►
From YouTube: 2022-03-04 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
B
A
Yeah,
this
might
have
gotten
deleted
before
the
morgan
switched
us
back.
B
A
C
Yeah
right
now,
I
think
I
have
a
mixture,
some
mixture
of
copying
to
my
own
calendar
and
actually
accepting
meeting
invites
from
the
official
calendar
I
can't
figure
out,
which
is
which
so
I
always
double
ch.
I've
just
started
double
checking
by
opening
up
the
open,
telemetry
calendar
and
making
sure
the
meeting
id
is
the
same
each
time.
B
C
B
C
A
So
I
did
ask
about
kind
of
the
questions
we
had
before.
I've
been
not
taking
good
notes.
A
Oh
yeah
so
kind
of
explain
the
bridging.
What
would
what
would
work
there?
I
think
she's,
like
okay
with
that.
You
know
it's
going
to
be
pulled
in
you
know
it's
going
to
be
still
a
single
api
artifact,
but
as
long
as
they
don't
configure
or
pull
in
the.
A
I
think
she
liked
the
fact
that
the
sdk
metrics
sdk
was
separate
artifact,
so
that
if
people
don't
pull
that
in
it's
no
op,
and
so
that
mentioned
that
you
know
the
the
bridging
the
metric
reader,
bridging
kind
of
talk
through
that
briefly
hoping
they
do
that,
but
they
and
then
asked
about
the.
Why?
Why
a
metrics
but
no
tracing
abstraction
and.
A
It's
like
the
so
they
were
using
open
tracing,
so
it
kind
of
made
sense
to
go
to
open
telemetry
on
the
metric
side.
Yeah.
I
don't
know
why,
like
I
mean
it
kind
of
it
sounded
like
there
was.
You
know
the
I
mean
the
metrics
group
is
a
different
group.
It
sounds
like
she
doesn't
have
sort
of
direct,
that's
not
as
much
her
group,
but
I
did
say
that
you
know
we
would
way,
prefer
them
to
use
micrometer
directly
than
to
create
a
third.
You
know
another
abstraction.
B
A
A
It
sounded
like
one
of
the
reasons
was
that
customers
like
the
api
or
something
about
their
api
and
yeah.
I
just
don't
think
that's
a
compelling,
and
I
mean
I
think
micrometer
is
fine.
A
A
A
C
Yeah,
it's
going
to
be
an
annoying
world
where
there
need
to
be
two-way
bridges
for
between
all
of
the
different
metrics
libraries
metrics
apis,
but
that's
probably
the
world
we
live
in.
C
C
A
What
did
we?
There
was
some
other
topic
that
oh
yeah,
we
talked
at
the
end
about
the
jvm
metrics,
the
cpu
metrics
jonathan's
spec
pr,
and
so
I
owe
him
a
comment
on
that
pr,
but
I
think
we're
pretty
close
to
I've
been
spamming.
A
The
spec
repo,
with
trying
with
prs
to
get
to
get
clarifications
on
how
we
should
move
forward,
and
I
think
mostly
no
now
so
the
if
you
didn't
catch
so
we're
not
the
they
don't
want
us
to
emit
system
metrics,
because
that
that
the
thing
that
will
be
confusing
with
the
collector,
although
I
did
find
that
the
go,
does
have
a
module
to
emit
system
metrics
for
people
who
don't
want
to
use
the
agent.
A
Yeah,
well,
it's
it's
very
much,
not.
A
No,
I
don't
think
they've,
I
don't
think
at
least
bogdan
and
tigrin.
Very
much
don't
want.
B
A
Oh,
that's
a
good
point.
I
did
not
make
that
connection
yeah,
so
this
was
my
pr
which
was
going
well
until
bogdan
chimed
in
that
he
agrees
with
the
the
change
but
disagrees
with
the
motivation
for
the
change
so
yeah.
B
A
A
B
A
So
yeah,
that's
a
good,
so
eventually
I
realized.
I
don't
even
care
about
this,
because
I
thought
that
I
was
interested
in
it
from
the
system.
Let's.
B
A
System
cpu
count
and
so
well
system
cpu
count
and
system
cpu
time,
but
I
forgot
that
we
don't
even
get
system
cpu
time
from
jmx
from
the
jvm
and
system.
Cpu
count
is
really
a
process,
cpu
count
the
runtime
available
processor.
A
B
A
A
A
So
like
there's
system,
cpu
utilization,
which
is
like
system
load,
but
this
is
very
vaguely
defined
as
recent
cpu
usage.
A
Whereas
the
system.cpu
utilization
is
defined
as
a
cpu
time
divided
by
elapsed,
time.
