►
From YouTube: 2022-10-20 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
A
C
A
D
D
A
Okay,
yeah
I
think
we
can
get
started
then
a
few
items
on
the
agenda.
If
anybody
else
likes
to
add
something
and
just
do
it
while
we
go
through
it
right
yeah,
let
me
I
think
you
added
in
the
first
one
so.
D
Yeah
I
I
just
noticed
since
I
started
doing
PR's.
That
was
a
number
of
terms
that
come
up
that
we
seem
to
repeatedly
have
to
tell
people
hey.
We
do
it
this
way,
like
you
spell
open,
telemetry
this
way,
capital,
O,
Etc,
and
just
wonder
if,
if
we're
at
the
point
now,
where
it'd
be
worth
creating
some
sort
of
guidance
page,
so
we
can
just
refer
people
to.
E
B
D
I,
don't
know
that
we
have
fully
fledged
standards
to
the
point
that
people
who
actually
do
this
for
a
living
would
appreciate
it
Patrice,
but
I'm.
Just
thinking
so
I'm
having
to
you
know,
save
having
to
keep
a
text
file
open,
so
I
can
just
copy
paste,
which
I'm
sure
others
are
doing
too.
Maybe
it's
worth
it.
B
I
was
going
to
say
to
start
off,
we
could
just
add
a
section
to
the
contributing
file.
How
does
that
sound?
B
If
we've
content
that
we
already
want
to
add,
we
know
we
want
to
add,
then
that
file
is
is
kind
of
there
for
that
purpose.
Although
the
present
time
it's
pretty
much
saying
how
to
build
the
site,
but
there
could
be
a
section
for
writing
guidelines
just
an
idea.
It
could
be
somewhere
else
as
well.
Are.
A
Maybe
something
like
from
I
mean,
even
if
we
say
like
I
mean
we
had
this
discussion
around
like
hey,
how
to
write
front-end
back-end,
whatever
I'm
just
wondering,
maybe
there's
even
something
from
from
the
cncs.
We
could
source
that
we
say
like
hey
if
you're
unsure.
This
is
how
the
cncf
is
doing
it,
so
we
just
so
like
just
taking
this
mentally
often
and
not
having
a
five-page
discussion
around
writing
certain
words
right
I
mean
this
is
important,
of
course,
for
words
like
open,
Telemetry
or
Auto
instrumentation.
A
B
B
But
if,
as
you
mentioned,
there
may
be
words
that
are
quite
a
bit
more
particular
to
open
to
this
website
to
Hotel,
if
we
can
only
start
with
that
to
say
if
you're
spelling
this
is
the
the
recognized
spelling
for
this
word
always
use
Ivan
or
don't
use
a
hyphen
or
whatever
I'm.
Assuming
that
list
will
not
be
very
long
and
we
could
just
add
it
to
contributing
for
now.
At
least
you
know
it
exists,
there's
a
place.
D
Yeah,
the
other
thing
that
is,
that
triggered
my
mind
on
this
was
styling
like
particularly
in
the
documentation,
not
not
the
rest
of
the
site.
D
A
You
mean
like
when
to
use
bold
courses.
B
So
when
I
came
into
the
project,
I
was
hoping
to
establish
some
uniformity
there
and
I've
tended
to
follow.
Google
has
a
writing
style
guide
for
technical
documentation
and
it's
the
one
that
I've
come
to
follow
because
I
I
like
it
it's
it's,
it's
modern
and
it's
simple,
so
informally.
B
That's
what
I've
been
following
I
noticed
that
if
we
just
take
the
capitalization
of
headings
and
and
titles
in
open
Telemetry
which
arose
in
part
from
open
tracing,
there
was
a
certain
convention
that
was
established
and
and
I
just
decided
not
to
go
there.
So
to
answer
your
question,
informally
I've
been
following
Google's
technical
writing
style
guide.
B
We
could
follow
that
and
say
yeah
and
and
describe
deltas,
but
that
it
that's
going
to
be
a
considerable
effort
to
apply
uniformly
and
to
try
and
and
force
and
at
this
speed
that
things
are
moving
now
and
there's
more
and
more
contributions
and,
in
particular
with
blogs.
At
some
point,
I
I
gave
in
and
said
it's
more
important
for
content
to
get
published
than
to
try
and
apply
style
guides.
But
that's
just
my
take
at
this
point
because
I
found
I
found
it
overwhelming
no.
