►
From YouTube: 2021-06-11 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
C
A
A
C
A
D
A
D
A
C
C
A
C
A
The
main,
actually,
I
did
try
at
one
point,
because
I
you
know
it
was
like,
but
I
could
not
do
the
key
bindings
for
at
that
point
eclipse.
My
fingers
didn't
like
didn't
like
it.
D
A
So
yeah
you've
been
busy
with
changing
all
of
our
tests.
Yeah.
A
Not
acceptable,
I
figure
it
was
just
a
matter
of
time
before
you
are
miriamized
us.
A
Https
test,
I
remember
for
some
reason,
I
thought
we
had
a
gap
in
coverage
because
all
we
test
is
get
a
simple
get
on
the
hdp
clients.
D
D
D
C
A
Okay,
but
not
the,
but
I
mean
that
one
and
that
yeah
that's
true
relevant
yeah.
A
From
they
did,
they
did
have
tests
that
ran
on
ibm.
I
don't
know
if
they've,
if
they
still
are
probably.
D
A
A
Yeah,
oh
yeah,
I
was
thinking
about
my
this-
is
helping
my
goal
towards
maxing
out
our
the
open
telemetry
org
limit,
github
actions
limit.
D
D
D
D
D
A
It
makes
sense,
it
just
hasn't
corresponded
to
us.
Well,
I
guess
I
gotta
check
the
actions.
Yeah,
that's
a
good
point.
I
gotta
I
mean
if.
B
John,
quite
today,
oh
I'm
eating
eating.
I
didn't
think
you
needed
to
hear
me
chewing
no
wonder
they're
having
delicious
cuban
food.
D
C
B
B
Yeah
he's
all
he,
he
gets
a
little
tiny
bowl
of
breakfast
up
here
and
then
he
comes
throughout
the
day
and
the
the
bag
is
hilarious
because
he
did
he
chews
on
the
bag,
all
the
time,
hoping
that
somehow
he
will
extract
food
from
it.
I
I
don't
know
he's
not
so
smart
anyway
yeah,
so
I
just
wanted
to
chat
about
the
schema
url
and
whether
it's
a
it's
an
annoying
addition
to
the
api.
B
And
I
just
don't
know
like
if
it
feels
like
if
we
were
to
builderize
it
now
and
we
never
added
anything
again,
it
would
be
a
little.
The
builder
would
be
a
little
anemic.
D
I
mean
there's
just
that
threshold
three
parameters
seems
like
a
reasonable
threshold
for
builder.
I
guess,
but
I
guess
there's
only
two
parameters,
so
the
name
is
required.
Name
is
required
yeah,
so
you
would
create
a
builder
with
the
name
yeah
right
and
then
it's
just
two
setters
on
the
builder
yeah,
that's
pretty
anemic.
I
guess
I
think
next
time.
D
A
C
D
I
mean,
like,
I
would
say
the
I
guess,
it's
sort
of
gross
that
we
have
been
available
on
the
version
only
in
that
version
of
the
method
yeah.
At
that
point,
I
guess
we
would
want
to
add
it
to
the
other
versions
or
something
and
that
that's
the
biggest
problem
with
our
current
approach.
I
guess
right
now
it's
sort
of
inconsistent.
B
D
D
B
B
B
D
C
D
B
Might
make
things
linkier
it's
actually
interesting.
I
wonder
if
we
would
take
the
the
second
method.
That's
on
there,
so
the
the
the
one
that
takes
the
name
and
the
version
would
we
implement
that
I
guess
we'd
have
we'd
implement
that
with
a
builder
call
internally
like
oh,
my
question
is:
would
we
do
it
with
the
default
method
when
we
make
it
into
a
default
with
the
builder
call.
B
B
A
Would
you
deprecate,
would
you
deprecate
no
leave.
D
B
D
A
Trust
what
are
we
expecting?
I
didn't
know
I
was
well.
I
thought
I
looked
in
the
java
repo
and
I
thought
I
saw
some
places
where
the
factory
methods
were
about.
A
D
D
B
D
I
don't
know,
I
don't
think
you
can
do
the
utility
versus
non-utility,
though
you'd
be
stuck
with
that
construct
from
the
top
right
right,
like
a
text.