B
B
Yeah,
I
agree,
though
the
number
is
not
ideal
but
yeah.
I
have
like,
if
there's
a
question
of
whether
java
apps
should
export
system
metrics
at
all
or
not
at
all,
I
think
exporting
stuff
is
better
than
not
exporting
stuff,
like
I've
always
appreciated
having
some
sort
of
symmetrics
from
my
apps
rather
than
having
to
run
a
collector.
B
A
I
did
ask
I
did
get
clarification
on
the
objection
if
it
was
just
about
the
system
namespace
and
yeah.
That
was
basically
the
it
was
because
they
don't
want
it
to
clash
with
the
metrics
coming
out
of
the
collector.
B
A
B
A
Yeah
tigran's,
he
finds
it
surprising
that
application
would
emit
metrics
in
the
system
metric
space,
okay,.
B
How
long
is
the
system
cpu?
Is
there?
Because,
when
I
was
running
apps
before,
I
knew
there's
a
downside
that
every
single
container
in
the
same
node
reports
the
same
system
cpu?
So
it's
sort
of
wasteful,
but
that
was
the
only
sort
of
downside.
Otherwise
it
was
very
convenient
to
have
those
without
having
to
run
extra
infrastructure.
B
A
Yeah,
no,
it's
a
really
interesting
what
you
said
about
the
being
collector
maintainers.
A
Yeah
I
didn't
want
to
so
on
this
one
bogdan
was
asking
for
some
maintainer
input,
and
so
I
think
some
folks
and.
A
A
A
At
that
one
later
now
that
I
kind
of
understand
the
landscape
of
what
we're
dealing
with,
but
I
think
we
have
almost
what
we
need
to
move
forward
with
the
jonathan's
cpu
pr.
After
a
while.
A
Okay,
so
yeah
matthias
was
oh
yeah
because
you
had
asked
honorary
about
this,
and
so
I
was
just
curious
when
like
if
it
was
gonna,
be
in
the
coming
sdk
release,
then
there
wouldn't
really
be
any
point
in
us
merging
this
now,
but
I
wasn't
sure
if
it
was
if
this
support
was
gonna,
be
in
the
next
release.
B
C
B
A
B
A
I
was
gonna
say
the
first,
no,
the
second
friday
of
the
month.
No,
I
don't
know
the
first
full
week
of
the
month
like
this.
This
week
started
in
february,
so
it
wasn't
the
first
full
anyway.
Basically,
you
could
make
it
you,
you
can
make
it
or
you
can
delay
it
if
you'd
like
to
delay
it
for
a
week,
it's
it
would
definitely
be
within
our
normal
release.
Yeah.
B
B
C
There's
nothing
I
mean
there's
nothing
urgent,
that
I
know
of
I
mean
that
and
that
android
fix
is
the
only
thing
that's
potentially
urgent,
but
I
don't
even
know
if
the
customers
confirmed
it
yet.
A
So
do
you
know
if
this
is
I
haven't
been
following:
if
the
duplicate
metric
identity
stuff
has
been
merged,
that
jack
was
working
on.
B
B
B
B
A
B
A
A
B
A
Cool
yeah,
maybe
I'll
I'll,
try
to
think
of
those
two
going
out
at
the
same
time,.
C
On
a
completely
different
topic,
I'm
gonna
be
out
of
the
country
from
the
16th
to
the
30th,
but
I
think
jack
will
be
back
by
then.
So
I
think
we
should
hopefully
have
maintain
your
stuff
covered.
C
A
Yeah
we
were
just
discussing
how
february
really
is
we
really
need
to
start
planning
vacate
like
southern
vacation
in
february
here
in
portland,
it's
like
by
that
point,
like
you're
just
done
with
the
gloom
and
dark
and
rain.
C
C
Well,
I
mean
my
younger
daughter
is
still
in
high
school,
but
she
already
is
super
grumpy
that
she's
gonna
miss
five
days
of
school
while
we're
gone.
So
next
year
is
like
sorry,
you
get
to
stay
home,
we're
going.
A
A
A
B
B
B
A
A
A
I
don't
think
we
need
no.
I
don't
think
we
need
to
revert
this
because
this
was
fixing
another
issue.
B
A
A
The
reason
why
I'm
suspicious
here
of
that
is
because
the
reason
why
it
that
duplicate
class
happens
and
why
we
was
because
when
something
is
loading
all
right,
so
if
we
go,
let's
look
at.
A
Is
because
when
this
is
loading
is
when
we
inject
this
and
we
inject
a
subclass
of
it
and
it
can't
reference
that,
and
so
there's
just
some
weird
ordering
there.
