►
From YouTube: 2023-04-04 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
A
Think
it
I
think
somebody
wore
a
hat
once
that
was
branded,
didn't
blur
it
out
like
you're
supposed
to
on
broadcast
television.
That's
how
that's.
A
We
could
Branch
out
if,
if
Tito's
hasn't
been
paying
up.
A
B
A
Yeah
I
just
tap
them
in
the
thread
there.
B
A
I
will
say:
I'm
enjoying
the
I
got.
A
new
work.
Laptop
I've
been
I've,
appreciated
the
docker
compose
file
that
we've
got
in
the
in
both
repos
core
and
contrib
for
running
the
examples
they
are
also
sufficient
to
run
these
tests.
So
awesome
I
have
not
had
to
shave
the
act
of
Ruby
on
an
M2.
A
B
A
Like
I,
don't
I'll
unplug
it
and
go
work
somewhere
else
in
the
house
and
and
not
pay
attention
to
the
battery
for
like
a
day
or
two
and
then
just
eventually,
maybe
on
day
three.
It's
like
you
should
probably
plug
me
in,
but
it's
like
casual
about
it
fans,
rarely
fire
up.
A
Despite
me
running
like
bananas,
things
and
the
coop
cluster
in
Docker
desktop,
don't
look
it
I
gotta
do
some
things
differently.
Every
once
in
a
while
or
I,
get
tripped
up
by
some
tool.
I
was
using
not
being
prepared
for
the
different
CPU
architecture,
but
usually
there's
a
workaround
at
this
point.
A
Yeah
previous
work,
laptop
was
I
was
just
coming
on
to
working
for
honeycomb
and
they
were
like
what
worked
laptop
and
it
was
just
when
silicon
was
released.
I'm
like
I'm
new
here
I'm
gonna
go
with
Intel,
because
I
don't
want
to
not
be
able
to
do
my
job
because
of
the
computer.
That'll
look
bad,
it's
a
bad!
Look!
B
I
have
an
M1,
I
think
but
yeah.
It
was
definitely
like
a
huge,
a
a
noticeable
upgrade
over
the
2019
MacBook
that
I
had,
which.
A
C
Yeah
I
just
have
questions
right
now,
I
mean
so
one
problem
I
recently
ran
into
or
not
recently
ran
into,
but
I've
been
working
on,
enabling
instrumentations
and
I
found
in
one
case,
for
example
like
we
have,
you
know
something:
Faraday
spans
right
and
I
think
it
was
something
like
48
000
spans
per
second
on
average
and.
C
So
I
don't
mind
the
volume
of
the
spans
doubling
if
it's
like.
If
we
can
remove
some
of
the
repetitive
data.
C
So
you
get
some
indication
of
the
overhead
between
the
time
that
the
Faraday
client
get
you
know.
Wrapper
function
is
called
to
the
time
that
it
streams
through
the
middleware
and
then
the
delay
when
it
hits
the
driver.
C
So
you
can
see
that
window,
which
I
think
is
valuable,
yeah,
but
I
think
that
repeating
the
attributes
isn't
necessarily
valuable
and
you
know
I
mean
that's
just
one
observation
I
mean,
and
it's
kind
of
like
also
like
the
batch
band
processors
buffer-
is
the
number
of
spans
in
the
queue
not
so
much
the
bite
size
in
the
queue
right.
So
it's
like
if
you,
if
we
double
the
number
of
spans
that
are
in
the
queue
for
whatever
reason.
C
Now
we
have
additional
adjustments,
yeah
and
they're
all
big
right,
we're
making
additional
adjustments
and
we're
doing
a
lot
of
network
transfer.
So
you
know
that
got
me
thinking
about
other
things,
but
that's
one
thing
that
I
was
thinking
about
and
I
was
thinking
and
I'm.
Also
thinking
about
this
in,
like
in
the
light
of
the
active
record,
you
know
instrumentation,
where
we
started
talking
about
like
a
when
combining
these
with
drivers.
How
much
of
these
are
duplicate?
How
do
we
you
know?
How
do
we
specify
that
we
want
to
suppress
them?
C
How
do
we
want
to
specify
that
we're
sharing
them?
What
additional
information
would
whatever
they
have,
that
the
driver
wouldn't.
