►
From YouTube: 2020-10-02 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
Which
is
actually
super
cool
like
I've?
Never
I've
never
used
splunk
before
myself
and
getting
to
see
like
architecture
of
the
the
logging
stuff
and
like
how
all
that
works
has
actually
been
was
pretty
interesting.
A
Yeah
yeah,
I
used
we
used
splunk
at
the
last
company.
I
was
at
for
log,
aggregation
and
searching,
and
so
yeah
it's
pretty
cool.
Like
I
mean
you
know
it's.
I
was
writing
some
pretty
crazy
queries
to
aggregate
stuff
across
all
different
servers
to
look
for
different
exceptions
and
different
stuff
that
yeah
it's
handy.
B
B
Well,
signalfx
has
its
own.
It's
it's
basically
still
kind
of
its
own
product.
The
there's
the
integrations
are
worth
trying
to
figure
them
out,
but
it's
still
kind
of
its
own
apm
product.
So.
A
So,
are
you
going
to
is
splunk
going
to
continue?
The
signal
fx
brand,
like
is
the
hotel
agent
they're
gonna
ship
gonna
be
signalfx
no.
B
B
C
C
By
the
way,
by
the
way,
I
think
I
think
it's
already
named
splunk
apm
or
something
I
don't
know
all
the
details,
I
I
always
struggle
with
this.
So.
B
Architecture
signal
effects.
Architecture
is
awesome,
it's
super
cool.
I
think
it's
it's
red
anyway,
so.
C
Anyway,
as
a
user
of
this
library,
I
want
to
ask
you
guys
as
maintainers
a
question
when
are
you
gonna
merge
my.
C
A
Oh,
oh
yes,
I
forgot
that
this
is
the
I
still
forget
this
is
now
the
combined
java
and
instrumentation
meeting
see.
C
It's
no
longer
everything
is
good
man,
it
still
fails
the
test
because-
and
I'm.
A
And
how
about
how
about
this
one?
What
what's
the
what's
going
to
be
the
decision
on
pre
or
post.
F
Yeah,
so
my
two
cents
here
was
also:
should
we
try
to
fix
another
problem,
which
is
we
expect
some
of
the
things
to
be
said
only
once
in
the
beginning
of
the
process
and
some
others
to
be
updated
at
the
runtime
like,
for
example,
the
config
can
be
updated
at
runtime,
but
we
recommend
everyone
to
set
the
spam
processor
and
everything
at
the
beginning
of
the
process.
Should
we
have
an
api
that
kind
of
force,
people
or
or
lead
people,
to
do
that
instead
of
doing
at
any
moment
and
have
unexpected
results.
B
B
F
It
is
yeah,
it
is
not
final
right
now,
it's
volapile
essentially,
because
because
I
need
to
always
get
the
latest
reference,
it's
just
volatile,
but
that
means
I
only
based
on
x86
cash.
I
always
have
to
to
check
and
look
for
for
any
barriers
and
for
for
a
lot
of
things,
so
I
may
have
some
cash
problems.
F
Indeed,
intel
did
a
very
good
job,
and
if
that
doesn't
change
at
all,
they
will
they
will
cash
it
as
as
shareable
or
whatever,
with
the
messy
that
yeah
shareable
state
and
and
they
will
be
able
to
to
to
read
it
without
generating
new
barriers.
But
it
may
cause
problems
on
some
processors.
B
Yeah
I
feel
like
this
is
something
we
could
it's
not.
It
doesn't
feel
super
urgent
to
me.
I
guess
at
the
moment
it
might
have.
B
There
might
be
some
long-term
things
to
think
about
there,
but
I
don't
know.
That's
just
me,
like
my
gut
tells
me
that
immediately
at
the
moment,
I
don't
feel
like
it's
a
super
urgent
decision,
but
but
having
a
having
a
good,
well-defined
life
cycle
for
the
sdk
does
feel
like
would
be,
it
would
be
valuable.
