►
From YouTube: 2022-12-01 meeting
Description
Instrumentation: Messaging
A
C
A
Let's
see
what
are
we
at
two
minutes
after
I
think
we
could
probably
get
started
here.
Let
me
start
sharing
my
screen.
A
Okay,
welcome
everyone
start
us
off.
If
you
haven't
already,
please
add
yourself
to
the
attendees
list
and
they
have
agenda
items
you
want
to
talk
about.
Please
add
them
to
the
agenda
here.
If
we
can
jump
in
Aaron,
you
added
the
potential
Fair
release
I'll
hand
it
over
to
you.
B
We
are
on
the
very
verge
of
a
release.
There
are
three
issues
that
are
kind
of
outstanding
for
the
112
and
the
3-4
release.
I,
don't
know
if
that's
the
right
number
I
feel
like
that's
wrong,
but
but
anyways
the
semcom
packages.
I
know
Tyler
has
added
the
version
1.13
and
I
figure.
If
we
get
1.13
and
the
1.14
should
follow
pretty
soon
after
but
needs
eyeballs
needs
reviews.
If
you
have
time,
please
go
and
check
it
out.
B
That's
it
for
that.
For
the
metrics
Milestone
there's
one
issue
in
PR
left
I
was
actually
going
to
approve
it
the
only
or
going
to
merge
it.
It
has
the
approvals
necessary
I
just
wanted
to
double
check,
Tyler
that
the
last
round
of
reviews,
you've
yeah.
A
That
I
wanted
to
address
on
here,
namely
around
this
error,
handling.
A
An
in
Focus,
it's
not
dropping
anything,
it's
just
telling
everyone
that's
reviewing
a
description.
This
is
I
think
needs
to
be
log
as
an
error,
but
it's
not
like
at
the
end
of
the
world
if
it
doesn't
get
logged
as
an
error.
B
Okay,
that's
that's
fine
yeah,
but
right
now
it's
starting
at
two
approvals
and
ready
to
merge
so.
A
Just
sure
okay
I
wanted
to
provide
a
little
more
feedback
from
that.
But
that's
that's
fine.
A
B
Why
I
took
took
a
second
to
give
that
so
with
that
outstanding,
even
still
more
eyeballs?
More
approvals
on
that
would
also
be
appreciated.
I
think
it's
just
about
ready
to
merge
mine,
less
the
feedback
from
Tyler,
so
that
puts
us
at
when
do
we
think
we
can
make
a
release.
A
A
If
you
look
at
what's
already
been
merged
for
this
release,
it's
all
really
just
bug
fixes.
So
I
think
this
would
be
released.
As
you
know,
One
Eleven,
one
if
we
didn't
include
the
semconf
stuff,
because
that's
really
like
the
additional
things
that
being
said
like
this,
has
been
open
since
September.
You
know
when
when
113
was
released,
yeah
I
think
115
might
even
be
out
right
now,
oh
boy,
that's
we've.
A
There's
another
one,
even
that
we
need
to
do
so,
I'm
fine,
with
pushing
this
out.
I
I
would
like
to
get
this
in.
So
one
of
the
main
things
that
was
really
a
problem
was
in
113,
there
was
a
big
refactor
to
the
network
in
HTTP
semantic
conventions
in
in
how
they're
defined
to
be
including
addresses
and
other
sort
of
attributes
associated
with
the
network
and
our
existing
functionality
around.
That
was,
it
was
it's
definitely
Legacy.
A
We
had
you
know,
functions
somewhere
that
were
I,
think
kind
of
ad
hoc,
and
they
weren't
really
well
I,
think
constructed
with
the
concepts
that
an
instrumentation
Library
would
need
included.
So
I
took
the
time
you
know
given.
A
This
is
a
time
where
these
would
all
be
to
change
some
of
them
kind
of
even
go
away
to
try
to
refactor
them
to
try
to
be
a
little
bit
more
distinct
into
what
the
purpose
of
they
are
or
of
the
of
the
function
is
and
what
the
specification
defines
so
say
here
like
we
have
now
client
responses,
client
requests,
server
requests
and
it
Scopes
things
appropriately.
