►
From YouTube: 2021-05-13 meeting
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Yeah
how's
how's
everything.
A
Yeah
man,
I
thought,
did
you
take
some
time
off
or
anything
or
or
just
like.
A
A
All
right,
we'll
just
give
it
a
couple.
A
A
D
E
A
Right
right,
yeah,
I
have
to
go
the
other
way
with
the
label.
Even
with
the
like
that
github
action,
you
were
using
right.
It
doesn't
do
that
everything.
E
D
E
G
E
A
Just
don't
allow
it,
but
why
man?
Oh
it's
a
feature.
What
do
you
mean?
It's
a
feature,
it's
a
feature
that
prevents
you
know
like
malicious
attacks
and
stuff.
A
E
E
A
E
A
E
It
right
add
a
label
if
you're,
okay
with
it.
If,
if
not
just
tell
this
to
measure
a,
we
don't
want
your.
H
E
H
G
G
F
E
Lets
you
do
it
right?
No,
I
mean
I
open
the
vr,
but
the
pr
it
come.
It
is
from
a
fork.
I
mean
if
I
open
up
vr
by
myself
in
in
an
organization
where
I
have
permissions.
For
example,
in
my
personal
yeah
repos,
it
works.
F
Okay,
I
think
I
think,
when
we
were
doing
the
performance
testing
stuff
within
the
nathaniel,
the
aws
intern
it
it
did
actually
work.
We
were
able
to
get
it
to
work,
but
it
was
flaky
because
it
would
like
it
would
try
to
write
too
many
comments
and
then
it
would
get
rate
limited.
But
I
have
no
idea.
A
That
was
the
that
was
the
scalene
one
right.
That
was
the
other
one
performance
test
right.
A
F
A
Didn't
even
think
about
it
like
that
yeah
cool,
while
we
all
right
nice.
Let's
start,
I
think
we
have
critical
mass.
Let
me
share
my
screen.
A
D
E
H
This
is
just
for
this
is
for
our
own
usage.
It's
not
for
users.
Man,
change
logs
are
only
applied
to
users,
yeah.
E
I
added
a
comment:
well,
no
because
it
doesn't
change
the
the
stuff
that
the
end
user,
I
guess
yeah.
H
A
A
Oh,
is
that
is
it
a
new
person,
I
think
so
they're
not
in
the
ch
and
they're,
not
in
the.
D
I
think
he.
A
Yeah,
okay,
anyways,
let's
just
get
started
so
the
first
topic
that
we
wanted
to
talk
about
alex
and
I
were
going
through
the
contrib
repo
issues
and
some
of
the
issues
relate
to
wow,
who
removed
that
some
of
the
issues
relate
to
inconsistencies
of
implementations
between
instrumentations
that
we
have-
and
I
think,
like
it'd-
be
nice
to
have
like
a
checklist
of
things
that,
like
we
have
to
make
sure
that
the
instrumentations
do
since
there's.
A
So
many
like
one-offs
that
people
have
created
for
certain
instrumentations
that
we
feel
that
should
apply
for
everything.
So
it's,
like
you
know,
check
for
semantic
conventions.
A
Does
the
instrumentation
require
like
actual
propagation
of
like
the
headers,
such
as
like
all
of
our
http
and
grpc
ones?
You
know
filter,
slash,
naming
convention
stuff,
like
you
know,
the
span
details,
callback,
the
excluded,
urls
thing:
you
know
all
these
like
features
are
like
sprinkled
across
the
instrumentations,
but
aren't
really.
We
don't
really
have
a
consistent
story
for
it.
Ideally
like
we.
A
We
should
have
like
a
abstract
way
of
doing
this,
but
I
think
we
all
realize
that
it's
like
it's
very
difficult,
since
all
the
instrumentation
libraries
are
like
fairly
different
and
and
the
ones
that
do
have
related
stuff
between
them.
It's
like
it's
not
common
enough
that
to
like
separate
out
into
like
a
different
package.
We're
already
doing
this
as
the
best
we
can
like.
We
have
the
base
instrumentation
class
and,
like
you
know,
http
utils,
but
you
know
that's
pretty
much
like
the
best
we
can
do
really
so
yeah.