C
A
A
It
depends
on
your
back
end
and
whether,
although
all
of
the
points
that
you
raise
are
good,
are
like
SDK
impact
issues
where
you
know
the
batch
Q
size
and
how
big
you
need
to
make
it,
depending
on
whether
you're
getting
two
spans
per
HTTP
call
I
can
think
about
back
ends
where
the
person
using
Auto
instrumentation,
maybe
would
want
Faraday
to
tell
I
just
wanted
to
know
that
there
were
HTTP
calls
and
I,
don't
really
care
about
the
driver,
and
so
they
would
want
Faraday
to
be
fat
and
then
maybe
just
turn
off.
A
A
So
I
see
I
could
see
levers
for
like
letting
people
like
by
default.
It'll,
be
you
get
both
levels
of
abstraction
with
all
of
the
attributes
that
they're
going
to
say,
but
if
we
give
it
levers
for
saying
I
want
you
to
be
thin
and
you
to
be
fat
or
you
to
be
fat,
and
you
not
to
be,
there
could
be
useful,
I,
don't
know
what
those
levers
would
look
like,
though,.
C
I
know
it's
makes
it
tough
to
add,
like
a
bunch
of
configuration
options,
it's
sort
of
it's
more
of
like
you
know
what
do
I
want
as
my
default
and
how
do
I
want
to
Signal
like
I,
want
verbosity
out
of
this
particular.
C
Yeah,
because
the
way
that
I
see
it
is
the
favorite
days
and
internals
man
and
the
driver
is
the
clients,
man,
and
so
the
client
would
have
the
semantic
attributes
on
it.
That
are
the
HTTP
client
symmetric
semantic
attributes,
because
you
know
when
pharity
goes
through
its
process
of
the
middlewares
and
each
one
of
the
middlewares
makes
an
adjustment
to
the
request.
C
Let's
say
you
want
to
record
request
headers,
for
example
right.
It's
like
you
want
to
record
them
when,
after
the
request
is
complete
in
the
Faraday,
because
the
middleware
is
like.
Oh
I've
completed
the
request,
but
I
don't
know
at
what
point
I'm
actually
making
the
request
right,
because
somewhere
down
the
chain,
other
components
may
have.
You
know
changed
the
requested
request,
headers
right
and
then
I
also
was
like
you
know.
You
have
a
retrial
middleware
within
Faraday.
So
far
already
itself
can
be
like
retrying,
multiple.
A
Rooms
already
yeah.
C
C
A
True-
and
this
is
fitbull
and
haven't
just
thought
about
it,
while
you
were
talking
what
a
faraday
defaulted
to
being
lean
and
then
there
was
a
configure
op
config
option,
which
was.
A
I
guess
successive
HTTP
attempts.
B
C
B
C
C
I
would
think
that
would
be
when
the
middleware
is
finished
or
like
when
the
connect
I
guess.
The
problem
is
that
the
connection
objects
can
be
yeah:
okay,
yeah,
yeah
yeah.
The
problem
would
be
that
we'd
have
to
evaluate
that
every
time
the
favorite
day
Builder
was
interpreted,
and
that
could
be
every
time.
A
new
object
is
instantiated.
B
C
A
So
if
you,
if
you
don't
want
drivers
fans,
disable
that
disable
the
x-con
or
HTTP
core
instrumentation,
like
look
at
your
look
at
your
interpretation,
you're
like
I,
don't
like
this!
There
are
ways
for
you
to
change.
It.
C
A
A
A
And,
and
do
you
have
are
you,
do
you
have
HTTP
client,
libraries
and
use
that
are
not
yet
instrumented.
C
As
an
example
right-
and
we
woefully
are
behind
on
you,
know,
tests
and
stuff
like
that
for
compatibility
testing,
but.
C
Okay,
well,
y'all
have
given
me
some
thought
for
food,
which
means
I'm
gonna
have
to
like
put
up
some
PRS
with
some
suggestions
here
and
be
like.
If
start,
you
know
default
to
lean
and.
B
I
had
one
thought,
while
we're
on
this
topic
and
parody
is
actually
quite
old
and
I
know
it
predates
this
client
context,
HTTP
client
context,
business
and
it
was
kind
of
added
for
koala
I.
Remember
and
it's
kind
of,
like
the
reverse
of
the
you
know,
of
our
rackspan
situation,
where
you
have
wrecked,
create
the
span
and
then
your
your
your
HTTP
server
framework
enriches
this
fan.
B
C
So
the
use
case
that
would
be
would
where
that
works
great,
is
if
I
have
instruments
of
the
drivers
fan
in
the
case
where
I
have
not
instrumented
the
driver
span.