A
Based
on
you
know,
if
you
have
a
request
coming
from
a
server
request
coming
from
a
client
or
something
like
that,
so
I
I
try
to
do
that.
One
of
the
things
that
was
kind
of
pointed
out
is
that
like
well
maybe
I'll,
step
back
to
one
of
the
things
is
also
in
the
structuring.
So,
instead
of
putting
everything
in
this
semcom
package,
it
was
also
there's
these
sub
packages,
HTTP,
confident,
netconf,
somewhere,
yeah
and
so
like
it
was
structured
in
a
little
bit
different
way.
A
I
I
went
back
and
forth
on
a
lot
of
these
I
think
that
this
is
probably
ideal
for
usage.
A
lot
of
the
issues
that
we
have
from
people
using
semcom
is
that
they
have
to
update
this
version,
and
then
they
have
to
like
Alias
their
Imports.
A
This
way
they
could
just
update
the
package
version
and
the
Alias
I
don't
know
the
package
name
should
remain
the
same,
but
there's
a
lot
of
questions
around
it.
I
think
it
is
a
new
approach.
A
I
think
one
of
the
the
nice
things
about
our
package,
structuring
in
semconf
is
that,
like
each
new
version
is
its
own
package,
so
we
can
kind
of
change
it.
However,
we
need
to
backwards
compatibility,
isn't
I
mean
we
are
remaining
backwards
compatible.
You
know,
because
we're
not
changing
the
existing
package
API.
A
So
the
new
API
for
the
new
package
is
how
that's
defined,
which
is
kind
of
nice,
but
it
also
you
know,
is
needs
to
be
kind
of
put
into
consideration
for
people
that
are
using
it
like
if
they
just
want
to
do
an
update.
They'd,
rather
probably
not
have
to
do
a
bunch
of
function.
Changes
and
just
do
this
update
so
I
think
Trying
to
minimize
that
that
amount
of
churn
is
ideal.
A
That
being
said,
I
do
think
that
this
is
a
positive
step
in
that
direction,
and
so
I
think
that
this
is
a
worthy
version
of
it.
The
net
conf
was
pointed
out
that
it
really
only
has
one
function
in
it.
That's
exported,
and
that's
transport
I
had
done
that
somewhat
intentionally.
I
had
a
lot
of
other
ideas
around
what
it
could
contain,
and
Aaron
pointed
out
so
I've
been
ever
since
I
saw
this
comment
and
working
to
try
to
add
them.
A
A
That
accepts
like
an
address
and
I
think
I've
defined
it
as
like
an
address
and
a
connection
or
an
address
and
a
listener,
and
then
it
will
annotate
all
of
the
attributes
that
you
need
for
those
which
is
I,
think
it
could
be
useful
in
the
grpc
library.
I
think
that
that's
somewhere,
where
we
have
a
client
and
we
have
an
address,
and
so
it
could,
you
know,
conform
with
all
of
the
specifications.
A
So
I
was
going
to
add
them
I'm
in
the
middle
of
writing
a
bunch
of
tests
for
it
I
have
it
all
together.
So
it
was
going
to
be
added
in
question.
A
little
bit
but
yeah
I
mean
the
original
intent
was
just
to
like.
Have
you
know,
compatibility
with
form
or
function,
and
so
transport
was
somewhat
of
its
own,
but
it
was
also
in
this,
like
distinct
net
package,
not
HTTP,
that's
why
I
put
it
in
a
different
one,
but
yeah,
so
that's
outstanding.
A
So
there's
still
a
little
bit
of
work
to
do
here
and
I
think
this
is
kind
of
like
coming
back
to
that
high
level.
Discussion
of
like
you
know,
this
is
kind
of
blocking
that
V12
Milestone
or
v112
Milestone.
So
I'd
like
for
this
to
be
included,
given
how
late
we
are
to
the
game
on
this
one.
But
I
could
also
understand
if
people
wanted
to
spend
some
time
reviewing
this
and
think
through
this
one
to
pull
it
out
and
instead
of
doing
a
v112,
we
do
a
v11.1
or
something
like
that
release.
