►
From YouTube: 2020-12-02 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
C
B
D
D
So
releases
releases.
C
C
I
was
like
no,
but
he
has
some
need
of
something
in
the
instrumentation
repo
that
you
all
are
publishing,
and
I
don't
know
what
it
is.
A
C
D
D
A
C
It's
mind
together
a
proposal
once
feedback
is
going
to
submit
an
otep
by
the
end
of
the
week
for
kind
of
what
the
what
the
process
is
for
getting
things
certified
or
marked
as
long-term
support
or
something
anyway
yeah.
It's
unclear,
I
don't
know
what
martian
needs
from
you
all.
It
sounds
like
he
needs
stable
apis
from
the
the
java
hotel
java
repo
and
I
he
mentioned
something
about
the
sdk,
but
I
think
he
was
using
sdk
to
mean
the
entire
repository,
not
the
actual
sdk
components.
C
A
C
Yeah,
I
mean,
I
think,
he's
just
gonna
I
mean
I
don't
know
he's
just
gonna
have
to
tell
whoever
he
promised
that
it's
not
gonna
happen.
I
don't
I
mean
I
just
don't
know
if
there's
any
other
solution,
like
he's,
he
said
we're
gonna
release
on
the
17th
and
I'm
like
we're
not
going
to
be
ready
by
the
17th.
D
Yeah
I
mean
there's
plenty
of
vendors
within
open
telemetry
who
would
love
to
have
this
released
and
dates
and
whatnot,
but
the
pride
the
the
community
has
been
pretty
intentional.
I
guess
about
not
providing
dates
and
not
and
saying
you
know
it's
going
to
be
ready
when
it's
ready.
C
Yeah
I
mean,
I
think
his
option
he
can
either
use
something
that's
not
stable
or
that
we
can't
guarantee
stability
on
wait.
There's
really.
I
don't
know
if
there's
anything
in
between.
If
we
try
to
force
stability,
I
mean
that's
just
we're
just
going
to
release
something
with
bugs
that
we're
going
to
have
to
break
anyway.
D
A
A
C
Stuff
right
I
mean
we've
been
trying
to
make
some
pretty
radical
changes
to
the
apis
to
get
them
cleaned
up.
I
mean,
I
think,
we're
all
in
agreement.
Yeah
just
be,
I
put
you
all
on
there,
except
for
mr
plum
just
to
bring
you
all
into
the
loop,
since
he
does
have
some
dependency
on
your
stuff
on
the
instrumentation
stuff.
D
D
Oh
speaking
of
the
the
all
the
the
radical
changes,
I
was
doing
a
bunch
of
refactoring
of
our
instrumentation
tracers
and
got
to
use
the
the
context
new
context,
api,
a
whole
bunch
and
like
it
it's
it's
so
much
more
like
memorable
like
I
didn't,
have
to
like.
Look
it
up
or
remember.
It
was
just
like
okay
yeah.
It
was
so
kudos.
B
A
C
Process
with
agent-
oh,
I
have
a
question
actually
before
that,
since
I
think
the
plan
was
to
release
again
this
week
as
the
first
week
of
december.
Do
we
need
to
have
a
pr
to
change
the
release
action
github
to
not
be
trying
to
attack
something
you
can't
commit
yeah?
Let
me
send
it
out
today:
that'd
be
cool.
You
know
what
that.
A
C
A
C
C
B
C
I
guess
the
only
other
thing
I
wanted
to
talk
about
is
just
your
last
little
pr
about
the
by
consumer.
I
mean
it's
it's
just
as
ugly
in
a
different
way.
C
It's
one
less
public
public
interface.
We
have
to
support
yeah,
yeah
yeah
and,
as
I
commented
on
the
pr
who
really
wish
we
had
union
types
because
it
would
make
all
of
this
so
much
easier.
I
mean,
of
course
it
would
make
everything
else
incredibly
more
confusing,
but
it
would.
A
C
A
C
A
D
A
D
Saw
you
were
you
were
pulling
it
in
shading.
It.
A
A
C
C
D
So
with
the
multi-release
jar,
on
the
I
mean
from
the
agent
side,
we
wouldn't
be
taking
advantage
of
that,
at
least
for
now,
because
I
think
it's
that's
in
the
sdk
right,
which
is
stuck
in
our
agent
class
loader,
which
isn't
going
to
be
multi-release.
Jar
aware
right.
D
A
A
C
Should
be
fine,
did
you
like
how
I
put
honorable
affiliation
last
week.