Spotless
no
check
style,
yes
and,
like
I
remember
looking
at
aws,
sdk
repo,
and
they
have
this
private
inspector
and
all
the
utility
classes,
and
I
really
don't
like
that,
but
it
does
mean
that
I'll
probably
send
vrs
that
accident.
They
have
the
wrong
method
order.
Still,
I
don't
think
the
method
error
matters
enough
for
us
to
care
so
much
about.
C
B
Speaking
of
which
I
did
reinstall,
the
google
java
format
plug-in,
and
it
does
work
much
much
better
now
than
it
used
to
than
it
did
like
a
year
ago
I
mean
the
big
issue:
was
import
ordering
didn't
work
right
and
now
it
looks
like
they
finally
got
that.
B
I
haven't
quite
gotten
bold
enough
to
check
the
box
an
idea
that
says
the
format
before
committing,
but
I
might
get
there
actually.
What
I
really
need
to
do
is
have
it
format
before
I
run
stuff
so
it'll,
so
that
error
prone
doesn't
yell
when
I
forget
to
remove
an
import
during
writing
tests.
B
D
A
I
do
it.
I've
had
to
exclude
there's
two
files
in
our
repo
that
don't
the
autofort,
the
google
format
messes
them
up,
because
we
use
some
c
style
comments.
So
I
I.
A
Oh,
maybe
you
could
maybe
you
fix
this.
I
feel
like
something
in
here
used
to
be.
Oh,
I
think
maybe
these
get
all
squished
or
something
weird
by
the
auto
form.
Yes,.
D
A
Stuff
yeah,
but
it
I
don't
know
it's
just
not
very
smart,
so.
B
D
C
A
Folks
from
their
java
agent
team-
and
so
they
are-
they
are-
sounds
like
they
are
going
in
on
the
commodity
agent
and
which
is
awesome,
which
is
exactly
what
we
want
everybody
to
do.
D
A
So
I
think
what
will
probably
happen
with
these
is
kind
of
these
companies
is
kind
of
the
path
that
new
relic
is
going
down
of,
supporting
either
through
the
collector
or
the
app
dynamics.
A
Folks
said
they
were
going
to
support
otlp
native
ingestion
to
their
back
end
and
then
start
once
the
our
the
open
telemetry
java
agent
has
feature
parity
with
whatever
their
customers
need,
then
they
can
start
having
their
customers
use.
That.
A
A
I
think
I
even
used
it
to
just
tried
it
out
once
like
five
years
ago.
I
thought
they
had
one
for
a
while.
Maybe
it
could
be.
D
A
I
think
maybe
you're
thinking
dynatrace.
A
I
feel
like
they
were
the
last,
because
I
feel
like
app
dynamics
has
had
a
cloud
solution
for
a
few
years.
B
A
If
I
remember
it
was
like
a
regular
cloud
thing,
but
it
may
not
have
been
like
a
multi-ten,
I
don't
know
if
it
was
a
fully
multi-tenant
solution
or
like
yeah,
you
click
in
their
website
and
they
deploy
it
in
the
cloud
for
you
and
they
they
manage.
So
it
is
fully
managed,
but
maybe
it's
not
fully
multi-tenant.
A
A
Yeah
I
watch
I've
been
watching
the
the
the
new
relic
repo
and
it
seems
like
they're,
also
kind
of
waiting
for
certain.
A
I
don't
know
certain
things
to
get
at
parity
in
the
open,
telemetry
java
agent,
although
at
the
same
time
so
we
did
discuss
this
morning
and
somebody
I
think,
jason
plum
brought
up
that
the,
and
this
will
be
interesting
to
see
how
if
this
becomes
a
stumbling
block
for
like
new
relic.
My
moving
customers
over
is
that
the
new
relic
is
supporting
dynamic,
attach
now.
A
And
but
I'm
not
really
sure
what
and
the
app
dynamics
folks
said
they
support
dynamic
attach
also,
but
they
weren't
really
sure
they
had.
You
know
what
the
customer
I
mean.
These
were
just
a
couple
of
engineers,
so
they
didn't
really.
They
didn't
seem
to
know
the
customer
use
case
why
somebody
would
want
that.