C
Mike
I'm
good,
either
way
like
if
you've
got
a
bunch
of
bug,
fixes
batched
up,
let's
get
them
out
there,
but
if
we're
also
comfortable
with
this
approach,
it
shouldn't
be
much
reason
to
hold
up
and
let's,
let's
get
to
do
functionality
there
too,.
A
Okay
yeah,
then
it
would
require
more
reviews.
Aaron's
got
some
feedback
I'm
trying
to
address
right
now,
but
other
another
set
of
eyes
on
that
PR
would
be
ideal.
A
I'd,
like
I,
mean
I'd
like
to
get
it
out.
I
think
it's
a
better
API
I
think
it's
more
in
line
with
the
specification
defines
and
it's
a
little
bit
more
clear
from
the
user's
perspective
and
instrumentation
authors.
A
So
I'm,
okay
with
it
I,
also
think,
like
kind
of
what
I
said
is
like
if
we
find
out
that
this
was
just
like
a
horrible
idea,
we
can
always
in
the
next
semcon
package,
make
a
change.
You
know
if
it
really
is
motivated,
but
I
do
think
that
that's
I
don't
know.
I
I,
like
the
API,
a
little
better,
so
yeah,
okay,
well,
I!
Think
if
that's
the
case,
so
maybe
then
also
to
kind
of
frame.
A
That
is
when
do
we
think
that
the
release
go
out,
and
so
I
think
this
could
probably
depends
on
the
author's
response
here.
A
But
I
don't
see
why
this
couldn't
get
merged
today,
if
I
can
get
a
suggestion
and
then
maybe
able
to
just
have
it
something
committed
and
then
I
think
once
this
gets
merged.
I
think
that
we
could
open
an
issue
to
try
to
to
get
the
release
out.
I
was
hoping
to
get
the
release
out
to
this
week,
but
it
was
been
a
busy
week,
so
it
didn't
happen,
but
yeah
I
would
like
to
at
least
get
it
out
early
next
week.
A
If
that's
the
case
so
I
guess
it
kind
of
depends
on.
If
if
we
can
get
this
merged
or
this
Milestone
complete
and
then
it
depends
on
then
what
the
review
side
state
of
this
is
I.
Think.
B
I'm
100
on
board,
so
would
it
be
fair
to
say
that
who's
available
to
do
a
release,
I
mean
I,
can
do
it
I
could
do
it
either
tomorrow
or
Monday
morning
pen.
You
know,
assuming
that
there
is
a.
A
Yeah
I'm
out
next
Monday
morning,
but
other
than
that
I
could
do
it
tomorrow
or
next
Tuesday
as
well,
so
I
mean
I'm
fine
with
either
okay
or
if
I'm
yeah.
If
you
want
to
take
it
on
that
sounds
good
too.
B
I'll
start
the
release
as
soon
as
that's
done
so.
A
So
also
just
to
kind
of
point
it
out
like
if
we
do
include
the
semantic
convention
stuff,
the
release
will
be
harder
because
contrib's
gonna
need
to
get
updated
and
the
HTTP
HTTP
Trace
instrumentation,
as
well
as
grbc,
probably
as
well
need
to
be
updated
to
use
the
new
functionality
of
the
semantic
convention
packages.
So
that
would
I
would
imagine
take
a
day
or
so
in
itself.
B
C
A
A
A
The
thing
that
I've
heard
feedback
from
the
community
on
as
well
as
I'm
sure
other
people
is
the
Prometheus
exporter
stuff,
is
really
needed.
The
semantic
conventions-
v113
I,
don't
know
if
it's
as
needed,
I
feel
guilty,
because
it's
really
far
behind
but
I'm
kind
of
leaning
towards
just
saying
like
we
pulled
that
out
and
then
maybe
in
the
next
release
cycle,
try
to
work
on
that
I
think
more
as
a
priority.
E
A
I
I
think
you're
right
Evan.
Oh,
we
already
did
One
Eleven
one
okay,.
A
Okay,
all
right
so
I've
updated
that
Milestone.