D
I
hadn't
seen
that
did
you
did
you
fix.
A
A
A
D
Yeah
I
mean
what
kind
of
do
in
that
if
it's
apache,
if
they
ship
with
a
notice
file,
then
you
have
to
also
ship.
You
have
to
also
ship
with
that
there's
sort
of
some
generic
things,
but
it's
it's
gray
enough
that
and
I
that
I
would
like
to
get.
A
D
A
D
So
probably
somebody
needs
to
open
a
cncf
help.
Desk
ticket
is
probably
the
end
result
or,
and
only
the
tc
members
are
allowed
to
do.
That.
C
D
Cool,
I
think,
honorable,
you
had
pasted
in
our
instrumentation
issue:
yep
yeah,
if
you,
if
you
don't
mind,
bundling
us
into
that
discussion
since
we
have
and
we
could
get
it
resolved
for
both
of
us.
D
So
the
weird
shading
versus
copying.
A
A
D
How
many
gradle,
how
many
modules
do
you
have
to
do
apply
the
weird
shading
in.
A
C
A
A
Yeah,
I
thought
yeah.
Actually
we
can
talk
about
that
a
bit
also
like
so
in
the
guava
stuff.
The
topic
did
come
up
where
we
have
these
internal
utils.
Like
czech
argument,
where
so
far
we've
been
avoiding
leaking
those
across
artifacts,
but
I
don't
think
there's
any
reason
to
avoid
that.
It's
on
the
same
repo.
A
A
A
A
A
A
A
Yeah,
definitely
something
I
would
avoid,
I
think,
repackaging
an
internal
package
that's
much
easier
to
like,
since
it's
just
copied
from
somewhere
out
anyways,
you
can
just
say
yeah.
This
is
going
to
be
relatively
stable,
it's
okay,
but
maybe
for
the
check
argument.
It's
still
a
bit
scary
to
link
that
across
artifact
boundaries.
A
D
With
the
weak
map,
does
it
make
sense
to
well,
you
said
yeah
since
it's
fairly
stable
and
we
would
then
kind
of
treat
it
stably.
A
A
C
A
C
C
Dependencies,
no,
I
don't
know
I
mean,
but
you
have
it,
you
use
it
in
the
agents.
Also
right.
A
That
we're
not
gonna
depend
on,
of
course,
agent
will
still
use
it
separately,
but
yeah,
so
so
we'll
use
it
in
the
context
package
we'll
also
use
in
the
jfr
package,
probably
also
in
the
trace
package
someday
and
rather
than
copying
into
every
single
package.
I
think
having
this
internal
intra
module
dependency
with
some
documentation
is
same.
I'm
sure
we
can
manage
that.
C
A
C
Oh,
I
did.
I
did
you
all
see
my
fun
story
of
figuring
out
why
my
local
builds
were
suddenly
taking
five
to
six
minutes
today.
D
C
C
C
C
By
the
way
on
that,
on
that
note,
I
haven't
seen
any
core
dumps
in
the
past
week.
A
I
mean
I've
been
meaning
to
add
gradle
version
like
the
version
test
shows
the
jvm
oh
yeah,
but
it
wouldn't
show
the
tool
chain
in
6.8,
they're,
gonna,
add
a
dump
of
tool
chains.
Also,
we'll
add
that
one
there,
let
me
upgrade
it
that'll,
be.
D
Useful
all
right,
let's,
let's
talk
about
instrumentation
repo
release.
D
A
B
B
A
B
B
So
my
my
my
arguments
are
just
the
follows.
I
don't
want
to
deal
with
sauna
type.
So
if
you
do
that,
that's
fine,
two!
We-
I
don't
know
if
you
want
to
be
present
in
j
center
at
all,
like
in
general,
in
the
future
and
some
of
our
clients,
probably
already
using
jcenter.
So
they
they
for
them.
It
will
be
a
breaking
change,
but
okay,
that's,
okay,
so
far
and
the
third
one.
B
A
C
A
A
D
B
I
have
seen
one
project
publishing
directly
to
maiden
central
well
via
sona
type,
I
mean,
and
it
was
like
the
huge
checklist
of
manual
checks.
Do
these
do
that
clicking
ui
this
clicking
ui
that
so,
if
so,
an
inventory
architecture,
what
not
we
can
do
to
totally
automatically
release
it.
B
A
I
don't
like
that
either,
of
course,
like
zipkin
one
is
fully
automated.
