►
From YouTube: 2020-08-21 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
Good
morning,
good
morning,
nikita,
while
we
wait
for
people
to
show
up,
don't
forget
to
put
your.
D
B
C
Cool
all
right:
well,
we
can
wait
another
minute
or
so
I
know
andrew.
I
put
a
link
in
the
agenda
to
our
current
p1
issues,
so
maybe
you
can
help
us
wrangle
what
to
do
with
some
of
those.
D
C
D
I
did
smell
smoke
on
the
weekend,
like
a
very,
very
strong,
actually
smelled
like
a
fireplace,
but
it's
not
too
bad
right
now,
but
I
think
the
air
quality
measurements.
So
I
can't
see
anything
obvious,
but
I
think
the
air
quality
in
the
middle
of
san
francisco
is
like
medium.
It's
like
it's
not
exactly
healthy.
D
Geez,
okay,
yeah
yeah,
I
heard
north
of
golden
gate
is
really
not
good
and
but
I
haven't
been
keeping
up
with
the
news
morning.
Everyone.
C
Hey
morgan
all
right:
well,
I
think
we
should
get
started.
We
don't
have
much
on
the
agenda
but
feel
free
to
add
things.
If
you
want
to
talk
about
stuff.
F
A
quick
update
about
the
salute
guy,
so
I'm
keep
talking
to
him
and
trying
to
bring
him
to
the
meetings.
Yes,
it's
not
gonna
be
thursday,
based
on
the
feedback
from
from
trask
that
we
cannot
hijack
that
meeting.
It
may
be
this
meeting
or
today,
like
next
friday,
at
9
a.m.
I'm
I'm
trying
to
negotiate
with
the
with
him
when
he
can
come.
C
Cool
sounds
good.
I
had
some.
There
was
some
communication
I
had
with
not
with
him
directly,
but
I
think
some
new
relic
folks
are
trying
to
set
up
a
conversation
with
him
as
well.
So
with
martin,
okay,
marcin,
I
don't
know
how
you
pronounce
his
name:
I'm
not
kidding
polish.
Yes,
I
don't
know
how
to
pronounce
polish
cool.
C
I
think
nikita.
Did
you
add
the
the
java
version
issue.
G
C
Yeah,
I
have
a
only
very,
very
tiny
update
my
meeting
with
trask
and
honorable
last
night.
Trask
is
going
to
reach
out
to
some
contacts.
He
has
around
grpc
a
conversation
that
he
was
already
involved
with,
so
I
think
he's
going
to
try
to
get
get
everyone
connected
together
to
chat
about
at
least
the
grpc
requirements.
B
F
I
think
I
think,
by
the
way,
trust
is
was
in
an
email
exchange
between
me,
morgan
and
eric
anderson,
the
techniques
from
from
java.
E
He
can,
he
can
think
eric
is
a
nice
person
guys.
E
What
he
was
he
was
going
to
reopen
that
those
emails
yeah,
so
we
already
set
up
the
connections
with
the
jrt
folks.
It's
just
either
to
follow
up
if
needed,
that's
exactly
what
what
traffic
is
going
to
do.
So
that's
the
plan.
C
I
did
my
work.
Thank
you.
Morgan
yeah,
so
trask
is
good.
Trask
is
running
with
that
for
right
now,
so
I
don't.
I
think,
until
that
conversation
happens,
there's
probably
not
much
more.
We
can
talk
about.
C
C
C
This
way,
I
think,
the
big
things
that
we
would
probably
want
to
do
as
soon
as
we
could
is
to
convert
a
bunch
of
the
api
classes,
the
interfaces
with
default
methods
and
some
static
builders
on
their
static
methods
to
for
creating
them,
so
make
the
api
more
interface-based
to
help
out
the
agent
auto
instrumentation.
F
Sense
that
makes
sense,
so
it's
more
like
the
default
methods
in
interface.
Have
we
consider
or
have
we
thought
about
times
so
the
times
will
be
also
something
that
we
think
about,
because
in
java
8
we
have
the
java
instant
yeah.
Don't
know,
don't
know
if
we
want
to
change
the
public
api
to
accept
java
instant
or
we
keep
nanoseconds
from
epoch.
F
That's
a
discussion,
but
definitely
we
need
to
change
the
way
how
we,
how
we
capture
the
timestamp
in
the
sdk,
because
java
instant,
if,
if
we
start
using
java,
instant
and
people
run
on
newer
than
java
9,
they
will
get
microsecond
granularity
instead
of
nanos
instead
of
millisecond
priority
by
default.
