►
From YouTube: 2020-07-17 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
A
A
B
B
Also,
I'm
I'm
jealous
that
you
can
that
the.
A
B
A
A
We
just
got
our
sort
of
like
I've,
been
ignoring
it
completely,
but
supposedly
you're
supposed
to
file
your
prs
with
open
source
team
to
get
them
approved
or
something
until
we
get
some
special
legal
approval.
So
we
just
got
it.
It
seems
to
still.
B
A
B
A
B
B
Yeah
yeah
they've,
embraced
I
mean
you
know,
I
mean
if
you
look
at
like
you
know,
10
years
ago
or.
A
B
Years
ago,
to
now
it's
like
different
company,
yeah
yeah,
but
by
the
time
I
joined
I
mean
so,
there
was
already
people
working
within
open.
Telemetry
like
sergey
kensalov
is
one
of
the
tc
technical
committee
members,
so
that
was
sort
of
in
place
before
I
started,
so
that
might
have
also
helped
me
smooth
my
onboarding
cool.
A
B
What
do
you
yeah
so.
A
A
A
B
A
B
You
lose
your
cycles
to.
B
B
It's
definitely
something
that
you
know
we
could
chat
with
like
the
open,
telemetry
governance
committee
or,
like
you
know,
just
sort
of
raise
it
up
a
level
of
people
who
know
more
about
how
to
facilitate
those
discussions.
Companies
yeah.
C
A
B
Cool
so
yeah,
so
this
morning
we
made
a
big
decision
where
we
got
to
move
off
of
java
7..
I
know
you
saw
that
and
yeah.
That
was
a
good
point
about
android
and
I
did
reach
out
this
afternoon
to
morgan
at
google
to
ask
him
if
there
was
somebody
in
open
telemetry
from
google,
who
could
sort
of
represent
the
android
perspective
and
pointed
him
at
that,
the
new
issue
to
and
if
we
could
chat
with
them
or
something
just
it
feels
like.
B
Since
we
have
google
in
the
yep
in
the
room,
we
should,
I
assume
they
have
well,
and
we
know
that
they
have
interest,
because
that's
the
whole
reason
why
the
api
is
stuck
at
an
older.
A
Yeah,
I
remember
one
issue
we
had
not.
We
like
there
was
one
google
cloud
library
and
they
made
a
weird
change
to
help
with
android
like
it
wasn't
really
the
right
change.
I
would
say
it
was
something
to
do
with
proguard
configuration
or
something
and
it
was
affecting.
The
non-android
builds
in
a
bad
way,
and
I
posted
to
the
issue
like
shouldn't.
You
talk
to
the
android
team
for
the
best
practice
here,
but
it
seems
like
it's
hard
to
have
those
feedback
loops,
even
when
you're
google,
so
I
don't
yeah,
it's
a
big
company
people.
B
Hey
is
that
frank?
B
Hi
frank:
did
you
wanna?
Did
you
have
some
questions
that
you
wanted
to
chat
about
or
you
just
wanna
hang
it
do
you
just
wanna
hang
with
us.
We
if
there's
nothing
else,
that
people
wanna
chat
about
anarag
and
I
kind
of
go
over
because
honorary
is
in
japan
and
the
morning
time
zone
doesn't
work
out
for
him.
So
we
kind
of
just
chat
about
things
that
were
going
on,
but
love
love
for
more
people
to
join
and
bring
questions
to
the
table.
C
Yeah
I
wanted
to
ask
about
debugging
how
to
do
it.
So
I
looked
at
the
issue
you
sent.
I
built
the
project
with
that
gradle
property.
C
B
Yeah,
so
when
you're,
so
that
part,
unfortunately
we
don't
have
so
the
advice
here.
Let
me
just
share
my
whole
screen
here,
so
that
we
can
pull
up
intellij,
which
advice
are
you
trying
to
debug
currently.
C
B
B
Awesome,
so
this
is
the
one
right.
Yes,
okay,
so
the
way
that
bite
buddy
works
well,
not
the
way
that
it,
but
the
default.
B
We
are
defaulting
to
we're
using
the
inline
so
for
bite,
buddy
advice.
You
can
also
do
inline
false.
I
have
not
tried
it
with
this
project,
though,
but
the
the
implication
of
the
default
inline
true
means
that
this
is
not
this.
