►
From YouTube: 2022-07-29 meeting
Description
cncf-opentelemetry meeting-2's Personal Meeting Room
A
C
D
C
A
C
B
B
A
A
D
C
Well,
I
don't
know
if
it's
only
java,
but
definitely
I
would
probably
not
have
tied
the
stable
releases
of
artifacts
to
conventions.
I
really
just
think
we're
going
on
forever
without
stable
instrumentation.
So
that's
the
most
important
thing
that
needs
to
be
solved
if
the
conventions
can
finally
come
out
of
thin
air
tomorrow
or
something
of
course,
that's
also
a
solution,
but
so
far
so
like
I
probably
would
not
have
tied
those
two
together
like
I
understand
some
other
reason
for
it,
but
I
don't
think
users
care
that
much
versus
having
something
to
use.
C
C
C
A
Yeah,
the
the
api
and
the
sdks
and
those.
C
A
Stable
that
just
leaves
you
to
kind
of
have
your
own
bespoke
instrumentation.
C
B
B
C
I
think
traces
tend
to
not
matter,
and
then
metrics
are
a
bit
more
important
than
the
stability
side.
Even
then,
though,
I
mean
people
have
been
living
with
some
changes
to
labels
for
years,
and
it
hasn't
been
that
bad,
so
like
micrometer
type
of
stuff,
or
sometimes
it
has
changed.
I
think
from
what
I've
seen
it's
okay,
but
yeah.
If
not
wanting
to
go
full
stable.
Then
at
least
something
that
says
some
sort
of
stability
that
is
like
of
the
apis
of
the
conventions.
D
To
that
would
I
mean
we
had
discussed
at
one
point
using
the
term
beta
or
you
know,
for
stable
apis.
I
don't
know
how
well
that
really
communicates
that.
B
And
the
the
instrumentation
right
now
is
lockstep
version
with
the
repo
as
a
whole
right,
so
just
thinking
we
could.
I
guess
I
mean
I
guess
the
real
question
is:
do
we
tie
stability
to
semantic
dimensions
and,
if
we're
required
to
then
we
don't
have
a
lot
of
options
because
many
conventions,
don't
even
I
mean
they
have
sort
of
a
a
sort
of
ad
hoc
hacky
way
to
say
this
particular
part
of
the
conventions
are
stable
right,
but
it's
not
really
like
it's
not
programmatically
accessible.
I
don't
think.
D
Well
in
to
that
point,
the
the
schema
versioning
mitigates
a
lot
of
that
the
problems
of
breaking
the
of
changing
attributes,
which
it
seems
to
seems
to
me
like
once
we
introduce
schema
versioning.
D
Well-
and
I
think
I
don't
know,
would
you
do
you
see
it
other
different
in
other
languages
onra,
because
I
know
in
java
we
had
instrumentation,
we
were
pretty
early
with
instrumentation
and
I
feel
like
a
lot
of
or
a
lot
of
the
effort
has
opened.
Telemetry
wide
has
still,
in
recent
times
been
going
into
just
getting
out
the
the
metrics
sdk
getting
out
the
logs
sdk
with
this
idea
that
oh
we'll
get
to
instrumentation
after
we
get
the
sdks
stable.
C
D
C
D
B
C
B
D
I
mean
they,
they
were
asking
instrumentations
to
be
hosted
externally
or
by
the
projects.
The
instrumented
libraries
themselves.
C
C
A
D
C
C
C
B
A
There's
a
bunch
of
other
artifacts
in
the
registry
related
to
exporters
and
stuff,
and
those
are
those
are
stabler
in
hotel,
java
and
they're
out
of
date.
I
just
I've
always
been
curious
on
like
whether
that's
actually
used,
and
so
how
important
is
it
to
keep
it
up.
A
Yeah
like
if
nobody,
if,
if
everybody's
in
that
or
if
it's
efficient,
a
percentage
of
people
are
not
maintaining
it,
then
it's
not
really
useful
as
a
tool,
and
so
you
don't
want
to
be
like
the
only
person
maintaining
it
a
tool.
That's
otherwise
done
used.
Yeah.
B
I
feel
like
the
the
documentation
story,
and
this
is,
I
think
this
really
feels
like
it's.