B
I
mean
dyna
trace
dynatrace.
Is
the
big
dynamic
attached
company
right,
like
they
they're
like
a
do?
You
have
to
do
anything
at
all,
we'll
control
your
apps
for
you
right
and
we'll
deploy
the
agent
into
your
code
and
make
it
running,
make
it
run
with
the
agent
without
you
having
to
do
anything
at
all
central
control.
I
believe.
B
A
Yeah
and
maybe
they
they
use
that,
as
a
I
mean
people
are
probably
probably
they've
been
using
that
as
a
competitive
gap
thing
and
that
creates
in
the
apm
vendor
market.
Everybody
else
has
to
add
that
feature,
even
if
there's
not
very
much
use
case
for
it.
A
They
were
asking
about
that
concept.
That
jen
was
talking
about
about
the
kind
of
control
central
control
they
wanted
to
do
like
we're
thinking
of
like
a
control
plane,
something
where
they
could
remotely
configure
the
the
java
agents.
C
So
it
sounds
like
that's
something
they
have
maybe
a
plugable.
B
A
B
I
mean
the
main,
the
main
thing
it's
used
for
a
new
relic.
At
least
it
was
so
they
so
they
had
a
way
to
do
like
xml
based
configuration
for
for
instrumentation,
like
we
were
talking
about
this
morning
yeah.
So
I
think
that
required
a
restart
of
the
app
to
apply,
but
it
would
go.
The
app
would
go
query
at
startup
and
grab
that
instrumentation
and
do
apply
it
at
startup.
B
But
the
bigger
thing
was:
if
you
wanted
to
do
more
deep
profiling
or
focus
more
like
more
sampling
on
a
specific
method
or
a
specific
class,
then
that's
the
kind
of
thing
that
could
happen.
The
control
plane
could
tell
the
agent
to
do
that.
A
Yeah,
I
wonder
if
we
can
sort
of
somehow
lump
that
in
with
the
dynamic
the
the
sampler
stuff,
because
I
know
that's
a
popular
use
case
for
dynamic
configuration.
A
D
It
works,
I'm
a
bit
scared
if
we
have
to
have
all
these
asian
features.
That
then
requires
to
not
use
our
library
instrumentation,
because
they're
just
so
invasive
or
something
that's
one,
scary
aspect.
I
guess
I'm
not
saying
anything
in
specific
but
like
especially
like
dynamically,
we
know
dynamic
attach
well,
it's
definitely
not
going
to
work
with
library
instrumentation
the
way
we
register
those,
and
so
if
that
comes
required,.
C
A
D
A
A
The
types
of
configuration
and
it
sounded
like
the
the
biggest
one,
was
what
john
was
describing
of
like
this
xml
based
or
you
know.
In
our
case,
like
just
of
saying,
I
want
to
capture
a
internal
span
on
this
method
and
capture
these
parameters.
Along
with
that,
and
so
you
know
something
json.xml
whatever
then,
and
those
are
very
easy
we
could
we
can
dynamic.
A
I
would
be
fine
with
dynamically
re-transforming
for
that.
You
know
those
kind
of
a
very
specific
subset
of.
A
Yeah
yeah,
I
was
demoing
the
glow
root.
Had
that
feature
you
could
in
the
ui.
Even
you
could
auto
complete
like
your
class
and
method
and
parameters
you
want
to
pull
in
and
it
would.
You
could
hit
a
button
and
it
would
re-transform
and
start
capturing.
D
A
Yeah,
you
haven't
seen
the
it's,
of
course,
that
there's
no
data,
because
something
is
busted,
of
course,
a
content.
This
is
just
a
demo
just
so
it
it
runs
in
two
modes.
You
can
run
it
locally
with
a
embedded
h2
database
or
you
can
run
it
with
a
central
collector,
which
is
a
cassandra
based
and
it's
pretty
cool
with
the
the
central
collector.
A
It
has
the
same
ui,
but
you
have
a
drop
down
here,
of
which
agent
you're
looking
at,
and
so
you
can
still
configure
stuff
on
the
fly.
You
can
capture
thread
dumps
and
heap
dumps
on
the
fly.