The
body
of
this
method
gets
inlined
directly
into
the
method
that
it's
advising.
C
C
B
I
don't
have
a
solution
to
that
other
than
system
out
print
lines.
Here
I
don't
know
honor
do
you
have
any.
C
A
A
A
C
Yeah
yeah,
I
I
tried
the
print
line,
but
that
it
looks
like
that
whole
method
just
doesn't
run,
and
I
can't
really
see
why.
C
I
don't
think
it
is
yeah
and
it
happens
when
I
add
a
new
class
that
I
defined
in
that
same
package
and
if
I
try
and
instantiate
it
it
blows
up.
Rather
it
just
doesn't
do
anything.
So
I'm
not
sure.
B
Yeah,
so
what
you
are
probably
hitting
and
we
can
look
at
the
oh,
let
me
find
this:
have
you
tried
running
with
this
setting.
C
I
was
looking
for
that.
No,
I
have
not.
B
B
Okay
cool,
so,
if
you
run
with
this,
it
will
give
you
a
bunch
of
information,
and
let
me
just
cover
a
couple
of
things
that
are
confusing
and
possibly
happening.
B
Okay,
did
you
add
that
class
to
the
helper
class
names
here.
B
Okay,
so
what's
happening
the
purpose
of
this
well,
ideally,
we
would
be
able
to
auto
detect
that
we
needed
to
inject
those
these
classes,
but
we're
not
that
smart.
B
Yet
so
what
happens
is
so
the
servlet
class
itself
that
we're
instrumenting
is
in
you
know
some
class
loader
in
some
tomcat
class
loader,
for
example,
so
this
instrumentation
needs
to
live
in
that
same
class
loader,
because
we
need
to
be
able
to
reference
http
servlet
directly
in
our
instrumentation
and
so
all
of
the
classes.
Here
the
instrumentation
engine
will
take
those
and
inject
them
into
the
proper
class
loader
where
the
servlet
exists
or
well.
I
guess
actually
the
where
this
http
servlet
this
class
exists.
B
Yeah
and
that
without
that,
the
if
you're
are
you,
where
are
you
calling
that
class
from.
C
B
Oh
right
right,
okay,
so
yeah
that
I
think
that
may
take
care
of
it
for
you,
it
may
just
be
adding
the
helper
class
here.
B
The
other
thing
that
bites
all
of
us
more
often
than
or
from
time
to
time
is
this
thing
called
muzzle,
which
will
suppress
the
instrumentation
from
even
getting
applied,
which
actually
it
may
have
muzzled.
B
You
because
you
didn't
have
it
in
the
helper
class
names,
but
try
that
and
yeah
and
try
that
setting,
and
at
least
if
you
have,
if
you
still
have
issues
if
you
post
that
full
just
attach
that
full
log
file
from
this.
B
We
can
look
through
that,
and
I
can
look
through
that
and
get
a
maybe
a,
hopefully,
a
sense
of
where
it's
going
causing
problems.
Okay,.
B
It
sees
okay,
you're,
applying
this
advice,
and
it
looks
through
this
advice
and
basically
creates,
looks
at
all
the
classes
that
you're
referencing
from
this
advice
so,
for
example,
you're
we're
referencing,
http
servlet
response
and
say
that
we,
you
know,
we've
got
our
response
object
here
and
we
call
or
request
object,
and
we
call
get.
B
It
would
kind
of
look
muzzle,
would
look
through
all
of
these
references
and
kind
of
create
this
list
of
all
the
methods
that
need
to
be
accessible,
because
what
it's
going
to
do
right
is
we're
going
to
inject
this
code
directly
into
your
into
the
servlet,
and
so,
if
this
method
doesn't
exist
on
the
request
object,
that's
going
to
cause
your
that's
going
to
cause
it
to
to
blow
up,
and
so,
if,
if
it
finds
that
this
method
doesn't
exist
in
your
class
loader.
A
B
Http
request
object,
class
that
is
loaded.
Then
it
won't
apply
this
instrumentation
at
all.
This
whole
instrumentation
will
be
muzzled.
C
Okay,
I
think
that's
exactly
what
happened
makes
sense,
because
I
I
also
imported
the
servlet
wrapper
from
java
x,
and
I
think
I
need
to
add
that
in
there.
