►
From YouTube: 2022-07-14 Governance Committee private 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
I
am
in
a
different
house
than
normal
because
I
tested
positive
for
covid,
so
that
is
really.
A
A
D
A
She
got
positive,
I
isolated
from
her,
so
I'm
staying
at
a
friend's
house
and
yeah.
That's.
B
That's
a
good
idea:
oh
yeah,
it
looks
like
everybody's
getting
that
at
least
around
me.
B
My
sister
got
it
and
I'm
sorry,
no,
my
yeah,
my
brother-in-law
got
it
and
it
yeah
and
then
my
sister
got
it
as
well
yeah-
and
I
don't
know-
and
it's
so
scary,
because
we
have
a
baby
here
at
home
and
we
have
no
idea
how
it
goes
with
babies.
So
there
is
no
vaccination
for
them.
So
for
three
to
five
years
they
just
today
the
brazilian
government
approved
a
vaccine,
so
we're
kind
of
safe
with
our
youngest
ones.
But
for
the
baby,
it's
just
scary.
A
Cool
well,
someone
want
to
ping
the
slack
channel
cause
four
is
not
a
quorum
of
us.
Unfortunately,
although
I
do
have
something
I
wanted
to
informally
discuss
with
you
all
to
get
some
buy-in,
but.
D
Fantastic
liz:
did
you
complete
all
your
travels
back
and
forth
across
the
continent?
That's.
D
A
Yeah,
that
is
life
in
the
middle
of
a
pandemic.
It
turns
out
no
matter
how
friendly
precautions
you
take.
Your
family
can
be
your.
D
Flink
sometimes
yeah
yeah
totally
it's
like
there's,
you
know,
eventually,
as
the
saying
goes,
everybody
will
run
across
it
in
some
form.
The
other.
A
All
right:
well,
let's
give
people
another
minute
and
then
we
can
get
started
with
the
agenda.
We
do
have
a
quorum.
So
oh,
no,
I'm
sorry,
then.
D
E
I
did
I
mostly
lifted
it
up
from
last
week,
though,.
C
Well,
I
don't
either,
but
let's
see
well
I
mean
the
issue
itself
seemed
like
there
were.
You
know
some
suggestions
and
nuances
and
stuff,
but
no
strong
objections
or
anything.
I
guess
it
sort
of
seems
like
it
demands
some
sort
of
ownership
to
figure
out
the
details
and
actually
make
a
specific.
A
Yeah
in
this
particular
case,
you
know,
I
think
the
gc
is
the
right
set
of
approvers
right,
because
this
is
not
a
tc
thing.
Community
is
a
gc
responsibility,
yes,
and
you
know
until
we
otherwise
delegate,
we
are
on
the
hook
for
do
those
community
things,
but
I
think
I
personally
would
be
happy
to
delegate
to
austin.
I
have
a
couple
of
other
people,
I
think
would
be
great,
but
we
don't
have
to
block
unnecessarily
having
those
people.
A
D
Yeah
and
again
again,
I
I
agree
liz
and
we
should
definitely
add
comments
to
this
discussion
on
on
the
on
the
pr.
But
I
also
think
that
you
know
again.
Many
many
folks
are
interested
and
you
know
if
we
could
actually
coalesce
the
comms
sing
into
you
know
being
a
community.
I
mean
they
do
that
anyway,
right
to
some
extent,
but
to
have
more
of
a
community
maintain
a
me.
D
I
don't
know
what
community
maintainer
means,
but
community
manager
for
such
a
large
project
having
one
community
manager
is
a
bit
strange,
but.
B
Yeah,
but
I
guess
there
are
some
calls
that
have
to
be
made
by
one
person
or
in
case
of
conflicts.
You
know
between
the
multiple
people,
there
might
be
different
opinions
and
causes.
B
No,
I
mean
I'm
just
saying
that
you
know
when
you
have
like
three
people
handling
one
specific
subject.
Sooner
or
later,
you're
gonna
have
a
three
opinions.
Three
three
different
opinions
on
the
same
thing,
so
I
I
I
don't
know
I
guess
my
point
is:
I
think
there
has
to
be
a
way
to
appoint
people
and
I'm
for
allstream,
so
I
I
think
he
would
be
a
great
community
manager.