The
only
thing
I
think
then
blocking
us
is
this
PR
here
at
this
point,
but
other
than
that
then
I
think
from
there
we'll
have
Aaron's
gonna
plan
to
take
the
release
for
whenever
this
is
whenever.
A
Okay,
cool
Aaron,
anything
else.
You
want
to
add
to
that
or
be
ready
to
okay.
Next
up,
we
have
the
contributing
code
owners
file.
We
talked
about
this
a
little
while
ago,
if
I
guess
about
a
month
ago.
A
This
relates
to
this
PR,
where
we
wanted
to
have
an
ownership
model
defined
across
the
control
repository
for
any
sort
of
instrumentation
and
other
things
included
here
and
something
that's
also
blocking
the
code
generation
for
instrumentation,
that's
being
donated
from
the
auto
instrumentation
sake
is:
is
this
policy,
so
Arizona
theory
has
done
an
inventory
of
all
the
open
modules
that
we
need
to
get
assigned.
A
This
PR
is
updating
the
code
orders
file
to
actually
do
that
assignment,
So
the
plan
just
for
everyone.
That's
on
the
call,
is
anything
that
doesn't
have
well
I
guess
right
now
we
just
need
an
owner.
These
also
be
assigned
a
main
painter,
but
I
think
that's
between
Aaron
Anthony,
myself
will
kind
of
think
dish
that
out.
If
it
doesn't
happen,
organically,
that's
happening
here.
A
So
the
idea
is
anything
that
doesn't
actually
have
an
owner
is
going
to
become
deprecated
and,
that's
not
to
say
it
can't
be
undeprecated
in
the
future.
If
somebody
does
want
to
take
on
the
ownership,
but
then
after
that,
module
is
deprecated,
we'll
remove
it
from
the
repository
with
the
purpose
that
it
needs
to
be
I,
think
maintained
and
if
it's
doesn't
have
an
owner,
it's
not
being
maintained.
So
we
wanted
to
to
remove
that.
I
think
that
this
follows
in
suit
with
the
collector
contrib
policy.
A
So
historically,
it
will
always
be
there,
but
since
it's
not
actively
being
maintained,
it's
going
to
be
removed
and
anything
that's
added
to
the
repository,
also
as
a
new
module
needs
to
also
follow
suit.
In
this
policy,
this
PR
started
to
dish
out
those
ownerships.
A
Obviously,
there's
going
to
be
some
I
think
Auto
tagging
that
needs
to
be
done
in
Auto,
I
think
everyone's
going
to
get
the
reviews
that
are
needed
appropriately,
but
like
there's
still
some
I,
think
policy
and
and
sort
of
functional
niceness
that
we
could
add
to
the
repository.
But
right
now
this
is
kind
of
the
starting
point,
and
once
we
have,
this
I
think
we're
gonna
just
move
on
next
steps
as
being
deprecated.
A
So
there's
a
question
is
when
we
should
deprecate
things,
I
think
here,
I
said
we
wanted
a
two-week
cycle
and
I
think
just
like
waiting
two
weeks
to
merge
this
PR
or
just
before
two
weeks,
I
think
is
ideal.
The
reason
for
the
two-week
cycle
is
next
auto
transportation.
A
Sig
meeting
is
in
two
weeks,
I
guess
a
week
and
a
half
at
this
point
so
I
wanted
to
make
sure
that,
like
we
have
a
full
cycle
to
talk
about
here
and
then
go
back
and
talk
about
it
with
them
and
to
make
sure
that,
like
that's
understood
and
that
I
think
from
there
we'll
merge
that
and
start
your
deprecation
process,
so
yeah
any
feedback
on
that.
Others
want
to
contribute
to.
D
A
A
It
was
recommended
that
a
prb
open
in
this
repository
starting
that
process
with
a
in
addition
to
the
code
owners
file
that
would
conform
to
this
kind
of
policy
we've
laid
out
I,
think
the
blocked
part
is
that
officially
we
haven't
merged
something
describing
this
policy,
which
is
also
something
we
should
probably
do,
but
yeah
I
mean
they
were
blocked
in
the
sense
that
we
wanted
to
get
this
done,
and
this
is
kind
of
one
of
those
first
steps
in
we've
adopted
this
I
guess.
