►
From YouTube: 2020-08-31 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
A
A
C
B
B
D
E
By
the
way,
guys,
I
added
a
small
item
by
the
end
of
the
dock,
just
to
mention
what
are
the
important
or
breaking
changes
in
the
specification
yeah.
This
is
just
to
signal,
of
course,
maintainers
on
you
know,
so
they
more.
Let's
have
an
idea.
B
Oh
no
worries:
are
you
ready
to
go
over
status.
G
So
yeah
I'll
give
the
overview
of
the
ga
spec
burden
now
changes
since
last
week,
drill
down
to
p1,
and
then
I've
got
some
other
items.
I'd
like
to
go
over
as
well
we're
making
good
movement
on
the
done
column
like
stuff
still
keeps
getting
done.
My
cursory
view
of
it,
it's
not
all
p
ones,
of
course,
as
you'll
see
going
forward,
is
a
lot
of
p2s
a
few
p3s.
Something
like
that.
I
don't
have
the
exact
breakdown.
G
This
is
just
cursory
glance,
so
our
to
do
columns
going
down
in
general
in
in
progress
this
column
doesn't
seem
to
use
update
as
much
so
I
think
this
is
just
like
more
of
a
better
signal
as
to
like
how
far
we're
getting
through
as
a
ratio
to
the
two
do's
for
the
drill
down
into
the
p1
issues.
G
Labeled
with
spec,
we
still
have
four
items.
This
was
as
of
friday
when
I
put
this
together.
G
Today,
so
that
that
hasn't
changed
dunn
adams
is
11..
I
think
there
was
some
more
discussion.
I
had
a
quick
look
this
morning
on
the
arrows
hint
issue.
I
think
that
was
one
related
to
one
of
the
longest
polls,
the
the
issue
that
was
identified
by
the
tcs,
possibly
one
of
the
most
contentious
ones
in
order
to
get
in.
So
if
that
one
goes
in,
that's
like
a
big
hurdle
to
get
through
the
rest
of
the
to
do's.
G
G
I
think
we
ought
to
talk
about
at
some
point
what
this
means
for
the
remainder
of
this
week,
because
this
week's
the
first
week
of
september,
starting
tomorrow
and
the
implications
of
going
into
next
week,
whether
we
need
to
start
bringing
out
the
accelerated
action
plan
that
was
mentioned
in
morgan's
blog
post.
You
know
as
to
tc's
being
some
issues
or
just
make
some
like
hard
calls
and
certain
things
don't
know
if
that's
necessary
or
not.
But
would
you
guys
need
like
to
talk
about
that
related
to
this?
Now.
G
So
trace
state,
let
me
see
we'll
just
scroll
down
on
gita
carlos.
A
B
H
A
I
A
B
A
And
the
one
that
we
have
in
progress,
this
is
so
there
is
an
autep
for
that
right,
which
is
blocked
by
error
hint
conventions,
you're.
A
Yes,
so
what
I
suggest
to
do,
I
suggest
to
merge
the
error
in
asap
and
to
be
able
to
do
that
break
it
down
into
two
separate
prs.
There
is
a
part
that
everybody
agrees
with,
which
is
actually
the
existence
of
the
error,
heat
conventions
and
that
will
unblock
the
okay
about
removal
of
the
spawn
status
and
the
remaining
part
of
that
pr
is
about
the
what's
the
function
name
for
the
api,
whether
it's
record
exceptions
or
odd
event.
A
J
The
author
of
the
air
and
api
api
and
semantic
convention-
pr-
and
I
don't
know-
I
definitely
have
some
reservations
about
splitting
into
two.
I
feel
like
the
semantic
convention.
It
is
hard
to
use
that
without
an
api.
I
think
the
api
goes
a
long
way
to
making
that
easy
to
use
properly.
I
guess.
A
D
J
C
B
Yeah,
okay:
are
there
any
dependencies
on
spam
status
right
now
like
is
this
something
we
could
remove
and
then,
if
we,
if
we
decided,
we
need
it,
we
could
then
re-add
it
later.
D
We
we
need
a
at
a
minimum,
a
boolean
to
indicate
that
there's
something
errorful
or
exceptional
about
this
span,
and
we
want
something
like
that
on
the
actual
span,
so
that
you
don't
have
to
parse
the
attributes
in.
D
A
A
Okay
problem
is:
we
have
this
as
a
concept
in
open
tracing
conventions.