A
I
think
it'd
be
nice
to
have
a
checklist
for
implementing
these
I'll,
probably
sorry,
a
checklist
yeah,
it's
like
like
something
we
add
in
contributing
or
something
like
that.
That's
like!
Oh
make
sure
you
check
for
these
things
when
you're
implementing
a
new
interpretation
kind
of
thing,
okay,
but
that
can
can
that
be
automated.
A
Is
that
probably
yeah?
I
think
so.
It's
like
just
like,
depending
on
like
what
kind
of
instrumentation
you
have.
It's
like,
oh,
be
sure
to
check
this
be
sure
to
check
this
okay,
I
don't
know
if
it
needs
to
be
like
like
when
you
say
automated.
Do
you
mean
like,
like.
A
It
would
check
it
in
our
code
if
it's
missing.
If
that's
what
you
mean,
I
don't
know
that
that
might
be
kind
of
difficult.
I
think
we'd
still
have
to
like
manually
check
for
these
things
like
how
would
you
check
for
semantic
conventions
automatically?
A
H
Through
I
mean
it's
yeah,
reviewers
and
implementers
right,
like
anybody,
who's
submitting
a
new
instrumentation
should
should
know
if
they're
implementing,
like
an
http,
instrumentation
or
right
now
or
whatever.
A
Yeah
I'll
probably
create
an
issue
for
that.
As
I
go
through
like
finding
the
discrepancies
and
stuff
it
just
needs
a
combing
through
all
the
instrumentation.
That's
pretty
much
it
so.
F
Yeah
there's
another
thing
on
there:
that's
the
suppress,
instrumentation
stuff.
F
One
actually
from
the
spec
they,
I
don't
think
they
like
how
he
implemented
it.
So,
oh.
F
No,
they
they
pushed
it
off,
but
basically
they
were
saying
that
your
instrumentation
shouldn't
depend
on
the
sdk.
So
in
this
case
we
don't
have
a
heart
dependency,
but
that
string,
the
suppress,
instrumentation
string
is
only
in
the
sdk.
So
right,
there's
sort
of
like
a
coupling
but
okay
to
be
determined.
A
Yeah,
that
might
be
a
separate
issue
like
we
just
have
to
might
have
to
just
rework
how
we
do
this
all
right,
yeah
cool,
any
questions
from
that
anymore.
A
All
right
cool
does
anybody
have
any
other.
H
Top,
oh
sorry,
good,
sorry,
just
just
one
more
one
thing
you
did,
you
didn't
mention
that
you
know
this
checklist
could
also
help
us
by
going
through
all
of
the
existing
instrumentations,
just
to
ensure
that
they're
all
consistent
yeah,
like
one
of
the
things
that
actually
the
reason
why
this
came
out
was
because
we
were
looking
at
some
of
the
web
web
frameworks
like
web
server,
instrumentations
and
like
most
of
them,
do
things
like
implement
like
a
very
basic
propagation
test,
but
not
all
of
them,
and
it
feels
like
this
is
the
kind
of
thing
that
we
should
just
like,
not
not
enforce,
but
like
at
least
be
aware
of
when
we're
doing
reviews
or
whatever.
B
A
Just
a
w3c
propagator
to
make
sure
that
the
you
know
the
trace
response
looks
correct,
but
for
something
like
like
starlit,
we
don't
have
that.
You
know
just
stuff
like
that.
A
Yeah,
that's
just
an
example,
but
I
think
it
is
a
good
practice
to
be
consistent
and
in
the
in
like
in
the
heat
of
things
when
you're
implementing
it.
It's
like
it's
very
difficult
to
remember,
like
all
of
the
things
you
got
to
do,
and
even
for
reviewers,
it's
like.
Oh,
this
looks
good
enough
right.
So.
A
Yeah,
okay,
cool
any
other
topics
in
general
that
people
want
to
talk
about
before
we
get
to
issues.
A
All
right,
okay,
so
the
first
issue
that
I
put
alex
made
this
metrics
api
sda
prototype.