Now
there
are
there's
some
way
to
do
it
automated.
A
D
Cool,
that's
all
yeah
thanks
for
getting
access
so
that
we
can
try
that
out.
A
Yeah
and
so
today
I
pointed
my
local
build.gradle
at
their
snapchats
repo
and
uploaded
it
uploaded
fine,
as
you
would
sort
of
expect.
So
that
was
my
first
sanity
check
that
things
are
working
and
so
would
we
also
want
to
publish
the
snapshots
just
wanna
type,
just
so
they're
in
the
same
place
or
our
artifactory
publishing
hasn't
been
unstable.
So
there's
less
motivation
to
change
that
we
could.
B
I
think
the
only
motivation
maybe
is
that
if
we
start
advertising
that
we
are
made
in
central,
then
it
will
be
easier,
probably
for
people
to
grab
snapshots
from
the
same
vicinity
as
well.
Okay,.
D
I
wonder
how
long
those
the
sonar.
D
A
So
for
the
artifact
naming,
what
are
we
thinking?
Crafts
were
we
done.
Finally,
after
half
a
year
of
renaming,
artifacts.
D
Artifact
naming
so
things
are
going
like
the
change
a
little
bit
with
the
the
running
the
tests
in
the
agent
that
pr
what
1643
I've
like
memorized,
that
pr
number
now.
D
So
I
don't
know
if
we
care
or
not,
I
don't
really
care
too
much.
I
think
things
are
at
least
close
enough
that
we
could
go
ahead
and
you
know
we
draft
we
drift.
That's.
I
I'm
fine
with
that.
A
A
D
Yeah,
let
let's
push
I
mean
a
part
of
let's:
let's
push
for
0
12,
the
next
release
going
to
maven
central
yeah.
B
D
Yeah
and
then
I
wanted
to
pull
out
as
because
I
mean
it
was
kind
of
huge
and
a
lot
of
the
stuff
was
good
like
standalone
anyway,
and
so.
D
Pretty
good
now
I
was
just
looking
through
the
the
diff
there's
still
up,
there's
still
some
tests.
I
got
to
go
and
look
through,
but
at
least
the
diff
looks
very.
It
looks.
D
D
Did
you
have
a
chance
to
look
at
this
nikita,
not
yet
okay,
cool?
I
will
just
summarize
so
in
the
so
I
added
the
should
start
span
and
call
that
each
time
first
so
that
we
don't,
because
now
that
we're
returning
context,
we
don't
know
whether
to
call
and
or
set
attributes
on
the
spam.
That's
in
that
returned
context,
and
so
on.
Rog
was
asking
about
potentially,
instead
of
doing
that,
instead
of
having
a
should
start
span
to
in
the
start
span
method.
D
If
we
see
that
we
don't
want
to
start
a
real
span,
we
could
instead
of
putting
an
invalid
span
in
the
context
right,
which
you
right,
you
pointed
out,
is,
was
a
big
problem.
We
could
put
like
a
a
synthetic,
a
wrap,
a
another
span
that
just
wraps
the
parent.
The
the
current
span,
but
doesn't
really
do
much
like
you
can't
call
end
on
it
like
end,
would
be
a
no-op.
B
A
A
D
Kind
of
question
that,
though,
because
is
it
a
database
spans
like
I
don't
know
like
I'm
confused
if,
if
the
semantic
conventions
mix
like
that
like
can
you
have.
C
C
B
A
D
I
also
I
like
what
would
you
do
like
if
your
database
span
had
to
make
two
http
requests
to
make?
You
know
beneath
it.
B
B
D
B
D
Think
I
mean
that
the
thing
that
seems
the
simplest,
at
least
like
like
the
least
problematic,
is
just
suppress
the
http
spans
and
you're
just
looking
at
the
database
client
and
if
it
has
to
make
multiple
you
know,
low-level
network
calls
to
to
you
know,
to
create
a
table
and
then
push
some
data
into
it.
That's
fine!
The
end
users
don't
tend
to
care
that
much
about
that
low
level,
because
all
they
know
is
they
called.
You
know
populate.
A
A
B
Yeah
but
okay,
coming
back
to
this
problem
on
on
the
screen,
I
don't
like
neither
of
them
to
be
honest,
so
I
don't
I
don't
like
this
shoot
starts
playing.
I
know
you
don't
like
should
start
spin
yeah
and
I
don't
like
the
strange
wrapper
that
we
use
to
just
suppress
some
operations,
but
not
others,
but
I
don't
have
any
well.