I
I
want
some
sort
of
a
client
representation
which
would
mean.
A
C
I
guess
yeah:
it's
the
situation
where
it's
like
is
the
uninstrumented
driver
present?
A
A
B
Yeah
yeah
I
I
could
see
a
configuration
option
for
Faraday.
That
is
like
create
a
span
or
decorate
a
span,
and
then
you
could
set
that
for
app.
That
works
for
you,
I,
don't
know
just
just
throwing
off
some
ideas.
C
Like
I'm,
not
able
to
say
unless
I'm
like
manually
configuring,
the
Faraday,
middleware
and
saying
okay,
if
I,
already
middleware,
this
is
what
you're
gonna
do
you're
gonna.
Have
this
peer
service
name,
which
I
think
is
something
that
which
is
something
I?
Don't
have
access
I'm
not
doing
right
now,
you're
gonna
be
a
internal
spam
and
you're
gonna.
B
C
C
So
Matt
have
I
heard
you
right,
you're,
saying
configuration
options,
probably
work
as
part
of
those
configuration
options,
try
to
Leverage
The
the
shared
context.
B
I'm
just
throwing
it
out
there
like
as
an
idea
like
I,
think
you
actually
have
has
some
use
cases
here
so
I.
Don't
know
that
that
my
ideas
will,
you
know,
solve
everything
but
I,
guess
you're
kind
of
thought
and
things
things
to
experiment
with.
C
Uh-Huh,
this
is
a
spicy
one.
Does.
C
Just
rack
events
I
mean
look
at
the
pull
request
list.
There
are
a
bunch
of
pull
requests
where
it's
like
changes
requested
or
the
author
responded
and
then
the
conversation
just
died
where
people
either
accepted
or
rejected
the
pr.
At
this
point,
it's
like
leaving
a
PR
open
in
limbo,
with
no
follow-up
comments
is
almost
the
same
as
rejecting
it.
So
why
not
just
reject
it
and
like
it's?
C
It
just
feels
weird:
it's
like
the
it's
like
when
the
conversation
is
paused
well,
I,
just
see
it
as
kind
of
like
obstruction
in
some
sense,
yes,
I
am
already
at
the
point
where
I
am
unnerved
about
waiting
a
month
and
a
half
almost
two
months
for
somebody
to
be
like
yes,
this
rack
events
thing
is
okay,
let's
merge
it,
but
I
realize
that
my
case
is
not
an
isolated
incident,
as
we
have
in
many
of
these
PRS
where
people
engage
and
then
they
follow
through.
C
So
it's
like,
we
have
that
I
kind
of
feel
like
I,
want
to
have
some
sort
of
like
a
clean
house
policy,
clean
desk
policy
or
whatever,
and
it's
like
come
on.
We
gotta
give
people.
We
can't
wait
three
months
to.
C
To
accept
or
reject
a
PR,
it
just
seems
a
little
I,
don't
know
maybe
I'm
making
too
much
of
it.
Maybe
others
disagree
and
say
that,
like
it's
just
part
of
the
game
and
it's
normal
but
don't
feel
right
to
me,
that's
all.
B
No
I
think
I
think
that
we
should
respect
contributors
time
and
try
try
to
get
reviews
in
and
if
they
are
addressing
things,
I
try
to
re-engage
so
you
know
I
think
from
from
our
perspective.
That's
what
we
want
to
do.
I
do
know
that
sometimes,
even
when
you
do
that,
a
few
who
open
the
PR's
end
up
gradually
disappearing
and
if
that's
the
case,
we
should
kind
of
just
ask
them.
Hey,
are
used
still
interested
in
this
and,
if
not
close
them,
so
I.
A
B
B
C
Well,
if
it
says
review
required,
it
means
no.
Why
that
that.
C
Yeah,
so
it's
either.
We
need
to
start
eating
into
the
mode
of
changes
requested
to
be
like
that
signals
to
the
author.
They
need
to
do
something
yeah
and
if
the
author
comes
back
with
a
rebuttal
like
we
have
to
like
close
the
loop
on
that
and
if
something
says
like
approved
and
has
been
sitting
there
for
a
while,
why
hasn't
emerged
just
yet?
C
You
know
what
I'm
saying
or
whatever
right,
like
I
I'm,
trying
to
figure
out
like
you
know,
for
for
folks
who
have
done
the
approval
you
know
or
whatever
the
case
may
be
right
like
we
need
to
figure
out
like
why
these
aren't
Sitting
In
Limbo
right
now.
