►
From YouTube: 2023-03-27 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
A
B
Vacation
rested
a
lot
and
it
was
great
that.
B
A
Let
me,
while
we
are
waiting
real
quick
since
you
ask
like
I,
don't
think,
there's
anything
urgent
but
but
I
think
we
we
maybe
need
to
to
talk
a
little
bit
about
the
the
workflows
and
in
those
auto
checks.
Right
I
mean
right
now
it's
a
little
bit
flaky
everything
I
have
to
feeling
right.
So
maybe
we
can.
A
B
Enabled
it
just
before
heading
out
and
I
saw
it
work
for
at
least
one
PR,
okay,
and
in
fact
it
worked
it
reformatted.
It
pushed
the
commit
and
re-triggered
the
format.
Checking
so
I
did
every
everything
that
it
was
supposed
to,
but
more
globally
I
I
mean
I'm
glad
to
work
on
that,
probably
in
in
the
upcoming
week,
I'll
be
working
on
doxy.
B
A
Hey
Philip,
hey,
Victor,
I,
don't
think
it's
an
it's
a
priority,
I
think
there's
just
a
few
things.
We
should
just
streamline
a
little
bit
so
this
thing
I
run
into
today.
Is
this
ref
cache
now
I
think
everybody
is
stumbling
upon
that
one
yeah.
So
so
that's
that's
the
one
and
I
said
I
mean
if
this
fixed
format
is
working,
I
think
that's
helping
already
so
I
think
that's
the
two
I
had
in
mind
right.
So,
okay,
so.
B
What
I
could
do
is
is
bring
it
bring
the
ref
cache
update
into
the
automation
setup,
and
then
we
can.
How
does
that
sound.
A
A
A
Awesome
I
I
put
a
a
few
items
into
into
the
agenda,
but
but
maybe
someone
someone
has
to
to
add
something.
Additionally,
so
please
add:
I'm,
not
part
of
Slack.
A
Real
quick
on
kubecon
I
hope
you
all
saw
the
the
blog
post
draft
I
created
I'm,
still
looking
for
all
kind
of
feedback
right
so
did
I
miss
anything.
A
Lolita
suggested
that
that
we
add
some
images
there.
I
don't
have
any.
So
if
anybody
has
some
some
good
pictures
from
from
kubecon
North,
America
or
hotel
slavery,
we
can
add
there
and
also,
if
you
think,
like
hey
I,
I,
missed
mentioning
anything
on
that
on
that
blog
post.
Then,
then,
just
let
me
know
we.
We
can
add
this
in.
A
Yeah
and
I
think
the
other
thing
was
just
to
to
highlight
this.
Once
again,
we
wanted
to
have
a
few
docs
ready
by
kubecon,
I.
Think
Philip.
You
did
a
lot
of
just
sampling
already
right,
so
we
could
probably
already
scratch
this
off.
C
I,
don't
think
it's
enough
to
scratch
off.
I
want
to
get
more
more
languages
documented.
The
python
one
is
still
outstanding.
Oh.
C
So,
currently
the
go.
One
also
needs
to
be
Rewritten,
because
it
it's
basically
like
says,
Hey
sampling
exists.
It
doesn't
actually
tell
you
how
to
do
it,
but
like
all
the
all
the
major
languages
that
have
tracing
as
stable,
which
is
basically
all
of
them,
should
have
a
sampling
dock.
That's
like.
B
B
C
A
Yeah,
yeah
and
I
think
we
also
should
not
try
to
deploy
the
ocean
with
that
right.
So,
let's,
let's
have
one
in
and
then,
if
more
more
than
it
makes
sense,
then
then
we
should
should
add
it
there,
but
yeah.
C
Yeah
they
they
have
a
page
in
the
read
the
docs.
That's
like
reference
material
and
they're
like
oh
well,
this
one's
super,
complete
I,
don't
see
why
anybody
wouldn't
want
this
and
I'm
like
well,
but
have
you
ever
worked
with
a
user,
but
it's
I
mean
this
is
not
really
a.