If
we
don't
want
to
put
http
attribute
on
db
span,
then
one
more
option
is
just
to
return
now
from
start
span
which
trust
doesn't
like.
A
A
D
So,
let's
also.
D
D
Are
kind
of
dead
end
anyway,
but
on
a
server
span,
if
you
had
a
nested
server
span,
I
would
I
wouldn't.
I
would
want
to
suppress
that
completely,
because
anything
downstream
from
that
should
have
complete
access.
I
think,
to
the
I
mean
it
should
be
able
to
update
the
name,
it
should
be
able
to
yeah
even
potentially.
B
But
but
how
does
this
this
use
case
differ
from
the
call
stack
that.
B
B
Yeah,
I
I
mean,
but
what
we
do
with
call
stackdubs
we
we
do
some
if
and
then
we
return
and
then
in
the
end
we
check.
Okay,
it's
now
so
we're
doing
doing
nothing.
D
C
A
A
D
D
A
D
A
D
B
D
B
So
I
will
review
that
pull
request
today,
but
I
don't
I
don't
like
either
of
the
approaches
and
I
don't
have
any
better
one.
So
I
would
say
whatever
reason
the
pull
request
right
now.
It
will
go
for
now
and.
D
D
D
Okay,
any
other
prs
that
we
need
to
chat
about.
B
I
I
don't
have
pr,
but
I
have
one
small
question
to
to
ask
you,
so
we
have
this
issue
about
better
testing
wildfly
regarding
that
boot
class,
prefixes
and
class
loader
issue.
If
you
remember
that,
I
thought
you.
D
Closed,
I
thought
you.
Oh,
let's
see
what
this
is,
would
have
caught.
D
B
B
The
important
thing
about
we
want
better
tests
for
wildfly
and
I
have
the
very
same
problem
with
glass
fish
as
well,
due
to
osgi
the
same
problem
with
strange
class
loaders
so
and
we
are
developing
a
bigger
testing
matrix
in
splunk
for
to
test
combinations
with
different
java
versions
and
different
operations,
application
services
and,
like
that
awesome,
wi-fi-
and
I
I
want
to
upstream
that
into
hotel,
which
essentially
means
that
inside
smoke
tests,
we
will
have
like
a
new
subproject
which
that
which
builds
like
20
different
docker
images,
with
different
combinations
of
everything
and
put
some
small
web
up
in
into
them.
B
B
C
A
D
Yep
yeah,
I
think
it's
very
complimentary
with
the
the
testing
with
the
full
agent
which
is
going
to
give
us
the
detailed
level
testing
of
all
the
instrumentations
yep
yeah.
B
B
B
D
B
D
I
like
issues-
I
I
like
not,
I
mean
well
stack
overflow.
Maybe
I
like
issues
because
they're
you
know
that's
it's
the
place
of
record.
It's.
B
A
I
so
I
tend
to
prefer
chat
a
little
bit
for
support,
but
it
shows
also
fine
and
if
the
user
likes
it
better,
I'd
like
to
support
it,
but
we
could
also
consider
having
a
dedicated
repository
for
support
issues,
because
it
can
get
a
bit
messy
when
the
support
issues
are
mixed
up
with
the
development
issues
and
they're
also
spamming
everyone,
that's
watching
the
repository,
even
though
really
it's
only
the
people
that
should
be
helping.
That
should
be
watching
dishes.
A
A
A
A
A
A
B
Yeah,
well,
I
don't
watch
it
currently
either,
but
if
we,
if
we
like
start
sending
people
there
and
and
it's
it's
again-
I
I
wanted
to
say
it's
more
searchable
than
github
repo
and
then
I
say:
yeah
not
really.
A
A
A
A
C
A
D
D
And
the
last
I
heard
somebody
from
the
guy-
I
don't
remember
who
was
talking
about
it
but
said
that
you
know,
or
maybe
it
was
just
that
issue
of
when
or
even
if
we
move
yeah
the.
A
D
Yeah
yeah
ben
ben
from
lightstep,
doesn't
I
don't
know
he
doesn't
seem
to
be
very
into
slack?
I
don't
know
if
it's
a
broader.
A
D
A
A
D
John
you're,
our
our
chat.
D
C
A
B
D
D
A
A
D
Pops
up
in
the
like
it
makes
a
noise
and
a
little
like
like
the
actual
message.
The
first
yeah
line
of
the
message,
pops
up
and.