Jaeger
uses
that
open
sensors
has
this
equivalent
concept
of
the
status,
which
we
also
use
for
error
indication.
This
is
used
by
the
collector
when
it
is
doing
the
translations
between
formats.
If
we
remove
the
span
status
and
we
don't
have
the
error
in
then
I
don't
know
how
to
do
this
in
collective
got
it.
It's
just
going
to
break
the
error,
translation.
D
Okay,
it's
much
more
like
we're
turning
span
status
into
something
else,
and
the
main
reason
we
want
to
do.
That
is
the
conventions.
The
the
labels
in
the
status
currently
are
subjective
and
people
are
struggling
with
that.
The
most
that's,
I
think,
the
main
objection
right
now
to
the
current
span
status
and
then
the
bike
shed
is
around.
D
B
A
Yeah,
okay,
so
when
is
the?
When
is
the
next
error
seek?
Can
you
guys
maybe
make
the
last
final
call
on
this,
and
we
move
forward
with
that
because
it
seems
like
you
still
have
two
parallel
suggestions:
open
the
one
is
the
error
in
the
other
is
keeping
spam
status,
but
I
think
you
you
need
to
make
a
call
on
this.
D
Right
it's
thursday,
thursday
9
a.m,
pacific,
but
I
believe
basically,
no
one
showed
up
to
the
last
one
so
yeah,
I'm
happy
to
say
that
on
thursday,
we're
going
to
make
a
call
on
this
and
okay.
B
J
K
A
So
is
it
the
position
of
the
error
c
that
if
the
current
proposal
is
not
accepted,
as
is
exactly
as
it
is
specified
in
the
pr,
then
you
you
say
that
okay,
we
don't
move
forward
with
that.
We
keep
this
status
span.
Is
that
the
position.
J
I
mean,
as
complications
keep
coming
up
with
this.
I
sometimes
like
just
ask
myself
is:
is
the
thing
that
we're
making
better
than
span
status
and
like
sometimes
the
answer
to
that
is
no?
I
guess
in
in
my
mind
so
like
I
I
personally
feel
like
you
know,
I
don't
want
to
lobby
for
something
that
isn't
an
improvement.
I
guess
is
where
I'm
coming
from
I'm
not
trying
to
like
be
a
roadblock
to
progress,
because
we
have
plenty
of
that.
I
just
want
to
make
sure
we
end
up
better
off
than
when
we
started
if.
C
C
A
As
as
one
of
the
persons
who
did
comment
on
the
improve
request,
how
do
you
feel
about
I
come
to
the
error
seek
and
we
resolve
whatever
is
remaining
between
us,
because
I'm
absolutely
willing
to
accept
it
as
a
concept.
I
just
don't
want
to
move
forward
without
with
in
a
way
that
disregards
the
opinion
of
people
about
the
apis,
because
the
api
is
also
about
other
things
right,
so
the
these
agreements,
I
see
they
are
not
about
the
concept
of
here.
This
agreement
is
about
what
the
api
that
is.
H
A
About
other
things
going
to
look
like,
so
it's
not
one-sided.
We
cannot
just
say
that
air
seek
has
the
full
authority
because
they
worked
on
this.
They
have
full
authority
to
make
the
final
call.
I
I
completely
agree
with
you
guys.
You
spend
a
lot
of
effort
on
this.
You
have
you
have
the
experience
and
you
had
all
the
discussions
to
make
a
decision
about
what
the
errors
are
going
to
look.
H
C
C
A
On
myself
that
maybe
it
took
too
long
for
me
to
come
to
the
final-
maybe
my
final
position
on
this,
but
I
I
want
to
help
you
guys
to
actually
not
just
move
forward,
but
this
is
the
week
that
we
agreed
on
having
the
final
specification
for
the
trace
on
first.
They
have
the
final
decision.
A
On
friday,
the
latest
merge
the
pr
and
then
I'm
happy
to
give
an
approval
to
the
otep
about
the
stock,
spawn
status,
removal
and-
and
we're
done
right,
does
that
work.
Can
we
make
make
this
happen
on
thursday,
because
I
I
don't
think
this
is
the
right
time
to
do
it
today.
I
could.
H
A
Invite
christian,
if
he's
willing
to
come
to
voice
his
opinion,
we
make
the
final
call
on
thursday.
A
B
C
There
is
a
very
similar
situation.
There
is
a
very
simple
solution
to
this
to
to
this
ticket
and
there
is
a
more
elaborate
solution
to
this
ticket.