A
piece
of
documentation
has
been
very
scattered
and
like
the
documentation
team,
if
you
want
to
even
call
on
that
kind
of
just
do
stuff
their
way
and
maintainers
have
to
lump
it,
or
I
mean
it
just
kind
of
gets
whatever
they
want.
They
get
and
maintainers
don't
have
a
lot
of
voice
as
much
as
I've
complained
about
that
it
hasn't
done
anything.
B
It
hasn't
changed
anything
so
I
mean
when
they
decided
they
were
going
to
move
all
the
documentation
over
to
open
telemetry.iio
and
like
all
right.
Well,
you
all
are
maintaining
it.
I'm
not
going,
and
you
know
looking
at
dr's
over
there
don't
barely
have
time
to
do
stuff
in
the
java
repo.
D
There
certainly
seems
to
have
been
a
pickup
in
the
like
in
sort
of
the
dev
advocate
the
people
interested
in
open
telemetry.
So
I'm
hoping
we
can
mobilize
them.
You
know
in
some
way,
as
opposed
to
the
the
the
the
way
that
they
tried
to
mobilize
that,
instead
of
mobilizing
us.
B
B
I
don't
know
who
you
think
you're
asking
here,
but
I'm
not
working
on
that
stuff.
That
was
a
very
odd.
I
don't
know
it
was
a
misunderstanding.
I
mean
I
think
I
understood
that
motivation
was
like
here's
where
the
java
community
lives.
I'm
like.
Well,
that's
not
true
at
all,
that's
where
jack
and
I
live.
B
Basically,
it's
not
really
like
a
big
community.
I
mean
the
community
effort
is
not
happening
in
the
sdk
very
much,
which
is
good.
It
shouldn't
be
right.
This
should
be.
It
should
be
like
the
core.
It
should
be
stable.
It
shouldn't
change
very
much.
It
shouldn't
just
have
random
community
people
coming
in
and
hacking
on
stuff,
but
yeah.
I
think
it
might
have
just
been
a
misunderstanding
from
the
demo.
The
demo
people
I
have
a
practical
question
is
not
going
to
be
attending.
Can
we
change
the
time
of
this
meeting?
B
Oh
and
I'm
sure
jack
would
be
in
favor.
Even
it's
like
what
8
o'clock
you're
8
15
your
time.
B
C
B
To
move
it,
I
don't
know
when
the
right
time
is,
but
I
don't
have
a
huge
amount
of
flexibility.
B
I
was
thinking
about
it
this
morning
and
then
it
turned
out
that
I
was
very.
I
thought
I
was
very
busy,
but
it
was
mostly
that
I
forgot
that
it
was
going
on.
I
actually
think
I
could
make
the
night
the
9am
meeting.
Is
it
nine
or
eight
nine?
No,
I
think
I
could,
if
I
just
put
it
on
my
calendar,
I
seem
to
put
it
on
my
calendar
and
remember.
B
D
Awesome
yeah
because
our
our
agendas
have
been
slowly
waning
yeah,
so
we
typically
can
it's
not
like.
We
need
the
extra
time
lately,
we'll
see
after
you
know
the
summer
holidays.
Maybe
things
will
pick
up,
but.
C
B
Related
topic,
both
jack
and
trask,
do
you
is
there?
Anybody
from
your
respective
companies
is
interested
in
becoming
more
active
and
working
towards
maintainership
on
java
and
the
core
repo.
A
I've
I've
asked
my
my
immediate
manager
if
we
can
get
more
resources
on
it
and
it's
it
hasn't
gone
anywhere.
I
can
like
cast
a
wider
net
and
just
try
to
appeal
to
people
directly
and
see,
but
that
would
be
kind
of
you
know
not
not.
D
C
A
I
went
to
the
hotel
community
day
in
austin
back
in
june
and
I
think
some
folks
that
used
to
work
with
onorag
were
there
and
they
were.
They
were
asking
some
probing
questions
about
how
much
time
and
effort
it
takes
to
be
a
maintainer
on
java,
and
so
I
want
to
be
surprised
if,
if
aws
offered
some
resources.
B
C
B
To
be
a
maintainer
is
a
lot
more
than
I
think.