D
I
I
agree,
I,
think
I.
Think
I
like
the
idea
of
starting
with
something
in
the
contributing
file
and
like
with
everything
we
can
always
review
and
change
it,
not
suggesting
that
once
we
decide
on
a
style,
we
then
go
through
every
single
page
and
make
sure
it
conforms.
But
perhaps
starting
as
we
mean
to
go
would
be
a
good
idea
and
then
we
can
maybe
even
do
things
like
in
the
pr
template
say:
hey
have
you
checked
that
you've
followed
our
best
practices
sounds.
B
D
B
B
A
A
I
I
mean
I'm
completely
with
you
on
that
that
content
throughput
is,
is
much
more
important
than
having
discussions
around
ad
character
line
length
and
if
everything
is
in
tactics
and
written
in
the
right
form,
right
I
mean
yeah,
yep,
yeah,
but
but
I
think
I
think
we.
We
should
really
do
it.
That
way
that
that
we
put
it
at
some
places.
So
if
people
are
uncertain,
then
they
can
look
it
up
and
and
because
right
now
I
mean
it's
true.
A
We
don't
give
any
guidance,
so
people
just
do
what
they
want
to
do
and
then
that's
that's
even
worse.
Right
I
mean,
and
at
least
they,
let's
call
it
and
best
effort
style
guide
right.
So
that's
that
that's
what
we
should
try
to
accomplish.
Does
this
make
sense.
A
No
okay,
cool
a
very
important
topic
is
that
we
had
this.
This
chat
I
think
already
a
few
weeks
back,
that
we
want
to
have
good
document,
not
at
least
documentation
around
metrics
for
kubecon
right
I
mean
there
were
announced
big
last
year
and
discussed
speak
last
year
at
kubecon,
and
so
it
would
be
good
now
to
let's
say,
not
only
deliver
the
code
but
also
the
documentation
so
Philip
and
I.
We
we
attended
the
maintainers
meeting
on
Monday
we
reached
out
to
to
everybody.
A
There
are
a
few
language
seeks
that
are
just
not
there
yet
so
where
we
said
like
yeah.
Let's
not
put
something
in
that
will
then
change
at
some
point
so
go
especially
called
out
that
they
don't
want
to
do
that.
We
mentioned
the
same
to
the
Swift
Team,
but
I
think
we
are
now
in
a
really
good
state
with
Java
Python
and
C,
plus
plus
node.json.net
are
still
I.
Think
there's
something
pending
for
net
and
I.
Think
the
node.js
team
has
a
few
people
that
wanted
to
write
something.
A
So,
let's
see,
maybe
we
can
get
something
by
Monday.
If
not,
if
anybody
of
you
says
like
hey
and
I
still
have
a
little
bit
of
a
bandwidth
for
for
looking
into
that,
just
just
raise
your
hand
during
or
after
this
meeting,
and
we
can
see
how
to
to
get
this
through
so
just
to
maybe
call
out
what
I
mean
by
yeah
having
some
documentation
and
it
does
not
have
to
be
perfect.
So,
for
example,
I
added
a
little
bit
of
stuff
in
the
C
plus
plus
documentation
right.
A
So
there's
just
like
hey.
This
is
how
you
initialize
things.
This
is
how
you
create
certain
metrics.
It's
not
really
going
much
into
detail,
and
it's
really
not
in
a
state
where
you
say
like
hey,
just,
can
copy
this
out
and
it's
it's
working
without
any
issues,
but
but
it's
it's,
let's
say
giving
people
an
idea
of
what
is
possible
and
then,
for
example,
here's
this
metrics
example
where
they
can
try
out
things
and
yeah.
We
are
in
different
states
on
that.
A
I
think
python
is
quite
good,
so
they
have
it
here
in
the
manual
section
as
well,
and
they
give
a
very
specific
example
what
you
can
do.
So
you
also
can
are
told
how
to
initialize
the
the
meter
provider
and
those
things
so
yeah
I
said
any
any
any
thoughts
on
that.
Any
comments
on
that
on
metrics.
A
Yes,
no,
maybe
okay
I
said
if
anybody
knows
someone
who
wants
to
to
do
something
there
yeah
another
topic,
we
discussed
very
lengthy
last
week.
Let
me
bring
it
up
again
two
weeks
ago
and
I
got
distracted
from
that
a
little
bit
because
I
looked
into
this
metric
stuff.