B
D
Yeah
nikita,
have
you
been
following
in
the
the
sampling
group,
how
they
would
like
access
to
trace
state
read,
write
access
to
trace
state
in
the
sampling
api,
I'm
afraid
I
wasn't
following
okay,
I
think
just
fyi.
That's,
I
think
the
most
critical
thing
when
looking
at
trace
state,
not
what
the
end
user
needs
to
look
at,
but
sampling
is
an
area
where
we've
shown.
B
Okay,
I'll
put
that
at
the
bottom
of
the
issue,
next
go
ahead.
B
E
Yeah,
so
basically
we
are
we're
trying
to
it's
kind
of
two
things,
although
one
of
them
is
having
whether
to
have
or
not,
the
child
often
follows
from
concept
from
open
tracing,
which
is
something
I'm
working
on.
That's
the
important
part
which
I
will
have
a
pr
for
this
later
today.
The
second
part
is
just
making
it
more
a
formal
how
to
handle
multiple
parents,
which
we
had
in
open
dressing
as
well.
E
Now
in
open
telemetry,
we
have
one
single
parent
and
then,
if
you
want
to
define
more,
you
can
do
that
through
links.
So
it's
formalizing
this.
So,
as
I
said,
the
first
part
about
the
child
off
and
follows
from
is
something
we
don't
have
an
infant
telemetry
at
this
moment,
and
this
is
the
hard
part.
Oh,
this
is
a
complementary
fear,
the
one
that
you
are
scrolling
through
now,
but
the
previous
one
is
relatively
straightforward
and
I
hope
I
will
publish
publish
appear
for
this
one.
E
Yes
pr
today
and
let's,
let's
see
how
it
goes
from
there.
D
E
Yeah
yeah,
but
I
I
feel
I
mean
I
will
have
prs
for
both
of
them
this
week
and
I
will
of
course,
try
to
iterate
very
fast,
but
I
will
yeah.
I
think
jury
is
not
around
but
yeah.
I
you
know.
I
basically
need
help
from
people
from
open
tracing
world,
which
is
death,
who
usually
helps
a
lot
and
jury,
but
he's
usually
busy,
but
I
will
try
to
poke
him
but
yeah
anyway.
This
is
this
yeah.
I
hope
this
is
done
this
week.
E
D
E
Actually,
it's
not
it's!
It's
not
related.
You
will
explain
that
to
me
well
to
everybody
in
the
discussion
I
will
we
can
follow
up.
I
think
it's
okay,
this.
B
G
Thanks,
let
me
see
spec
compliance,
my
language
stick,
so
this
one.
I
think
this
is
fantastic.
This
is
like
breaking
down
tabular.
What
this
is
great
has
been
implemented.
I
just
have
a
few
questions
as
to
how
we
can
apply
this
and
rely
on
this
across
all
the
different
stakes.
First
and
foremost,
of
course,
is
if
the
trace
rows
are
all
there
and
do
they
represent
what
we're
trying
to
shoot
for
on
the
first
week
of
september.
A
So
I
added
the
rules.
I
think
the
trace
rules
are
complete,
but
I
don't
think
they
are
sufficient
to
declare
traces
completely
right
for
the
for
the
traces.
We
also
need,
I
think,
contest
context,
propagation
and
error
handling
tables
as
well
right
and
for
especially
for
context
propagation.
I
could
use
somehow
to
add
the
remaining
tweets
there.
A
A
A
B
I
have
not
not
lately.
I
can
get
back
involved
in
that,
though,
who
is
just
because
I
I
haven't
been
some,
I
don't
know
carlos
did.
I
see
you
raise
your
hand.
Oh
okay,
great.
E
D
And
keegan,
do
you
have
to
shop
this
around
to
the
spec?
Sorry,
the
the
sig
meetings
this
week
to
ask
them
to
to
fill
out
the
compliance
matrix.
L
Yeah
javascript
is
mostly
done.
There's
I
think,
a
couple
of
blank
rows
that
I
wasn't
100
sure
about,
but
yeah
I
mean
for
the
most
part,
it's
filled
in
yeah.
D
Actually
a
quick
side
question
while
we're
on
the
subject
of
javascript
daniel.
Sometimes
we
need
to
differentiate
between
web.js
and
node.js
and
other
times
we
just
say
javascript,
I'm
wondering
what
the
highest
order
bit
should
be
like.