Most
managers,
a
expect
and
b
are
willing
to
give
up
unless
they
really
feel
like
open
telemetry
is,
is
strategic
for
them
right
which
clearly
new,
relic
and
swunk
do
mike
roberta,
doesn't
not
relevant
for
our
business,
at
least
not
at
this
very
moment
so,
but
it
feels
like
it's
aws
and
like
I
I'm
actually
a
little
disappointed
that
nobody
from
splunk
has
stepped
up.
B
Mateosh
occasionally
will
look
like
put
in
a
pr,
but
I
mean
what
we
really
need
is
we
really
need
careful
reviews
more
than
anything
else,
and
a
lot
of
people
just
want
to
like
write
some
code.
They
don't
want
to.
You
know,
put
in
the
on
the
unpleasant
work
of
doing
deep
reviews,
which
is
kind
of
the
most
important
thing
a
maintainer
can
do.
B
I
think
it's
actually.
It
is
a
relatively
rare
skill
set
also
like
you
not
only
have
to
be
very
kind
of
a
deep
java
expert,
relatively
cheap
job,
but
you
have
to
really
care
about
user,
apis
and
and
design
of
apis
and
making
and
thinking
forward
about
how
they're
going
to
be
used
and
that
sort
of
stuff-
and
it's
not
a
very
it's,
not
a
skill
that
a
lot
of
people
have
trained
themselves
in
and
I
don't
think
there's
any
like
off-the-shelf
training.
B
C
D
C
They
don't
even
have
that.
I
think
that's
where
it
gets
a
bit
hard
to
even
step
up
to
maintainership,
but
if
they're
very
good
at
java,
like
obviously
all
of
us
had
that
stage
like,
I
think
for
me
as
work
on
army
or
something
I
hadn't
really
done
libraries
before,
but
I
was
good
at
java,
so
I
could.
C
A
Have
to
do
the
code
reviews
so
they
have
to
care
a
lot
about
code
reviews,
yeah.
B
Passion
so
finding
kind
of
that
combination
where
you
something
you
care
about
and
want
to
learn
about
and
having
the
job
expertise.
Is
it's
hard
to
find
it's
a
bummer
like
why
won't
plum
step
up
that
dude
cares
deeply
about
apis
and
is
really
good
at
java.
I
keep
on
prodding
him
every
time
I
see
him,
but.
C
D
Couple
contractors:
what
do
you
think
our
big
sdk
side
thing
is
coming
up,
I
mean,
is
it
just
this?
Is
it
just
the
kind
of
steady
stream
of
little
things
or
just.
A
Yeah,
that
shouldn't
be
a
huge
lift.
That's
that's
kind
of
what
I'm
targeting
next
is
an
api
implementation
for
events
and
log
openers
per
the
changes
that
recently
got
merged
to
the
spec
and
yeah
I
mean
the
log
sdk
is
the
simplest
of
the
sdks,
so
assuming
the
spec
moves
along
on
that
we
shouldn't
have
a
dramatic
number
of
changes
related
to
that,
and
you
know
I
I
would
be
I'll-
be
ecstatic
once
just
we
have
stable
apis
and
sdks
for
the
three
signals.
A
I
think
that's
like
a
really
solid
point
to
get
to,
and
it's
it's
like
within
view.
You
can
see
a
path
to
that
where
I
can.
B
Yeah,
my
my
thought
has
always
been
that
once
we
hit
that
point,
the
amount
of
very
active
work
in
the
sdk
will
drop
off.
It's
probably
going
to
be
auto,
configure
like
it's
going
to
be
really
doing
configuration
work
and,
like
all
of
the
thoughts
we've
had
about
you
know,
configuration
file
formats
and
that
sort
of
stuff
and
like
there's,
I
think
it's
going
to
be
mostly
on
that
end
of
things.
On
the
configuration
side.
A
Yeah
and
we'll
so
you
know
I
think
java
has
been
kind
of
at
the
forefront
of
some
languages
in
terms
of
prototyping
stuff
at
the
ideas
and
things
like
that,
but
I
think
you
know
I
think
it's
probably
up
to
some
other
language
to
drive
the
story
of
file
based
configuration,
and
you
know
if
that
is
specified
at
the
spec
level.
It's
pretty
easy
to
implement
once
it's
like
just
there
and
it's
you
know,
there's
a
document
describing
it.