B
Too,
so
you,
if
it's
from
java
x,
you
will
you
don't
need
to
add
it.
So
we
are
smart
enough
to
build
up
this
reference
list
of
things
that
muzzle
needs
to
check
at
runtime.
C
B
Or
that
I
guess
no,
I
didn't
say
that
right,
helper
classes
are
classes
that
we
need
to
inject
into
the
class
loader
versus
the
java
x
classes
are
already
there.
C
I
can
give
this
a
shot
in
the
background,
while
you
guys
continue
to
discuss.
B
Yeah,
let's
see
what
else
did
we
chat
about?
That
was
interesting,
so
required
for
ga
I'm
gonna
go
through,
so
we
decided
that
the
specification
label
I
mean
the
the
things
that
we
mark
as
back
are
were
probably
required
for
ga,
but
there's
probably
other
things.
B
Obviously
bugs
some
way
for
vendors
to
create
distributions
ted
was
on
the
call
and
mentioned
he's
been
trying
to
sort
of
make
this
a
term
make
this
a
an
official
like
open,
telemetry
distros,
whether
it's
a
collector
distro
or
the
the
java
agent
distro.
B
So
he
liked
that
we
were
thinking
about
that
and
then
there
are
the
maintainers
group,
ted
and
or
some
some
people
are
working
on
a
list
of
like
requirements
around
documentation.
B
B
Probably
a
lot
of
these
things
will
just
apply.
I
would
guess
to
the
sdk,
but
some
things
like
perf
testing
for
sure
basic
docs
apply
to
us,
so
I'm
gonna
go
through
and
kind
of
try
to
take
a
first
shot
at
tagging,
things
that
are
required
for
ga
trying
to
just
a
very
minimal
sub
set,
though
there's
like
a
ton
of
things
that
I
would
be
very
sad
if
we
don't
have
by
ga,
but
I
don't
want
to
mark
those
as
required.
B
You
probably
saw
I
started,
I
don't
know
if
you
saw,
I
started
tagging
some
things.
I
created
a
new
label
called
discussion
because
I,
like
I
really
and
thanks.
It's
been
very
helpful
having
you
in
now
like
having,
I
feel
like
it,
takes
three
people
to
have
a
discussion.
A
B
So
now,
with
you
and
nikita
and
myself
talking
through
all
these
issues,
it
feels
like
you
know,
like
we're,
making
good
progress
just
in
the
issues
on
the
repo
and
whereas
before
I
would
have
to
bring
a
lot
more
of
these
issues
to
the
weekly
sig
meeting,
just
to
get
people's
attention
and
opinions.
B
So
I'm
thinking
with
the
discussions
of
I
mean
like
if
it
looks
like
it's
just
a
discussion
label
at
that
and
then,
when
once
we
resolve
that,
then
I
kind
of
like
the
idea
of
creating
a
new
issue
with
the
actual
action
items
from
that,
as
opposed
to
just
a
forever
long
issue.
B
That
I
mean
at
least
we
can
track
back.
If
somebody
wants
to
go
back
and
see
why
we
decided
xyz,
but
those
long
discussions
get
long
and
it
a
lot
of
not
really
that
relevant.
Once
we
start
once
we
make
a
decision
and
start
working
on
the
implementation.
B
Oh
yeah,
we
brought
up,
we
discussed
also
your
yeah,
so
we
all,
I
know
nikita
and
I
went
and
thumbs
up
your
proposal
so
yeah
and
nikita
is
going
to
open
a
spec
repo
issue
about
this
and
probably
again
knowing
him.
He'll
probably
include
this
also,
oh.
B
I
remember
that
I
asked
about
ask
people
what
they
thought
about
the
proposal
to
just
auto,
install
based
on
oh
yeah,
based
on
these
environment
variables,
and
the
concern
was
like,
if
you
have
in
your
environment,
you
have
a
bunch
of
these
set
that
it
could
be
a
little
bit
mysterious
and
I
mean
certainly
you
could
unset
them
for
a
particular
process
and
I'm
not
really
sure
that's
a
use
case.
A
B
Yeah-
and
so
it's
worth
I
mean
bringing
that
up
in
this.
B
A
A
B
B
B
B
B
We
put
a
marker
on
the
the
span
that
we
report.