Should
we
be
like
for
tables
like
this,
for
example,
is
calling
it
javascript
in
general
good
enough,
or
should
we
actually
start
training
our
users
to
differentiate
between
the
two
flavors.
L
L
L
Different
default
context
manager
and
that's
that
great.
G
Okay
and
then
for
the
error
handling
section:
is
there
a
suitable
person
to
fill
that
out.
D
B
B
For
sure
cool
and
then
just
fill
in
the
the
matrix.
D
G
And
then
the
next,
I
I
think,
like
in
sequence,
is
what
I
had
in
mind
in
order
to
have
this
useful
for
the
rest
of
the
community
is
like,
after
these
have
been
lined
up
for
the
like
all
right
now,
we've
got
everything
down.
That
represents
what
we're
trying
to
shoot
for
for
the
first
week
september
for
to
find
the
spec
for
the
first
week
of
september
then
can
follow
on
after
it's
been
locked
down
to
ask
the
ga
languages
java
javascript
python
go
erlang
and
net.
G
I
took
these
languages
from
our
ga
announcement
draft
google
doc
that
morgan
you
put
together
so
for
them
to
fill
out
and
go
like
okay,
now
double
check
these
and
see
that
they
line
up
with
the
spec
that
we've
locked
down
in
order
to
know
what
it
takes
in
order
to
get
these
these
out,
yep
perfect.
G
So
we
already
talked
about
distinction
javascript;
no,
no,
no!
No!
Okay!
Thank
you
needed
rose
for
blah
blah
blah,
okay,
we
already
talked
about
that
and
I
guess
metrics
resource
can
follow
on
as
lower
party.
This
will
be
the
next
phase
right.
G
Okay,
related
to
that,
I
I
had
a
chat
with
alex
or
alex.
I
actually
brought
this
up
and
I
had
a
peek
into
the
python
sick
latest
happenings
and
I
think
it
was
called
out
on
the
sig
meetings.
There's
less
people
joining
the
sig,
so
this
is
just
like
looking
down
the
line
after
the
specs
been
locked
down
in
implementing
the
changes
in
order
to
fill
out
the
table
with
plus
signs
alex.
G
I
was
wondering
if
you
want
to
speak
towards
what
the
current
level
current
participation
of
python
said,
whether
there's
anything
that
could
be
helpful.
M
Yeah,
I
think
I
just
wanted
to
bring
out.
I
was
chatting
with
some
of
the
other
folks
during
the
lastic
meeting
and
we
noticed
there's
a
lot
less
folks
that
attend
the
meeting.
So
for
the
past
I
don't
know
three
months
or
so,
we've
you
know,
we'd
have
between
eight
and
ten
people
show
up
at
the
sig
meetings.
Now
we're
down
to
like
three
to
five
makes
a
good
sig
meeting
and
there's
a
lot
of
change
happening
in
the
spec.
M
So
we
kind
of
all
agreed
that
we
would
focus
all
of
our
energies
on
the
core
like
api
sdk
implementations
and
everything
else
is
going
to
kind
of
get
left
behind
until
the
the
tracing
api
is
completed
because
we're
getting
a
lot
of
we're
getting
a
lot
of
people
that
are
creating
pull
requests
for
things
like
instrumentations
or
or
other
components
that
we
just
don't
have
any
bandwidth
to
focus
on.
So,
if
there's
there's
any
folks
out
here
that
you
know
are
know
that
there's
people
doing
python
that
are
interested
in
helping
out.
B
G
I
also
know
that
they
took
a
quick
look
at
the
code
owners
file
for
the
python
seg.
There's
some
names
on
it
from
before.
I
don't
know
whether
more
participation
of
ex
people
who
have
an
approvers
or
maintainers
name
on
it
but
haven't
participated
in
the
past,
would
be
helpful
to
jump
in.
I
saw
yuri's
name
on
the
list.
I
don't
want
to
call
him
out
specifically
whether
that's
like
the
best
person
to
call
out
or
but
yeah,
I'm
not
sure
whether
that
would
be
an
easy,
helpful
way
to.
M
Yeah
I
mean
I,
I
think,
we've
we've
just
had
a
lot
of
trend
as
far
as
even
the
approvers
like
we
have
ten
approvers
on
the
list
and
I
think
less
than
five
of
them
are
still
involved
in
the
project.
So
if
we
could
find
more
people
to
to
send
to
the
project,
that
would
be
great,
but
I
suspect
this
is
true
of
almost
every
sig
at
this
point.
