►
From YouTube: 2023-02-08 meeting
Description
cncf-opentelemetry@cncf.io's Personal Meeting Room
A
A
I
have
on
my
agenda
later
to
try
out
that
Cosmos
I,
don't
understand
why
the
we
read
database
and
then
why
is
the
read
item
underneath
the
read
database.
B
It's
it's
their
well
I'm,
not
completely
sure,
but
yeah.
We
even
use
it
in
service
bus
and
event
hub.
A
The
project
board,
but
first
maybe
just
or
yeah
so
the
whole
the
whole
ECS
question
mark
of
aligning.
It
sounds
like
we
have
kind
of
two
paths:
either
we
align
and
we
make
all
these
changes
or
we
don't
and
we
don't
make
probably
any
of
them,
but
we
should
think
about
that
and
so
I
think
what
our
job.
Ultimately
it's
going
to
be.
The
community
is
going
to
have
the
broad
Community
is
going
to
have
to
decide
on
which
of
those
two
paths
to
take.
A
C
On
do
you
have
any
expectations
about
like
a
broader
Community,
when
it's,
when
some
decisions
can
be
can
be
done?
My
impression,
like
I,
was
running
through
the
documents
through
the
meeting
notes
from
Tuesday,
and
my
impression
was
that
it
looks
like
we
want
to
go
this
way,
but
yeah.
A
I
think
there's
there's
you
know
the
we
presented
it
on
Tuesday.
There
was
definitely
you
know
some
shock
and
awe,
which
was
completely
expected
of
kind
of
the
point
of
presenting
it
of
like
hey.
This
is
gonna
break
stuff,
even
though
it's
not
marked
stable,
so
I
think
that,
from
a
timeline
perspective,
Riley
and
I
were
chatting
trying
I
think
we're
gonna
try
to
push
to
have
a
decision
by
within
two
weeks.
A
Okay,
because
that
would
allow
us
to
still
stay
on
schedule
with
the
HTTP
semantic
conventions
and
part
of
the
pushback
or
part
of
the
request
on
Tuesday's
spec
meeting
was
to
provide
more
motivation
of
why
merging
with
ECS
is
beneficial
for
open
Telemetry
users
and
I,
like
like
real,
concrete,
like
like
okay,
these
I
I,
don't
know
what
that
really
even
means
I
mean
it
seems
pretty
clear
that
the
ECS
is
pretty
widely
adopted
within
the
logging
community,
and
so
there's
benefits
to
us.
A
D
A
Ecs
folks,
as
far
as
helping
to
make
their
their
case,
yeah.
E
So
so
Alex
will
need
some
like,
like
in
general,
I
talked
to
the
TC
and
PC
and
also
Alex
who's
from
ECS,
and
he
has
to
talk
with
his
elastic
to
see
it's
open
country
is
willing
to
make
the
offering
or
the
elastic
we
want
to
take
the
offering
or
not.
Maybe
that's
not
the
intention,
maybe
elastic
would
say
like
we
would
want
to
have
the
ultimate
control
ECS.
E
We
don't
even
intend
to
merge
that
with
open,
so
I'm
sure,
then
that's
a
totally
different
conversation,
so
we
should
give
Alex
some
time
and
ask
like
if
he
could
get
back
to
us
on
Friday
or
early
next
week,
so
I'll
like
I'll,
probably
schedule
a
meeting
with
tcmcc
if
I
start
back
from
Alex
and
talk
to
ECS
next
week
like
early
next
week
and
then
before
the
Valentine's
Day
I
feel
like
we
got
the
general
consensus.
Is
it
if
this
is
something
we
should
push
for
then
great?
E
We
don't
have
to
solve
everything,
but
at
least
we
know
for
HT
here
what
to
do,
then
we
can
take
the
thing
like
the
matter
you're,
making
great
changes.
We
can
take
the
work
and
and
because
those
things
are
mostly,
we
always
tablets
in
ECS
for
years.
So
we
have
relatively
good
confidence.
We
don't
have
to
just
bake
it
for
another
month
or
something,
then
that
should
help
us
to
get
stable.