Is
this
Persona
based
new
landing
page
for
for
the
website
right?
A
So
it
was
the
health
of
Patrice,
I
I
restructured
it
again,
a
little
bit
so
I
just
wanted
to
show
this
to
you
again,
I
think
we
should
not
spend
another
full
half
meeting
on
discussing
that,
but
let's
just
quickly
look
into
that.
So
the
idea
is
to
replace
those
language
buttons
here
with
something
more
obvious
where
we
can
say
like
hey,
get
started
based
on
your
role
and
then
you
can
say
something:
hey
I'm
a
developer,
so
what
I
want
to
do
is
I
want
to
add
open
Telemetry
into
my
code.
So
what?
A
A
I
want
to
accomplish
here
are
the
certain
tasks
you
need
to
do
for
that,
or
can
do
for
that
and
then
the
same
for
Ops
where
or
someone
goes,
maybe
a
different
way
where
they
say
like
hey,
actually,
I,
don't
want
to
touch
the
code
so
I'm
more
caring
about
having
some
some
means
of
automatically
instrumenting
it
how
to
set
up
the
collector,
maybe
even
how
to
do
something
with
the
operator,
etc,
etc.
A
So,
yeah
this
this
is
still
pending
and
last
time,
I
think
we
spent
a
lot
of
time
on
discussing
this
year.
So
I
don't
know
if
this
is
already
in
a
good
shape
or
if
there's
any
major
complaints
around
how
this
looks
like
foreign.
B
Having
worked
with
you
on
the
landing
page,
I
I,
like
it
I
like
this
role-based
approach,
I,
don't
have
enough
of
a
domain
expertise
to
to
figure
out
if
you
know
for
all
the
rest
of
you.
If
this
appeals
to
you
as
well,
I
think
I'd
I'd
vote
for
for
this
to
the
land,
I
think
it's
an
improvement.
D
Yeah
I,
I,
agree,
I,
think
it's
an
improvement.
I
think
you
know.
Following
the
80
20
rule,
we
could
start
to
dig
into
details
and
and
have
long
conversations
over
whether
it's
100
perfect,
but
it's
still
much
better
than
it
was
and
I
think
we
should
go
forward.
F
A
Should
we
push
this
out
before
kubecon
to
to
get
some
good
feedback
on
it,
or
should
we
do
it
after
kubecon
to
not
distract
people
too
much
okay
go
for
it,
yeah
yeah
then
I
will
yeah.
Then
then
I
will
maybe
not
right
now,
but
then
maybe
tomorrow,
I
will
just
merge
it
in
I
think
I.
We
have
all
the
approvals
right.
So
here's
again
the
link.
A
If
anybody
wants
to
give
it
a
last
few
and
maybe
it's
missing
or
finding
any
any
minor,
typos
or
whatever,
then
then,
let
me
know
otherwise:
I
will
merge
it
in
tomorrow.
My
morning,
hours
yeah
I
had.
B
A
E
Yes,
I
just
wanted
to
mention
that
we
can
take
over
adding
peers
to
make
changes.
Sever
and
I
noticed
that
you
have
been
kind
of
taking
over
that,
and
you
know
I
definitely
don't
want
you
to
feel
like
you
are
responsible
for
making
those
changes.
E
I
think
I
think
Libby
and
I
added
some
comments
to
the
newest
iteration
of
that
page,
which
I
think
we
both
think
look
good.
Just
a
couple
other
things
to
add,
but,
like
I
said,
we
can
submit
peers
for
that
and
then
someone
I
had
a
question
so
for
the
metrics
for
the
remaining
ones,
node
and
Dot
net.
Are
you
wanting
to
get
those
done
by
Wednesday
like
out
of
kubecon
or.
A
Ideally
I
mean
if,
if,
if
we
could
pull
this
off
this,
this
would
be
great
I
mean,
if
not
that's
I
also
have
only
parcel
control
over
that.
But
ideally
it
would
be
good
if,
if
at
least
the
languages
that
are
in
a
ga
or
RC
State
have
something
on
metrics
right.
A
So
maybe
just
to
give
you
a
very
simple
example,
so
Java
has
something
on
on
it
already,
but
but
a
minor
change
I
did
for
Java
and
python
as
well
is
the
following
that
when
you
go
into
the
getting
started
that
right
now,
you
not
only
get
traces
right.
You
go
also
get
metrics
out
of
the
box
now
right,
because
grpc
is
Auto
instrumented
and
it's
not
only
emitting
traces.