For
some
of
these
and
it's
perfectly
fine,
I,
think
another
state.
That's
perfectly
fine,
as
the
switcher
to
draft
means
I'm,
not
gonna.
Look
at
it.
C
You
know
what
I
mean
like
I'm,
the
person's
not
ready
for
it
to
have
gotten
in
a
review
and
I
try
not
to
look
at
draft
PR's
in
general.
Just
to
save
myself
time.
You
know,
but
anyway,
I'm
just
kind
of
I
am
expressing
a
little
bit
of
frustration
because
of
the
the
time
is
taken
to
try
to
deal
with
this.
You
know
rack
events
one,
but
then
I
realized.
No,
that's
not.
It's.
C
It's
not
just
me,
so
that's
all
yeah.
B
B
A
B
B
We
need
we
need
to
figure
out
how
to
address,
but
I
feel
like.
Maybe
these
things
are
old
enough,
that
we
can
make
a
point
next
week,
I
think
to
come
up
with
a
plan
to
clean
out
stuff
from
last
year
and
I.
Think
I
will
make
it
my
goal
this
week
to
make
it
through
everything,
from
February
16th
above
and
at
least
get
some
feedback
and
go
from
there
and
then
yeah
I
think
we
need
to.
We
definitely
need
to
figure
this
out
going
forward.
I
think.
B
C
I
appreciate
that,
thank
you
one
suggestion
I
can
make,
for
that
is
for
us
to
use
the
stalebot.
Stillbot
can
just
say
yo
anything
over
three
months
old,
we're
going
to
Auto
close
them.
If
we
see
no
activity,
it.
A
A
A
I'm
I'm
100
on
maybe
stalebot
can
help
us
be
more.
C
A
C
B
A
I
will
try
to
do
the
same
thing
of
getting
on
reviewing
stuff
from
this
year
and
aspirationally.
Looking
at
some
of
the
older
stuff,
but
I
put
my
face
on
rack
events,
because
I've
already
got
it
on
my
new
laptop
and
have
I
think
conquered
all
of
the
mini
bosses
to
get
through
the
good
luck,
the
level
boss
to
review
it.
A
C
Just
like
there's
a
refactoring
that
was
in
there
where
I
bet,
I,
didn't
refactor
I
copy
and
pasted
and
knowledge
duplication
of
the
config
options
in
the
middleware
because,
like
the
middleware,
was
doing
all
the
config
options
instead
of,
like
the
installation
doing
the
configurations
so
I
copied
it
over
and
I
made.
The
changes
and
I
did
that.
Instead
of.
A
C
C
Your
face
on
it,
you
did
it
thanks,
I,
appreciate
it
and
then
the
last
thing
that
I
was
gonna
say
was
because
of
I
kind
of
want
to
Leverage,
The
expertise
of
people
who
are
rupees,
but
not
necessarily
Hotel
experts,
and
so
like.
We
got
a
group
of
people
here
called
the
Ruby
architecture
team
and
they
work
on
the
you
know
they
work
on
yarov
and
they
work
on
rails
and
I'd
like
I'm.
Gonna
engage
them.
C
So
if
you
see
them
like
making
comments
on
PRS
and
whatnot,
it's
it's
because
because
I
asked
them
for
help,
you
know
what
I'm
saying
so.
A
C
Nice
so
yeah
so
like,
for
example,
I
composed
internal
right
he's
like
a
maintainer
trilogy,
and
these
changes
that
came
in
and
I
was
like.
Oh
man,
Daniel
could
you
you
know,
take
a
look
at
this
I
think
or
Jay
Hawthorne.
C
You
know
so
like
just
getting
them
in
to
like
give
me
a
hand
to
look
at
stuff
just
to
make
sure
I'm
not
doing
anything
inefficient
or
if
they
have
recommendations
for
us
on
better
ways
to
instrument
things
in
general,
because
they're
more
familiar
with
like
the
internals
of
rails
and
the
and
the
VM
and
whatnot,
but
they're
not
prop
they're,
probably
not
going
to
commit
PRS
so
that
doesn't
make
them.
So
in
the
past
we've
we've
told
people
yo,
please
do
not
unless
you're
an
approver.
C
Yeah,
what's
what's
kind
of
interesting
is
like
there's
like
those
those
things
when
you
talk
when
you
look
at
what
the
guidelines
are
for
like
how
a
person
becomes
a
reviewer
or
contributor,
even
it's
like
they
have
to
attend
this
meeting
and
do
these
other
things
I,
don't
know
if
those
constraints
apply
to
us,
you
know
a
triage
or
whatever
I
don't
know.