E
And
furthermore,
if
we're
going
to
work
on
anything
like
database
or
message
queue,
ECS
already
has
something
it
makes
it
easier
for
the
future
work.
If
the
answer
from
ECS
is
they
don't
even
think
they
want
ECS
and
open
climate
tree
eventually
to
become
the
same
product,
easy
I
should
stay
separate
and
under
the
elastic
governance,
then
there's
no
point
to
continue
this
discussion
and
then
we
know
the
proposal
like
the
old
type
of
thing,
and
we
couldn't
even
rely
on
this
thing.
Then
the
old
type
would
make
no
sense.
E
So,
let's
forget
about
it,
but
still
from
open
empty
perspective.
We
think
ECS
is
well
established,
there's
still
value
for
us
to
collaborate.
So
if
there's
certain
things,
we
think
we
can
make
the
change.
It
makes
sense
we
should,
but
then
it's
a
lower
priority
than
the
compact
given
like
open
country
is
also
well
established.
Let's
just
take
the
existence
and
Mark
that
as
stable,
because
I
already
see
some
feedback
like
from
from
GitHub
they're
using
that
in
production.
So
if
that's
the
case,
then
it's
even
more
as
far
as
to
make
breaking
changes.
E
E
Did
I
explain
that
clearly,
okay,
it
people
feel
like
in
the
end
we're
still
struggling.
We
cannot
make
a
decision
to
align
with
these
series
or
not.
Let
me
know
first,
it's
going
to
reply
to
his
blog
post
and
say
this
Project's
failed.
We're
going
to
waste
another
year
for
semantic
convention
and
I
feel
like
people
won't
want
to
pick
the
third
option.
So
we
only
pick
like
one
from
the
two.
C
Okay,
so
basically
the
this,
this
items
that
just
created
like
a
couple
of
days
ago
for
renaming
a
bunch
of
attributes
we'll
park
them
until
the
decision
for
ECS,
will
be
made
or
will
be
made.
Basically
in
two
weeks.
D
D
A
Yeah
I
may
open
one
more
issue,
because
kind
of
my
goal
here
is
to
queue
up
what
we
would
so
that
we
can
make
quick
progress
if
we
do
want
to
align
and
so
blood
Mill
and
I
met
with
Alex
this
morning
and
kind
of
worked
out.
The
network
level
attribute
mappings
for
http,
so
I'll
probably
open
one
more
issue
for
these
changes
and
then
yeah
the
only
one
that
I
think
we
might
want
to
that.
We
might
want
to
move
forward
without
ethius
on.
A
Is
this
one
because
it
had
already
come
up
in
the
open,
Telemetry
Community
in
this
Cosmos
PR
and
that
they
have
a
user
agent?
And
then
you
know
we
have
a
browser.user
agent
and,
as
Nev
mentioned,
this
gives
us.
The
user
agent
is
gonna,
is
being
kind
of
deprecated
anyways,
and
so
this
gives
us
then
a
namespace
under
which
to
add
sort
of
the
the
newer
stuff.
A
So
I'm
I
may
send
a
PR
just
to
start
pushing
this
along
getting
another
convinced
people
look
at
things
until
there's
a
PR.
E
Yeah
that
just
to
make
sure
like
we're
aligned,
so
in
case
the
ECI
is
open.
Climate
tree
merge
wouldn't
happen.
Then
I
think
we.
We
still
have
the
motivation
to
explore
ECS
and
see
if
there's
any
great
concept,
like
this
user
agency
that
we
should
take,
but
for
anything
critical
like
the
hd.url,
because
we're
not
going
to
along
with
ECS
and
and
sending
that
is
going
to
introduce
another
huge
discussion
about
breaking
existing
customers.
We
should
we
should
just
stick
with
the
current
version
and
Mark
that
as
well
right.
A
That's
the
as
I
measure,
the
wind.
That's
the
way,
I
see
the
wind
blowing.
A
A
Okay,
this
one:
okay,
that's
fine!
That's
this
one!
Oh
yeah!