F
So
the
only
problem
that
I
have
is
in
order
to
achieve
the
next
step
that
we
just
discussed
if
we
need
yet
another
breaking
change,
maybe
to
think
right
now
about
that.
That's
that's
my
only
concern
and
that's
where
I'm
coming
from
with
the
mindset
that
we
we
do
have
the
opportunity
right
now
to
break
this
and
to
do
it
correctly
in
in
nice,
and
we
will
not
have
that
benefit
that
opportunity.
Let's
say
in
six
months:
that's
that's
the
only
mind
that
the
only
thing
that
I
have
in
my
mind
so
yeah.
B
It
does
put
it
on
that
management
interface,
which
would
enable
us
to
add
life
cycle
life
cycle
things
to
it.
After
the
fact
it
wouldn't
impact
any
of
the
tracing
or
any
of
the
the
actual
instrumentation.
G
So,
just
completely
randomly
yesterday
I
started
brainstorming
a
bit
on
maybe
a
related
topic,
so
when
bogdan
mentioned,
that
context
could
be
injectable.
I
was
thinking
that
one
thing
like
open
telemetry
right
now
isn't
object.
Injectable,
because
open
telemetry
is
just
a
three
static
method,
so
I
thought
maybe
that
should
be
itself,
be
an
object
and
that
can
have
then
that
can
be
managing
the
life
cycle.
Also,
one
thing
I
noticed
is
the
resources
on
the
tracer,
but
since
resource
is
a
global
objection
that
being
open
telemetry,
so.
F
G
G
D
A
Marcin
brought
that
up
this
morning
in
the
instrumentation
meeting,
so
he
said
he
was
discussing
with
you
already
onorag,
because
I
I
said
he
should
really
discuss
with
you
that
I
would
I've.
G
F
A
Yeah,
it's
he's
exploring
a
new
approach
now
bogdan,
which
is
having
a
an
abstraction
in
spring
cloud
sleuth
that
can
talk
to
both
open
telemetry
and
brave.
F
I
know
he
he
mentioned
that
to
me
and
mostly
he
was,
and
I
agree
somehow
with
him.
He
has
a
hard
deadline
in
a
month
or
something-
and
I
don't
think
we'll-
have
a
stable
api
back
that
point.
So
he
may
better
do
that
for
the
moment,
because
that
allows
him
to
upgrade
the
api,
even
if
we
break
a
bit
and
then
later
later,
whenever
we
stable
to
replace
that
abstraction
with
us
that
that
was
what
I
told
him.
A
So
the
date
they
mentioned
was
there
was
also
who
was
the
other
person.
Spencer
was
on
the
call
and
the
dates
you
because
my
marching
did
wasn't
quite
sure
about
the
dates
but
spencer
mentioned
so
spring
boot
to
four
is
mid
november,
so
that
puts
their
time
frame
at
end
of
november
early
december
for
sleuth.
F
To
give
him
also
the
flexibility
to
not
break
things,
because
also
I'm
creating
right
now.
Two
issues
that
I
promised
looking
at.
A
Okay,
do
you
want
me
to
go
over
this
morning's
meeting.
A
All
right,
oh
so,
we
spent
pretty
much
the
whole
meeting
talking
with
marcin
and
spencer,
and
so
these,
let's
see
one
of
the
topics
was
about
dupli
like
should
we
should
we
even
have
spring
library,
instrumentation.
A
In
in
open
telemetry
or
if
they're
going
to
you
know
assuming
they're
using
open
telemetry
directly,
they
would
provide
they
would
provide
that
instrumentation
yeah.
I
know
right,
I
was
we
were
nikita
and
I
were
all
in
on
that
we
were
like.
We
are
more
than
happy
to
defer
that
expertise
to
you,
so
we
would
still
have
right.
We
would
still
have
spring
auto
instrumentation
in
the
agent.
They
don't
care
about
that
so
much,
and
they
would
then
basically
be
the
official
place
for
like
spring
mvc,
sprint
web
flex.
B
They
would
they
would
they
take
over
the
starters,
or
would
that
be
something?
Yes.