So
basically
for
those
who
haven't
been
caught
up
in
the
metric
space,
so
we've
we've
been.
The
metrics
community
has
been
like
like
revamping
the
api
and
sdk
specs.
A
A
So
there's
this
new!
This
is
the
api.
The
metrics
api
has
been
pretty
much
already
created
in
the
specs.
This
task
right
here,
is
to
so
the
sdk
specs
hasn't
been
created
yet,
but
it's
been,
there's
been
an
ask
from
the
python
community.
If
there's
any
interest
to
prototype
certain,
you
know
various
metric
scenarios
that
customers
will
run
into
in
order
to
drive
the
sdk
specs.
A
I
believe
they're
already
doing
this
in
dot.
Net
java
and
it'd
be
good
to
have
like
a
language
like
python
to
pick
this
up
as
well.
A
So
if
anyone
has
any
interest
in
trying
to
do
that,
like
no
one
has
made
any
progress
for
the
other
languages
yet
so
we're
all
like
still
on
the
same
page
for
this,
so
you
know
it'd
be
good
to
maybe
get
python's
name
in
there.
If
anyone
has
time
to
pick
this
up.
A
H
H
C
I
don't
know
if
it's
not
a
bad
time
to
introduce
myself,
I'm
sorry
for
joining
late.
I
was
kidding
it's
so
awkward.
A
A
Yeah
yeah:
that's
why
they
they
swapped
it
yeah.
I
think
the
community
has
a
wrong
link.
You
want
to
follow
the
one
the
calendar
yeah,
the
calendar,
cool.
A
Right,
you
want
to
not
to
put
you
on
the
spot
and
we
weren't
really
talking
about.
We
just
finished
the
topic,
so
yeah.
C
C
You
know
yeah
sure,
so
I'm
working
with
alolita
this
summer-
I
don't
know
if
you
guys
all
know
her
or
not
yeah.
C
Okay
cool,
so
I'm
just
interning
at
aws
this
summer
and
although
I
love
python
the
most
so
I
wanted
to
see
if
I
could
just
get
started
with
contributing
on
some
of
the
you
know:
grunt
work
issues.
A
Get
started
you
were
able
to
choose
which
cig
you
wanted
to
participate
in.
C
Yeah
I
mean
this
is
my
first
week,
so
we
just
got
started
and
there's
not
really
been
much
direction
given
to
us
besides,
like
just
find
some
issue
or
something
open,
telemetry
and
then
just
get
started.
So
that's
what
I'm
trying
to
do.
A
Sure,
okay,
I
guess
we
could
just
quickly
go
around
the
room,
so
I'm
layton,
I
work
at
microsoft,
I'm
one
of
the
maintainers
for
noting
python,
so
welcome.
Eddie
and
yeah
feel
free
to
go
through
our
like
contributing
docs.
Just
like
learn
like
how
our
architecture
is
set
up,
learn
more
about
open,
telemetry
and
like
if
you
need
help,
you
know,
feel
free
to
reach
out
to
me
or
any
of
the
other
guys
in
slack.
B
A
D
H
Eddie
welcome
I'm
I'm
alex.
I
work
at
lightstep
and
I'm
one
of
the
I'm,
the
owner,
maintainer
and
yeah.
It's
exciting
to
have.
You
join
so
welcome.
Yeah.
C
C
A
Thanks
erin,
you
wanna
go
next
yeah.
F
Yeah
yeah
sorry,
hey
eddie,
I'm
aaron.
I
work
at
google.
I've
been
working
on
open
source
for
about
a
year,
I'm
a
an
approver
here
so
yeah.
I
can
also
like
assign
issues
and
things
so
feel
free
to
reach
out
yeah
and
yeah.
Let
me
know
if
you
have
any
questions
about
the
spec
and
stuff
too.
I've
been
following
that
yeah.
That's
really
cool
thanks!
So
much
guys,
yeah.
E
Yeah,
well,
I'm
diego.
I
also
work
at
lightsaber,
also
another
maintainer
and
yeah
sure
contact
me
if
you
need
any
help.
G
I
mean
I
don't
know
that
approver.
G
H
H
B
Hey
I
I
I
work
on
open
telemetry
in
my
free
time.