M
N
Well
for
python,
I
I
think,
there's
a
bigger
trend:
yeah
used
to
be
like
chris
from
google
and
me
spending
most
of
the
time
driving
the
product
and
both
of
us
left.
So
we
have
later
from
microsoft
and
it's
relatively
junior,
and
also
we
have
someone
like
yusuke
recently
during
google,
so
it's
becoming
hardhaven
project.
F
B
Yeah,
just
unlucky
man
yeah.
I
think
it's
a
lot
of
people
just
moving
around
chris
chris
moved
to
spanner
and
it
sounds
like
you
keep
moving
to
google.
A
M
Yeah,
I
think
I
think
it
might
be
tight.
I
think
we'll
probably
be
like,
maybe
a
month
behind
when
the
tracing
spec
is
completed.
I
think
that's
kind
of
expected
and
some
of
the
interesting
as
well,
but
I
you
know,
I
do
think
that
we're
we're
going
to
have
to
pause
everything
else
that
we've
been
kind
of
doing
around
the
the
other
pr's
that
we're
getting
in
order
to
achieve
that.
So.
E
K
Yeah
I
mean
I
don't
have.
I
can't
speak
for
python
in
particular,
but
like
a
month
after
seems,
like
a
good,
I
mean
unless
a
specific
language
has
a
lot
more
people.
Adding
code
than
java
does
a
month
seems
like
a
reasonable
lag
close
to
spec.
E
Yeah,
I
would
say
yeah
based
on
well
on
these
that
at
least
one
month
should
be
allow
at
least
one
month.
You
know
of
you
know,
delay
yeah.
I
agree.
A
A
B
There's
there's
no
timeline:
there's
there's
no
requirements
on
when
the
the
implementations
go
right,
like
the
the
whole
thing
that
we're
doing
this
week
is
primarily
for
the
spec
not
for
the
implementations
and
the
implementations
will
follow.
B
I
assume
there
would
be
a
delay
just
because
some
implementations
may
hit
ga
before
others,
because
even
if
they
all
hit
release
candidates
on
the
same
date,
there
might
be
some
implementations
that
need
to
make
major
changes
or
a
larger
set
of
changes
based
off
of
feedback
from
users
and
others
where
it
turns
out
that
their
first
release
candidate
is
pretty
much
good
to
go,
and
so
this
this
is
fine
right.
This
isn't
throwing
a
huge
wrench
in
anything.
F
F
A
I
guess
the
final
decision
here
proposed
decision
that
I'm
proposing
is
that
we
move
forward
with
the
change
and
use
x
encoding
instead
of
base
64
for
trace,
ids
and
span
ids.
A
If
you
disagree,
if
you
have
any
other
new
information
that
is
not
in
that
thread,
please
comment
there.
Otherwise,
let's
give
it
some.
I
don't
know
a
couple
days
and
I,
if
I
don't
hear
anything
new,
I'm
going
to
make
a
decision
that
it's
final
and
I'm
going
to
submit
a
pr
on
the
stream
change
to
the
spec.
B
Okay,
all
during
well,
while
we
discuss
the
next
one
I'll
just
put
that
in
oh
teagan,
you
already
wrote
a
comment
on
that
spec.
Just
saying
that
great
we're
on
the
issue
saying
that
carlos
latest
change
is
to
apply
from
the
specification.
So
this
is
just
a
rollup
that
carlos
mentioned
earlier
in
the
meeting
of
these
basically
spec
changes
that
have
been
already
been
approved.
That
sdk
maintainers
need
to
take
note
of
so
for
sdk
maintainers
go
ahead.
E
Yeah,
probably
the
one
I'm
a
little
bit
curious
about
the
first
one.
It's
approved
and
everything
there's
this
possibility
of
having
multiple
head
entries
for
trace,
trace
state
so
probably
will
change
a
little
bit,
the
first
one
together
keys.
So
instead
of
keys,
it's
going
to
be
getter
items
like
with
both
the
key
and
the
value.
Okay.
Yes,
so
that
may
change,
I'm
still
researching.
E
L
Yeah,
if
you
just
click
on
that
pr
you'll
see
that
the
cla
check
ever
returned.
I
just
happened
to
notice
this
this
morning,
so.
B
H
We've
had
that
issue
before
we've
just
tried
it
multiple
times
and
sometimes
it
works.
Sometimes
it
doesn't.
K
B
Okay!
I
will
mention
this
at
the
governance
meeting.
