►
From YouTube: 2021-09-09 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
Stuart
is
going
to
be
taking
over
for
me
as
kind
of
the
data
dog
representative
here.
A
Awesome
well
and,
and
not
awesome
sad
said
to
see
you
are
you
yeah,
leaving
the
tracy.
B
A
Congrats,
thanks
for
all
of
your
work
over
the
years
on
the
datadog
code
base
that
we
are
benefiting
from
in
this
group.
Thanks.
A
A
There
you
go,
that's
how
you
do
it
like:
how
do
we
parallelize
this
cool?
Well,
let's
start
with
like
sdk
question.
F
Who
added
was
this?
This
was
this
was
added
by
someone
who's
not
here,
and
he
asked
me
to
represent
the
needs.
Basically,
the
splunk
would
really
like
to
make
sure
that
the
logging
sdk
gets
implemented,
asap
and
would
like
to
know
how
antoine
would
like
to
know
how
he
can
help
make
it
move
forward
more
quickly.
F
E
Yeah,
I
actually,
fortunately,
we
had
a
few
jewish
holidays
here
in
israel,
so
I
didn't
have
that
much
time
to
look
at
it,
because
we
have
some
duplication
in
parts
of
the
I
created.
I
now
created
a
small
pr
containing
only
the
log
adapter
implementation,
and
I
just
went
over
quickly
and
pretty
briefly,
but
I
went
over
the
pr
and
I
saw
that
the
actual
real
I
responded
there
on
this
pr
and
q
linked
so,
and
you
should
be.
E
This
is
my
pr
only
containing
the
logo
adapter,
and
my
response
was
that
the
actual
real
point
of
duplication
is
the
load,
adapter
class
and
so,
and
also
I
created
a
pretty
small
and
short
pr.
That
only
does
that
and
adds
resource
and
instrumentation
library
info
for
the
log
record
file
so
and
the
build
work
locally.
So
it
seems
like
it
can
also
even
solve
some
of
the
failed
tests
in
the
in
a
wheels.
Pr
and-
and
my
suggestion
is-
if
it's
okay
with
him-
this
is
the
only
part
of
duplication.
E
Actually
so,
like
the
the
other
parts,
he
did,
the
log
exporter
that
they
implemented
was
grpc,
which
we
did
not
implement.
Yet.
So
it's
it's
great
that
he
already
did
that
and
and
if
he
is
okay
and
you're.
Okay
with
merging
the
short
pr
that
already
to
to
my
best
knowledge,
goes
through
the
tests
and
everything.
E
But
basically
the
suggestion
if
we
merge
the
the
short
pr
that
I
created,
which
is
very
like
precise,
only
the
log
adapter
and
what
we
needed
to
to
have
that
will
can
continue
from
there
with
this
exporter,
and
it
will
actually
solve
a
few
of
his
a
few
of
his
breaking
tests
in
the
build.
F
Yeah
pr
is
kind
of
a
mess
it's
got.
It's
got
weird
stuff
in
it
that
I
think
I
might
my
advice
would
be
to
close
that
thing
and
start
over,
because
it
has
a
whole
bunch
of
strange
things
like
deleting
it
claims
that
the
context
root
method
has
been
deleted
and
there's
some
very
strange
stuff
in
there.
So.
E
You
mean
the
the
one
that
will
create
it,
yeah
yeah,
so
you
can
take
a
look
and
later
on
at
my
pr.
Is
it's
very
like
pretty
precise,
it's
only
the
lobe
adapter,
the
logo,
doctor
tests
and
the
log
record
enhancement
that
we
needed
to
add
the
resource
and
the
instrumentation
library
info,
and
maybe
some
really
small,
small
stuff
that
I
needed
to
fix
in
other
tests
to
support
the
add-on
of
this
resource
and
instrumentation
library
info.
But
it's
very
it's
very
short.
E
So
if
you
want
to
take
a
look
at
that-
and
I
think
it
will
also
fix
a
few
of
the
breaking
tests
in
wheels
pr,
because
I
saw
that
some
of
the
stuff
are
missing
there,
but
I
did
see
that
he
implemented
the
grpc
log
exporter,
which
is
great
because
then
we
we
can
implement
other
stuff
for
us
like
different
types
of
exporters
yeah,
I
will
try
to
get
a
chance
to
take
a
look
at
today.
E
F
It
was
less
working
on
it
and
more
desperately
wants
it,
and
that
is
antoine.
Pullman.