A
A
A
I
wish
I
could
show
it
like
actually
working
but
yeah.
I
have
a
gatling
script,
that's
supposed
to
send
traffic
to
it
to
generate
charts,
but
the
gatling
thing
dies
constantly
and
I
have
to
log
into.
I
have
to
log
into
ec2
to
restart
it.
A
Anybody,
oh
I
asked
about.
I
asked
her
a
question.
Nikita
had
new
thoughts
about
capturing
lower
level
things
via
nettie,
so
you
know
maybe
it's
something
that
we
could
have
opt-in,
yeah.
A
D
A
A
C
D
A
A
C
A
Let's
see
release,
we
talked
about
release,
we're
going
to
really
try
to
release
our
project
early
next
week.
We'll
get
nikita's.
A
Two
extension
prs,
I
think,
are
the
only
things
that
we
really
want
in
there
oh
yeah,
and
we
can
update.
B
And
honor,
I
had
to
give
you
a
hard
time,
because
the
pr
with
the
api
disk
was
already
open
and
ready
to
be
merged.
If
you
had
approved
it
and
merged
it,
I
put
it
into
the.
I
put
it
into
my
change
log
updates,
because
I
saw
that
it
was
there,
so
you
could
have
just
merged
that
and
it
would
have
just
been
done.
B
D
D
B
A
Oh
yeah
and
and
john
was
trying
to
claim
that
the
but
the
release
button
it
worked
perfectly
and
I
guess
it
did
but
because
I
always
say
I
don't
want
to
hit
the
release
this
time
on
our
repo
I'll
work
on
the
release,
I'll
try
and
get
the
release
note
in
by
monday.
But
somebody
else
gets
to
hit
the
release
button
when.
B
I,
when
I
did
the
release
this
week,
it
was
actually
incredibly
fast
like
I
was
surprised
like,
especially
since
the
last
release,
when
sona
type
with
oss
summon
type
was
down
like
it
was
like
a
six
hour
process
to
release,
and
this
was
like
10
minutes
and
done.
C
A
A
A
He
was
just
mentioning,
somebody
was
asking
him
about
scala
how
to-
and
he
was
you
know,
wasn't
sure
what
kind
of
scholar.
C
D
C
A
B
And
in
particular,
it's
more
not
about
like
how
the
agent
might
do
it,
but
how
we
can
design
an
api
in
the
sdk
to
support
this
working
like
we
can't
just
say:
hey
here's,
a
here's,
a
place
to
send
your
logs
and
or
here's
like
a
logging
pipeline,
because
somehow
we
need
to
be
able
to
support
getting
the
context
into
the
mdc
from
the
call
site
and
not
from
the
and
not
once
you're,
actually
doing
the
appender
work.
So
not
not
sure
how
that
problem
gets
solved
in
general
or
if
it
even.
A
B
I
assume
that
they
they
take
a
copy
of
the
mdc
and
send
it
along
to
the
across
the
thread
boundary.
So
so
we
just
need
to
make
sure
we
can
get
the
co
somehow
get
the
context
into
the
mdc
before
they
do.
D
B
Yeah,
so
this
is,
I
think
this
is
a
big
question
that
I
think
has
been
kind
of
glossed
over
in
tigrin's
otep
about
how
you
actually
should
get
the
context
into
the
into
the
appender
world,
which
is
almost
certainly
going
to
be
running
on
a
separate
thread
anyway.
It's
it's
a
problem
to
be
solved,
and
maybe
there
is
no
well.
I
mean
there
has
to
be
some
sort
of
solution
to
it,
like
you
said,
because
mdc's
are
generally
just
stirring,
string
right.
D
D
B
D
D
D
D
A
B
Avoid
that
we
might,
we
might
want
this
actually,
so
I
was
thinking
I
was
thinking.
We
can
almost
certainly
get
rid
of
that
check
all
together.
That
is
happening
that
the
one
that
nikita
optimized
and
you
worked
on,
I
don't
think
it
can
ever
fail
like
the
way.
That's
only
called
during
export
and
we
can't
possibly
have
gotten
bad
ids
in
there
not
indexed
yeah,
so
I
think
we
could
we
can,
or
we
could
at
least
provide
a
different
method
that
completely
buys
that
check
right.