A
That
no,
no,
I
might,
I
might
have
it
so
if
your
instrumentation
in
adot
was
doing
this
and
then
were
you
injecting
a
subclass
of
this
and
that's
all
inside
of
the
transform
method.
A
And
so
you
injected
this
and
remember
that
transform's
not
not
recruit
like
nested
like
so.
You
load
this
inside
of
transform
because
you're
injecting
a
subclass
of
it,
and
then
that
triggers
this,
but
this
class
to
load
but
we're
inside
of
a
transform.
So
our
instrumentation
doesn't
get
applied.
Then
doesn't.
B
Is
there
a
way
to
make
it
more
safe?
Because
I
mean
someone
could
do
it
right
like
if
their
custom
instrumentation
did
this
like
injected
a
subclass
or
something
then
instrumentation
like
completely
unrelated
stuff
gets
disabled?
Basically,
so
if
there's
a
way
to
have
some
safety
there,
that
would
be
nice.
C
A
Work
but
at
least
it
feels
like
if
we
could
jeez,
I
don't
even
know
how
we
would
like
capture
that
a
lot
like
I
would
at
least
we
could
log
that
something
bad
yeah
I'll
I'll
comment
on
your
issue
and
yeah
we'll.
I
agree
that
you
know.
B
B
A
Yeah
I'm
worried
about
this
state
also
because
I'm
worried
about
yeah.
This
kind
of
also,
since
we
haven't
clearly
haven't
completely
solved
that
and
and
because
it
doesn't
repro
yeah,
we'll
see
if
they
come
back
with
the
repro
but
yeah.
I
I'll
probably
revert
this
change
before.
A
B
C
Yeah
I'm
going
to
I'm
going
to
drop
off
unless
you
have.
B
B
C
B
And
that's
that's
the
straw
proposal,
so
you
are
willing
to
maintain
micrometer
in
the
curriculum
yeah.
I
mean
I'm
happy.
B
C
What
what
package
is
that
being
published
out
of
right
now
and
are
we
gonna
have
to
change
it.
C
A
C
Right
well,
have
a
good
week
see
you
later.
A
B
A
Normally,
it's
always
been
harmless.
B
A
Ones,
I
don't
think.
B
I
mean
I'm
quite
used
to
ignoring
it
for
random
classes.
It
was
most
surprising
because
it
was
referencing
our
class
right,
yeah.
B
A
Yeah
find
out,
and
then
also
if
sometimes
the
ordering
of
the
two
java
agents
makes
a
difference.
A
That's
what
the
open
telemetry
collector
is
for,
if
you
want
to
send.
B
B
If
only
it
was
open
source,
then
I
could
just
search
the
log
message
and
find
exactly
what
java
whatever
is
creating
these
classes,
but
I
don't
think
it
is
customers,
like
probably
this
app
dynamics,
I
think,
like
app
dynamics.
Blog
shows
up
when
I
search
some
things,
but
it's
not
clear
if
it's
actually.
B
A
A
A
Yeah
yeah,
we
had
a
discussion
about
that
so
yeah.
I
think
we
just
agreed
what
that
we
won't
officially
support
it,
but
we'll
accept
pr's.
A
I
think
we
could
probably
change
our
position
on
that
in
a
couple
years
once
open
to
limit
everybody
like
this
is
just
standard
to
go
through
the
collector.
If
you
want
to
report
to
multiple
places.
B
B
A
Yeah,
just
because
we
just
introduced
support
for
recently
so
yeah
yeah.
B
A
B
Oh
man,
look
at
your
pr,
I'm
still
much
more
okay
with
java
tests
like
it's
to
mention
that
that's.
B
A
B
B
A
B
A
B
B
A
Because
I
don't
I
mean,
I
guess
we're
already
in
a
state
of
we
have
some
in
both,
but
I
would
feel
better
about
being
like
I'd
like
them
to
all
be
one
or
the
other,
and
I
think
we
agreed
to
start
doing
java
ones,
because
eventually
we
want
them
all
to
be.
B
Java
yeah,
I
I
think,
given
that
even
nikita
is
sort
of
okay
with
that
philosophy,
and
I
don't
know
if
I
would
say
it's
a
priority
to
rewrite
all
the
current
tests
in
java,
like
we
could
say,
we
want
them
to
be
in
java,
but
on
the
flip
side,
if
we
don't
stop
ex
accepting
groovy
tests,
then
that
becomes
impossible
because
we
just
keep
on
getting
more
and
more
groupie
tests.
So
it's
actually
more
important.
I
think,
on
the
new
side,
once
we're
ready
for
that.