
experimental
or
like?
Is
it
something
that
I
could
reasonably
put
in
production.
D
A
D
Spec,
oh
my
god,
is
that
oh.
D
D
A
A
B
A
Yeah,
that's
the
alpha
one.
A
But
this
is
something
that
we've
been.
I
mean
if
you
look
at,
if
you
kind
of
go
through
and
look
at
the
stuff,
there's
a
lot
of
things
from
you
know,
early
days
where
we
were
trying
to
figure
out
how
to
represent
status
on
the
website,
and
we
eventually
eventually
got
to
what
we
have
now,
which
is
this
idea
of
try.
You
know
you'd
have
each
signal
and
then
it
would
be
stable,
experimental
experiment
or
whatever,
or
these
can
be
stable,
beta
or
experimental.
I
think
or
stable
experimental
not
yet
implemented.
There's.
B
Could
you
go
to
the
dot.net
page
because
somebody
had
started
changing
things?
So
that's
one
example.
If
you
go
to
the
python
page
as
well,.
A
A
If
you
look
at
the
thing
here,
there's
actually
some
there's
like
one
plus
that
isn't
filled
in
for
python.
What
else
anything
else
so
python
actually
does
seem
like
it
is
pretty
good.
What
is
optional.
A
Anyway,
baggage
metrics
actually
yeah,
so
you
have
all
this
data
here,
and
the
thing
is
this
is
all
very.
This
is
all
like
the
very
low
level
specific
stuff
right
like
for
something
to
actually
be
stable
done,
whatever
it
does
need
to
meet
all
these
acceptance
criteria
and
there's
a
lot
of
them,
and
I
mean
actually
here's
a
perfect
example
like
so
tracing
is
mark
stable,
but
then
there's
like
resource
features
that
aren't
stable
or
that
aren't
implemented.
A
There's
propagators
that
aren't
implemented,
there's
environment
variables
that
are
like
not
implemented
right.
So,
if
you
looked
at
this
and
said
like
well,
is
it
actually
state,
like
you
know,
there's
kind
of
an
interpretation
gap
like
I
know,
philip
has
talked
about
the
I
guess
in
a
philosophy
with
this,
there
are
people
that
if
they
don't
see
stable,
1.0
whatever,
then
that
means
it's
not
going
they're
not
going
to
use
it
because
they
don't
want
to
change
it.
B
If
we
go
back
to
the
can
we
go
back
to
the
the
website
to
the
python
page,
for
example,
and
what's
what's
either
wrong
or
missing
here,
that's
not
satisfying
users
fit
it.
A
B
A
B
A
D
A
B
B
B
And
so
how?
How
do
you
think
we
can
help
inform
people
about
that?
The
second
issue
seemed
to
be
austin
that
you
mentioned
non-uniformity.
If
you
go
to
the
python
page.
B
What
do
you
consider
the
status
to
be
and
make
sure
if
you
say
it's
stable,
that
it
kind
of
conforms
to
what
the
spec
says
stable
is,
and
I
think
third
then
well.
I
would
rather
not
get
into
this
issue
of
trying
to
automate
the
figuring
out
what
the
status
is.
I
think
it
should
be
up
to
each
sig
to
say
this
is
what
we
think
the
status
is
unless
philip
says.
Well,
you
know
users
are
saying
the
status
is
bogus,
but
then
that
would
be
something
else
to
address.
D
Every
once
in
a
while
somebody
comes
up
with
like
oh
was
this.
One
thing
I
was
trying
to
do
is
like
blah
blah,
but
for
the
most
part
I
think,
there's
actually
a
pretty
good
like
when
a
when
a
sig
marks,
something
as
stable
or
alpha
or
beta.
It
tends
to
match
expectations
pretty
well
from
what
I
could
see,
which
is
good,
but
yeah.
We're
definitely
gonna
have
people
who
have
a
very
specific
thing:
they're
gonna,
try
that,
like
just
happens
to
not
be
implemented
and
they're
not
gonna,
be.
A
Can
we
libby
has
her
hand
up.