I
would
just
really
want
to
see
that
role
been
officially
an
official
way
of
getting
into
that
role.
B
You
know
so
it's
not
just
you
know.
How
do
we
appoint
a
community
manager
or
how
do
we
appoint
going
to
managers.
A
Should
be
a
process
right
like
one
of
which
is
governance
committee
appoints
the
other
approach
is
similar
to
any
sig
the
maintainers
themselves.
A
point
right
maintainers
can
vote
in
other
maintainers.
Although
I
think
in
the
past
we've
said
you
know,
subject
to
tc
approval,
but
in
this
case
you
know
be
subject
to
gc
approval
right.
D
I
agree,
but
I
I
mean
my
my
only
point
being
that
I
mean
this
is
such
a
large
project
that
I
don't
know
if
one
community
manager
is
enough
in
one
sense,
it's
more
a
group
effort
and
and
typically
the
way
even
the
sigs
have
run.
Is
you
know
that
many
maintainers
and
many
contributors
work
together?
So
I'm
just
wondering
I
mean
again.
This
is
just
discussion.
Point.
E
I
mean
I'm,
I'm
generally
supportive
of
this.
The
my
biggest
concern
is
there's
more
detail
in
the
pr
description
than
there
is
in
the
actual
pr.
This
essentially
just
adds
a
heading
in
his
name.
It
doesn't
at
all
define
the
role.
What.
D
E
Supposed
to
do
how
he's
supposed
to
achieve
those
things
and
what
resources
are
at
his
disposal.
I
also,
I
wonder
if
this
is
like.
Is
this
really
how
you
go
about
creating
an
entirely
new
role,
create
a
pr
against
the
community
repo
and
wait
for
the
gc
to
approve
it?
That
seems
kind
of
informal
yeah?
Maybe.
A
I
think
what
it
boils
down
to
as
far
as
the
specificity
or
lack
thereof-
I
think
you
know,
as
austin
himself
has
admitted
to
me
right
like
this-
is
a
codification
of
things
he's
already
been
doing
so
in
that
sense,
you
know
yes,
it's
a
little
bit
looser
than
it
could
be,
but
we
do
have
the
ability
to
look
at
the
kubernetes
at
the
similar
kubernetes
as
inspiration,
so
this
is
not
completely
whole
cloth.
D
C
I
think
we
could
also
have
some
kind
of
intentionally
short.
You
know
sort
of
allow
list
of
definite
responsibilities.
This
person
has
that
are
like
a
subset
of
what
they're
probably
doing
as
a
starting
point.
I
mean
in
my
mind
it's
more
just
like
there
are
a
few
things
where
they
might
need
the
authority
to
make
certain
decisions
kind
of
on
their
own
right.
A
The
main
one
that
was
like
the
main
one
that
was
discussed
is
you
have
to
be
a
project
level.
Maintainer
quote
unquote
to
be
able
to
interface
with
the
cncf
to
get
their.
C
Yeah
I
caught
it
by
like
a
very
short
list
of
things
and
then
say
so.
These
are
the
very
specific
you
know:
responsibilities,
slash
rights,
this
person
has
or
whatever
and
then
and
then,
broadly
speaking,
these
are
the
mushier
goals
for
the
role
in
terms
of
what
they're
supposed
to
be
thinking
about,
but
that
would
be
more
qualitative
and
and
conceptual
abstract.
C
That
seems
I
I
mean
I
don't
actually
mind
the
informality
of
a
pull
request
that
gets
merged.
As
long
as
we
have,
I
would
say,
like
an
appropriate
level
of
discourse
and
discussion
about
it
and
that
it's
not
just
like
a
rubber
stamp.
I
mean
that's
the
thing
that
would
disturb
me
if,
if
it
just
got
the
post
and
then
merged,
I
mean,
but
as
long
as
we're
debating
it
a
little
bit
and
getting
at
the
nuances,
I
I
don't
feel
like.
D
But
I
mean
again,
I
think,
we're
trying
to
create
a
role
to
serve
a
purpose
right.
That
is
we're
not
getting
cncf
access
to
assets
or
you
know
some
kind
of
cncf
responsibilities.