A
Unless
this
PR
emerged
is
merged,
so
yeah
I.
D
Just
just
on
the
probability
sampler
and
to
throw
out
a
name
I
know:
Josh
probably
had
some
interest
in
that
right,
but
the
other
thing
is
maybe
the
is
doesn't
have
to
be
just
a
single
person.
There
right,
you
can
also
I
presume,
say
something
like
open,
Telemetry,
sampling,
SQL
or
something
like
that.
If
there
are
people
who
or
do
you
want
it
to
be
an
individual
person
there
well.
B
A
A
fair
question:
the
issue
with
assigning
a
group
to
it
is
that
the
group
can
have
zero
members
and
the
accountability
is
easy
to
assume.
Somebody
else
is
accountable
for
it.
I
I,
don't
know
if
that's
always
going
to
be
the
case,
but.
B
I,
don't
I
I
would
suggest
that
or
owner
responsibility.
B
There
should
be
people
whether
or
not
it's
like
an
additional
to
that.
There's
also
the
other
people.
Remember
that
this
this
I,
don't
believe,
gives
you
any
additional
right
capabilities.
This
is
just
who,
who
do
we
need
to
go
talk
to
when
there
is
a
PR
like
who
do,
users
get
should
go
to
when
there
is
a
PR
and
additionally,
who
gets
tagged
in
the
PRS
so
having
a
person
as
a
point
of
contact
or
multiple
people
like
that?
That
is
ideal.
B
C
I
think,
though,
for
something
like
the
Z
pages
extension,
which
is
you
know,
needed
by
The,
Collector,
that
it
would
be
reasonable
to
have
the
collector
maintainers
be
the
owners
of
that
as
a
whole.
There
will
always
be
Collective,
maintainers,
I,
think
and
I.
Don't
know
that
we
should
be
wanting
to
keep
up
to
date
with
changes
in
that
group
membership
and
rather
to
just
specify
the
group,
but
I
also
want
them
to
accept
that
responsibility.
B
A
A
If
it
is,
we
have
other
problems,
but
I
also
think
that,
like
yeah
to
to
I
I,
do
like
the
idea
of
being
able
to
communicate
to
a
person,
because
it
holds
that
I
think
responsibility
I,
who
are
The
Collector,
maintainers,
Anthony,.
A
E
A
As
it's
not
like
10
people,
you
know
because
then
you
can
go,
paying
I,
don't
know
one
or
three
of
them
or
something
like.
C
That
right
right,
which
is
why
I
would
say,
maintainers
and
not
approvers,
but
although
I
don't
think
there
are
that
many
core
approvers
either
yeah
maintainers,
to
keep
it
a
small
and
focus
group
and
the
people
who
are
ultimately
responsible
for
that.
A
Yeah
I
do
also
to
Aaron's
point
we
had
talked
about
in
the
past
if
GitHub
allows
it
eventually
allowing
these
owners
to
merge
PRS
in
those
directories
which
I
still
am
in
favor
for
but
I
I
don't
know
if
it
allows
it
yet
I
think
that
was
a
a
feature
that
GitHub
didn't
have.
C
Yeah
collector
contrib
still
hasn't
come
up
with
a
solution
for
that.
The
best
that
I
think
that
has
been
proposed
so
far
is
basically
a
bot
that
has
merged
permissions
and
can
validate
comments
from
owners.
A
Yeah,
that's
a
little
scary,
but
yeah
yeah,
okay,
okay,
yeah,
I,
I,
think
I
think
especially
Z
pages.
There
may
be
a
reason
to
just
say
you
know
a
strong
reason
to
just
say
that
the
maintainers
but
I
think
also
for
the
instrumentation
like
I,
would
be
a
little
hesitant.
If
somebody
came
along
and
said,
like
you
know,
company
XYZ
wants
to
own
this.
So
just
put
our
group
here
because
then
there's
really
nobody.
A
You
know
that
company
that
you
are
I
think
directly
in
line
with
you
know,
one
of
the
other
things
is
I
know
and
The
Collector
contribute
like
once.