I
think
liz
or
someone
was
the
person
who
set
it
up
or
sergey
daniel
looks
like
it
just
passed
after
you
kicked
it.
L
It
looks
like
it,
it
looks
like
it
worked.
B
H
I'll
leave
it
yeah
sure
sure.
So
I
mean
I
just
wanted
to
bring
this
up
to
the
maintainers.
I
know
everyone
is
super
busy
and
also
you
know
doing
a
lot
of
work,
but
also
we
have
new
developers
joining
in
to
the
project
all
the
time.
H
H
You
know
built
into
the
maintainers,
and
this
is
you
know
again,
an
issue
that
we've
found
with
some
of
the
sdks,
as
well
as
the
collector
code
that
we've
been
working
with
on
different
projects
and
and
it
is
that
it
you
know,
our
our
maintainers
are
just
completely
overwhelmed,
sometimes
right.
So
how
do
we
first
of
all
increase
that
pipeline
in
terms
of
graduating,
more
folks
to
become
maintainers
and
also
the
second
aspect
is
driving
up?
H
You
know
quality
of
even
with
the
maintainers
work
on,
for
example,
specifically,
there
are
a
lot
of
otlp
changes.
Bogdan
has
been
doing
in
the
recent
couple
couple
of
weeks.
You
know
again
great
credit
to
him
he's
taken
it
off
to
actually
make
significant
changes
in
the
collector
code
to
company.
H
You
know
handle
the
issues
that
the
prometheus
exporters
uncovered
in
the
collector,
but
on
the
other
hand,
there
are
a
couple
of
things
that
you
know
are
very
hard
for
new
developers
to
understand
one
that
how
does
the
spec
get
updated
right?
Does
the
otlp
the
new
changes
that
are
being
made
in
otlp?
H
Where
are
they
reflected?
How
do
we
kind
of
understand
them
if
you're
you
know
landing
in?
And
secondly,
there
is
an
imbalance
in
the
you
know.
Maintainers
themselves,
in
terms
of,
for
example,
collector,
which
is
a
very
core
component,
has
maintainers
only
from
splunk
and
great
credit
to
them,
but
on
the
other
hand,
sometimes
it's
good
to
have
code
reviews
across
you
know,
engineers
from
different
orgs
to
just
step
up
the
quality
of
you
know
what's
being
built
and
also
just
to
ship
more
robust
code.
H
So
again,
just
bringing
these
up
as
areas
that
we
should
look
at,
because
it's
just
that
aws,
for
example,
will
be
continuing
to
build.
A
lot
of
you
know,
support
for
the
collector
as
well
as
other
providers
will
be
so
I
mean
you
know
how
do
we
make
these
components
more
and
more
robust
and
not
not
evolving?
I
mean
I
understand
that
we
are
not
at
ga
yet,
but
it
still
is,
you
know
putting
a
lot
of
pressure
on
the
maintainers.
A
So
yeah
on
the
collector
part,
I
would
love
to
see
more
maintainers,
especially
not
from
splunk
in
the
project.
I
completely
agree
with
you.
We
we
discussed
this.
I
think
yeah
last
time
in
the
collector's
sig
meeting
as
well.
Mark
marco
mark
carter
was
there
as
well
from
aws,
and
we
were
wondering
if
specifically
people
from
aws
would
be
able
to
help
more
on
the
collector's
side
and
eventually
that
would
be
the
path
to
becoming
the
maintainer
right.
The
involvement.
E
H
H
A
A
Sorry
in
the
product
repository
which
is
considered
to
be
part
of
the
specification,
then
the
corresponding
changes
are
done
on
the
collector's
side.
That
is
usually
how
we
work.
So
I'm
not
sure
I
fully
understand
what
is
your
concern
here.
Is
that
that
the
changes
are
happening,
that
they
are
not
visible
or
can
you
please
clarify
what.
H
No,
no
they're
absolutely
visible.
It's
more
that,
for
example,
if
josh
has
questions,
I
mean
I've
been
working
with
both
josh
and
bogdan
and,
and
you
know
both
are
very
prolific,
maintainers
and
contributors,
and
and
often
you
know
they
again,
even
when
they're
reviewing
the
prs
which
they're
they
are
filing
or
you
know
bogdan
is
filing
and
josh
is
reviewing
there
is
there
are
discussions
of
in
the
metric
say
of
what
you
know
does
what
right
and
and
and
while
I
understand
that
the
code
is
evolving
as
it
goes.