Hence
we
are
trying
to
create
a
position
which
is
not
a
typical
maintainer
or
a
typical
role
that
is
carved
out
on
these
projects
already
and
and
and
it's
okay-
I
mean
again-
I
it's
just
going
to.
We
should
go
through
that
process,
just
even
if
it
has
three
steps
in
that
whole,
you
know
workflow
and
then
go
from
there.
D
B
All
right,
so
I
guess
because.
E
D
C
C
Fog
but
originally,
but
I
actually
thought
it
was
an
issue
until
just
now
like.
Oh,
that
was
a
pr
yeah.
No,
it
should
definitely
be.
I
mean
I'm
happy
to
do
it
right
now
and
be
like
hey
austin.
You
know,
I
think
the
gc
talked
about
this
we're
more
or
less
on
board.
But
can
you
make
this
into
a
more
formal
proposal
with
you
know
some
of
the
specific
requirements
of
like
things
you
actually
need
like
access
to
cncf
whatever
and
write
it
up,
and
then
we
can
do
a
more
formal
review
of
that.
B
If
there's
a
pressing
need
to
get
it,
I
don't
know
formalized
rather
sooner
than
later,
then
what
we
we
can
consider
doing
is
just
write
down
what
he's
already
doing
and
we
and
and
define
the
role
and
you
know
to
find
the
process
of
getting
into
that
role
and
whatever
else
is
needed
we
can.
We
can
do
this
on
a
second
step.
B
I'm
finding
just
you
know,
writing
down
what
he's
doing
already.
That's
fine,
and
but
the
only
thing
that
I
think
we
should
have
is
a
what
is
the
process
to
get
there?
So
if,
if
someone
wants
to
become
a
community
manager
tomorrow,
what
is
it?
How
do
the
process
look
like
how
it
is.
E
E
A
Yes,
the
problem:
the
problem
is
that,
as
far
as
the
project
level
maintenance,
we
as
the
gc
are
viewed
as
the
project
level
maintainers.
Okay,.
A
That
authority,
but
we
need
to
explicitly
do
that
and
say
that,
hey
by
the
way,
a
member
of
the
community,
you
know
the
a
community
manager
is
for
the
purposes
of
cncfa.
C
D
Okay,
so
ben
you
have
an
action
item.
I
think
you're
just
going
to
ask
austin
to
kind
of
write
down
what
he's
doing
right
now
and
then
we
can
just
take
a
look
at
it.
I
guess
and
finalize
based
on
the
time
urgency
right.
That
is,
if
is
he
if
he
is
blocked
on
stuff
that
he
cannot
access
right
now
or
if
it.
C
C
I
don't,
I
don't
think
it's
super
urgent,
just
based
on
talking
to
austin
every
weekend,
but
okay.
B
So
looking
at
the
gender-
I
I
believe
I
don't
know
I
might
be
judging
it
the
wrong
way,
but
I
think
that
the
last
item
is
actually
the
most
important
one
for
the
next
items
and
I
think
the
other
ones
can
be
discussed.
Yeah.
B
D
Is
a
bit
more
needs
discussion.
C
Sure
I
mean
I
actually
wouldn't
mind
hearing
from
you
all
I
mean
I
have
to
admit
that
I
found
out
about
all
this
a
little
bit
like
after
the
fact,
and
mostly
from
people
who
don't
work
at
lifestyle
when
I
dug
into
it
more
so
the
issue
is,
I
understand
it
is
that,
like
essentially
that
that
josh
mcdonald
ended
up
creating
a
lightstep
repo,
that's
a
fork
of
the
gometrix
sdk
in
order
to
pursue
experimental
features
around
histogram
support
and
some
other
things,
and
that
this
was
perceived
as
like
the
opposite
of
good
practice
and
open
source,
where
you
typically
do
all
of
your
development
upstream
and
so
on
so
forth,
and
then
and
then
people
got
very
agitated,
like
in
various
circles,
including
some
things
that
have
since
been
revised
in
the
public
record,
but
were
initially
pretty
spicy
and
that
and
then
that's
the
situation.
C
First
of
all,
is
that
accurate
summary
I
mean
genuinely.
I
might
be
missing
something:
that's
just
what
I
pieced
together.