Maybe
we
can
chat
about
this.
One
briefly,
I
sent
a
draft
PR
and
the
only
reason
why
the
reason
why
I
care
about
this
is
that
in
the
Java
instrumentation
we
are
already
are
like
taking
the
forwarded
exported
Proto
into
a
fat
into
consideration,
because
we've
had
customers
who
panic
when
they
see
that
they're
we
their
Telemetry
says
the
server
is
listing
on
HTTP
instead
of
https,
and
so
my
concern
is
when
we
Mark
gets
stable.
A
Without
this
consideration
than
were
kind
of
stuck,
although
I
can
always
introduce
an
experimental
attribute
in
the
Java
instrumentation.
So
I
guess
probably,
is
not
that
much
of
a
priority.
But
I
did
like
this.
I
mean
it
seemed
like
capturing.
A
new
attribute
seems
to
be
the
preferred
way
to
go,
and
we
already
have
client
IP,
which
is
basically
the
forwarded
for
so
we
could
have
client
scheme.
A
A
So
I
don't
couldn't
find
if
we
have
any
good
guidelines
for
when
we
do
something
generally
I
see
like
it's
nice
to
do
something
on
the
you
know
on
the
SDK
side,
because
then
you
know
each
back
end
doesn't
need
to
implement
it,
but
also
in
this
case
you
it.
You
have
to
multiply
that
across
all
each
potentially
each
instrumentation.
A
B
A
Yeah,
so
when
I
I
erased
this
in
the
spec
meeting
two
weeks
ago
and
the
I
think
the
the
concern
was
like
lying
like
not
like
the
Telemetry
line
about
what
it
really
is
there,
although
there's
some
precedence
in
like
our
net
host
name
being
like
logical,
yes,
and
so
you
know
it's
sort
of
The
Logical
scheme,
if
you
will
but
not
the
sock
scheme,.
B
C
My
two
sense
here
is
like
a
you
know:
if
we
have
less
conventions
implemented
by
SDK
or
even
by
back
end,
it's
better
right,
because,
as
a
user,
I
would
probably
like
to
see
like
a
road
telemetry.
C
For
example,
if
I
go
to
some,
you
know,
UI
or
something
I
would
I
would
ex
like
I
would
expect
to
see
exactly
what
what
was
there
without
any
kind
of
substitution
or
wrapping
something,
because
basically,
if
I,
like
I,
just
need
to
remember
less
right
and
like
I
need,
if
I,
if
I
see
some
something
that
I
need
to
understand
like
well,
it
was
substituted
by
SDK
or
by
back-end,
or
it
was
kind
of
replaced
by
visualization.
C
So
in
this
case,
it
will
be
probably
better
like
to
see
both
right
like
original
thing
and
forward
it
forwarded
scheme
if
they
are
kind
of
different.
So
it
will
just
bring
you
the
understanding
that
you
have
this
this
guy
in
the
middle,
and
you
know.
A
So
how
about
the
other
the
other
question
of
the
just
capturing
them,
since
we
already
have
these
attributes
under
headers
and
just
capturing
it
here
and
deferring
that
to
the
back
end.
C
Yeah
I,
like
it
better,
actually
just
because
that's
exactly
what
I'm
probably
talking
about
so
like
we
have
this
heater.
We
have
this
like.
We
put
this
to
expectable
place
and
they
can
just
or
like
even
the
UI,
just
visualize
it
visualize
it
as
is,
and
when
we're
thinking
about
these
kind
of
scenarios,
when
you
need
to
query
traces
where
this
kind
of
attributes
equal
something
or
contain
something
I
believe
it
will
be
also
a
good
case
for
it.
Just
because
you
will
be
kind
of
composing.
A
C
A
B
Not
just
to
clear
rates
or
some
metrics,
if
you
want
to
use
this
the
product
part
in
metrics,
if
it's
interesting
to
you
at
all
right,
you
would
have
to
parse
it
for
metrics,
which
is
much
worse,
so
I
think
the
question
is:
do
we
want
this
on
metrics
and
if
not
it
doesn't
matter?