It's
also
emitting
metrics
out
of
the
box
right,
and
this
is
already
a
big
win.
A
I
mean
that's,
that's
not
even
telling
you
like
hey
how
to
get
metrics
yourself,
but
it's
telling
you
for
certain
Frameworks.
You
just
get
it
out
of
the
box.
I'm
not
sure
if
this
is
true
for
net
or
JavaScript
in
any
in
any
place,
but
but
for
example,
in
the
in
the
JavaScript
getting
started.
If
there's
anything
where
metrics
are
are
Auto
generated
for
something
like
Express,
then
then
this
would
be
already
a
great
start
to
say
like
hey
by
the
way,
this
is
also
dumping
out
metrics
for
you
right.
E
Okay,
yeah
I'm
interested
in
working
on
the.net,
metrics,
stuff
yeah,
but
I
was
curious
about
the
timeline.
Okay,
yeah.
A
I
I
was
I
was
reaching
out
to
the
dotnet
folks
and
they
said
they
have
something
so
on
the.net
Repro.
There
is
already
something
on
metrics
documented,
so
maybe
it's
also
only
just
something
that
needs
to
be
copied
over.
I
can
share
this
later
on
the
on
this
Slack.
E
A
A
A
C
Hey
new
friends
and
familiar
friends:
I'm
Jason
I
work
a
lot
on
Java
but
kind
of
recently,
I've
also
put
in
a
spec
change
to
deprecate
the
Jaeger
tracing
exporters,
and
that's
because
Jaeger
supports
otlp
natively
now
and
Yuri,
who
is
the
founder
of
Jaeger
tracing,
is
also
on
board
with
us.
But
we
need
to
go
through
a
deprecation
phase
and
the
question
of
how
long
do
we
have
it
be?
C
Deprecated
came
up
and
one
idea
was
to
ask,
via
a
tweet
for
people
to
chime
in
on
this
issue
and
another
I
think
Severance
suggested.
Maybe
a
blog
post
could
be
helpful
here.
So
I
just
wanted
to
like
re-re-up
that
topic
and
see
where
we're
at
and
if
there's
anything
actionable
that
I
can
do
to
help
that
along.
A
So
I
I'm
still
for
for
the
idea
too,
to
have
something
like
a
blog
post
or
something
like
that.
I
mean
the
the
Jager
exporters,
also
in
some
places,
still
in
the
docks
right.
So
we
have
them
here
and
there
no
I
think
yeah,
not
in
JavaScript
anymore,
as
think.net
head
still
has
it
yeah.
So
no,
it
also
does
not
have
it
anymore
anyway.
Some
some
of
them
have
it
I'm,
not
sure.
A
If
this
is
also
something
we
should
do
to
to
remove
them
one
by
one
and
only
call
out
Jager
SSD
sdotelp
back
end
right,
yeah
but
yeah
I
said
I
mean
if,
if
a
blog
post,
saying
like
hey
Jager,
is
now
doing
otil
p
as
their
default
and
and
that's
why
we
encourage
you
to
do
this
and
here's
a
survey
and
whatever
I
I
would
yeah
I
would
think
that
the
hotel
block
would
be
a
great
place
for
them.
A
Yeah
there's
in
the
open
Telemetry
I
o
directly
on
the
readme
to
see
how
to
submit
a
blog
post.
Oh
perfect,
okay,
yeah,
cool.
C
D
So
there's
a
a
working
group
focused
on
end
user
research.
Rhys
is
actually
one
of
the
co-ledos
of
it.
Yeah
I
would
I
would
ask
Grace
your
question
specifically
since
she's
here.
C
C
C
Yep
yeah,
because
if
people
say
look,
it's
a
real
big
pain,
it's
it's!
It's
really
difficult
for
us!
We're
like
we
already
require
so
much.
We
can't
switch
then
we'll
give
people
like
I,
don't
know
18
months
or
something
right,
but
if,
like
only
if
everybody's,
like
no
it's
easy,
we'll
just
switch
no
big
deal,
then
we'll
give
them
six
months
right.
D
I
mean
if
we
have
just
off
the
top
of
my
head.
If
we
have
one
of
those
survey
tools
like
the
online
quiz
tools,
we
could
put
one
up
and
eventually
put
it
on
the
banner
of
open,
telemetry.io
and
just
get
people
to
answer
that
we
could
probably
do
the
same
sort
of
thing
in
Twitter.