C
I
was
just
wondering
if
there
are
a
few
things
built
in
here
that
users
are
looking
for
and
if
we
can
maybe
tease
them
apart,
so
I
think
they're
looking
for,
or
at
least
what
I
would
be
looking
for
is
how
stable
is
is
the
code.
You
know
if
I
was
to
deploy
it
in
production.
Would
that
be
cool?
Or
would
I
be
really
risking?
My
my
end?
C
Users
experience,
I
think,
there's
another
aspect,
which
is
how
much
of
the
spec
has
been
implemented,
and
then
I
think,
there's
probably
a
third
aspect
that
it
sounds
like
users
are
looking
for
to
what
you
were
saying
patrice,
which
is
from
the
experts,
so
the
sig,
the
relevant
sig,
how
good
they
think
it
is.
A
Well,
thinking
back
to
the
what
was
let
me
open
up,
you
did
a
tree
said
a
pole
or
not
patrice.
Sorry,
philip
did
a
poll.
Do
you
mind
if
I
open
up
that
image.
D
D
A
D
B
I
am
okay.
I
think
you
raised
a
good
point
in
terms
of
how
individuals
probably
navigate
the
website.
Is
they
see
status
in
the
top
menu
bar
go
to
that
page
and
don't
see
any
link
to
language
say
status,
so
maybe
that's
what
we
should
do
is
somewhere
towards
the
top
of
that
page
say
hey
by
the
way.
B
D
Think
that
would
be
good
or
just
something
that
like
yeah,
we
could
figure
out
the
right
solution
there
yeah,
but
two
other
things
come
to
mind.
The
first
is
there:
there
is
a
collector
status
and,
to
my
knowledge,
there's
no
like
good
place
for
that
there
we'll
want
to
make
sure
that
that
is
in
a
good
place.
Yes,
it
is
an
intro
thing
that
there's
nothing
here
about
the
status,
which
is
still
technically
not
ga.
D
The
second
one
is
the
spec
maintains
this
notion
of
spec
sdk,
and
that
has
a
status
and
that's
different
from
a
language
sdk
and
that's
highly
confusing.
I
think
we
might
want
to
figure
out
a
way
to
wordsmith
that,
like
when
the
word
sdk
is
used.
In
this
context,
it
actually
means
like
the
specification
of
an
sdk,
not
like
actual
sdks.
A
A
C
I
was
wondering
if
perhaps
the
status
page
should
almost
be
a
dashboard
with
you
know:
maybe
not
red,
green
yellow,
but
something
like
that
for
every
single
item
so
that
visually.
If
you're,
looking
for
the
rust
sdk,
you
know
it's
towards
the
bottom
of
the
page
on
the
right
and
you
can
just
go
check
there
every
time,
but
each
can
be
a
link
to
the
relevant
section
in
the
documentation.
A
Yeah
or
even
even
cards,
right
like
what,
if
this
was
just
a
bunch
of
cards
like
so
at
the
top,
you
had
each
language
as
a
card
and
then
you
click
it
and
it
takes
you
to
like.
I
don't
necessarily
want
to
get
into
the.
I
don't
want
us.
I
feel
like
there's
a
variety
of
more
and
less
complex
ways
you
can
solve
this,
but
I
feel
like.
B
A
B
C
I
mean
I,
I
agree
status,
it's
something
that
that
you
would
want
to
find
when
you,
when
you
come
to
the
website,
but
I
think
there's
been
a
bit
of
a
bleed,
at
least
in
the
documentation
where
status
and
other
pieces
of
information
are
included.
D
Also,
this
is
just
totally
supplementary
info,
but
looking
at
the
the
public
data
set,
the
status
page
in
terms
of
like
number
of
hits
is
actually
pretty
high
over
the
past
28
days,
it's
like
almost
8
000
yeah.
A
It's
also
top
level.
I
do.
I
think
that
it's
good,
I
mean
this
was
kind
of
my
my
thought
right
like
if
each
of
these
is
a
link,
you
know
you
just
have
a
little
card
with
the
logo,
the
language
logo,
and
then
it
takes
you
directly
into
where
you
want
to
go.
I
don't
like
you
could,
probably
group,
maybe
you
don't
have
to
split
collector
out.