B
D
D
B
Yeah,
I'm
a
tracing
person-
I
I
don't
know
metrics
that
much
and
I
I
won't
pretend
I
understand
the
the
underlying
issue,
but
from
where
I'm
standing,
I'm
hearing
the
problem
with
histograms
since
at
least
six
months
and
the
problem.
As
I
see
it-
and
I
know
this
probably
is
older
than
that-
the
problems
I
see
it
is
the
prometheus
wanted
to
do
some
exponential
histograms
or
I
think
they
have
a
different
name
for
that
and
they
they
had
a
plan.
B
They
had
a
a
vision
on
how
to
do
that
and
the
way
things
work
at
well.
Prometheus
is
stipulating
implement
things
and
people
use
it
and
they
change,
and
then
they
they
people
use
it
again
and
you
know
get
better
tests
before
it
is
part
of
any
specification
before
it
becomes
a
standard
that
people
implement
now
at
the
same
time.
B
Around
the
same
time,
we
hotel
wanted
to
have
exponential
histograms
as
well
on
our
site,
so
we
added-
or
we
started
some
work
on
specification
on
that
there
was
a
call
for
for
comments.
Some
people
answered
those
calls.
Some
people
from
the
prometheus
community
missed
the
call,
and
then
this
pack
got
approved
got
merged.
B
Some
work
was
done
on
at
otlp
level.
I
think,
and
things
were
progressing
from
there
and
at
some
point
at
the
beginning
there
were
some
people
from
primitives
that
started,
showing
concern
about
the
way
that
things
were
were
done
in
hotel
because
it
was
incompatible
or
seen
as
incompatible
by
the
prometheus
folks
and
one
of
the
now
that's
fine.
I
mean
if
you
have
two
specs
running
parallel,
it's
bound
to
happen
that
they're
going
to
be
incompatible
or
at
some
point
most
for
most
things
they
are
compatible,
but
the
problem
is
well.
B
I
think
one
of
the
contention
points
is
we
hotel.
We
promise
that
we're
going
to
be
compatible
with
the
standards
of
of
today
and-
and
today
you
know-
is
from
the
time
that
we
made
this
this
commitment.
So
we
said
that
we
would
be
compatible
with
open
tracing.
We
would
be
compatible
with
provisions
as
well
now
prometheus
community,
then,
is
seeing
that
promise
as
broken
if
we
don't
implement
in
a
way
that
is
compatible
with
prometheus.
B
So
that's
the
the
cruise
of
the
issue
and,
of
course,
over
the
past
few
months,
I
guess
a
couple
of
months.
B
There
was,
I
don't
know
people
wanted
to
deliver
on
that,
because
it's
a
topic
that
has
been
happening
for
some
for
a
long
time
now
and
from
where
I'm
standing
josh
made
a
very
practical
decision
of
forking
a
the
sdk,
which
is
fine,
I
mean
he's,
got
all
the
right
to
do
that,
either
in
name
of
lightsaber
or
not
and
publishing.
B
What
is
what
is
his
exponential
histograms
on
on
his
opinion
or
his
take
on
on
exponential
histograms.
B
B
So
I
guess
that's
the
state
of
the
state
of
the
things.
As
of
right
now,
so
I
guess
people
are
seeing
this
fork
as
an
official
light
subfork
of
the
metrics
sdk
and
by
not
by
not
having
that
made
in
a
collaborative
way
in
the
community
because
of
reasons
it
is
seeing
them
as
a
whole
style
move.
B
C
Know,
for
sure
I
mean
lightstep
much
to
our
commercial
detriment,
has
made
every
effort
to
participate
in
hotel
upstream,
I
think
at
some
level
we
do
have
agreements
on
customers
to
deliver
things,
and
in
this
case
this
is
one
of
them.
So
we
have
to
do
something
for
very
large
metrics
customer
and
that's
just
what
it
is.
I
don't
think
josh's
messaging
on.
C
This
is
very
clear,
but
in
my
mind
it's
like
an
experimental
fork
to
provide
functionality
that
we
need
to
provide
commercially,
and
I
think
it
would
be
a
huge
huge
no-no
to
like
you
know,
force
something
through
the
official
hotel
go
sdk
with.