So
that
would
be
an
improvement
so
anyway,
what
I'm
trying
to
say
is
you
should
look
also
into
java
instant
support.
C
F
But
I
would,
I
would
put
all
these
things
if
you
have
a
issue
or
something,
I
would
just
record
all
these
ideas
into
that
to
not
forget
about
what
things
we
should
change
before.
If
we
agree
on
java
8.,
I
think
trucks
provide.
A
C
F
A
Yeah
for
us,
we
would
like
to
still
have
support
for
java,
seven,
at
least
on
the
api
level,
so
we
can
implement
our
own
sdk.
If
you
want
to
move
this
sdk
to
java
8,
it's
not
a
problem
for
us,
but
at
least
the
api.
So
we
can
be
compatible.
K
What
about
instrumentation?
Because
I
think
that
at
the
moment
we
always
talk
or
mostly
talk
between
the
sdk
and
api
thing.
But
what
about.
L
Instrumentation,
I
think,
on
the
instrumentation
side,
we
decided
that
we
are
going
to
drop
java
7
for
most
of
it,
but
still
have
java
7,
compatible
library
based
instrumentation,
interesting
or
particular
gradle
pro,
for
particular
builds
that
require
compatibility
with
android.
G
Library,
manual
library,
instrumentation
from
instrumentation
repo,
which
are
required
or
interested
for
android
projects,
will
still
be
well,
as
android
currently
have
can
have
java
8
language
features.
So
that's
that's
new
information,
so
probably
so
the
point
being
we
auto
instrument,
library
instrumentations,
which
are
interesting
for
android,
should
continue
to
work
on
android.
G
C
There's
another
question
that
I
think
hasn't
been
asked
directly,
and
that
is
how
much
actual
active
support
do
we
want
to
give
for
the
sdk
running
on
android,
because
we
definitely
have
not
been
focusing
on.
We
have
no
nobody's
like
testing
anything
on
android
or
running
benchmarks
or
android
or
even
testing
whether
our
current
implementation
is
particularly
viable
on
android
is
my
guess
that
the
sdk
is
not
really
going
to
work.
C
G
Oh,
but
when
we
talked
with
trust
in
honor
rock
about
the
gel
7-8
and
we
talked
about
android,
we
discussed
that
we
actually
need
some
some
users
to
advocate
for
the
whole
android
experience,
because,
from
my
point
of
view,
the
whole
android
story
is
currently
just
a
speculation.
G
C
We
did
have
one
person
login
issue
about:
are
the
random
id
generator
not
working
correctly
on
android?
I
think,
because
in
the
java
7
implementation
at
least
whatever
was
being
used
at
the
time,
the
there's
a
there
was
a
bug.
I
think
it.
Maybe
it's
been
fixed
in
later
versions
of
android,
with
thread
local
randoms
not
actually
being
created
initially,
with
a
with
a
a
random
seed,
they're
always
created
with
exactly
the
same
seed,
which
means
that
you
get
duplicate,
trace,
ids
and
span
ids,
which
is
very
bad.
C
F
It
I
think,
I
think
in
general
our
sdk
will
work.
The
only
parts
that
may
not
work
is
the
process,
the
spam,
processor,
exporters
and
stuff,
where
we
may
have
to
do
something
smarter
in
terms
of
how
do
we
batch?
When
do
we
send
the
request
to
save
battery
and
stuff?
But
besides
that,
I
think
the
sdk
that
accumulates
things
into
the
span
and
then
pass
them
to
the
spam
processor
should
work.
It's
not
something
that
I
know
we
don't
have
so
this
is
based
on
the
open
sensors.
F
Open
sensor
was
used
a
bit
on
on
android
and
the
only
problem
we
had
was
using
disruptor
because
we
were
having
a
batcher
with
disruptor
and
stuff.
So
the
recommendation
was
to
use
the
the
simple
spam
processor,
which
just
passed
everything
to
the
grpc,
but
that
caused
a
bit
of
a
battery
problem
because
we
are
sending
every
span
individually.
F
C
Or
if
we
have,
if
we
do
have
interest
in
android,
my
guess
is
a
better
or
another.
Viable
solution
might
be
just
to
have
a
separate,
android,
spam
processor
and
exporter
pipeline,
like
rather
than
trying
to
have
one
that
meets
all
those
needs.
Let's
just
have
a
separate
like
have
someone
who
knows:
android
really
well
build
a
spam
processor
and
contribute
it.
