►
From YouTube: 2022-02-22 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).
B
B
Well,
this
one
was
early
to
ask
matteo
sure
nikita.
It
makes
sense
to
me,
but
I
think,
but
I
want
to
wait
to
get
one
of
them
to
weigh
in
before.
I
re
do
another
automation
to
to
clean,
clean
up
the
automation.
It
will
clean
up
the
automation
a
bit,
or
at
least
I
don't
need
this
pr,
but
I'll
need
to
clean
up
the
automation
to
do
the
latest,
instead
of
or
instead
of
bumping
it.
B
B
Yeah,
oh
yeah,
I
wanted
to
ask
about.
I
saw
you
the
crane
pr.
I
was
curious,
so
that
was
about
the
point
of
copying
the
github
images
over.
To
I
mean
sorry,
the
docker
images
over
to
ghcr
is
because
of
that
pull
docker
pull
limit,
and
I
was
just
wondering
I
was
surprised
we
haven't
hit
that
in
the
instrumentation
repo.
A
A
A
B
Yeah
yeah,
that's
interesting,
I'm
curious
which
yeah,
because
certainly
we
pull
we
always
pull.
We
do
pull
this.
Where
I
see
well,
we
pull
it
for
I
see,
but
we
don't.
We
pull
it
for
building
our
smoke
test
images
right,
but
then
we
put
our
smoke.
B
B
The
tell
me
if
my
understanding
is
right
of
the
the
jmx
metrics
in
this
repo.
Is
that
it's
for
like
if
you
have
an
app
that
exposes
jmx
metrics
already,
and
you
basically
want
to
scrape
from
jmx,
as
opposed
to
like
scraping
from
prometheus
or
yeah
collector
nut
collector.
B
What
do
you
call
and
prometheus
endpoint
and
it
it's
certain
known,
metrics?
It's
not,
or
I
assume
it
says,
it's
groovy.
It's
also
configurable.
If
you
want
to
grab
your
own
groovy
great
jmx,
metrics.
B
A
A
B
Yeah
I
got
at
first,
I
thought
they
were
asking
which
repo
to
contribute
to,
but
then
I
then
I
thought
that
they
were
just
asking
like
which
project
they
should
use.
A
B
Oh
yeah,
just
as
far
as
where
sue
like
I
was
kind
of
imagining
it
kind
of
like
the
jaeger,
remote
sampler.
Does
that
make
sense.
A
A
B
A
A
I
see
I
mean
first,
I
was
thinking
this
either
way
more.
Okay,
I
think
it'd
be
good
to
finish
the
review
and
then
move
rather
than
losing
the
pr
oh
cool.
A
A
B
Okay,
I
didn't
I've,
read
the
spec
more
or
less.
B
B
Yeah,
I'm
kind
of
I
I
have
some
interest
in
this
also
because
it
we
in
we
our
ingestion
service.
We
send,
like
an
item,
count
the
inverse
of
the
sampling
probability.
Basically,
so
it
does
have
a.
I
have
some
hackery
in
the
java
agent
today
to
deal
with
that
that
sort
of
mimics
it
was
a
like
dumping
stuff
into
this
tray
state
to
propagate
down
and
then
not
propagating
the
trace
state
over
the
wire.
A
B
A
A
B
A
B
It
totally
makes
sense
to
me
from
for
a
library
as
a
library.
I
think
this
is
a
great
solution.
I,
what
I
don't
know
is
in
for
the
java
agent.
A
A
B
A
B
I
just
know
that
in
practice,
when
java
9
happened
and
they
introduced
the
platform
class,
loader
jdbc
was
the
big
one
that
was
no
longer
in
bootstrap
right.
B
Yeah,
except
I
have
I
hadn't,
run
into
this,
because
I've
never
tried
to
put
this
in
into
an
agent
before
yeah.
This
is
another.
A
B
B
No,
like
m-beam
stuff,
like
platform
server,
that.
B
A
B
B
A
A
A
I
wouldn't
expect
the
xml
stuff
to
be
on
bootstrap
either.
Maybe
it
is
but
I've
added
like
do
you
think
it's
anything
in
bootstrap
should
be
available
right,
like
all
the
ad
opens.
All
of
that
stuff
is
handled
by
the
platform
class
letter
like
there's
no
way
the
bootstrap
cluster
can
restrict
what's
inside
it,
so
it
must
only
be
the
base
stuff.
Is
that
a
accurate
way
of
thinking
about
it.
A
B
I
think
it
might
only
be
I'm
trying
to
think
like
nothing
here
that
at
least
that
I've
used,
because
the
logging
I
always
have
done
essentially
the
patch
logger,
the
shading-
to
to
internal
logging,
because
it
causes
other
problems
with
agents.