That
would
be
unacceptable
in
my
mind
from
a
commercial
standpoint,
I
think
we
have
a
need
to
provide
something
for
people
and
we
don't
want
it
to
be
like
in
someone's
personal
repo.
So
it's
a
lightsaber
repo.
C
I
think
it
should
be
framed,
because
this
is
the
way
I
want
it
to
be
as
like
an
experimental
but
production
like
you
know
stable
from
like
a
not
gonna
crash
standpoint,
but
like
an
experimental
branch
to
try
and
explore
a
scaled-out
histogram
workload
for
metrics
or
whatever
that's
the
way
I'd
like
to
perceive
it
and
then,
if
it
works,
then
we
can
take
that
back
much
in
the
way
the
prometheus
community
does
that
too,
and
works
with
the
prometheus
end
users
and
then
like
figures
out
what
the
right
technical
solution
is
and
then
comes
back
to
the
spec.
C
But
I
don't
think
it's
been
very
clear
that
the
end
goal
is
to
have
the
light
step.
Sdk
can
reconverge
with
the
hotel
go
sdk,
and
I
think
that
that
that
end
goal
should
be
clarified
in
some
way,
because
that,
frankly,
is
what
I'd
like
to
see
and
I'm
speaking
with
a
lifestyle
pal
but
yeah
I'll.
Let
it
go
ahead.
D
Yeah,
so
I
mean
dressy
again
thanks
for
going
through.
You
know
the
understanding
with
the
larger
prometheus
community.
So
I'd
like
to
clarify
a
couple
of
things.
I
agree
with
then
that
you
know
the
messaging
that
josh
did
could
have
been
a
bit
better,
because
I
think
that
that
just
created
a
bit
of
a
fair
bit
of
anxiety
amongst
a
lot
of
folks
because
they're
like
okay,
what
what
happened
here
without
you
know,
people
actually
really
understanding.
Even
the
go.
Maintainers
were
not
really
very
aware
of
it.
D
So
I
think
that
a
discussion
and
communication
there,
because
josh
is
such
an
integral
part
of
the
go.
You
know
sdk
and
api
has
been
so
deeply
involved
in
it
for
a
long
time,
so
that
would
have
been
actually
very
useful
and
we
could
have
helped.
You
know
in
terms
of
any
kind
of
you
know,
discussion
or
mediation
or
whatever,
because
there
are
ways
of
doing
implement,
experimental
implementations,
that's
all
good,
and
but
there
could
also
be
development
branches
where
this
is
done
on
the
project.
Not
necessarily
you
know
just
given
his.
D
You
know
position
as
an
expert
on
the
project.
It
was
kind
of
tough
and
and
the
other
other
area
where
I'd
like
to
call
out
the
prometheus
work.
Given
you
know,
I've
worked
on
the
interop
of
prometheus,
with
hotel
on
the
metric
side.
Is
that
hotel
has
you
know
and
josh
included,
has
been
having
a
lot
of
discussions
in
the
prometheus
web
group
and
richie
richard
as
well
as
others
from
the
prometheus
community
have
joined
in
for
many
discussions
and
that's
an
ongoing
process
right.
D
There's
nothing
but,
however,
the
you
know
prometheus
community.
Obviously,
as
they
work
towards
a
more
complex
histogram
implementation
need,
you
know
we
should.
We
should
work
with
them
or
they
should
also
work
with
us
in
order
to
kind
of
make
sure
that
you
know
the
customer.
The
end
user
at
the
end
of
the
day
has
compatibility
across
all
the
projects
right,
because
you
could
have
multiple
implementations
for
time
series
based
histograms,
but
does
that?
D
E
Yeah,
I
just
wanted
to
say
in
the
beginning
of
the
project
we
made
a
big
deal
about
the
api
and
the
sdk
separation,
and
the
whole
point
of
that
was
so
that
companies
and
and
third
parties
could
develop
their
own
sdks
specifically
for
these
things.
I
don't
everybody's
talking
about
this,
like
it's
a
problem
and
I
this
is
in
some
ways
the
initial
goal
of
the
project.
E
If
we
didn't
want
this
to
happen,
then
why
did
we
split
the
api
in
the
sdk?
In
the
first
place,
I
mean,
I
would
say,
the
messaging
of
the
project
from
day.