Yeah
feel
free
to
reach
out
to
me,
like
this
wonderful
community,
yeah.
B
A
Yeah
yeah
eddie
also
we
alex,
and
I
have
worked
with
a
lolita
before,
especially
with
the
previous
interns
so
like
if
he
needs
to
like
correspond
with
her
or
anything.
You
know
to
coordinate
like
stuff
that
you
can
work
on.
You
know
just
feel
free
to
ask.
A
H
Put
in
a
request
to
update
the
zoom
link
in
the
community
repo,
so
thanks
for
pointing
that
out,
nice,
okay,.
C
A
Okay,
nice,
so
I
think
I
think
we've
addressed
the
metrics
issue
yep,
so
just
putting
that
this
on
everyone's
radar.
That's
what
we're
going
to
be
working
on
next,
so
conf
support,
retry,
timeout,
configurable.
B
If
you
remember
we,
we
decided
to
change
the
retry
timer
to
a
reasonable
number,
so
somebody
was
asking
the
if
they
can
have
a
chance
to
configure
it
because
the
for
them,
the
updated
one,
is
still
pretty
high,
so
they
they
were
asking
for.
Unfortunately,.
B
No,
it
does
not
do
that
so
so
like
if,
if
we
provide,
if
we
give
them
a
chance
to
do
that,
anything
that
that
like
if
in
future,
spec
formalizes
it
or
something
like
if
it
does,
it
will
come
as
a
breaking
chain.
So
it's
it's
our
call.
A
B
A
Okay,
so
like
yeah,
so
the
question
is
whether
or
not
we
should
support
the
configurability
of
the
timeout
mechanism.
No,
not
the
time
of
the
the
amount
of
seconds
in
timeout
for
the
otlp
exporter
right.
B
A
Okay,
so
chican's
concern
is
that,
like
this
is
not
actually
specked
out,
I
believe
and
like
so
in
the
future.
If
this
ever
was
changed,
I
guess
it
would
be
a
breaking
change,
which
is
true.
B
Yes,
so
you
want
to
do
that.
Something
then
something
that's
very
something
that
will
be
specific
to
python.
B
A
Have
that
whole
paradigm
of
like,
I
think,
like
I
also
message
this
with
away:
it's
that
whole
like
hard-coded,
so
it's
like
explicitly
configured
and
then
you
know
environment
configured
and
then
like
a
default
value.
That's
like
the
basic
architecture
that
we
always
follow
in
terms
of
like
a
configuration
so
like
to
be
complete
and
to
be
consistent.
I
guess
we
would
have
to
implement
that
too
right.
B
A
F
H
Just
like
this
is
this
is
a
problem
that
would
be
common
across
all
implementations
of
the
exporter
right.
H
B
I
don't
know
there
is.
There
is
not
any
recommendation
about
the
that
like
so,
there
is
the
there
is
no
concept
of
like
number
of
retries.
It
says
that,
like
it
should
retry
with.
You
know
exponential
backup
on
the
like
the
recommended
that
there
is
there's
also
no
recommended
timeout
like
like,
let's
say
30
seconds
or
15
seconds,
or
something
like
that
right.
It's
entirely
left
to
like
the
implementation,
so
some
some
some
time
back.
I
I
joined
the
specifications
and
asked
like
what
do
they
recommend.
B
So
the
answer
I
got
then
was:
we
do
not
recommend
anything
because
they
want
something
they
want
to
spend
some
time.
You
know
researching
it
like
the
back
end
like
the.
How
does
the
back
end
responds
back
to
this
like,
so
they
want
to
do
some
groundwork
and
then
then
recommend
the
timeout
any
time
out.
So
they
didn't
want
to.
You
know,
specify
anything
so
it's
up
to
the
languages
and
then
we
had
around
900
seconds
like
a
900
second
area.
B
H
B
A
A
Yeah,
I
think
maybe
like
for
now
you
can
like
set
this
to
a
valley
that
would
probably
like
make
this
guy
be
like
all
right.
This
is
chill,
but
also
ask
it
like
for
them
to
like
create
an
issue
like
as
like
a
feature
request
or
something
like
that,
and
we
might.