B
So
I
mean
it
could
be
driven
from
the
java
instrumentation
side
as
well,
though
like
if
mateo
should
want
to
pick
that
up,
or
I
mean
the
agent
is
certainly
going
to
end
up
using
it.
However,.
B
B
I
mean
I
could
go
my
it's
funny.
My
I
just
got
a
splunk
t-shirt
in
the
mail
today.
I
think
my
manager,
the
t-shirt
for
the
project
last
project
I
worked
on
and
I
think
that
was
funny.
I
was
just.
B
D
I
think,
looking
out
for
new
work
on
the
in
the
stk
repo,
like
like
a
configuration
file,
configuration
thing
or
like
a
profiling,
you
know
some
time
down
the
road
and
looking,
I
think
we
might.
We
might
have
better
luck,
getting
roping
somebody
in
sort
of
specifically
to
drive
one
of
those
things
versus
you
know
the
the
all
the
sd,
all
the
other
review
and
all
the
other.
You
know
not
really
like
a
maintainer
track,
but
to
offload
that
stuff
to
give
the
two
of
you
who
are
left
time
to
do
the
other
stuff.
A
Yeah
yeah
and
I'd
be
like
so
if
somebody
wanted
to
drive
that,
I
mean
I'd,
be
I'd,
be
really
happy
about
it,
and
that
would
give
them
very
good
pr
reviews
and
I'd
suggest
that
they
become
an
approver.
D
Yeah
we
should,
but
we
we
should
proactively
like
if
we
see
those
things
coming
proactively.
You
know
reach
out
and
try
to
get
somebody
as
opposed
to
necessarily
just
waiting
for
somebody
to
show
an
interest
in
it.
Yeah.
A
D
Talk
about
it
in
the
sig
yeah
because
you
know
like
I
you
know
we
might
you
know
mathish
might
be
interested
in
like
file
based
configuration,
say.
A
D
B
A
Nice,
I
didn't
see
that
she
found
some
an
issue
where
jappy
didn't
catch
a
compatibility,
change
or
issue.
C
B
D
A
B
B
C
B
C
D
Yeah,
I
guess
I'm
just
when
I
update
our
stuff,
our
distro
and
things
I'm
used
to
things
you
know
having
to
breaking
changes
because
of
the
instrumentation
and
java
agent
apis.
I
probably
just
you
know,
threw
in
an
extra
method
or
something.
C
B
Yeah
dependable
just
came
in
this
morning
and
updated.
We
just
turned
to
pentabot
on
one
of
our
repos
to
try
it
out
updated,
tried
to
update
the
sdk
to
from
1.13
to
1.16,
but
it
depended
by
it
wasn't
smart
enough
to
figure
out
that
our
alpha,
our
dash
alpha,
is
a
different
art,
different
set
of
artifacts.
It's
not
just
like
an
alpha
version
of
things,
so
it
tried
to
make
everything
be
dash
alpha
because
clearly,
that's
the
latest
version
right
and
I
had
to.
B
D
Does
depend
about?
Does
it
not
understand
that
gradle
our
gradle
dependencies,
or
did
we
choose
not
to
use
it
for
gradle
dependency?
We.
B
A
C
D
I
was
thinking
of
about
because
I
added
a
step
into
the
instrumentation
release
now,
where
it'll
send
a
pr
downstream
to
the
contrib
repo
and
update
the
hotel
version
down
there,
but
like
another
option
would
have
been
if
we
had
dependebot
downstream.
It
would
just
automatically
notice
when
the
new
release
hit
maven.
D
But
yeah
I'm
hoping,
I
think
at
some
point
I'll
if
this
one,
I
just
added
that
so
if
that
goes
smoothly
in
the
next
release,
I'll
add
it
to
the
java,
the
core
repo
to
have
it
open,
a
pr
in
the
instrumentation
repo
to
do
the.
C
C
C
Yeah
yes
not
easily,
because
it's
it
just
reports
and
otherwise
it's
just
strings
inside
a
cockpit
file.
So
there's
no
mechanism
for
it
to
actually
update
the
file.
I
think.
C
C
B
C
B
Me
I
saw
a,
I
saw,
an
alert
that
the
version
of
okay
should
be
that
we
were
using
inverta
had
a
cve
via
the
kotlin
standard
library.
I
don't
know
how
recent
that
was.