D
E
E
C
F
Perfect,
this
might
also
be
a
question
for
the
Jaeger
community.
You
know
like
that
that
Community
and
their
ecosystem
really.
C
F
Give
you
some
insight
into
to
what
the
current
usage
is.
Yeah.
C
A
C
C
F
B
Just
by
the
way
on
the
in
the
website,
repo,
we
do
have
discussions
activated
so
I
mean
that's
what
it's
therefore
to
get
feedback,
and
maybe,
if
you
want
to
start
a
discussion
thread
rather
than
a
an
issue
or
both
yeah.
C
A
Lead
them
to
that
issue,
and
then
they
can
comment
there
and
then
still
if
people
say
like
hey
for
whatever
reason
in
the
universe,
they
don't
have
a
GitHub
account
and
they
can
still
I,
don't
know,
send
you
an
email
or
whatever
cool,
yeah,
okay
golden.
Then
we
expect
to
to
get
some
some
blog
posts
from
you
and
then
we
can
see
I
think
it
would
be
great
to
to
have
it
published
during
kubecon
I
assume
or
do
you
have
any
timeline
for.
C
E
Okay,
so
it
looks
like
the
next
one
is
me
again
and
this
morning's
end
user
discussion
group.
One
of
the
end
users
brought
up
a
question
around
whether
we
have
any
plans
to
document
things
like
when
basically
use
cases
on
auto
versus
male
instrumentation.
E
They
brought
us
specifically
that
some
Auto
instrumentation
libraries
like
GRE
is
super
chatty
and
we're
just
curious
around
where
they
might
be
able
to
find
information
on
like
when
to
use
manual
versus
mental
used.
Auto.
C
I
think
in
the
case
of
java,
it's
even
worse
because
we
also
have
Library
instrumentation
so
that
came
up
earlier.
Well,
I
brought
it
up
earlier
in
the
JavaScript
meeting,
because
I'm
curious,
if
users
even
understand
what
library
instrumentation
is
and
how
that
really
might
relate
to
manual
or
automatic
instrumentation
foreign.
That's
not
helpful.
It's
just
a
data
point
yeah.
F
But
for
for
what
it's
worth,
Reese
I
spent
a
couple
years,
trying
to
Define
a
difference
between
manual
and
auto
instrumentation,
in,
like
the
light
step,
documentation
and
like
working
with
our
users
and
I,
eventually
concluded
that
this.
F
This
just
added
confusion
to
people
to
try
to
to
essentially
use
terminology
and
like
categorizing
these
things
and
with
our
own
docs
I,
switched
to
an
approach
of
just
just
talking
about
installation
right
for
each
language,
there's
one
or
two
different
ways
to
install
all
the
stuff,
and
there
isn't,
generally
speaking,
like
a
universal
way
to
install
stuff.
That's
the
the
one
part
of
open
Telemetry,
where
it
just
really
really
varies
by
language
and
so
rather
than
having
like
generalized
turns
like
like
with
our
own
docs.
F
We
just
switched
to
to
just
being
like
you're
trying
to
install
in
Python.
You
can
like
use
these
like
help
helpful,
like
Auto,
installer
things
or
here's
how
you
do
it.
You
know
by
hand
if
it's
Java,
we
only
talk
about
the
agent
pretty
much,
because
we
couldn't
find
a
user
who
wanted
to
do
it
any
other
way.
F
So
I
don't
know
how
helpful
that
is.
But
that's
just
a
thing
I
found
is
that
it's
it's
so
language
specific
that
that
approach
seemed
to
be
a
little
less
confusing.
E
A
Yeah
I
I
mean
maybe
to
Addison
just
this
Persona
thing.
I
wanted
to
introduce
this
also
touches
kind
of
onto
this
topic
right
because,
if
you're
more
an
Ops
person,
you
don't
really
care
about
manual.
Instrumentation
right
I
mean
the
code,
for
you
is
a
black
box.
Maybe
you
have
to
maintain
300
applications.
So
the
thing
you
want
to
accomplish
is
get
observability
out
of
this
black
box
and
maybe
you're
lucky
and
your
developer
took
care
of
that.
A
Maybe
not,
and
then
you
have
to
find
a
way
to
chat
it
in
and
then
it's
of
course
different
for
a
developer
and
there
at
least
this
is
my
point
of
view
they
like
to
have,
of
course,
something
automatic
to
get
started
right.