That
would
probably
be
the
better
option.
F
I
think
you're,
muted,
bargain
the
exporter
does
not
need
to
change,
because
you
just
make
a
call
it
to
the
to
the
back
end
or
something
it's
more
one.
There
are
two
things
credentials
that
people
ask
me
so
some
some
kind
of
authentication
features
for
the
exporter.
That
was
only
thing
but
the
the
processor.
F
Yes,
that's
that's
where,
where
a
lot
of
people
complain
about
and
dealing
with,
the
fact
that
apps
goes,
for
example,
in
the
background
the
the
processor
should
get
triggered,
and
you
should
push
all
the
things
flash
all
the
things,
then
there
were
a
lot
of
things
related
to
android,
but
all
of
them
were
related
to
batching
and
when
to
call
the
exporter.
So
yes
to
answer
your
thing,
specific
spam
exporter
for
android
is
probably
the
only
thing
that
we
need.
F
We
will
need
the
similar
thing
for
metrics
and
we
need
probably
the
capability
of
flashing,
our
batches
and
stuff
whenever
the
app
goes
background
and
a
couple
of
other
things.
But
yes,
let's,
let's
move
forward,
I
think
I
I
like
the
approach
of
building
some
custom
components
for
android
when,
when
needed,.
C
New
relics,
android
team,
is
starting
to
think
about
getting
interested
in
open
telemetry.
So
if
they
do,
then
they
will
certainly
hopefully
be
good
resources
for
that
excellent
I'll.
Add
some
notes
to
the
agenda
just
about
this
discussion
after
work
so
yeah,
I
don't
know
andrew.
I
put
a
link
into
the
current
p1
issues.
I
don't
know
if
you
want
to
kind
of
try
to
lead
us
through
maybe
getting
some
assignments
or
figuring
out
what
to
do
with
some
of
those.
D
Yeah,
I
just
had
a
scrub,
while
you
guys
were
talking.
I
I
like
the
weekly
digest
in
the
julia
tracking
and
putting
into
the
document
john.
Like
that's,
I
think
that's
the.
C
C
C
D
Friday
mornings,
man
I'm
totally
wrong:
google,
doc,
okay,
so
yeah
this
this
link
still
applies,
hope,
telemetry,
java,
yeah.
Okay,
there
are
four
of
the
seven
p1
issues.
I've
noticed
like
all
of
them
has
been
triaged.
That's
been
marked
required
for
ga,
which
is
great
four
of
the
p1
issues
for
the
seven
p1
issues
needs
owners.
Two
of
them
have
a
label
for
bug.
D
One
of
them
is
related
to
release
so
that
one
like
a
gradle
artifact,
not
not
making
it
up
to
the
right
place,
yeah
or
maybe
an
artifact,
so
that
one,
I
think,
probably
needs
to
be
addressed
sooner
rather
than
later.
So
that
way,
any
changes
to
the
code
can
make
it.
I'm.
C
Looking
for
somebody
who
knows
gradle
and
maven
way
better
than
me
to
help
out
with
that,
which
is
why
I
put
the
help
one
label
on
there
just
fyi.
This
number
1566
has
there's
two
competing
pr's
to
fix
that
one
right
now
and
we're
having
some
discussions
in
her
about
how
we
want
to
approach
solving
that
bug.
So
I
think
there's
actually
kind
of
two
people
maybe
assigned
to
it
I'll
update
that
after
the
meeting.
C
All
right
so
we'll
get
three
three
contributors,
three
assignees
for
that
one
issue
which
is
good
yeah,
so
the
first
one,
the
populate
implementation,
spec
compliance,
metric,
metrics
or
1570.
I
say
first
one,
I'm
looking
at
my
screen
the
compliance
matrix.
I
will
take
that
on
I'll.
C
I
can
assign
that
to
myself
right
now
and
I
will
give
a
first
cut
at
that
at
getting
that
pr
into
the
spec
repo,
the
other,
so
there's
some
things
that
are
already
assigned
one
of
them
to
who's
this
assigned
to
oh,
that's
right!
Do
we
have
an
update
on
the
grpc
stuff
from.
K
Sergey
yeah
he's
still
digging
into
that,
but
he
told
me
basically
that
at
this
moment
well
he
feels
that
the
microprofile
api
is
too
complex
for
a
used
case.
It's
very
thrift
and
thread
factory
related
oriented.
So
that's
one
of
the
things.
The
other
thing
is
that
they
use
java
8
right,
yeah.