A
B
Yeah
this
I
mean
this
is
the
kind
of
the
dream
right.
This
is
the
dream.
Have
the
library
authors
maintain
the
instrumentation
so.
F
You
need
to
make
sure
you
need
to
make
sure
the
starters,
because
that's
where
you
may
have
overlapping
and
circular
dependencies
between
your
project
and
their
project.
So
anyway,
just
I
don't
know
who
owns
the
starter,
but
that
may
be
the
the
the
pain
point
because
of
the
circular
dependency,
because
you
need
to
depend
on
them
for
for
embedding
the
instrumentation
into.
A
Oh,
so
we
wouldn't,
we
wouldn't
depend
on
them.
We
would
probably
just
keep
our
same
auto
instrumentation
for
the
like
web
flex,
web
mvc
spring
scheduling
that
kind
of
stuff,
and
but
we
would
not
try
to
extract
those
into
common
library.
Instrumentations.
A
F
So
so,
anyway,
you
need
to
have
a
coherent
story
and
I
would
love
some
code
to
leave
there
some
code
whatever
needs
to
leave,
but
avoid
duplicate
code
if
possible.
I
just
yesterday
I
just
had
to
deal
with
the
duplicate
code
in
splunk
and
now
I'm
I
hate
duplicate
code
as
much
as
you
can
imagine
so.
F
A
You
know
it's
interesting.
I
asked
that
so
sleuth
because
I
was
like.
I
always
thought
that
sleuth
was
spring
cloud.
A
But
they're
like
yeah
the
the.
I
really
think
they
should
bite
the
bullet
and
rebrand,
because
if,
if
we
all
think
that,
like
everybody,
thinks
that
but.
A
That
sleuth
is,
is
really
it's
spring
boot.
So
as
long
as
the
users
are
spring
boot,
they
support
them
so
that
that's
a
you
know
potential.
We
would
not
have
coverage
for,
say
somebody
who's
using
spring
outside
of
spring
boot,
but
you
know,
if
spring
doesn't
really
care
about
those
users,
then
it
doesn't
seem
like
it's
something.
We
would
need
to
worry
about
that
much.
A
I
don't
remember
I
did
not
ask
that
specifically
marcin
did
mention
that
he
seemed
to
like,
like
he
seemed
to
indicate
that
there
had
been
some
issues
with
that
decoupling.
G
G
A
Yeah
yeah
yeah,
I
was,
let's
see
what
else
yeah
they
were
just
covering
sort
of
how
it
works
currently
and
sleuth
is
auto
configuration
for
that
and
that's
why
we
have
this
old
version
of
spring
web
and
we
explained
okay.
A
This
was
the
library
stuff
statics,
so
yeah
I
was
asking
if
there
was
any
chance
that
he
could
make
our
10
p.m.
Tuesday,
10
p.m.
Pacific
meeting
to
chat
about
these,
but
that's
7
a.m
for
him
and
he
was
like.
G
G
A
Like
I
don't
think
there
we're
going
to
be
able
to
do
the
all
of
us
meet
at
the
same
time,
so
I
think
it's
best
if
he
talks
to
you
on
a
rag
about
this
topic.
A
A
So
these
are
the
ones
that
they
would
like,
and
the
slf
for
j
log
back,
oh
yeah,
just
log
back
yeah,
I
asked
he
was
happy
that
we
were
already
injecting
the
same.
We
were
using
the
same
naming
pattern
they
were
already
using
so
that
matches
up
nicely
for
them.
A
I
asked
for
one
of
these
that
would
be
sort
of
like
a
good
test
bed
for
them,
since
we
don't
have
library
instrumentation
for
any
of
these,
yet
that
we
could
prioritize-
and
he
mentioned
grpc-
would
be
a
good
one,
because
in
the
the
brave
world
this
like
provides
almost
I
mean
it's
like
a
super.
He
said
it's
super
easy
for
them
to
wire
up.
