►
From YouTube: 2022-09-02 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
B
B
C
Let
me
get
it,
I'm
not
sure
if
it
helps
to
have
me
in
these
meetings.
I
forget
half
the
time
because
it's
friday
morning,
but
I'm
I'm
here.
A
Yeah,
I'm
glad
you're
here
I'm
waiting
on
riley
he's
having
audio
issues
so
he's
went
on
and
off
a
couple
of
times
and
his
his
audio
isn't
working
so
now
he's
decided
to
reboot
his
system
and
and
then
we'll
try
again
so.
A
I
can
appreciate
that,
especially
as
as
the
weather
changes.
C
A
Not
raining
right
so
google's
a
little
confused
or
at
least
not
sensitive
enough.
Yet
so
technology
will
start
to
improve.
Hey
riley
good,
to
see
you,
let's
try
again
audio
is
it?
Are
you.
A
A
Okay,
all
right
enjoy
your
weekend.
I
guess
there
was
one
comment,
riley
that
maybe
you
can
address
offline
listed.
Let
me
show
you
something.
There
was
a
comment
from
john
gracie.
A
Regarding
needing
more
information
that
we
had
triaged
last
week,
okay
and
and
he
was
wondering
well
gee-
we've
we've
provided
he
believes
extensive
explanation
and
so
forth
in
the
area
and
so
he's
trying
to
throughout
you.
So
what
more
information
is
there
that
he
needs
to
provide?
A
And
so
I
couldn't
specifically
remember
okay,
what
we
had
decided.
I
thought
you
or
carlos
perhaps
could
provide
a
feedback
to
them
specifically.
C
I
can
say
my
impression
of
this
issue
from
the
distance
is
that
we
ran
into
concerns.
I
think
coming
from
bogdan,
mainly
or
the
kind
of
like
I
wanna,
call
old
guard
of
tracing
saying
there's
I
mean
like
this.
This
proposal
threatens
to
create
these
traces
that
are
like
very
different
structure
than
the
normal
structure
of
traces,
and
I.
A
C
It
it
it
worries.
People
is
my
impression,
I
truly
don't
think
lightstep
has
product
support
ready
for
such
traces
either.
These
are.
This
is
pushing
tracing
in
a
direction
where
no
one's
ready
to
go
kind
of.
That's
that's
what
I
see.
A
C
I
I
yeah,
I
think
another
way
to
put
it
might
be,
that
there
are
there's
fear
and
uncertainty
which
causes
nobody
to
want
to.
Some
people
are
just
like
staying
away
from
it
fear
and
uncertainty,
not
a
technical
problem.
A
And
so
from
a
from
a
triage
standpoint,
are
we
what's
our
recommendation
to
the
individual?
Are
we
saying
that
they
should
raise
us
in
a
different
forum
for
discussion
purposes
or
you
know,
what's
the
action
to
them
or
what's
the
action
to
the
cc
or
whatever
in
terms
of
this
area,.
C
C
A
C
Know
we
have
a
customer,
won't
name
them
who
has
expressed
desire
to
see
kafka
traces
that
are
more
dynamic
than
you
get
from
the
status
quo.
I.
A
C
It
means
this
stuff
needs
to
be
done,
but
it
also
means
sometimes
some
sort
of
like
model
for
how
you're
going
to
use
those
choices
in
reality
in
practice
that
I
don't
think
that
a
lot
of
people
have
we've
seen
people
write,
I
mean
it's
just
like
not
a
standard
application
for
the
legacy
trace
data.
A
C
A
A
Probably
things
that
we
need
need
another
category
called
need
discussion
to
clarify
and.
A
A
A
This
is
all
right
and
we
should
just
remove
the
triage.
We
need
more
information
yeah,
we
probably
should
remove
the
triage
messaging.
Even
it
doesn't
promptly
categorize
this.
A
Yeah
all
right,
great
and
then
and
the
other
aspect
in
terms
of
the
the
triage
aspect.
What
we
do
here,
I
think
they're,
the
the
need
more
information
whenever
we
do
decide
in
categorizing,
something
as
as
needing
more
information
or
when
we
reject
something.
A
We
need
to
provide
a
rationale
for
that
right
and
then,
therefore,
we've
captured
our
thoughts
and
that
can
be
laid.
That's
communicated
to
to
the
original
author.
There.
A
All
right
and
so
josh
we
would
normally.
A
Go
through
the
new
ones
that
have
occurred
over
the
last
seven
days
and
try
to
decide
what
we
want
to
do
with.
I
guess
we
can
go
ahead
and
try
that
if
you'd
like.
C
I
I
mean
I
I
feel
this
is
a
post,
1.0
issue.
It's
true.
We
could
clarify
the
current
spec
and
straighten
up
the
language,
but
what
people
are
really
asking
for
is
more
work
on
histograms
that
have
different
types
I
put
together
with
engage
histogram
too.
C
And
I
have
a
theory
about
it,
but
it's
just
like
not
the
time
for
it.
I
think.
C
There's
a
pre
1.0,
spec
clarification
being
requested
quite
clearly,
but
there's,
but
the
implication
here
is
a
post
1.0
issue
to
develop
something
to
allow
negative
issue
numbers
like
and
we're
in
a
place.