C
C
K
Oh
yeah
you're
here,
okay,
please,
if
you
have,
you
can
mention
the
rest.
M
Yeah
yeah
yeah,
it's
yeah
yeah.
I
just
read
and
try
to
to
understand
how
I
can
implement
it
and
how
I
can
run
it,
and
I
couldn't
understand
they
mentioned
several
times
about
java
ee
java
e
concurrency
that
they
should
compatible
with
it
and
main
point
is
to
have
like
executor
on
executor
and
then
propagate
context.
But
I
still
couldn't
couldn't
repeat
just
simple
example
when
I
can
propagate
from
one
turn
from
one
thread
to
another
thread.
C
M
M
C
M
C
All
right,
I
don't
think
we're
going
to
make
any
progress
on
that
today.
I
don't
want
to.
I
don't
want
to
yeah.
C
I
don't
know,
I
think
that
I
haven't
done
any
of
that
research.
Sergey
would
be
the
only
one
who
who
would
have
the
details
there.
Okay,
do
you
have
a
document
where
you've
collected
any
of
that
info.
C
Yeah,
if
you
could
do
that,
that
would
be
very
helpful.
Okay,
thanks
nikita
on
1314,
you
had
an
action
item
to
talk
to
the
instrumentation
group
about
rapper
objects,
yeah.
G
So
the
the
trust
promised
to
think
and
come
up
with
to
try
out
if
we
can
get
away
without
shading
and
bridging
he
promised
that
week
and
a
half
ago
when
I
asked
him
on
when
tuesday
your
tuesday,
he
said
he
he
didn't
done
it,
he
didn't
do
it,
yet
he
hasn't
done
it
yet.
C
Yeah,
we
have
a
lot
of
different
ways
to
use
the
past
times.
Don't
we.
G
C
K
Yeah
yeah,
that's
correct!
I'm
planning
to
iterate
on
the
on
the
specification
side
on
this
pr
today,
so
yeah.
Once
it's
done,
it's
gonna
be
super
straightforward.
C
So
if
I
look
at
the
the
actual
bug,
we
need
someone
to
to
verify,
and
maybe
I
can
probably
do
this,
but
it
looks
like
it
was
getting
published
locally
but
didn't
end
up
in
maven
central,
so
maybe
something
to
do
with
the
way
we're
doing
our
publishing.
I
don't
I
don't
know
this
is
where
I
was
hoping
we
would
have
somebody
who
was
more
of
an
expert
in
that
could
take
a.
C
C
Honorable
is
he's
not
able
to
make
it
to
the
sick
meetings,
no
he's
in
tokyo.
So
it's
like-
I
don't
know
like
one
in
the
morning
for
him
all
right.
C
Yeah
we
so
trask
and
I
meet
with
him
on
thursday
evening
and
I
think
nikita
you
meet
with
him
on
tuesday
night
tuesday
night,
our
time
tuesday
morning,
wednesday
morning.
Your
time.
C
Yeah,
I
don't
think
I'd
do
that
without
asking,
but
I
can
ask
whether
he'd
be.
I
can
ask
him
next
time.
I
talk
to
him.
Okay
or
we
can
ask
on
getter
I
can
I
can
do
that.
I
will
make
a
note.
G
G
Yeah
bomb
is
not,
there
is
a
bomb
is
not
okay,.
F
Let
me
let
me
you
can
assign
to
me
let's,
let's
let
me
do
something
for
this
community.
C
K
K
C
Was
my
impression
at
the
moment
our
correlation
context?
Implementation
also
doesn't
match
the
spec
like
there's,
there's
yeah,
there's
some
stuff
in
this
back
that
we
have
not
yet
implemented,
and
we
also
have
not
implemented
like
a
propagator
for
that
to
actually
pull
the
baggage,
header
and
populate
the
correlation
contents.
Yeah.
K
For
the
part,
I
think
I
will
do
that
in
my
pr
draft
pr.
I
didn't
include
that,
but
that's
just
suggested.
John
probably
I
will
put
a
basic
implementation
and
then
I
will
wrap
it
up
once
specification
changes
are
done,
and
this
is
a
request
for
bulk
and
to
take
a
look
at
my
pr.
I
forgot
the
name,
but
it's
about
getter
keys.
You
know
so
together
not
only
will
return
a
key,
but
also
you
know
sorry
the
value
based
on
the
key
but
return
all
the
keys.
B
C
K
Yeah
yeah
exactly
yeah
yeah,
so
basic.