One
has
been
very
like
supportive
of
this
type
of
behavior,
so
to
turn
around,
and
then
say
this
is
a
problem
is
sort
of
weird.
A
Yeah,
I'm
going
to
say
like
when
we
came
together
with
open
sentences
and
open
tracing
right
like
this
was
the
whole
debate.
Ghana,
jana
and
I
went
back
and
forth
with
open
tracing
folks.
Many
years
back
about
you
know,
hey,
like
you
know,
open
tracing
doesn't
specify
the
the
api
right
like
it
specifies
the
api,
but
not
the
implementation
right.
That's
an
interesting!
That's
an
interesting
decision
right
like,
but,
on
the
other
hand,
having
an
opinionated
default.
Implementation
is
good
too,
which
was
the
open
senses
side
and
we
got
the
best
of
both
worlds
right.
A
So
I
think
that
it's
really
important
that
we
continue
to
facilitate
people
doing
alternative
implementations.
As
long
as
you
know,
it
is
clear
that
this
is
an
experiment.
This
may
or
may
not
make
it
into
the
final
bill
of
hotel,
and
I
think
that
it
seems
asinine
to
penalize
lightstep
to
almost
for
doing
this
development
out
in
the
open.
D
D
A
B
Yeah,
I
I
agree
with
all
that
he
said,
and
I
think
and
whenever
I
talk
about
open
telemetry,
I
I
do
make
the
point
of
saying
you
know,
there's
an
api
and
then
there
is
the
case
and
there
is
the
the
people
should
implement
an
alternative
sdks
if
they
see
the
need.
So
if
I
have
a
real-time
processing
situation,
I
do
need
another
sdk.
I
cannot
use
one
that
exists,
but
I
guess
the
the
what
is
causing
so
much.
I
don't
know
discussion
about
that
is.
It
is
an.
B
It
is
a
gold
metrics,
sdk
maintainer
that
is
taking
one
specific
feature
or
that
is
working
the
project
to
implement
a
sdk
in
a
way
that
is
likely
not
to
be
the
same
as
in
the
future,
because
people
are
rewriting
or
re-implementing
the
gold
metrics
sdk
right
now.
So
you
know
the
code
base
is
not
going
to
be
the
same,
and
even
perhaps
worse
than
that
is
there
is
no
functioning
gold,
metrics
sdk
at
the
moment.
B
So
people
who
want
to
use
open
telemetry
with
goal
for
metrics
they
they
are
going
to
use
josh's
experimental
repository
instead
of
any
official
one,
because
there
isn't
an
official
one.
The
collector.
C
Cannot
use
metrics.
That
is
the
reason
why
I
think
josh
needed
to
do
that
because,
like
we
do
need
to
have
something
that
can
be
used.
I
think
the
messaging
really
should
be
improved.
I
have
asked
was
for
us
to
have,
and
I
think
we
created
some
sort
of
issue
to
discuss
this.
I
think
it
deserves
a
discussion
because,
like
there
is
a
combination
of
human
beings
who
don't
represent
either
their
company
or
open
telemetry
who
are
doing
work
and
then
and
then
it's
being
various
motivations
are
being
attributed
to
that.
C
I
can
go
over
by
the
way.
Well,
that's
more
important
than
what
I'm
supposed
to
do
in
30
seconds,
but
I
thought
this
had
to
go.
We
probably
should
continue
this
discussion
and
whatever
forum
gets
created
as
part
of
that
instance,
I
think
there's
a
lot
to
learn
on
this
one.
I
just
want
to
mention
one
thing.
I
I
appreciate,
as
with
my
lights,
I've
had
on
the
support
for
life,
steps
right
to
write
their
own
stk
or
whatever,
okay
yeah
sure.
C
The
only
issue
here
is
that
I
think-
and
I
actually
frankly
don't
understand
the
technical
details.
Enough
to
even
have
an
opinion,
but
I
think
the
sdk
work
that's
being
done
here
has
implications
back
to
the
protocol
and
I
think
that's
probably
why
there's
so
much
sort
of
consternation
about
it
like
this
isn't
just
a
matter
of
like
internal
memory,
representation
or
something
like
this.