A
Right
right-
and
I
I
recognize
that
it's
just
that,
like
you
know
this,
this
like
would
solve
the
immediate
problem,
maybe
not
his
problem,
but,
like
I,
don't
think
it's
within
the
scope
of
this
pr
to
address
the
configurability
of
it.
F
B
So
I
I
looked
at
the
glowing:
go
implementation,
the
two
diamonds
that
he's
talking
about,
like
the
python,
also
has
them
so,
but
the
thing
is
so
the
batch
timeout
that
he's
talking
here
is
the
like.
The
delay
between
the
two
export
that
the
span
processor
does
and
the
export
timeout
is
the
maximum
timer.
This
span
processor
waits
for
the
export
to
finish,
but
we
are
not
using
this
export
timeout
in
our
fan
processor.
We
are
using
it
only
when,
when
flushing
like,
when
the
sdk
is
getting
shutdown.
F
Okay,
so
so,
if
they're,
if
their
exporter
is
retrying
for
like
two
minutes,
does
it
start,
does
the
batchman
processor
start
dropping
spans,
or
does
it
wait
until
just
like
a
shorter
period
and
then
drop
them.
B
F
F
All
right,
so
maybe
we
should
just
look
into
what
other
languages
do.
That's
like
a
first
step.
F
A
Handle
it
first,
if
there's
a
kind
of
a
general
consensus,
maybe
we
can
revisit
actually
just
implementing
it
yeah,
but
for
now
like
we
shouldn't
just
implement
stuff
like
really
nearly.
I
guess.
B
Cool
yeah
that
that
sounds
good,
so
I'll
like
if
it's
out
of
scope
for
this
pr
I'll
I'll
ask
him,
do
you
know
create
this
update
issue?
A
Oh,
I
think
we
covered
this
alex
right:
yep,
yeah,
okay,.
C
C
Basically,
I
think
before
it
would
keep
so
like.
If
you
pass
none
to
the
git
tracer
function,
it
would
just
like
give
an
error
and
not
return
anything
or
I
guess
it
would
throw
like
an
error
in
the
logs,
and
I
pretty
sure
it
would
that
I
have
like
a
couple
links
in
the
like
the
page,
the
google
doc.
C
So
I
think
like
if
you
click
that
first
one
it's
line
1000,
which
is
pretty
funny,
but
anyway
it
would
set
it
to
error,
missing
module
name,
and
so
I
think
right
now.
They
want
to
change
that
to
an
empty
string.
C
C
I'm
not
sure
if
there's
anything
else,
I'm
missing
here
so
as
far
as
I
can
tell,
I
think
this
would
stay
the
same
these
lines
and
then
this
way
how.
A
C
H
C
And
yeah
anyway,
so
okay
go
ahead,
so
yeah,
so
I
think
let's
see,
I
think
so.
This
also
gave
the
same
thing
where
it
would
give
like
the
name
error
missing
module
name.
So
I
think
these
will
be
the
same.
This
would
be
the
same.
I
just
add
tracer
2
to
this
and
do
another
test
like
that,
and
I
think
this
would
just
need
to
be.
A
Changed
yep,
I
think
that's
pretty
much.
It
just
be
sure
to
read
up
of
what
it
means
to
be
an
invalid
name
right.
C
A
C
A
This
is
going
to
be
my
first
like
pull
request
ever
nice
man.
It's
always
it's
always
exciting.
The
first
one
I.
A
Next
time
also
like,
we
should
not
actually
link
to
the
pr
we
should
link
to
what
the
most
recent
specs
are
like.
You
should
refer.
B
A
C
C
Yeah
they
want
us
to
work
from
like
the
aws's
organization's
fork.
I
see.
A
Now
so
yeah
exactly
yeah
cool
any
other
issues
that
people
want
to
sneak
in
all
right.
I
guess
I'll
just
move
right
along
okay,
so
this
is
an
issue
or
like
a
pr
that
I
submitted
wanted
to
talk
about
last
week,
but
I
wasn't
here,
I
think,
like
real
quick.