Basically,
those
values
that
come
in
that
baggage.
We
will
have
to
create
a
correlation
context,
value
and
put
it
in
the
co.
In
the
context.
That's
okay,
okay,.
C
K
Well,
in
general,
we
we
need
to
use
this
propagator
for
for
specifically
well,
if
you're,
it's
like
the
jagger
tracing
the
part,
but
you
know
do
we
have
this
notion
of
baggage
in
open
tracing
but
which
used
to
be
the
same.
You
know
trace
and
baggage,
but
now
it's
separated
because
of
open
telemetry.
C
Right
yeah,
that's
the
one
where
I
think
we
have
a
little
bit
more
work
to
do
even
beyond
what
you
you
had
in
on
the
getters,
and
I
marked
that
as
blocked
on
spec.
Just
because
the
change
to
use
the
baggage
header
has
not
been
approved.
Yet
I
think,
there's
still
an
open,
spec
issue
on
that.
C
C
Yeah
spec
spec
issue.
J
C
K
C
Perfect
awesome
and
then
once
that's
once
that's
done,
then
I
think
my
guess
is
I'll.
Be
able
to
pick
up
the
correlation
context.
Api
changes
next
week
to
at
least
get
ourselves
into
like
in
spec
compliance
from
correlation
context.
That's
my
plan.
At
least
I've
started
digging
into
that
a
little
bit.
K
C
F
Yeah
one
thing
that
we
need
to
do
before
that
is
probably
one
of
the
a
lot
of
the
meetings
that
we
have.
Maybe
one,
including
unrug,
maybe
go
over
the
specs
and.
C
Yeah,
so
when
I
take
on
when
I
fill
in
the
matrix
that
that
was
this
put
together,
that'll
hope
anything
where
there's
a
hole.
I
will
definitely
log
an
issue
for
you,
as
a
part
of
that,
so
that'll
get
us,
get
us
a
lot
of
the
way
there.
I
think
I
think
we're
pretty
close,
though
that's
correlation
context.
I
think
we're.
F
C
K
C
Know
I
know
I
know
all
right
well,
I
think
unless
we
have
any
other
issues
I'll,
let
you
two
go
and
we
can
be
done
a
little
early.
Okay,.
N
I
I
actually
did
join
a
little
bit
to
talk
about
vlogging
and
make
sure
that
I'm
not
bugging
you
guys
too
much
wow
trying
to.
B
C
C
If
some
other
person
on
the
project
wants
to
start
digging
into
reviewing
logging
prs,
that
is
up
to
them,
but
I,
as
kind
of
the
maintainer
who's
tracking
things,
the
most
and
doing
the
most
reviews,
I
don't
have
the
bandwidth
for
it
and
because
we're
waiting
for
post
ga
to
really
do
that,
I'm
going
to
drop
those
on
the
floor.
Apologies,
but
that's,
I
think,
just
the
realistic
thing
that
for
me
at
the
moment.
F
Okay,
I
can,
I
know
I
know.
I
know
this
is
something
important
for
for
splunk
and
stuff.
Do
you
mind
if
I
review
those
and
and
merge
them?
No,
it's
normally.
It's
totally
fine.
Okay,
perfect!
I
will
make
sure
david
just
ping
me
and
we
can
work
around
as
long
as
they
are
not
part
yet
because
we
decided
to
be
part
of
the
extensions
for
the
moment.
So
we
can
play
a
bit
and
even
if
we
make
some
mistakes
we
can
fix
them
later.
F
That's
we
already
discussed
that.
I
think
I
already
said
that
in
one
of
the
issues
to
make
sure
we
make
it
extension
so
just
pinging
on
internal
slack,
we
can
figure
it
out
a
way
to
to
make
progress
on
this,
at
least
as
experimental
thing,
and
then
we
can
figure
it
out
how
how
we
merge
them
in
the
api
after
the
ga,
for
the
other
things.
N
C
N
N
No,
absolutely
and
and
kind
of
my
ability
to
participate
is
you
know,
splunk
sponsoring
me
for
logging,
so
like
we're
kind
of
by
nature
at
loggerheads,
and
I
didn't
want
to
be
stepping
on
and
making
things
worse
for
you
loggerheads.
Did
you
just
say
longer
actually
didn't
mean
that
one,
but
take
it.
C
F
Okay,
see
you
guys,
I
have
to
go
with
carlos
on
triaging
specs,
because
you
know
what
you
don't
have
too
many
notification
from
the
specs
repository.