I
think
that
the
way
that
exponential
histograms
are
implemented
in
that
branch
probably
has
some
effect
on
otlp
at
some
level,
I'm
guessing
and
I've.
C
Versions
so
much
I
mean,
I
still
think
it's
you
know
well
within
the
right
of
anyone
to
write
code
if
they
want
to
and
put
it
where
they
want,
as
long
as
it's
not
in
hotel.
But
I
think
that's
probably
why
there's
so
much
weight
on
this
like
if,
if
this
ended
up
getting
steam,
that
kind
of
tips,
the
skills
on
the
protocol
discussion,
which
is
not
ideal
right
and
we
still
haven't
resolved
that
protocol
discussion.
C
C
If
you
adopt
this
like
that
kind
of
stuff
right
like
I
think
we
could
do
more
to
message
that
and
still
be
true
to
the
spirit
of
all
this
stuff
and-
and
I
do
think,
there's
a
lot
of
confusion
because-
and
I
don't
want
to
throw
stones
at
josh-
he's
amazing.
I
just
think
in
retrospect.
I
think
this
got
really
messy.
A
A
You
know
being
clear
about
what
our
policies
is,
the
project
of
what's
official
and
what's
not
and
when
a
maintainer
is
not
acting
as
their
maintainer
hat,
but
instead
is
their.
You
know:
private
hack,
dan.
E
Yeah
I
just
wanted
to
so
I
have
been
following
this
for
a
while.
As
far
as
I
know,
this
does
not
affect
the
protocol,
the
actual
structure
of
the
protocol.
E
It
affects
the
interpretation
of
the
data,
so
that
the
question
is,
if
you
have
a
bucket
with
a
boundary
on
some
exact
power
of
two
like
two
four
eight.
Sixteen
do
you
put
it
in
the
bucket
above
or
the
bucket
below,
so
the
that,
like
that's
the
core
incompatibility
difference
and
it
only
matters
for
exact
powers
of
two,
it's
a
it's
a
very,
very
small
incompatibility,
but
it
doesn't
actually
affect
the
proto
itself.
E
C
E
Josh
did
make
a
spec
pr
which
went
the
prometheus
route
on
this
and
he
did
implement
his
version
using
the
prometheus
version.
So
the
only
the
only
thing
is
that
he's
ahead
of
the
the
open
source
version
of
the
metrics
sdk.
E
C
D
I
think
that
context
is
very
helpful,
daniel
and-
and
I
I
just
think
that
you
know
again-
maybe
communicating
that
on
the
hotel
blog
would
be
a
good
thing.
It's
just
communication.
I
think
well.
C
A
D
No,
no!
It's
just
that.
I
think
the
go
maintainers
need
to
be
in
sync
in
terms
of
a
way
forward
right,
even
if
it
is
saying
that
hey
you
know,
when
we
have
done
refactoring
of
the
metrics
sdk,
we
will
make
sure
that
you
know
we
keep
the
suggestions
that
josh
is
making
in
mind
in
order
to
you
know,
consider
that
right
to
say
it's
just
communication.
B
E
Yeah
like
three
weeks
ago
to
resolve
the
incompatibility
he
he
just
decided
that
the
compatibility
difference,
the
incompatibility
was
small
enough
that
it
was
easier.
The
prometheus
people
seemed
to
feel
strongly
about
it,
so
it
was
easier
to
just
go
their
way
on
that.
Okay,
so.
B
Perhaps
it
it
perhaps
it
would
be
if
there
is
a
blog
post
about
anything,
would
be
nice
to
have
a
a
joint
blog
post
with
the
prometheus
community
and
hotel
saying
you
know,
we
are
ready
for
yeah.
B
E
C
So
does
it
seem
okay
to
have
a
blog
post
from
a
combination
co-authored
by
open
telemetry,
go
maintainers
and
sort
of
the
a
few
of
the
folks
in
the
prometheus
side?
You
know.
D
E
Sure
sure
I
also
did
just
check
the
pr
that
josh
made
to
be
compatible
with
prometheus
is
still
not
merged.
He
opened
it
17
days
ago.
It's
not
merged
yet,
but
maybe.
E
It's
waiting
on
reviews,
it's
2633
in
the
spec,
okay,
okay,.