B
B
A
B
And
they
had
open
telemetry
as
a
dependency,
and
so
our
agent
was
so
in
our
shaded
on
stuff.
It's
supposed
to
find
sdk
is
not
shaded,
but
the
api
is
shaded,
so
it's
supposed
to
find
that,
but
it
was
finding
the
sdk
on
the
user's
class
path.
B
B
B
Not
doing
that
like
if
java
on
startup
is
creating
a
sort
of
custom
platform
class
loader
that
includes
the
users
stuff
and
anything
that
the
user
any
modules
that
the
user
referenced.
A
So
that's
if
they
created
a
custom
jre,
not
even
I
don't
think.
A
A
A
B
A
B
Yeah
we
have
almost
no
jpms
tests.
We
have
one
jpms
test.
I
was
trying
to
mess
around
with
that.
One
smoke
test
it
basically
piggybacks
on
jetty
having
a
jpms
mode
to
run.
A
A
A
B
A
A
A
Mean
yeah
and
I'm
sure
we
don't
want
to
limit
ourselves,
I
mean
if
we
wanted
to
use
the
java
http
client
for
some
reason
it
seems
like
just
because
it's
a
java
agent
that
can't
be
a
reason
to
not
use
these
things
right,
so
there
must
be
some
way
to
work
yeah,
even
if
we
don't
know
it
yet.
I
hope
we
can
figure
it
out.
A
B
A
B
Or
if
java
agent
users
have
to
add,
I
mean
it's,
I
don't
think
it's
so
bad
for
because
it's
not
like
you
as
long
as
you
don't
have
to
my
mind,.
B
A
A
A
A
B
A
B
A
B
B
I
would
expect
11,
I
always
go
to
11.
A
A
A
A
A
A
B
A
A
Yeah,
if
it
was
there,
then
it
would
probably
factor
it,
and
if
it's
not
there,
then
I
have
a
feeling
the
job
agents
being
treated
like
any
of
those
external
external
module
name
like
when
you
add
that
to
the
manifest,
and
so
it's
not
actually
modularized,
and
so
that's
why
stuff
gets
exposed
to
the
java
agent.
So
that's
why
I'm
starting
to
feel
a
bit
more
optimistic
that
maybe
users
don't
have
to
add
stuff
for
job
agent
dependencies
because
they're
probably
being
treated
as
one
of
those
non-module
dependencies?
A
B
A
A
B
They
see
it
looks
like
we
say
that
it's
http,
but
you
know
it's
getting
terminated
at
iis,
for
example,.
A
B
B
Confusion
in
going
in
this
direction
than
in
that
other
in
the
other
direction.
A
Yeah,
I
think,
like
scheme,
if
this
feels
like
one
part
of
the
url,
the
url
never
changes,
but
even
if
it's
handled
internally,
somehow
so,
if
the
urls
being
forwarded
along
this
to
make
sense
to
reflect
them
to
scheme,
if
it
was
net.tls,
something
like
that,
then
of
course
that
would
be
the
true
thing.
The
http
scheme
is
just
a
part
of
the
url.
I
think.
B
A
A
A
B
Well,
I've
got
a
new,
a
a
new
thought
that
I'm
mulling
over
of
I
started
poking
around
in
the
node.js
repo.
A
B
And
I
kind
of
like
if,
if
all
the
reap
of
all
the
instrumentation
reapers
do
the
same
thing,
then
that
becomes
the
de
facto
spec.
B
B
Is
there
anything
I
thought.
B
A
B
It's
yeah,
it
totally
makes
sense
from
from
instrumentation
ecosystem
perspective,
and
we
really
do
want
to
push
that
and
that's
what
the
whole
instrumentation
api
is
about,
and.
A
B
On
the
metrics
sdk,
I
thought
there
was
one
more
thing
that
so
josh
had
split
out
from
his
pr.
There
was
the
identity
callbacks,
the
multiple
callbacks,
one
of
those
I
thought
got
split
out,
and
so.
A
A
A
We
would
be
ready
on
friday,
I'm
also
going
to
discuss
what
our
plan
would
be,
because
whether
we
want
to
try
to
release
in
march,
which
would
mean
probably
one
rc,
with
all
this
breaking
changes
before
that,
and
so
it
probably
wouldn't
be
able
to
get
released
on
the
first
week
of
march
or
if
we
just
target
for
first
week
of
april.
A
B
A
A
B
All
right
well
have
a
good
good
tuesday,
hey
you
saw
the
the
patch
release
also.
B
We're
getting
there
that's
why
I
had
to
go,
make
some
more
release.
Automation,
changes
to
increase
the.