This
is
not
really
a
debate.
That's
different
from
any
other
debates
that
we
have
in
this.
A
C
A
C
Yeah
so
I
mean
it's
basically,
fine,
as
is
I,
can
make
some
changes
to
sort
of
get
them
to
at
least
sign
off
on
it.
But
then
similar
similar
changes
need
to
be
made
for
go.net
Plus,
a
few
others.
A
A
C
So,
there's
there's
that
the
the
other,
the
other
sampling
topic
or
I
guess
set
of
topics
is
I'm
still
not
satisfied
with
the
conceptual
document
that
that
I
that
I
modified
because
it
it
goes
into
like
head
sampling
versus
tail
sampling,
at
least
at
a
high
level
of
like
what
the
hell
these
things
are
and
what
you
could
potentially
do
with
them.
C
But
it
doesn't
really
go
into
sort
of
what
you
get
and
what
you
give
up
and
much
just
because
because,
like
that,
that's
kind
of
that's
kind
of
part
of
it.
Like
there's,
there's
a
there's,
a
handful
of
reasons
that
most
people
want
to
sample
their
data,
but
I
think
and
then
I
think
a
lot
of
people.
In
my
experience,
don't
even
really
know
the
difference
between
head
and
tail
sampling.
But
then
beyond
that
they
like
a
lot
of
them.
C
I,
don't
expect
to
solve
this
problem
between
now
and
kubecon,
but
at
least
have
like
some
scenarios
listed
out
where
maybe
it's
it's
illustrative
to
a
lot
of
people
that,
like
okay,
you
know
I
know
based
off
of
this
scenario
that
I
should
be
doing
tail
sampling,
and
these
are
the
sorts
of
like
rules
that
I
should
put
in
my
tail
sampler.
C
That's
kind
of
it
and
I
need
to
also
do
a
scan
for
tail
sampling,
processor
docs
in
The
Collector,
which
is
kind
of
weird,
because
for
my
job
we
tell
people
not
to
use
it
because
there's
a
whole
bunch
of
like
there's
like
an
order
of
magnitude
in
your
operational
complexity,
jump
once
you
start
using
it.
It
has
several
other
issues
that
are
like
solvable,
but
there's
there's
a
whole
thing
there
that
I
can
elaborate
on,
but.
C
At
least
something
they're
written
for
people,
because
flawed
as
it
may
be,
it
is
still
helpful
for
it's
more
helpful
than
not
having
a
sampler
in
the
first
place.
But
what
it
means
is
is
now
you
you
can
no
longer
have
just
like
a
collector
that
you
send
data
to.
You
need
to
have
like
a
stage
set
of
collectors
that
load
balances
across
them
and
the
operational
characteristics
of
these
things
are
going
to
be
completely
different,
because
now
you
have
this
stateful
system
in
place
and
there's
really
kind
of
no
predicting.
C
What
is
gonna
work
for
you
and
like
how
you
should
provision
your
resources,
but
like
these
are
the
sorts
of
things
that
people
are
going
to
run
into,
which
is
why
I've
been
a
little
bit
hesitant
to
to
write
up
to
write
it
up.
A
Oh
right
yeah,
but
yeah
no
I,
I,
agree,
I
mean
this
is
super
important,
but
but
definitely
not
something
we
can
rip
off
before
kubecon
and
maybe
we
shouldn't
so
I
mean
yeah
but
I
totally
agree.
I
mean
this
is
one
one
of
the
the
topics
that
creates
a
lot
of
haddock
and
at
a
certain
point
when,
when
people
get
into
this
discussion
like
hey,
we
have
way
too
many
traces
and
how
do
we
manage
them
right?
So
yeah
totally
makes
sense
soon.
A
Awesome.
Okay,
but
but
do
you
still
think
that
that
we
can
improve
on
the
concepts
page
until
Cube
corner?