A
B
Yeah
I
mean
I
feel
like
non-reviewers,
should
be
able
to
review
PR's
I'm,
not
sure
that
that
I
think
one
way
to
actually
become
a
Reviewer
is
to
review
PRS
so
saying
that
people
should
not
review
if
they're,
not
a
reviewer
I
think
maybe
is
a
little
too
extreme,
but.
C
B
B
B
Recognition
right
getting
these
roles
in
hotel
and
I
think
they're
just
made
so
that
they're
not
like
subjective.
You
know.
Some
people
are
just
not
subjectively
getting
assign
things
and
yeah
I
think
I
think
they
vary
over
time.
I
think
people's
I
think
to
the
project
and
ability
to
commit
to
the
project
and
I
have
varied
over
time.
So,
while
people
such.
B
Done
enough
work
to
get
certain
like
roles,
I,
don't
know
that
it's
have
definitely
not.
You
know
continued
at
those
levels.
B
Yeah
and-
and
this
seems
to
happen-
it
seems
to
happen
a
lot
actually,
it
seems
to
happen
and
all
the
segs
and
it
seems
to
happen
to
to
multiple
people.
It's
just
like
we're
kind
of
at
the
whims
of
our
employers
and
Things
become
higher
priorities
over
time
and
but
yeah
I
think
talking
about
this
and
definitely
like
raising
a
flag
when
things
are
like
becoming
gridlocked
and
terrible.
B
I
think
makes
sense,
but
I
think
you
know
yeah,
there's,
there's
only
so
much
that
I
think
we
can
do
about
freeing
up
our
own
schedules,
but
we
should
try
to
do
that
as
much
as
possible.
I
think
there
needs
to
be
maybe
a
renewed
effort
to
try
to
find
people
who
are
willing
and
able
to
kind
of
pick
up
some
of
these
other
roles.
So
I
don't
know
if
we've
seen
people
who
are
in
the
community
who
are
doing
a
lot
of
work
that
we
like
and
kind
of,
has
some
respect
for.
B
But
if
we
do,
we
should
talk
about
trying
to
vote
people
into
like
approval
roles.
If
you
know.
C
Yeah
I
think
what
scares
some
people
off
about
that
was
that
the
meeting
requirements
that
were
described
in
the.
A
A
B
A
C
So
I
haven't
read
the
latest
on
this.
Maybe
there
were
some
changes
that
were
on
this,
but
I
feel
like
at
least
the
people
that
have
been
contributing
recently.
C
You
know,
and-
and
even
it's
been
a
little
while
but
like
I
feel
like
their
engagement
is
pretty
high
and
I've
tried
reaching
out
to
them
before
I'll.
Give
you
a
list,
you
know
you
can
see
them
like
like
Tristan
is
pretty
engaged.
You
know,
Ricky
is
pretty
engaged.
C
I
have
high
hopes
for
Memorial
with
this
great
contribution
in
the
past
you
know:
Nicholas,
volcker
and
Chris
right,
Crystal
chimes
in
every
once
in
a
while,
so
like,
even
if
we
can
get
a
little
bit
of
their
time.
You
know
I
know
I,
know.
Rick
is
out
there
trying
to
figure
out
why
Pete
threads
are
dying.
C
A
I
would
be
interested
in
their
opinions.
We
should
encourage
them
to
share
them
all
right
man,
yeah
and
also
do
we
have
a
definition
of
contributions
here.
It
says
active
contributor
in
the
community
review
or
PR's
requirements,
multiple
contributions
to
the
project,
I'm
of
the
opinion,
unless
this
thing
says
otherwise,
if
you're
having
conversations
on
issues
in
PRS
and
in
cncf
slack
and
the
sync
meetings
that
you
can
come,
those
are
contributions.
A
C
C
B
I
I
think
we
covered
some
good
ground
here.
I
think
we
missed
last
week
also,
so
this
stuff
has
probably
been
been
needing
to
be
aired.
So.
C
B
Least
we
got
like
a
partial
airing
today
and
I
think
we
can
continue
these
discussions
next
meeting
and
hopefully
we
can
have
a
look
at
least
at
this
year's
PRS
and
try
to
get
those
moving
forward.
C
All
right
and
then
I
appreciate
it
I
like
the
Airing
of
Grievances
Park
onto
the
Feats
of
Strength.
Indeed,.