We
we've
always
been
in
place
where
the
otlp
protocol
is
a
superset
of
what
you
can
actually
do
with
the
apis
and
what
we're
asking
for
is
to
make
the
apis
a
little
stronger
so
that
we
can
do
what
the
data
model
allows.
C
C
B
C
Guess
what
I'm
saying
is
the
the?
What
did
you
expect
to
see?
We
can
do
that
now.
I
agree
accept
that
the
additional
context
is
very
suggestive
of
another
issue
it
may
be
already
filed.
It
is
not
a
it's,
not
a
1.0.
It's
a
post,
1.0
issue.
This
additional
context
is
asking
for
something,
so
I
guess
we
say
accepted
willing
to.
A
A
A
A
A
B
A
C
A
C
I'm
looking
at
2759
typing
a
comment.
It
looks
like
a
mistake
to
me:
it's
it's
a
bug.
I
I
believe
that
our
that
we're
just
missing
a
number
we
should
support
up
to
ten
thousand
as
default
and.
C
Anyway,
this
is
a
small
bug.
We
can
get
it
fixed
all
right.
I
can't
tell
yeah
because
because
remember
the
defaults
and
prometheus
were
10
seconds
and
we
decided
to
make
integers
be
a
thousand
times
what
the
default
square
floating
point
so
that,
if,
if
you
measured
time
in
milliseconds,
which
is
pretty
standard,
you'd
get
the
same
numbers.
C
Thus,
the
upper
boundary
is
10
for
floating
points
and
10
000
for
integers,
and
I
would
accept
that
change.
I
don't
think
riley
would
disagree
either.
B
Yeah
you
could
assign
this
to
me.
I
I
I
consider
a
fix,
because
I
also
got
people
pinged
regarding
why
this
was
designed.
So
I
I'll
just
add
10
seconds
and
also
a
little
bit
like
why
we
choose
this.
A
B
There
are
some
places
we
use
seconds
and
there
are
places
where
we
use
different
units
for
time.
So
the
things
were
inconsistent
and
the
trouble
here
is
people
are
saying
if
I
want
to
define
a
view
and
take
a
particular
thing
that
that
needs
to
be
a
counter
or
something.
But
I
want
to
convert
that
to
histogram,
then
I'm
I'm
screwed.
C
Yeah
no,
I
I
also
have
seen
this
I'm
working
on
some
runtime
instrumentation
for
go
and
I
prefer
the
seconds
unit.
I
mean,
there's
probably
a
compression
argument,
integers
compress
a
little
bit
better,
but
I
so
that
we
do
have
respect
for
system
cpu
time
process
cpu
time
in
seconds
and
you
know
yeah.
I
I
also
feel
like
post
1.0
I'd
be
glad
to
talk
about
units
conversion,
and
maybe
that
is
connected
with
a
histogram
issue.
B
C
Oh
yeah,
I
think
that
it
would
be
cool
if
we
were
in
a
place
where
the
spec
allowed
you
to
convert
between
them
and
like
that'd,
be
nice.
So
we
know
how
to
convert
nanoseconds
to
seconds
and
that,
I
think,
is
kind
of
sort
of
allowed
in
the
data
model.
Thing
like
you
might
want
to
convert
units
and
but
we're
not
saying
how
to
do
it
so
question
mark.
I
don't
feel
that
we
should
be
mandating
order
of
magnitude
in
our
units.
C
I
think
it
should
be
okay
to
mix
time
and
units,
but
I
know
that
as
a
vendor
that
my
back
end
team
hates
that
idea,
the
collector
could
be
configured
to
convert
units
on
the
right
path.
I
think,
but.
C
I
guess
that's
my
answer
finally
found
it.
Yeah
choosing
your
units
so
that
the
numbers
are
measured
near
one
is
actually
a
pretty
good
plan
and
I
don't
know
milliseconds
near
a
thousand.
That's
not
so
bad.
It's
only
three
orders
of
magnitude
off
or
whatever,
but
yeah.
I
know
I
think
people
should
be
allowed
to
change
it
and
that's
it
because.
A
Okay,
cool:
let's:
let's
go
on
to
this
next
one,
then
if
we
can.
C
C
Yeah
I
mean
the
data
model
made
the
breaking
change.
I
guess,
if
what
he's
saying
is
true,
because
we
define
anyway,
this
isn't
gonna
help.
My
words
are
not
gonna
help
with
logan's
concern.
C
C
C
We
could
specify
our
way
out
of
this
too.
Here's
a
here's,
an
idea,
define
a
conversion
from
oh
shoot.
No,
I
can't
do
that
nevermind.
I
wanted
to
find
like
a
legacy.
Conversion
like
if
you
have
scope,
attributes
append
them
to
the
name
and
make
the
name
unique,
but
the
consumer
will
not
know
about
the
name.
C
C
Yeah
the
data
model
allows
conflicts,
defines
conflicts
and
says:
when
you
have
them,
you
should
pass
them
through.
In
that
sense,
it's
a
breaking
change
that
the
spec
allows
to
happen.
It's
a
creation
of
a
conflict.
C
C
B
A
A
But
it's
not
something
that
we're
going
to
reject
right,
yeah,
it's
not
something
that
we
say
needs
more
information.
I.
B
C
C
Let's
discuss
this
in
tuesday's
meeting,
I
will
comment
on
it.
It's
not
a
quick
one.
It's
a
big
one
and.