A
This
was
brought
up
because
instrumentations
like
fast
api
and
like
stuff,
that's
that
is
built
on
ascii,
have
like
a
kind
of
this
mechanism
where
multiple
spans
are
generated
for
like
a
single
or
a
single
request,
and
I
didn't
really
understand
how
that
worked
so
like,
if,
like
in
terms
of
like,
let's
take
a
look
here,
real
quick,
they
look
like
fast
api
right,
everyone
wait:
is
it
fast
api
or
is
it
ascii
yeah?
So,
like
you
can
see
how
like
there's
like
this
main
parent
span
with
the
span
kind?
A
That's
a
server,
and
then
here
we
like,
we
wrap
the
receive
and
send
with
another
span
that
serve
as
like
children
for
them,
and
I
was
wondering
if
anyone
has
yo
alex
what
the
that's
crazy
yeah
right,
like
people
see
that
who
saw
that
right.
Wasn't
chirping.
A
Yeah
but
but
yeah
like
so
like
there's
like
children
span
that
are
generated
that
are
generated
based
off
of
this,
and
I
was
wondering
how
like
this
translated
in
terms
of
like
how
the
behavior
would
be
like
in
terms
of
like
setting
the
parent
status
code
like
what,
if
one
of
them
fails,
you
know
like
how
the
interaction
works
and
what
would
what
this
would
look
like
in
terms
of
a
trace
so
like,
for
example,
like
apparently
like,
if
you
send
a
request
like
to
the
to
the
server
like
multiple
children
span,
can
be
generated
somehow,
but
I
don't,
I
don't
really
know
like
when
those
cases
occur
and
then
when
they
do
like
what,
if
one
of
them
fails
right,
will
that
be
reflected
in
the
parent
span
like?
A
A
I
believe
I
think
I
think
away
was
the
one
with
more
context
on
this,
but
like
from
just
like
just
looking
at
it.
I
would
think
that,
like
even
if
something
were
to
fail
or
to
error,
so
like
the
first
era
case,
it's
like
it
fails
but
like
it
sets
a
status
code
says
something
right.
So
that's
fine.
The
question
is
like:
should
it
should
the
parent
also
reflect
that
status?
H
Oh
yeah,
I
think
you're
right.
Oh,
I
have
more
context
on
this,
but
from
from
my
understanding
was
that,
like
it
wasn't
necessarily
true
that
the
that
the
channel
should
reflect
the
status
of
a
single
request
inside
that
channel,
but
at
least
that's
kind
of
my
takeaway
from
the
conversation
last
week.
I
don't
know
if
anybody
else
has
got
other
thoughts
here.
I'm
not
actually
sure
what
the
spec
says
about
this
either.
A
Oh,
it's
in
the
other
one
right,
anyways,
the
specs.
I
don't
think
it's
clearly
defined
on
like
in
terms
of
what's
the
status
or
or
as
well.
We
know
like
what
the
kind
should
be,
which
we
also
don't
do
right,
but
in
terms
of
like
setting
the
status,
I
don't
think
that
specs
talks
about
that.
So,
if
anything,
what
do
you
think
this
would
be
in
semantic
conventions.
H
It's
like
the
the
thing
is
so
so,
when
you
create
a
channel
like
your,
your
span
might
be
done
before.
All
of
the
messages
that
you're
going
to
be
passing
in
to
this
channel
are
done
right
like
right,
like
yeah,
you
want
to
have
a
one
operation
for
kind
of
wrapping.
All
of
the
all
of
the
messages
that
are
being
sent
back
and
forth
right.
A
But
what
these
would
these
child
spans
like
possibly
be
finished
before
the
parents.
A
Man
right
yeah,
it's
totally
possible
right
and
then
yeah
and
then
like.
If
something
goes
wrong
yeah,
I
guess
I'll
be
fine,
yeah,
I'm
not
sure
like
I
just
don't
I'm
not
too
familiar
with
ascii.
That's
why
I'm
not
really
sure
like
why
we
have
this
like,
like
no
other
instrumentation,
does
something
like
this.
So
it's
kind
of
unprecedented.