D
G
F
I
don't
think
this
is
something
we're
going
to
be
releasing
this
week
in
one
dot,
whatever
we're
on
I've,
totally
forgotten
six
one:
six
yeah,
that's
six,
I'm
getting
confused
between
like
the
splunk
version
and
the
android,
my
android
library
version,
and
this
version
I'm
like
anyway,
oh
yeah,
one,
it's
not
gonna
happen
for
one
six,
but
this
is
hopefully
we
can
target
this
at
the
end
or
at
the
beginning,.
F
E
Yeah
cool
and
also,
we
also
started
to
implement
the
like
a
very,
very
simple
kind
of
like
sort
of
a
poc
for
the
exporter
which
only
just
logs
it
after
the
the
adaptation
like
you
suggested
back
then,
when
we
talked,
we
do
something
very
simple
to
see
in
the
logs
that
the
actual
like
adoption
to
the
resource
logs
that
works
correctly,
and
we
will
create
a
new
pr
for
that
as
well
and
and
and
again,
we
try
for
now.
A
All
right,
jack
stale
span.
F
Yeah
jack,
I
actually
thought
I
had
merged
this,
so
apologies.
I
will
connect
with
honorag
and
make
sure
that
he
is
okay,
since
he
didn't
comment
again
after
you
responded
to
his
stuff,
but
I'll
talk
to
him
this
evening
and
hopefully
get
this
merch
tonight.
F
F
G
A
A
All
right
instrumenter
api
progress,
so
we
made
a
little
bit
of
progress.
We
emerged
got
a
bunch
of
jacks
pr's
from
last
week,
reviewed
and
merged,
and
this
one
is
lori
on
the
phone
here.
A
A
Is
here
hey
so
this
one?
Am
I
right
in
reading
that
this
one
is
going
to
close
10
of
10
10
of
our
to-do
items.
C
F
A
And
iv
vines
you
start
pulling
and
the
iv
vines
just
keep
coming
and
coming.
A
Now
I'm
gonna
have
that
green
day,
no,
not
green
day.
F
A
All
right,
yeah
release
release
week.
F
Yeah,
so
if
there
are
things
that
people
need,
I
put
I
put
jack's
pr
in
there.
That's
the
one
we
already
looked
at.
If
there
are
things
that
people
feel
like
need
to
go
in
this
release
and
can't
wait
a
month,
please
let
me
know
asap
so
and
I
can
discuss
tonight.
F
F
So
just
heads
up
is
a
little
bit
of
a
spicy
release
because
it
does
completely
rework
all
of
our
grpc
otlp.
Sorry,
not
grpc.
In
general,
the
otlp
protobuf
binary
manual
binary,
optimization
protobug
yeah.
E
Where
would
that
changes
be
like
where,
where
exactly
would
that
take
effect
like
what?
What
places
in
the
project
should
that
affect
every
place?
We
use
a
protobuf
object
or
specific
places.
F
And
yeah
there's
a
whole
bunch
of
very
carefully
handcrafted
crafted
manual
by
byte
writing
code.
That
honorable
has
put
together.
E
F
F
New
exporter
code
or
the
marshalling
code
doesn't
use
the
protobuf
library
anymore,
so
that's
yeah.
In
fact
it
does
some.
It
does
some
interesting
tricks
with
taking
the
proto
definitions
and
generating
different
classes
based
on
them
to
enable
some
stuffing
out
so,
but
we
still
have
the
existing
protobuf
module
like
the
optometry
proto
module
is
unchanged
and
will
still
be
there
as
a
dependency.
You
can
use
cool
cool,
great.
J
Yeah,
the
exemplar
pr
does
nothing
right
now.
It's
it
it
wires
in
the
do
nothing
example,
our
support
the
I
if,
if
we
want
exemplars
in
this
release,
if
you
give
me
five
days,
I
can
get
you
exemplars,
but
it
it's
not
going
to
hit
this
week.
Yeah,
let's,
let's
hold
up,
I
don't
want
to.
I
don't
want
to
rush
it.
J
Yeah
yeah
I
I'd
prefer
not
to
either
just
given
given
the
fun
that
is
exemplars.
J
A
Yeah,
there's
some
there's
a
few
nice
optimizations
in
one
six,
I'm
looking
forward
to
pulling
that
in
and
re-running
some
benchmarks
of
the
java
agent
yeah.
I.
K
F
Couple
other
things
right:
there.
F
Receiver,
that
has
been
owned
by
ryan
fitzpatrick,
mostly
oh,
this
is
in.
J
Collector
control
yeah:
this
is
collector
contrib,
so
yeah.
Basically,
I
looked
at
collector,
contrib
jmx
receiver
and
it
was
using
the
old
1.0
release.