That
says
that
this
is,
you
know,
multiply
this
by
10
to
get
to
generate
metrics
on
the
server
side,
but
we
would
really
like
to
be
able
to
aggregate
metrics
on
the
client
side
that
don't
aren't
affected
by
sampling
that
are
more
accurate.
B
A
A
B
Well,
in
theory,
we
could
have
you
know
the
like
a
span.
In
that
case,
the
no
op
span
would
be
just
like
yeah,
something
that
was
tied
into
the
timer.
A
A
B
B
Timer
api
and
he
was
showing
creating
the
span
and
creating
the
timer.
It
was
just
a
little
a
little
rocky
because
you
have
to
capture
the
start
time.
B
First
I
mean
you,
could
you
he
was
capturing
it
in
both
places,
but
it
could
have
used.
The
span
builder
set
start
at.
A
A
Timing,
clock
yeah,
so
we
do
something
similar,
so
our
miria,
we
use
the
libraries
times
so
that
allows
metrics
and
spans
to
be
synced
up.
I
expect
most
good
libraries
would
have
something
like
that,
but
actually
yeah
I
don't
know
I
should
look
at
grpc.
That's
a
good
question
whether
it
has
timings
built
into
the
library
itself.
A
B
A
A
A
B
C
B
A
A
A
Like
the
number
of
active
requests
in
the
server,
I
guess
that's
also
spams
dns
resolution
time
connection
acquisition
time.
These
would
need
to
be
events
on
the
span,
but
they
could
be
right.
A
A
B
A
A
A
A
B
Yeah
you're
not
you're,
not
so
yeah.
It's
unfortunate
from
a
performance
perspective.
A
B
Yeah
well
it'll
be
interesting
to
see
what
john
proposes,
because
so
far
he
was
not
generating
them
in
yeah.
C
A
A
B
Was
yeah
in
the
maintainers
meeting
on
monday,
he
was
giving.
He
gave
everybody
a
hard
time
about
the
spec.
He
was
like.
How
are
the
spec?
How
is
the
spec
ever
going
to
get
to
ga
yeah?
We've
got
80
open
issues.
B
Required
for
ga-
and
you
know
we
closed
like
a
couple
a
week,
so
it
was
good.
I
think
he
lit
a
fire
under
I've.
Seen
yeah
some
more
because-
and
I
think
it's
the
the
technical
committee
is
sort
of-
I
think,
owns
that
repo,
so
kind
of
need
more
for
them
to
be
more
involved
and
yeah
pushing
those
things
through
resolution.
A
B
B
B
A
B
There
yeah-
I
looked
at
it,
maybe
like
last
week
when
you
had
first
put
it
together.
B
A
B
B
A
B
We
have
yeah,
let's
take
a
quick
look
at
our
open,
pr's
yeah
I
went
on
the
issues
were
up
to
almost
a
hundred,
so
I
went
on
a
issue
closing
spree
last
night
and
yep
got
it
all
the
way
down
to
90.
yeah.
B
A
A
A
C
A
A
A
B
A
B
B
Yeah
there
it
is
we're
using
it
for
just
like
the
release
by
release.
A
B
Just
kind
of
which
that's
kind
of
been
the
zero
six
zero.
Let's
see
it
was
like,
it
was
kind
of
helpful.
I
thought
yeah
just
to
pop
in
a
few
things
and
sort
of
have
some
semblance
of
target.
A
B
I
think
informally,
it's
best
for
to
have
a
to
not
merge
something
from
your
own,
or
it
should
have
at
least
an
approval
from
somebody
from
another
company.
A
B
A
B
No,
I,
like
I
don't
know
when
the
last
time
I
had
tried
like
I
still
thought
that
there
were
some
like
notepad
or
I
thought
there
were
some
things
where
I
had
had
bad
experiences.
A
B
The
crlfs,
but
I
tested
it
out
and
yeah
I'm
happy
with
it
and
yeah
that
will
that
will
save
us
some
headache
from
random
windows
contributors
accidentally
changing
the
lining.
A
So
speaking
of
windows
is
the
github
actions.
A
B
Yeah
yeah
yeah,
so
I
so
the
I
had
raised
the
so
my
motivation.
Well,
I
don't
know
I
don't
really
care
that
much.