H
H
In
in
proto,
as
well
as
in
the
collector
I
can,
I
can
send
you
the
pr's,
if
I
mean
they're,
very
specific
changes
that
are
that
are
being
made
and
again
it's
it's
more
so
an
observation
on
the
process.
I'm
not
again.
This
is
not
any
any
criticism.
It
really
is
trying
to
figure
out.
You
know,
what's
a
better
way
of
actually
ensuring
that
the
code
is
consistent,
has
consistent
quality
is
robust
and
can
be
actually
used
for
ga.
A
Okay,
I'm
not,
I
don't
know
which
pr's
were
the
the
actual
calls
for
concern.
If
you
could
forward
the
links,
please,
I
would
I'd
like
to
have
a
look
and
understand
this
better
sure.
H
H
In
general,
I'd
like
to
kind
of
see
a
bit
more
support
for
maintainers.
It's
you
know
again,
riley,
for
example,
footsie
plus
plus,
you
know
he
is.
He
is
overwhelmed.
H
N
Any
other
priority
and
how
many
people
are
currently
focusing
on
this
report
so
for
c
plus
plus
I
can
speak.
Definitely
we
would
want
some
like
more
maintainers,
so
my
plan
is
to
get
another
dev
from
microsoft
to
run
the
maintainer
job
and
it'll
be
great.
If
we
can
have
someone
from
amazon
to
to
help
out
as
well.
B
B
N
And,
and
like
I
I
mentioned,
the
topic
like
has
covered
another
issue,
so
regarding,
like
the
part,
so
I
I
my
basic
understanding
is
we
probably
need
to
establish
some
contract,
for
example,
if
I'm
committed
on
a
particular
like
mentoring
thing,
it's
it's
like
just
unfortunate,
like
a
13
interns
like
asking
me
questions
on
gator
and
I
cannot
handle
them
all,
so
I
started
to
step
back
and
set
this
contract.
N
So
if
I'm
committed
to
become
the
mentor
for
this
particular
work
stream
and
we
have
six
work
streams
in
the
class
class,
so
I
volunteer
myself
to
be
the
mentor
for
two
work
streams
and
the
promises
part
is
just
something
that
I
haven't
committed,
because
I
I
need
to
do
my
own
plan
and
I
I
think,
moving
forward
probably
need
to
follow
similar
things.
So
if
we
can
commit
on
this,
normally
people
will
keep
the
promise
and
they
will
dedicate
the
time
if
they
cannot
promise
on
this.
N
I
I
think,
like
keep
pinging
people
wouldn't
help
and
probably
would
make
this
even
more
problematic.
In
this
case.
It's
just
unfortunate
that
the
community
doesn't
have
enough
capacity
to
take
x.
Number
of
interns-
and
we
probably
should
reduce
the
number
by
like
x,
divided
by
two
or
something
like
that-
yeah
the.
B
The
the
and
speaking
for
google
is
much
appreciate
your
work,
the
the
40
google
interns
joining
this
summer
was
probably
a
one
time
thing,
so
that.
B
B
Well,
I
I
one
hopes
that
the
google
intern
program
next
summer
is
prepared
for
covent,
if
it's
still
with
us,
but
but
yeah
we'll
we'll
see,
but
I
suspect
it.
That
was
a
a
very
special
event.
H
It's
just
that
there
will
be
new
new
developers
who
join
and-
and
it's
just
very
important,
to
kind
of,
have
a
process,
a
graduation
process
and
a
mentorship
process
for
degree.
New
developers.
H
Ted
we
have
built
a
new
higher
guides,
like
we
contributed
a
new
hire
guide
to
the
c
plus
plus
repo.
You
know
documenting
some
of
the
issues
that
you
know
all
our
interns
kind
of
discovered
and
uncovered
and
learned
from,
but
yes
happy
to
help
you
on
adding
more
detail
there.
G
Hello
lita,
I
also
had
I
put
some
links
in
the
maintainers
meeting,
notes
related
towards
the
open,
telemetry
protocol
and
also
the
specification,
and
I
I
don't
know
whether
I
need
to
call
it
out
or
whether,
like
it's
already
been
known
as
implicit
or
whether
that's
the
way
we're
running,
but
I
also
had
a
little
bit
of
question
as
to
like
the
process
of
how
the
pro
was
updated.
G
In
that
repo
of
the
open,
telemetry,
proto
repository
and
then
later
on
it
once
it's
coming
towards
a
beta
or
stable
phase,
then
an
update
to
the
spec
repo
under
the
protocol
subdirectory
would
follow
suit
to
say
all
right.