It
had
never
been
updated,
so
I'm
making
a
fix
where
it
can
actually
test
collector
against
multiple
releases
of
our
jmx
receiver,
but
two
things
that
are
important
to
me.
Just
one
is:
if
we
make
a
release
that
someone
checks
to
make
sure
it
works
with
the
collector
contrib
repo.
J
I
don't
care
who-
and
the
second
thing
is
that
there's
a
communication
path
of
we
make
a
release
and
whoever
owns
that
thing
gets
notified
to
go
like
update
their
stuff.
J
F
G
J
So
if
we
have,
if
we
have
a
general
otlp
collection
thing,
that'd
be
good,
but
there's
like
this
weird
groovy
mapping
in
the
jmx
receiver
of
like
this,
you
know
make
this
metric
basic
make
this
metric,
and
so
just
an
integration
test
to
make
sure
that
what
you
say
on
the
one
side
turns
into
otop
on
the
other
side
that
the
collector
can
read
would
be
ideal.
F
Yeah
that
whole
java
java,
contrib
jmx
stuff
has
not
been
well
maintained.
It's
kind
of
was
dropped
at
1.0
and
then
left
to
fester
into
decay.
A
It's
good
to
hear
that
people
are
using
it,
though
I
think
that
was
one
of
our
questions
and
if
it
is
being
used,
then
certainly
that's
motivation
to
make
sure
that
it's
well
supported.
J
Yeah
well
at
a
minimum
if
there
are
bugs
send
them
my
way,
I
don't.
I
can't
guarantee
I'll
get
to
them
quickly,
but
we're
we're
looking
at
it
right
now.
J
Are
you
suggesting
it
should
be
rewritten
in
scala?
Scala
is
better
than
maybe.
J
F
J
Do
you
do
you
know
what
I
did
in
scala
to
counter
monorepos.
J
Quit
doing
scala
was
that
it
no
well
I
mean
I,
I
never.
I
never
wanted
to
quit
doing
scala,
but
I
did
find
a
job
that
paid
better.
What
what
I'll
say
is
I
made
a
build
system
which
composes
other
build
systems,
so
you
can
compose
together
microrepos,
so
it
builds
about
two
million
lines
of
scala
code.
It
builds
the
entire
open
source
ecosystem
every
night
against
the
latest
version
of
scala,
because
I
absolutely
hate
mono
repos,
but
we'll
talk
about
that
later.
J
If
we
need
to
they're
useful
but
they're,
not,
I
don't
know
if
they're
the
right
company
to
hate
monterey-
bros,
oh
I'm
at
the
right
company
to
because
you
get
to
complain
then
think
about
it.
J
Anyway,
yeah,
so
so
in
terms
of
jmx
yeah,
if
you
want
more
ownership
on
that,
let
me
know,
and
if
everybody
hates
groovy
and
doesn't
want
to
maintain
it
well,
everybody
can.
A
J
But
you
need
it,
you
need
a
scripting
language.
What
are
there
job?
What's
the
most
popular
javascripting
language
right
now,
written
in
column,
script,
kts.
J
Yeah,
but
the
only
difference
between
kotlin
and
groovy
is
they're,
sorry,
kotlin
and
scala.
They
just
took
away
features
like
that.
I.
H
J
J
A
J
Cool
so
it
sounds
like
I
should
try
to
contact
ryan
to
figure
out
what
the
plan
is
on
the
collector
code.
Can
I
add,
at
a
minimum,
I
can
update
the
collector
code,
so
it's
relatively
easy
to
do
this.
Can
we
add
a
thing
in
our
release
process
where
we
notify
the
collector
of
a
new
version
and
like
bump
their
integration
test
in
a
pr?
Is
that
reasonable
to
do
here?
J
I
can
put
it
in
the
readme
and
instructions
on
releasing
or
whatever.
J
J
That
one
that
one
I
already
tried
to
dive
into
and
yeah
that
solution
is
a
long
time
coming,
but
we'll
get
there.
I
mean
there
was
a
proto
release
this
week
right
or.
F
J
The
so
that
one
you
you
should
be
able
to
pull
in
that
version
of
the
proto.
Just
don't
use
the
new
feature
not
doing
that
until.
F
J
F
F
Not
directly
at
least
yeah
anyway,
I
don't
want.
I
have
just
I'm.
I
have
a
blanket
policy
right
now
for
for
java,
we're
not
touching
the
proto
until
the
collector
supports.
I
don't
want
to
get
into
that
dance
of.
Maybe
the
plucker
won't
work
with
this.
Maybe
it
won't
the
collector
claims
it
will,
but
then
it
doesn't
actually
etc.