A
So
this
is
the
kind
of
thing
where
it's
like
like
this
is
why
we
need
to
kind
of
have
like
this
code
owners
thing
where
it's
like
who
can
whoever
contributed
this
instrumentation
would
have
more
context
on
like
the
behavior
of
it
right.
So
we
get
stuck
like
stuff
like
this,
so
this
is
a
separate
topic,
but
like
this
is
why,
like
you
know,
we
kind
of
have
to
define
what
it
means
to
like
own
a
product
now.
A
So
it's
because,
like
not
now
we're
responsible
in
like
figuring
this
out,
so
it
just
so
happens
that
ascii
is
popular
enough.
That,
like
we,
would
have
to
figure
this
out
eventually
but
like
if
this.
If
this
issue
happened
for
like
freaking
like
sk,
learn,
or
something
like
that
like
like
who
the
hell
out
of
all
of
us
like
us,
six
people,
would
they
be
able
to
solve
this
so
yeah,
just
something
to.
A
H
So
I
guess
the
the
one
thing
to
keep
in
mind
here
is
that
this
isn't
like
this
isn't
so
much
an
ascii
specific
ascii
specific
thing
as
it
is
a
tracing
semantic
conventions
scenario
right
like
ascii
is
not
the
only
web
server
out
there.
That's
going
to
open
up
a
channel
and
have
a
bunch
of
things
happen
on
the
channel.
It
doesn't
isn't
just
how
grpc
does
its
communication
anyways
like
do
we
do?
F
I
think
the
g,
our
grpc
instrumentation,
is
not
operating
at
that
level,
so
and
also
grpc
doesn't
necessarily
use
the
http
transport.
I
think
it's
like
a
considered
implementation
detail
so
like
what
our
instrumentation
does
is
it
basically
operates
at
like
the
rpc
level,
so
it
it
only
will
report
spans
like
I
don't
think
it
goes
lower
level
than
that.
It
just
reports.
Okay,
somebody
made
this
rpc
and
then
they
would
all
be
separate.
If
that
makes
sense.
A
Right,
so
if
that's
the
case,
then,
if
we're
not
making
the
parent
span
reflect
the
status
of
the
children
span,
then
this
pr
is
actually
incorrect.
Then.
H
H
A
A
A
Yeah
cool,
okay!
Let
me
leave
a
comment.
A
A
I
don't
know
okay,
nice,
so
it
seems
like
I
guess,
that's
our
last
pr,
that's
pretty
much
it
does
anybody
else.
Have
any
other
topics
or
pr's
or
issues.
Do
you
want
to
talk
about,
or
maybe
just
the
this
pr.
E
Anyways
yeah,
so
basically,
the
idea
behind
spr
is
the
every
time
anyone
opens
up
a
pr.
E
D
E
When
that
is
once
that
is
done.
The
pr,
I
think,
needs
to
be
ex
run
again,
but
it
should
pass
now
because
it
will
skip
the
check
and
that's
pretty
much
it
so.
The
soon
we
get
disapproved,
I
think,
the
better,
because
that
will
make
sure
that
we
will
not
be
adding
unnecessary
public
symbols
to
our
code.
A
Sure
diego
question
is
this:
does
this
only
apply
to
public
symbols
that
are
like
newly
added
like
what
happens
to
the
current
ones
that
we
have.
E
No,
no,
it
also
detects
anything
that
has
been
changed
because
it
pretty
much
only
looks
for
classes
variables
or
constants
or
functions
that
are
at
the
module
level
and
have
not
been
underscored
so
whether
it
is
a
change
to
current
classes
or
current
public
symbols,
or
if
they
are
new,
it
will
take
them
both
into
consideration.
E
A
Awesome.
Anything
else.
E
Do
this
for
instrumentation
for
contributing.
A
Like
that'd
be
good,
yeah
yeah
right
I'll,
do
it
I'll
do
so?
I
would
yeah
wait
until
this
is
merged.
You
know
just
to
apply
all
the
comments
and
stuff
so.
A
Nice,
okay,
so
I
think
that
is
it
if
nobody
else
has
any
other
pressing
matters,
I
will
see
you
guys
next
week.
Thank
you.