Should
we
just
postpone
it
after
after
kubecon
entirely
I.
C
Think
we
can
and
at
the
very
least,
there's
more
scenarios
that
can
be
elaborated
on
in
there.
Okay,.
A
A
So
maybe
someone
volunteers
for
that
yeah
I'm,
not
sure
anymore,
of
what
we
were
about
collector
receivers.
I
just
copied
this
over
from
last
week.
B
C
B
C
Think
the
the
discussion
last
week
was
about
there's
like
a
handful
of
receivers
that
are
used
more
than
others
and
I.
Oh,
oh
right!
This
that
this
is
in
the
context
of
logs,
because
we're
considering,
since
the
logs
data
model,
is
stable.
If
you
have
a
like
a
log
aggregator
of
something
that
you're
using
you
can
just
pull
up
one
of
the
receivers
that
probably
exists
for
it
in
in
the
collector
and
say
here
you
go!
That's
where
my
logs
go
and
that's
seems
like
kind
of
an
easy
win.
C
Page
and
sort
of
in
how
to's
that's
the
more
operational
side
of
it
there's
a
whole
other
piece
about.
You
know
on
the
languages
side
like
how
you
do
like
in-app
structured
logs
and
get
them
in
the
context
of
traces
and
all
that
kind
of
stuff.
C
But
if
you
have
a
whole
bunch
of
existing
logs
coming
from
like
a
Splunk
log,
aggregator
or
or
fluenty
or
fluent
fluent
bit
one
of
the
fluents
you
can,
you
can
point
it
at
The,
Collector
and,
and
there
there's
there's
really
like
the
only
reason
why
I
know
this
is
because
I've
worked
with
customers
who
like
needed
to
do
this
and
I
looked
it
up
for
them.
C
A
To
sell
like
hey
you
I
mean
that
that's
one
of
the
big
promises
we
have
similar
for
open,
Telemetry
right
that
that
you
can
use
what
you
have
already
so
so
maybe
we
should
document
this
out.
A
little
bit.
Yeah
I,
know
totally
makes
sense.
You
know,
and
the
same
for
the
conceptual
pages
I
think
nothing
has
changed
since
last
time.
So
I
think
we
still
have
this
very
sparse
page,
so
yeah,
maybe
I,
think
it's
a
very
important
one
right
for
for
kubecon
to
to
have
that
one
updated.
A
If,
if
anybody
fancies
doing
that,
yeah,
let
us
know-
and
even
if
it's
only
incremental
improvements,
we
should
work
on
that
cool
anything
else.
On
kubecon
anything
we
we
need
to
discuss
a
little
bit
after
out
of
the
loop
with
a
lot
of
things
with
kubecon,
but.
A
Okay,
anything
else
on
on
cook
John,
no
okay;
let's
maybe
move
that
one
to
the.
If
we
have
time
for
them,
calendar
entry
has
separate
notes
that.
B
Probably
should
have
been
last
is
just
that's
why
it
was
last,
but
if
you
go
to
the
calendar,
entry
I
usually
open
up
the
dot
that
this
meeting
notes
from
there.
But
if
you
click
on
that,
there
seems
to
be
another.
B
No,
no
in
the
open,
Telemetry
Google
calendar
entry.
B
A
B
A
E
Question
on
the
blog
post,
I
haven't
had
a
chance
to
look
at
it
yet,
but
is
there
any
mention
of
the
maintainer
circle.
E
Take
a
quick
look
I'm
just
pulling
it
up
as
well.
There's.
A
Okay,
if,
if
you,
if
you
can
so
so,
if
you
can
just
go
to
the
blog
post
and
and
let's
say,
comment
on
it
or
or
even
make
a
suggestion
where
to
put
it,
then
then
I
easily
can
easily
can
add
it
in
so,
okay.