I
think
we
can
move.
B
D
B
To
understand
first
right-
and
I
think
that
nothing
will
call
like-
I
don't
think,
there's
actually
even
any
callers
that
need
to
protect
the
the
safe
version
like
at
the
moment.
There's
almost
no
callers
that
thing
at
all.
It's
basically
only
called
that's
true
during
the
export
pipeline.
So,
okay,
I
think
we
can.
D
B
C
D
B
Anyway,
very
excited
to
see
what
the
guanodon
says
about
this
fix.
B
D
A
A
A
Cool
yeah,
no
you're
right.
We
don't
because.
D
A
Remember
was
that
in
the
main
extension
issue,
or
was
that
somewhere
else.
A
A
We
didn't
discuss
it
with
the
only
isolation
that
we
discussed
was.
The
that
is
implemented
is
the
isolation
of
different
extension
jars
from
each
other.
D
D
A
Very
oh,
inefficient,
yeah.
D
A
Yeah
yeah,
I
have
a
feeling
it
would
well
yeah,
because
there's
somebody
wanted
and
if
somebody's
using
nettie
shouldn't
they
end
with
their
own
exporter,
shouldn't
they
not
use
the
dash
all
and
not
pulling
our
exporters.
D
But
if
like
there
could-
and
we
know
some
guava-
I
think
it's
still
there
in
the
non-dashboard,
which
could
have
a
compatibility
problem.
I
think
in
general
we
do
want
to
make
sure
that
conflicts
can
happen
with
our
agent
cluster
yep.
I'm
sure
we
can
get
rid
of
guava,
so
that
would
of
course,
be
a
good
thing
to
do
anyways,
but
this
imprint
support
seems
nice
to
have
that
insulation.
D
A
B
B
A
B
D
B
A
global
reflector,
assuming
you
already
have
a
dependency
on
semantic
attributes
and
you
can
pull
the
url
out
of
there.
B
We
still
have,
I
I'm
been
holding
off
and
putting
the
schema
into
the
resource,
although
that
is
the
next
step.
Only
because
there's
some
there's
a
big
open
question
about
how
what
happens
if
you
merge
two
resources
that
have
different
a
different
schema
version
like
what's
the
outcome
like
what
shouldn't
happen,
I
don't
think
anyone
knows
yet
I
mean.
Actually,
I
think
it's
written
in
the
spec.
That
is
undefined
right
now,
it's
behavior
is
undefined.
If
you
try
to
merge
two
resources
with
different
schemas
anyway,
I
don't
have
it.
B
A
Speaking
of
resources,
this
was
something
else.
The
app
dynamics
folks
brought
up.
They
want
to
do
on
a
detection
of
container
info,
and
do
we
already
have
a
kubernetes
detector
resource
detection.
B
D
A
They
said
they
were
going
to.
C
A
B
D
I
think
long
term
I
mean
anything
in
the
collector
should
be
in
from
the
language
also
just
to
not
require
the
collector
it's
more
of
a
bandwidth
thing
of
who's
going
to
implement
that
we
have
no
motivation
really
because
we
expect
the
collector.
But
if
someone
that
doesn't
want
to
use
the
collector
wants
this
information,
they
can
implement
the
resistor.
B
It
knows
what
it
knows:
the
source
of
the
data
where
it's
flowing
from
and
can
look
that
data
up
in
kubernetes.
I
think
I
assume
I'm
not
a
kubernetes
expert,
but
I
assume
that's
how
it
works.
D
B
Probably
was
it
something
you
can
call
on
is
this?
Is
it
something
that
you
can
call
on
your
local
host
to
get
that
data?
That's
why
I'm
just
like
wondering
how
I
mean
sure,
if
you,
if
it
knows
if
the
resource
sector
knows
how
to
talk
to
the
kubernetes
system,
then
sure,
but
I'm
not
sure
how
we
build
a
generic
resource
detector.
To
do
that,
I'm
not
too
sure
about
the
detail.
I
don't
think.
A
A
For
azure,
vms
or
some
services,
there
is
something
like
that,
like
just
from
localhost,
you
can
hit
a
port
and.