C
Right
and
just
like
a
capture,
this
specific
part
of
it
basically
just
part
one
parse,
one,
one
part
of
it
related
to
scheme.
D
D
A
So
Ludmila
do
you
think
it's
worth
exploring
factoring
it
into
the
existing
HTTP
scheme.
B
Let
me
go
through
the
issue
and
think
about
it.
Okay,
you
can
assign
it
to
me
if
you
want
to.
B
D
A
If
it
doesn't
affect
this,
then
I
feel
comfortable,
removing
it
as
a
blocker,
because
I
I
can
deal
with
it
on
the
Java
instrumentation
side.
A
Yeah
this
one
I'm
thinking
of
just
I
think
probably
I
should
just
close
it,
given
that,
if
it
will
happen
naturally
with
the
ECI,
if
we
go
ECS
and
if
we're
not
going
with
ECS
we're
not
going
to
do,
we
wouldn't
do
it.
C
Yeah
I
believe
it
makes
sense.
So
basically,
ECS
pass
solves
this
this
one,
but
they
feel
like
I,
don't
want
to
go
with
GCS.
Basically,
we
are
not.
We
don't
want
to
break
any.
We
want,
don't
want
to
do
any
breaking
changes,
then
it's
better
to
leave
it
as
is
and
think
about
any
kind
of
changes
like
this
on
the
next
iteration.
A
Okay,
cool
I
will
do
that
copy.
This
foreign.
A
A
I
think
four
of
these
are
going
to
get
closed,
or
at
least
the
the
they're
no
longer
going
to
be
blockers
on
HTTP.
Once
these
first
four
Once
lead,
Millis
shared
at
attribute
groups,
spec
PR
goes
through.
A
A
Or
this
Riley
do
we
just
need
to
put
in
a
PR
to
up
or
to
say
that
they
should
be
and
then
to
update
all
of
the
existing
metrics
descriptions.
E
Oh
okay
and
it's
not
a
blocking
issue
at
home
like
why?
Why
do
we
look
at
the
cycle?
Do
we
think
description
is
something
we
cannot
change.
I
I
can't
remember
the
details,
so
only
if
we're
saying,
if
later
you
change
the
description,
it's
considered
as
a
different
metric
stream,
then
this
is
something
might
be
breaking
otherwise.
I
would
suggest
just
ignore
it.
For
now.
A
Okay,
yeah,
so
that's
okay
to
change
descriptions,
I.
A
A
E
A
E
A
A
A
Okay,
let's
see
what
else
we
have:
okay
defining
what
constitutes
breaking
changes
is
important,
but
we
will
I
think
we'll
push
on
the
General
working
group.
For
that,
oh
did
people.
Let
me
see
if
we
got
I
put
out
a
doodle,
because
I
I
asked
for
that
group
to
meet
weekly
so
that
because
I
know
we'll
have
a
lot
of
things
that
are
important.
A
A
E
E
E
D
A
D
E
D
E
E
A
And
do
we
think
this
is
a
blocker
I
guess,
probably
not,
because
we
can
Mark,
we
can
make
our
doc
mixed.
B
So
assuming
who
will
ever
have
a
dedicated
field
saying
that
this
is
HTTP
semantic
sorry,
this
is
an
https
plan,
all
right,
so
it
will
be
so
the
like.
If
we
push
the
score
tab,
if
it
goes
further
ever
then
it
will
be
an
additive
change.
So
some
new
thing
that
says:
okay,
it's
http
and
it
doesn't
seem
to
be
blocking
and
it
does
not
affect
anything
directly.
From
my
perspective
and.
A
A
Okay,
cool
I
think
I
feel
like
we're.
Looking
good
we're,
pushing
forward
stands
the
ECS
stuff
so.
B
A
B
I
feel
I
have
two
provocative
questions,
but
we
are
out
of
time.
I
can
stay
a
bit
longer,
but
if
you
need
to
go,
we
can
discuss
them
some
other
time.
A
I
I
should
talk
to
her
okay
meetings.
Yes,
yes,.