E
A
Yeah
and
anything
else,
I
mean
I
just
wanted
to
to
quickly
see
if
we,
if
we
have
any
need
to
to
follow
up
on
the
on
the
stack,
Overflow
discussion
that
that
we
had
last
week
two
weeks
ago,
so
I
spend
at
least
a
little
bit
of
time
in
it
looking
at
stack,
Overflow
and
then
there's
indeed
a
bunch
of
open
Telemetry
questions
that
nobody
is
is
answering
them
yet.
E
B
E
B
E
That
was
what
I
got
out
of
that
okay,
like
before
they
were
looking
at
tagging
tagging
the
questions
to
to
direct
to
the
to
the
proper
cigs
and
stuff.
B
D
A
Yeah
cool,
no
I,
I,
just
I,
said
I
just
also
looked
at
it
the
other
day
and
it
looks
like
there's
like
600
something
questions
and
then
there's
many
many
unanswered
questions
I
mean
some
of
them
are
really
complex.
Right,
I
mean
some
of
them
are
really
like.
Yeah
I
could
answer
that,
but
it
would
take
me
like
I,
don't
know
half
an
hour
to
to
even
come
up
with
something,
but.
D
So
we
were
actually
going
over
some
of
these
in
the
maintainer
Sig
this
morning,
so
I
think
generally
the
ones
that
are
tagged
per
language.
The
guidance
from
the
hotel
governance
committee
and
the
community
is
going
to
be
that
the
language
sigs
should
tackle
those
and
then
the
the
generic
ones,
maybe
about
the
spec
or
how
to
use
open
Telemetry.
Generally,
things
like
that,
we
can
tackle
in
the
comms
and
maintainer
six:
okay,
okay,
cool
yeah.
D
The
maintainer
stick
really
just
to
make
sure
that
they
have
the
right
tags.
The
challenge
is
only
a
handful
of
us
actually
have
enough
rep
on
stack
Overflow
to
add
tags
to
things.
That's
probably
a
challenge
here
too,
at
least,
if
anyone's
like
me,
you
stack
or
flow
up
overflow
a
lot
as
a
reader
and
not
as
much
as
a
contributor
and
and
but
gerasi,
and
a
few
other
people
in
the
maintainers
they've
had
enough
reputation.
They
could
add
text
and
things
to
make
sure
the
right
language
is
protect.
Okay,.
D
D
Anyways,
so
that
that's,
we
should
also
go
over
them
with
this
thing
as
well,
the
more
eyes
on
the
scenario.
A
Okay,
okay,
so
it
just
means
like
for
us:
it
remains
more
like
the
the
generic
ones
to
see
like
hey.
How
can
we
exactly
map
them
back,
or
also
see
like
I
mean
that
that's,
maybe
something
also
still
interesting?
What
we
discussed
last
time,
maybe
there's
some
eventually
that
are
maybe
the
ones
still
not
even
that
I
can
sort
them
like
that.
We
at
some
time
at
some
point,
maybe
even
started
to
look
at
the
let's
say
most
upvoted
one
or
the
ones
with
the
most
views
and
then
say:
okay.
A
This
should
maybe
be
part
of
the
docs
as
well
right.
So
yeah,
that's
more
than
our
duty
to
to
look
at
that.
This
one
here,
for
example,
is
definitely
an
interesting
one.
Yeah.
A
Yeah,
but
but
and
there's
a
bunch
of
them
right
so
yeah
now,
okay,
that
totally
makes
sense,
cool.
A
Yeah
I
think
that's
it.
I
didn't
I
didn't
have
to
add
anything
to
to
the
agenda
today.
C
A
B
A
C
So
I
actually
I
have
a
I.
Have
a
fun
question.
Do
y'all
y'all
want
to
see
some
really
radical
that
I
think
we
should
get
our
find
a
way
to
get
the
hotel
docs
incorporated
as
yeah
sure
all
right,
so
I'm
gonna
share
my
screen
here.