I'm
super
happy
on
circle,
ci
minus
the
cost
issue,
but
I
had
raised
the
cost
issued
to
sarah
novotny,
who
is
at
microsoft
and
is
on
the
the
governance
committee.
B
She
was
sort
of
when
open,
telemetry
and
open
tracing.
I
mean
when
open
census
and
open
tracing
came
together.
The
governance
committee
was
like
half
of
people
from
each
project
and
then
she
was
added.
She
came
over
from
the
kubernetes
project
and
as
like
a
tiebreaker
or
somebody,
so
she's
not
super
involved
in
the
project,
but
she
was.
B
She
raised
it
at
the
governance
committee
meeting
today
as
far
as
pulling
moving
the
circle
ci
cost
over
to
cncf,
so
I'm
waiting
fingers
crossed
because
if
we
can
get
that
moved
over,
I'm
like
I
don't
really
care
at
that
point.
If
somebody
wants
to
play
around
with
it,
it.
B
A
Like
the
one
thing
is
that
I
found
it
very
easy
and
nice
to
set
up
the
windows
and
mac
build
on
github
actions,
and
maybe
circle
ci
supports
it.
Nowadays
I
haven't
checked
their
latest
state,
but
that's
my
favorite
feature
of
actions
and
even
if
we
don't
migrate
the
primary
build
to
it,
having
one
windows
and
mac
build,
I
don't
know.
Maybe
that
doesn't
make
sense
if
it
just
takes
too
long
to
run
though,
but
that
was
one
like.
A
A
B
A
B
A
A
B
Okay,
I
would
guess
it
seems
like
the.
We
will
eventually
be
on
github.
Actually,
I'm
just
not
personally,
I
don't
personally
care
enough
to.
I
was
kind
of
curious
to
try
it
out
to
see
if
it
would
what
it
was
like
and
if
it
would
resolve
my
cost
circle.
A
B
B
Internally,
we
prime
we
use
a
lot
of
the
azure
devops.
A
A
A
B
Cool,
well,
I'm
glad
that
you
joined
that
we
were
able
to
help
with
that
yeah
yeah
and
yeah
yeah
keep
reaching
out
over
the
github
issues
and
welcome
to
join
what
what
time
zone
are
you
in.
C
I
am
in
pacific,
california,.
B
Okay,
yeah,
if
you
ever
want
to
join
in
the
the
the
thursday
morning,
meetings
are
at
nine,
it's
not.
It
tends
to
be
more.
I
mean
not
like
this
meeting
is
a
great
place
to
come
and
you
know
troubleshoot
stuff,
but
welcome
to
join
the
morning.
Meeting
and
yeah.
Just
issues
are
awesome.
C
Yeah
yeah
sounds
good.
I
I
I
did
make
another
one
well
yesterday,
I
don't
know
if
you
saw
it
about
concurring.
B
This
looks
like
something
that
would
be
good
to.
I
bet
we
can
pop
this
into
a
into
a
test,
one
of
our
tests
and
see
what
happens
and
tdd
it.
C
Yeah
it's
a
little
bit
non-deterministic,
but
they
should
all
be
under
the
parent.
It's
pretty
likely
that
one
of
them
won't
be.
A
B
B
Yeah,
I've
never
tried
it
with
with
streams,
but
this
still.
C
B
C
C
B
B
C
If
you
look
on
the
the
white
list
in
the
abstract,
execute
your
instrumentation
fortune,
executor
is
in
there.
C
A
C
B
B
B
Cool
yeah
yeah
thanks
for
code
and
repros,
are
super
helpful.
B
Cool
anything
else.
Well,
we
were
looking
at
the
prs.
Was
there
any
other
prs
to
that?
Okay,
so
this
is
the
docs.
B
Yeah,
let's
just
see
I'll,
take
a
peek
at
this
and
yeah,
hopefully
we'll
get
nikita's
blessing
and
we
can
just
start
trying
it
out.
Yep.
B
Yeah
yeah
yeah
it'll
be
good
to
have
a
test
there
to
make
sure
that
we
don't
break
the
java
7
support
there
when.
A
A
B
Yeah
and
that
will
that
will
motivate
us
to
split
out
the
that
auto
bootstrap
yeah,
that
too
the
massive
auto
bootstrap
yeah.