This
is
awesome
to
throw
something
in
like
in
like
in
Java
agent
or
whatever,
so
is
it
is
what
they
maybe
are
used
to
from
from
an
APM
vendor.
They
just
throw
it
in
and
then
Telemetry
drops
out
and
they're
happy
with
that.
A
But
then
over
time
they
realize,
like
hey
I,
mean,
as
you
said
like.
Maybe
it's
too
chatty,
maybe
it's
not
giving
the
right
information
and
then
they
can,
let's
say
peel
the
onion
and
see
like
oh
I
can
do
this
manually.
Oh
I
can
go
there
go
go
down
there
step
by
step,
but
I
mean,
as
as
Tad
said,
this
is
a
a
story.
I
think.
As
all
this
open
Telemetry
itself
right,
I
mean
I.
A
I
kicked
off
another
lengthy
discussion
on
that
myself,
just
a
while
back
and
I
think
this
also
led
to
to
a
few
discussions,
but
at
the
end,
where
I'm
standing
right
now,
I
mean
what
I
want
to
have
for
the
dogs,
and
we
have
it
now.
Here's
the
issue
as
well.
A
It's
exactly
something
like
that
where
you
say
like
hey
dear:
this
is
for
the
developer:
hey
here's
something
all
in
one
you
can
use
like
a
Java
agent
or
a
python
distro
or
whatever,
and
then
they
can
go
down
step
by
step
and
say
like
oh
actually,
I
don't
need!
All
of
that.
I
can,
let's
say
deconstruct
this
step
by
step
and
maybe
at
the
end,
to
do
a
little
bit
more
manual
stuff.
A
This
is
a
long
way
to
go
and
I
think
the
the
one
thing
that
that
you
may
be
brought
up
additionally,
that
maybe
needs
to
be
somewhere
there.
Why
should
I
not
use
automatic
instrumentation
for
everything
right
I
mean
that's,
maybe
definitely
something
that's
missing
here
then,
then,
at
the
lighter
steps,
where
you
say
like
hey
by
the
way,
adding
in
something
like
a
Java
agent
just
adds
overhead
or
it
just
adds
a
lot
of
things.
You
maybe
don't
need
for
your
microservice.
A
It's
called
a
microservice
for
a
reason
right.
So,
let's,
let's
figure
out
a
way
how
to
to
make
it
easier.
So
so
that's
that's
how
I
try
to
get
around
it
somehow,
but
it's
a
long
way
to
go.
E
F
F
Simpler
right,
like
I'd
like
to
this
year,
kick
off
an
effort
to
add
like
a
configuration
file
and
and
like
if
you
aren't
trying
to
install
a
bunch
of
custom
plugins
for
a
bunch
of
different
providers
or
do
something
fancy
you
just
want
basic
open
Telemetry
with
you
know,
w3c,
headers
and
otlp
exporter,
and
all
of
that
there
should
be
like
a
a
much
simpler
way
to
install
that
in
a
lot
of
languages
like
we
don't
we
don't
have
that
yet,
but
if
for
each
language
we
kind
of
picked,
this
is
the
primary
way
we
expect
people
to
to
do
it
in
this
language.
F
D
F
You
could
just
be
like
Step
One
is
use
the
Java
agent
and
there
might
be
some
subset
of
users,
so
it's
like
for
some
reason
that
doesn't
work
and
here's
some
alternate
docs
for
what
you
can
do,
but
if
we
just
decide
to
go
hard
like
for
Java,
it's
the
agent.
This
is
what
you
should
do
for
python.
It's
like
there's
like.
F
F
Help
clarify
the
docs
if
we
just
just
pick
one
of
these
paths,
for
each
language
and
and
just
try
to
make
sure
the
docs
for
that
one
path
are
very
front
and
center
yep.
F
Because
I
found
Yeah
in
our
own
docs,
it
was
like
the
choice
that
confused
people
like
how
do
I
started
like
you
got
a
couple
options
here
and
they're
like
I
I,
don't
want
to
learn
I
just
want
to
copy
paste
something
and
have
it
work.
Yeah.
E
Okay,
we
do
have
the
getting
started:
docs
I'm,
not
sure
where
those
are
what
the
state
of
those
are.
A
A
It's
also
not
consistent
if
it's
real
Auto
instrumentation
in
the
sense
of
yeah,
don't
add
a
single
line
of
code
or
if
it's
a
mix
of
that
or
is
its
pure
manual
instrumentation,
but
this
has
also
to
do
with
the
state
of
where
things
are
so
PHP.