I
didn't
dig
into
it
too
deeply,
but
I
don't
know
whether
that's
something
that
we
we're
not.
D
B
A
C
B
B
Anyway,
we're
probably
gonna
bail
out
of
here
yeah.
It's
been
a
very,
very
been
a
pleasure.
C
C
D
Does
it
is
there
a
go
equivalent
of
brave.
C
Yeah,
by
the
way,
I
don't
with
this
thing-
it's
actually
not
that
related
to
them,
like
I'm,
trying
to
open
source
our
api
photos,
and
so,
as
part
of
that,
I
wrote
this
tool
to
make
it
easier
to
run
proto
compiler.
So,
if
anyone's
struggling
to
run
the
portal
compiler,
because
you
have
to
manage
its
dependencies
and
plugins
and
stuff,
I've
been
trying
to
make
my
life
easier
with
this
tool.
A
Did
you
switch
your?
Are
you
operating
under
a
different
handle
or.
C
C
On
the
maintainer
thing,
of
course,
also
to
keep
in
mind
like
once,
logs
become
stable.
I
don't
know
if
this
is
the
intent
to
trust
question,
but
the
sticker
repo
will
have
a
much
smaller
stream
of
issues.
Presumably
so
at
some
point
you
could
just
merge
the
instrumentation
sdk
team
into
one,
maybe
each
side
a
bit
more
involved
in
the
other
repo
just
because
what,
under
the
assumption
that
sdk
will
probably
won't
have
as
much
work
anymore.
C
B
C
Yeah-
and
maybe
that's
like
the
approvers-
can
be
both
these
java
provers,
if
there's
still
any
approvers
left
after
that
trimming
or
and
the
java
instrumentation
team
can
just
by
default,
be
the
approvers
in
the
sdk
report.
Maybe
that
would
help
with
the
code
reviews.
D
A
D
I
would
thumbs
up
that
because
also
then
that'll
give
us
all
right
permission
to
both
repos
so
that
we
can
transfer
issues
as.
D
B
C
B
Anyway,
we
should
we
should
bring
this
up
on
next
thursday,
but
I'm
not
going
to
be
here-
I'm
not
going
to
be
on
vacation,
so
I
won't
be
around
next
week,
won't
be
around
next
week
at
all.
A
And
I
will
open
an
issue,
I
think
to
try
to
you
know,
propose
that
we
trimmed
down
the
sdk.
So
just
like
name
a
bunch
of
you
know,
artifacts
that
I
think
are
decent
candidates
to
move
to
contrib
and
just
let
discussion
take
place
and
then
you
know
we
can
go
from
there.
B
Yeah,
I
think
the
the
aws
ones
are
probably
the
ones
there's
not
that
much
left
right.
There's
like
an
id
generator-
and
I
don't
remember
oh
the
resource
yeah-
that
resource
detectors.
Unless
we
can
get
somebody
from
aws
to
sign
up
for
maintenance
to
be
maintaining
that
stuff.
We
should
move
that
to
contrib
for
sure.
A
I
mean
even
if
we
could
get
someone
to
sign
up
like
I
mean
it
kind
of
breaks,
a
pattern
to
put
it
in
the
core
sdk
repo,
and
it's
not
a
precedent
that
you
know.
There's
any
other
examples
of
so.
C
I
think
the
only
thing
was
that
so
id
generator-
that
is
definitely
x-ray
so
and
that's
why
it
was
already
moved
and
so
the
resource
detectors
we
have
still
always
wanted
to
actually
include
them
in
the
upstream
java
agent.
I
think-
and
so,
and
not
only
database
of
course,
and
so
there
is
sort
of
awkward
if
it
was
only
going
to
be
aws.
We'd
expect
it
to
have
gcps
or
there
also-
and
we
don't
have
any
pattern
yet
of
depending
on
contrib
from
job
agent.
C
C
C
D
D
C
In
communities
yeah
I
mean
definitely
maybe
I'll
still
see
the
pr
or
something
like
there
was.
One
thing
I
was
waiting
like
I
do
want
to
fix
the
windows
build
right
now.
The
wire
compiler
has
some
issues,
so
I
was
waiting
on
the
new
version
of
gradle
for
that.
So
I'll
probably
do
that
at
some
point
you
should.