You
can
just
have
sdk
and
then,
like.
A
A
D
A
A
I
think
for
the
status
page,
but
what
this
ends
up
looking
like
is
dot
net
c
sharp
is
like
stable
for
everything,
because
the
way
they
did
it
and
then
everyone
else
looks
like
they're,
not
which
I
don't
think
is
actually
an
accurate
representation
of
maturity.
D
A
But
I
also
think
that
it's
one
of
those
things
that
it
like
it
only
counts,
because
none
of
the
ergonomics
are
the
same
and
none
of
the
developer
experience
is
completely
different.
Like
it's
literal,
you
know.
If
you
go
into
net
and
you
say
like:
oh
I'm
going
to
use
open,
temptation.net
like
what
that
means
is
you're
using
system.diagnostics
and
you're
just
telling
it
export
open,
telemetry
format
right
like
yes,
it's
cool
that
they
did
that.
But
it's
not.
A
Or
even
just
the
like
on
itself
like,
but
I
I
feel
like
if
we
think
of
the
status
page
as
less
a.
I
want
to
go
here,
and
I
want
to
see
the
status
of
the
entire
project
or
some
slice
of
the
project
and
more.
A
C
A
And
that
that
kind
of
solves,
I
feel
like
the
biggest
problems,
which
are
it's
hard.
If
you
don't
know
where
to
go,
it's
hard
to
find
the
status
you
care
about,
and
it
lets
us
have
a
simpler
answer.
I
feel
like
for
maintainers
to
be
like
hey
as
long
as
you
make
sure
all
you
have
to
care
about
as
a
maintainer.
Is
your
status
right
here
like
on
your
page
on
your
your
index
page.
C
A
A
Update
six
updated
language
success
section
add
collector
status,
collect
page
professor
sas
page
respect,
sas
elsewhere,
yeah,
I'm.
I
feel
okay
about
that
thumbs
up.
B
So
in
terms
of
reworking
the
individual
language
pages,
are
we
agreeing
to
use
the
python
page
as
the
template
template
right.
A
B
Remember
it
somebody
else
had
proposed
they,
philip,
I
don't
know
if
your
members
somebody
had
started
wanting
to
update
each
of
the
git
repos
for
each
of
the
language
stakes
to
include
some
sort
of
status
info,
and
I
remember
pick
you
that
there
were
peculiarities
in
some
cases
for
some
languages
anyways.
We
can
always
reformat
things
as
fidep
said.
A
A
D
A
B
Think
it
it
may
be
that
I
don't
know
if
it
could
ever
happen,
that
the
statuses
would
be
different
for
one
or
the
other,
in
which
case
then
that
top
level
entry,
let's
say
traces,
you
get
two
sub
entries
one
for
api
and
then
what
the
status
is
one
for
sdk
and
with
the
state.
I.
D
I
would
probably
make
the
case
that
for
most
end
users,
that
kind
of
thing
is
probably
not
valuable
to
distinguish
because,
like
if
your
means
of
using
the
api
is
still
not
stable,
then
like
okay,
the
fact
the
api
is
stable
doesn't
really
help.
You
like
it
could
be
helpful
for
instrumentation
library
authors,
but
I
feel,
like
those
people
at
that
point
would
probably
be
looking
at
a
github
repo
for
the
the
thing
a
lot
more
than
they'd
be
looking
at
dogs.
D
Yeah,
maybe
yeah-
I
I
I
don't
know
if
this
like
the
content
here
is
necessarily
like
actively
harming
people,
so
I'm
less
opinionated
about
that.
A
Yeah
I
want
to
be,
I
have
another
meaning
by
the
way,
and
I
want
to
be
respectful
of
everyone's
time.
Thank
you.
This
is
a
really
good
conversation,
if
you
aren't
already
in
hotel,
comms
on
slack
feel
free
to
join,
and
we
can
continue
this
discussion
there,
but
I
feel
like
we
have
a
path
forward
for
this,
and
I
think
this
will
be
a
really
great
thing
to
kind
of
get
landed
here
in
the
next
short
little.
While
it
sounds.