For
example,
I
mean
they're
still
early,
but
they
are
also
now
thinking
about
providing
some
some
ways
of
of
Auto
instrumentation.
A
So
yeah
we
have
a
bunch
of
getting
started,
but
they
said
it's
not
it's
not
consistent
across
languages.
I
think,
let
me
see,
I
think
go,
for
example,
is,
is
extremely
complicated
right
because
they
go
down
the
path
of
doing
manual
instrumentation,
which
makes
sense
yeah.
So
so
it's
much
more
complicated
in
Java
I
mean
Patrice.
A
You
you
created
this
one
and
I
I
made
it
myself
a
competition
with
myself
how
quickly
I
can
can
run
through
that
I
think
you
can
do
this
in
less
than
two
minutes
and
then
have
something
up
and
running
and
and
yeah
I
think
at
the
end.
What
we
want
this
this
is
at
least
something
I
have
we
have
talked
before
is,
for
example,
in
Python.
A
We
have
this
roll
the
dice
web
app
right
and
at
the
end,
what
would
be
really
cool
if
we
would
just
have
an
instance
of
this
application
in
every
language
and
then
you
can
just
go
ahead
and
say,
like
Okay,
add
tracing,
add
metrics
ad
locks
and
then
here's
some
additional
things.
How
you
can
do
research
detection
whatever,
but
yeah,
there's
there's
a
long
way
to
go.
A
F
I'll
try
to
think
about
how
to
help
here,
because
definitely
the
thing
we
have
here
is
like
literally
I
used
to
have
this
in
our
docs,
where
you
have
like
automatic
and
manual
and
users.
It's
like.
Is
this
a
car
like
what
I
like
I?
Don't
like
I,
don't
know
what
the
I
don't
really
know,
what
the
sidebar
means,
even
and
so
I
can
I
can
try
to
help
think
about
like
what,
what's
a
a
little
bit
better
way
of
rearranging
the
information.
It's.
F
Of
the
information
is
bad.
It's
just
this
kind
of
layout
leads
to
the
user
having
to
like
kind
of
learn
a
whole
bunch
of
stuff
before
they
can
even
kind
of
like
understand
the
the
lay
of
the
land
yeah.
F
Has
one
of
the
highest
success
rates,
at
least
when
we've
measured
it
because,
even
though
it's
all
manual
instrumentation,
it's
extremely
normal,
for
what
go
developers
do
right?
There
is
no
like
agent
concept.
All
you
do
and
go
is
like
look
up
the
godox
for
something,
and
you
know
copy
paste,
the
the
installation
stuff.
So
the
instructions
like
go,
install
all
these
packages
and
set
them
up
is
like
it
looks
like
a
lot
of
work,
but
it's
not
weird
for
go.
Developers
res.
D
F
E
It
does
I'm
curious
about
how
you
measure
like
this.
The
success
but
I
know
we're
running
up
on
time
and
yeah
yeah.
Let's.
A
Let's,
maybe,
let's
maybe
use
the
last
five
minutes
for
two
more
topics,
so
yeah
I
think
I'm.
Just
this
is
a
topic
we
could
spend
a
few
hours
on
it
and
not
get
through
to
real
conclusion,
but
I
think
that
the
point
I
wanted
to
make
is
like
we
have
a
few
things
in
Motion
in
the
docs.
We
want
to
do
to
improve
on
that,
but
you're
completely
right
this.
This
is
one
of
the
major
things.
What
confuses
people
a
lot
so
yeah,
let's,
let's
figure
something
out:
okay,
yeah!
A
D
It
may
be
a
longer
discussion
than
we
have
time
for,
but
tangentially
to
that
this
meeting
starts
at
quarter.
Past
is:
do
we
want
to
perhaps
move
it
to
start
at
the
top
of
the
hour?
G
A
G
Okay,
awesome
and
I'll
talk
fast
nice
to
meet
you
all
on
Melissa
I
work
with
Constantia,
where
we
work
with
open
source
communities
and
companies
I
used
to
work
at
Linux
foundation
in
fact
years
ago
and
did
marketing
for
LF,
and
so
one
of
the
the
ideas
that
we
had
was
as
as
open
television
is
kind
of
moving
toward
graduation
Etc.