Okay,
so
I
have
preview
access
to
GitHub
co-pilot
for
docs,
which
was
just
announced
and
released
last
week
as
like
a
super
early
private
beta
thing.
C
Basically,
it
uses
gpt4
to
summarize
a
corpus
of
text
or
index
I
guess
you
could
say,
and
so
it's
focused
on
two
things:
a
summarizing
content
from
like
a
q,
a
standpoint
and
then
B
in
that
summarization.
It
sources
it
based
off
of
like
like.
Where
did
it
like?
What
docs
did
this
get
pulled
from?
So
if
I'm
on
mdn,
like
as
a
JavaScript
developer,
let's
say
you
can
sort
of
tune
it
which
is
kind
of
neat.
C
Let's
say:
I'm
a
new
developer
and
I,
don't
know
pretty
much
anything,
but
I
heard
that
VAR
is
bad
and
I
should
use.
Let
but
like
that's
kind
of
confusing
that
talk
about
like
block
Scopes
and
timeouts
and
stuff
and
I.
Don't
know
what
the
that
is,
if
I'm
a
new
JavaScript
developer
instead
I
could
just
go
like.
C
This
is
probably
helpful
in
the
context
of
the
question
that
you
asked,
rather
than
you
having
to
like
hunt
for
code
samples
that
might
be
relevant
to
you
and
then
it
also
generates
a
set
of
follow-up
questions
that
you
might
have
based
off
of
what
you
just
asked
so
like.
What's
the
difference
between
let
and
const,
and
so
like
the
other
thing
is.
C
This
is
still
super
early
preview,
but
they're
working
with
entirely
open
source
stocks
sets
and
I've
tried
like
pointing
pieces
of
the
hotel
docs
to
like
chat,
gbt
and
stuff,
and
it's
super
hit
or
miss,
and
a
lot
of
that
is
just
I
mean
like
the
index
stuff.
Stopping
at
September
of
2021
and
like
2021
is
like
kind
of
the
start
of
when
we
started
having
a
lot
of
content
about
open,
Telemetry,
yeah.
C
So,
like
these
tools
are
really
bad
with
Hotel,
but
like
here's,
a
potential
way
to
like
sort
of
you
know,
I
I
mean
my
experience
working
with
users
like
they
they
come.
They
come
at
things
like
this
like
they
don't
they
don't
go
to
the
hotel
site
or
some
dock
site
and
say:
oh
I'm,
gonna
go
read
all
the
stuff
and
then
I'm
gonna
learn
everything
I
need
to
know.
They're,
like
no
I.
C
Have
this
thing
I'm
trying
to
do
right
now,
like
what's
the
like
the
minimum
I
need
to
sort
of
figure
that
out
so
I
think
like
this
might
be
a
direction
that
we
want
to
go
since
they
are
focusing
right
now,
they've,
obviously
GitHub
stocks,
because
it's
a
GitHub
product
but
react
ndn
and
Azure
docs
is
all
open
source,
stuff,
I
think
they're
they're
trying
to
work
with
open
source
repositories
right
now.
They
don't.
They
don't
have
a
like.
It's
still
super
early.
C
D
C
And
then
what
I
like
about
it
is
it
it
reinforces
that
you
need
to
have
good
content
because,
like
otherwise
it's
just
going
to
hallucinate
stuff.
Yes,.
C
A
C
I
think
it's
early
enough
that
they're
they're,
not
even
really
like
I,
could
see
like
for
them
a
path
to
monetization.
I
guess
could
be
that
you
know
for
every
business
that
has
their
own
dock
site.
You
could
sort
of
sell.
You
know
like
copilot
indexing
for
for
your
own
docs,
but
given
they're,
given
that
it's
GitHub
I
would
imagine,
they're
probably
going
to
have
something
for
like
projects
like
this,
where
you
know
it's
actually
free
and
like
they
maintain
it,
and
they
even
use
that
to
try
to
improve
the
service
itself.