J
F
J
The
simple
cases
beautiful
yeah,
if
you
make
sure
I'm
I'm
happy
to
help
there-
that
would
be
that'd,
be
awesome.
We
okay
cool
cool.
We
we
have
we're
working
on
integration
and
performance,
evaluation
tests
on
the
protocol,
library
itself
and
that's
taking
forever,
because
it's
a
huge
pain
in
the
butt,
but
all
that
code's
written
go.
So
I
assume
it's
just
not
useful
to
the
sig.
G
F
F
It's
called
stress
now,
that's
fast
test.
I
don't
know
anyway
remember
what
it's
called.
So
there
is
some
and
things
did
some
things
didn't
break
when
they
pulled
all
that
stuff
out
so,
but
they
broke
in
confusing
ways
they
weren't
clear.
What
was
going
on
so
some
better
specific
integration
tests
would
be
great
with
a
and
we
were,
I
think
we
may
still
be
even
pointing
at
the
latest
latest
tag
for
the
doctor
image,
which
I
think
is
a
terrible
idea.
We
should
pin
ourselves.
A
Java
instrumentation
repo,
all
of
the
we
have
a
bunch
of
tests
that
you
do
use
the
collector.
So
all
the
smoke
tests
go
through
the
collector,
and
I
was
thinking
about
that,
john
of
because
we
broke
yesterday.
We
got
broken
yesterday
again
by
another
collector
change,
but
on
the
plus
side
it
was
really
easy
for
me
to
go
to
the
collector
prs
and
see
which
prs
had
just
landed,
and
that
told
me
right
away:
oh,
they
dropped
the
old
legacy
port
and
so
that
made
it
easy.
So
there's
pros
and
cons
there.
G
Yeah,
that's
like
that's
tough.
Do
you
need
a
like
a
matrix,
a
support
matrix
for
different
versions
of
the
collector.
J
F
F
We
moved
ahead
before
the
collector
did
because
there
hasn't
been
a
good
way
to
know
when
you're
supposed
to
actually
move
okay,
so
two
cases
that
I'm
aware
of
one
of
them
was
attributes
and
labels,
and
there
was
a
another
one.
I
don't
remember
something
that
else.
Metric
related
we've
never
had
a
problem
as
far
as
I
know,
with
anything
traits
related
so.
J
Okay,
which
is
good
I,
I
suspect
this
latest
protocol
will
get
adopted
quickly,
because
it's
for
prometheus
compatibility
and
there's
dedicated
people
on
that,
as
opposed
to
labels
to
attributes
which
was
just
optimistic,
but
even
so
that
sounds
like
another
inverse
thing
where
we
need
the
collector
to
say
what
version
of
the
protocol
it's
supporting
at
any
point
in
time
and
advertise
that
to
all
of
us.
If
that's
not
happening
well,
that's
something
I
can
escalate
and
figure
out
how
to
make
that
happen
too.
Yeah
I
mean
collector.
E
I
actually
have
some
super
short
question
for
I
guess
for
john
or
for
someone
who
maintains
their
repo.
I
saw
that
for
the
first
pr,
there
needs
to
be
like
an
approval
from
a
maintainer
right
to
be
able
to
to
build
actually
yep
yeah
yeah.
F
E
I
F
A
Yeah
yeah
john
taught
us
all
about
this
amazing
github
feature
where
you
can
go
and
turn
off
your
email
notifications,
because
a
bunch
of
us
follow
a
bunch
of
repos
and
just
get
way
too
much
email
and
then
you
just
go
and
monitor
here
so
say
I
want
to
see
the
java
repo
and
then
I
can
go
through
and
kind
of
briefly
like
for
a
repo
that
I
sort
of
half
follow
I'll,
be
like.
Oh
I'm
not
really.
I
kind
of
understand
this,
I'm
not
to
follow
this.
A
I
say
okay
done
and
then
like,
oh
exemplars.
What
are
exemplars-
and
I
can
go
in
here-
read
about
it
done
here
once
I'm
done
and
then,
if
somebody
comments
again
it'll
pop
up
here
and
then
I
can
go
in
and
it
usually
will
cue
you
up
to
where
you
left
off,
it's
a
really
nice
it.
Actually
I
was
skeptical,
but
it
it
has
been
a
very
nice
workflow
and
the
vi
keys,
like
the
gmail
keys.
F
K
F
I
F
If
you
type
question
mark
it'll
pop
up
all
of
the
keyboard
shortcuts.
A
All
right,
y'all
well
happy
to
give
everybody
some
time
back
today.
Great
to
see
you
all.