G
How
do
we
continuously
inform
end
users
of
where
things
are
at
so
you
have
the
status
page
and
what
we
were
thinking
of
doing
was
creating
just
you
know,
a
visual
of
essentially
what's
on
the
status
status
page
and
doing
a
summary
that
could
be
shared
on
social
media,
just
kind
of
like
an
ease
more
easily
consumable
snapshot
of
where
things
are
at
in
open.
G
Telemetries
I
wanted
to
throw
that
out
there
as
an
idea
and
see
if
that
made
sense,
number
one
and
then
there's
a
doc
that
I
had
looked
at.
Let
me
see
if
I
can
find
it
and
I
was
curious,
that
there
was
an
updated
version
of
this.
I
cannot
remember
where
I
got
this
doc
from
it
was
somewhere
on
the
GitHub.
We
put
it
in
the
the
dock.
G
So
this
it's
like
a
state
of
open
Telemetry,
so
there's
the
status
page,
which
has
you
know,
here's
where
everything
is
at
so
we're
thinking
of
kind
of
summarizing
it.
But
the
idea
initially
came
from
this
dock,
which
I
found
like
back
in
the
spring
and
I'm
I.
Don't
know
if
there's
an
updated.
A
version
of
this
I
cannot
remember
where
I
got
it
on
GitHub,
so
I
wanted
to
see.
If
that
idea
made
sense,
would
that
be
useful?
And
if
so
is,
is
the
status
page?
G
F
D
F
S
of
having
a
status
page
has
come
from
in
general,
the
community
really
demanding
that
we
have
some
kind
of
road
map,
and
especially
it's
it's
becoming
over
this
next
year,
it'll
be
a
little
less
necessary
because
tracing
metrics
and
logs
will
be
stable,
and
that
was
like
really
the
big
confusion
for
people
was.
F
You
know
you
told
me,
you
had
all
these
things
and
I
showed
up,
and
it
was
just
tracing
and
I'm
irritated,
and
when
are
you
going
to
give
me
the
metrics,
but
part
part
of
the
issue?
Is
this
kind
of
has
to
be
fixed
at
the
spec
level?
A
thing
I
would
like
to
have
us
do
better
going
forwards,
but
we
need
a
little
more
like
TPM
or
product
management
support
is
organizing
the
spec
road
map
having
that
organized
well
enough
that
it's
reasonable
to
keep
a
page
up
to
date.
F
F
Of
of
the
status
page,
because
it's
just
trying
to
solve
this
like
consistent
piece
of
feedback,
we
get
from
users
that
that
they
they
don't.
They
know
open
Telemetry
moves
slow,
but
they
don't
know
where
it's
moving
too
right.
Yeah.
G
And
yeah-
and
this
is
intended
to
be
a
point
in
time
so,
if
like
let's
say
like
monthly
or
quarterly
right,
where
it's
like-
oh
here's,
everything
that's
available
now
and
here's
the
stat,
you
know
just
I'm,
just
thinking
like
a
visual.
How
can
give
something
like
a
bite
size?
You
know
social
media
tweet
or
something
so
yeah.
That
was
so
it's.
It
sounds
like
it
answers
some
of
that
question
assuming
what's
on
status,
is
up
to
date
and
accurate
yeah.
F
F
F
Yeah,
but
if
we
could
get
the
spec
into
a
place
where
it
has
a
stronger
road
map
where
things
are
kind
of
organized
more
into
projects,
instead
of
just
like
300,
open
issues,
then
being
able
to
funnel
that
information
onto
this
page
that
that's
where
I
would
love
to
see
something
like
this
end
up
and
we
could
stop
calling
it
status
and
call
it
like
roadmap
or
something
right.
D
A
A
I
mean
I
like
this
idea,
but
it
looks
like
we.
We
need
to
do
some
work
for
them
so
yeah.
Maybe
let's
follow
up
on
that
I
think
two
weeks
or
why
on
GitHub
issues,
so
you
don't
hesitate
to
to
raise
this
one,
and
then
we
can
maybe
also
use
that,
maybe
even
if
you
have
from
from
any
other
project
already
something
we
can
look
at
how
it
can
look
like.
So
that
could
also
be
helpful.
Sure
yeah,
awesome.
A
Cool,
so
we
are
running
not
out
of
time
there
over
time,
so
yeah
everybody
needs
to
jump,
including
myself.
So
thank
you,
everybody
for
for
this
meeting
and
talk
to
you
in
two
weeks
then
hi.