All
right,
we've
tried
it
out,
and
this
is
what
it
looks
like,
and
this
is
what
we
propose
to
lock
down
anything
like
fine-tuned
adjustment
ghosts
go
there
is.
Maybe
this
is
a
question
towards
tigran.
I
saw
you
participate
a
lot
in
the
otlp
stuff.
Is
that
kind
of
the.
G
A
Schema
right
now,
that's
in
progress
and
you're
right,
because
it's
alpha
it's
we
we
are
free
to
change
it.
I
think
any
time
we
are
allowed
to
be
breaking
changes.
If
I'm
not
wrong
alpha
is
that
yes,
anytime,
I
just
check
it's
not
even
better.
A
As
for
the
process,
I
I'm
not
sure-
maybe
maybe
some
improvements
are
due
here,
so
I
I
really
like
to
understand
what
what
are
the
concerning
parts
so
that
I
can
help
with
those
right,
so
I'd
love
to
see
if
there
is
any
there
is
any
pr
that
looks
like
it
was
not
done.
Well,
if
it
looks
concerning
I'd
love
to
see
that,
and
so
I
want
to
understand
better
what
is
the
issue
here
so
so
that
I
can
try
it
out.
H
During
my
understanding,
though,
was
that
you
know
metrics
is
almost
beta
right,
because
if
and
again
this
is
just
based
on
earlier
discussions
on
when
metrics
goes,
big
becomes
ga
ready,
and
I
thought
that
that
target
was
end
of
this
year.
So
is
it
really
at
alpha.
A
Well,
almost
I
don't
know
what
almost
means
here,
it's
appealing
we
don't
have
broke
down
here.
He
has
the
latest
information
about
how
close
we
are
to
having
complete
metrics
protocol.
I
see
the
last
change
was
done
two
days
ago,
so
it's
it's
and
we
still
have
open
pr's
yeah.
A
B
Yeah,
I
was
gonna
say
we're.
Probably
gonna
go
through
the
same
process
that
we're
just
finishing
up
this
week
for
tracing
right.
Where
we
said
we
set
a
date.
We
go
and
prune
down
everything
to
the
core
set
of
p
ones
that
we
need
and
we
give
people
three
weeks
and
if,
by
the
end
of
three
weeks
it
isn't
done,
then
the
cc
gets
to
weigh
in
on
on
the
final
decisions.
B
H
G
But
related
towards
the
point
you
made
to
the
the
trace
portion
of
otlp
is
considered
stable
and
I
can
speak
at
least
for
a
light
step.
We
are
very
con
very
interested
in
the
stability
of
otlp's
trace
portion
to
be
stable
because
there
are
active
real
users
on
it
and
they're
expecting
stable
behaviors
on
that.
So,
like
any
type
of
breaking
changes
that
have
to
be
like
very,
like
heavily
weighed
in
that
portion
of
otp.
A
Yeah
traces
are
declared
stable.
I
know
people
who
use
it
in
production
already,
so
we
cannot
make
any
breaking
changes
anymore.
Any
changes
that
we
make
have
to
be
backwards
compatible
in
the
sense
that
any
two
nodes
that
run
different
versions
of
the
protocol
have
to
be
able
to
interoperate.
A
Whatever
changes
we
introduce,
they
need
to
work
with
either
combination
of
nodes
right,
not
just
one-way
compatibility,
but
it
has
to
be
two-way
compatibility
between
different
any
different
versions
of
the
nodes,
so
that
and
any
proposals
about
changes
to
the
traces
has
to
go
through
this.
Through
this,
the
lengths
of
the
changes
right
matrix
no
matrix
is
still
alpha.
We
have
not
made
any
declarations
about
the
stability
of
matrix
portion
of
the
protocol.
A
A
I
agree
completely,
I
agree,
and
I
know
that
bogdan
is
focused
on
that
and
he
he
wants
to
complete
it
asap.
B
So
we
we
have.
We
have
our
action
items
which
for
the
tracing
spec,
we
know
that
we
need
to
go,
find
more
maintainers,
particularly
for
python
and
c
plus
plus,
and
I
think
I'll
lead
in
two
grand
underscored
how
important
it
is
that
we
keep
progress
going
for
metrics
as
soon
as
our
focus
moves
from
tracing
soon,
as
we
get
everything
lined
up
the
end
of
this
week
for
tracing
all
right.