Somebody
left
that,
like
the
ownership
like
they
took
their
name
off
as
a
way
to
signify
that
nobody's
there
anymore
and
I.
Think
that
that's
more
of
a
positive
signal
to
also
saying
like
hey.
This
is
in
Jeopardy
of
being
deprecated.
If
there's
going
to
be
no
owner
or
something
like
that,
so.
A
But
yeah
I
think
that
if
it's
an
internal
group
in
the
maybe
even
say
like
an
internal
maintainer
group
in
hotel
I
think
that
that's
that's
a
pretty
safe
bet
at
that
point,
to
add
them
as
an
owner
Anthony.
Is
there
a
policy
document
in
the
contrib
collector
contrib
that
we
could
copy,
or
should
we
just
start
generating
one.
A
Yeah
it'd
be
ideal
for
copy,
but
I.
You
know
I
just
want
to
make
sure
like.
We
set
clear
expectations
for
people
in
the
in
the
repository
and
what
they
know
when
they're,
adding
or
when
they.
A
Ownership
of
something
as
well
I
think
that's
also
true,
because
if
I
remember
correctly,
I'd
heard
that
like,
if
you
aren't
responsive
within
a
certain
amount
of
time,
then
you
also
are
yeah
yeah
Calhoun
can't
exactly
Okay
cool,
so
timeline
on.
That
is
we
need
we
can
have.
We
do
need
an
issue
I
think
or
maybe
that
initial
to
find
ownership
model
for
packages
that
live
in
this
repository
could
include
an
action
item
to
include
a
policy
PR
as
well.
I
think
that's!
That's
fair!
A
Yeah,
let's
see,
okay
becoming
calendar
cool.
All
right,
cool
I
will
add
that
to
here
a.
C
Bit
further
up
in,
there
is
also
a
section
not
adding
new
components
that
has
discussion
of
you
know
the
need
to
obtain
a
sponsor
someone
from
within
the
group
who
is
also
willing
to
become
a
code
owner
and
the
need
for
it.
You
know
if
it's
a
vendor
component,
someone
from
that
vendor
needs
to
be
one
of
the
sign
up
to
be
one
of
the
code
owners.
A
A
Okay,
perfect!
So
then
that's
also
a
PR.
We
need
for
this,
but
otherwise
I
think
if
that's
moving
forward,
I
am
anticipating
a
PR
for
the
auto
generation
of
instrumentation
whenever
the
contributor
has
time
so
we'll
we'll
see,
but
in
two
weeks
I'm
guessing.
That's
probably
also
the
timeline.
A
Okay,
cool
we're
at
the
end
of
the
agenda
pause
here
in
case
anybody
has
something
else
they
wanted
to
talk
about,
that
isn't
on
the
agenda.
C
A
Oh
yeah
I
usually
get
about
12
days
into
that:
okay,
cool
yeah.
B
B
C
I
think
there
was
recently
a
an
article
on
the
kubernetes
blog
about
tracing
in
kubernetes
and
or
the
the
integration
of
open
Telemetry.
And
let
me
see
if
I
can
find
that
it's
going
to
be.
E
Yeah
I
just
saw
that
it
was
talking
about
all
the
different
components
that
have
added
it
and
like
the
API
server,
tracing
status
and
stuff.
A
Yeah
I
hope
they
mentioned
David
by
name
in
there.
A
Okay,
yeah
Anthony.
If
you
want
to
post
that
in
the
slack
channel
that
also
oh,
perfect,
yeah
or
in
the
chat.
A
It's
worth
giving
it
to
read
I'm,
actually
pretty
excited
to
read
about
this
awesome.
Okay,
well,
I!
Think
with
that
we
could
probably
end
it
early,
get
a
half
hour
back
for
everyone,
so
yeah.
If
you
have
something
else,
you
want
to
talk
about
definitely
in
slack
or
via
issues.
Otherwise
we
should
show
we
should
still
be
meeting
next
week
right,
like
that's
yeah,
this
release,
okay,
so
yeah
I,
guess
we'll
see
you
all
next
week
at
the
same
place
same
time,
see.