A
So
if
we
could
accomplish
the
same
thing
for
them,
that
would
be
great,
and
and
for
that
wiring
up,
he
was
just
asking
for
sort
of
like
a
standard
creation
pattern
like
you
know
where
they
can
just
always
know
they
take
this
library
instrumentation.
They
instantiate
the
tracing,
object
and
wire
in
things.
A
And
spring
cloud
gateway
support,
so
not
from
brave
from
so
this
was
the
sleuth
team
other
spring
libraries,
oh
yeah,
we
were
explaining
tyler
was
explaining
how
I
mean
we
typically
just
we
will
support
back
as
far
as
we
can
that
we
get
for
free
that
before
we
find
a
incompatibility,
but
yeah
merchant
was
like.
Oh
we're.
You
know
they're
not
going
down
that
road.
A
Of
course,
they're
going
to
stick
with
the
spring
boot
versions,
bomb
and
whatnot,
and
then
I
mentioned
because
he
was
saying
that
he
was
looking
at
that
other
pattern
of
doing
an
abstraction
layer
that
could
talk
to
both
open,
telemetry
or
brave.
A
I
mentioned
that
if
they
did
the
original
thing
that
they
were
talking
about,
of
having
a
open,
telemetry,
sdk
api
implementation
that
talks
to
brave
that
we
could,
we
could
use
that
inside.
The
java
agent
then,
which
that
we
talked
about
this
on
tuesday,
has
just
a
selling
point
on
on
that
particular
a
benefit
of
that
particular
approach.
F
Well,
I
asked
for
both
things
yes
asking
about.
A
F
Yeah
but
that's
for
for
us
kind
of
a
starter.
A
Yeah
yeah,
you
know
I
I
I
think
I
I'm
all
in
on.
You
know
the
giving
back
to
the
to
supporting
this
with
sleuth
making
that.
F
F
Actually,
I
heard
that
a
lot
of
people
in
brave
they
are
seeing
us
as
their
killers.
We
are
here
just
to
kill
their
project,
so
that
being
said,
I
think
we
should
play
nice
with
that,
and
that's
one
thing
the
way
how
we
can
solve
this.
Secondly,
is
for
people
that
they
really
want
to
move
from
grave.
By
giving
this,
we
give
them
a
path
to
move
gradually,
not
have
one
pr
to
change
everything.
So
we
have
multiple
advantages
in
this
and
I
think
it's
important
for
us
to
do.
B
Yeah,
I
think
the
same
is
excuse
me.
The
same
is
probably
true
with
micrometer
around
metrics
yeah.
F
A
F
F
F
A
So
we
left
it
action
plan
was
so
marcin
is
still
spiking
figuring
out,
which
approaches
wants
to
take
timeline.
I
asked
what
strictly
like
instrumentation
is
required
and
he
said
really.
The
log
back
logging
instrumentation
is
the
only
one
that
they
couldn't
ga
without
and
then,
as
far
as
other
ones
from
this
list,
grpc
would
be
the
the
best
sort
of
sample
for
them.
A
So
yeah
we
didn't
really
get
to
most
of
the
other
agenda
items
which
was
totally
fine,
because
the
sleuth
was
a
really
good
discussion
for
zero
nine
next
targeting
end
of
next
week,
there's
just
three
issues
that
I've
marked
for
zero.
Nine,
the-
and
these
are
just
the
assignees
that
we
have
nikita-
was
going
to
check
with
jacob
see
if
he
was
able
to
do
this
or
not
otherwise,
we'll
find
somebody
else
or
or
we
won't
fix
it-
for
a
zero
nine,
hey.
A
F
A
E
E
H
Okay,
cool
yeah,
I
said
something
before
yeah,
but
you
guys
couldn't
hear
me.
I
was
wondering
okay.
Yes,
I
wanted
to
give
you
guys
an
update
before,
because
I've
to
go
soon.
Yeah
there's
been
a
bit
of
a
real,
so
I've
been
moved
to
another
team,
so
yeah,
I
won't
be
working
on
open
telemetry.
H
G
G
A
H
And
so
I'll
still
be
finishing
off
my
project
current
project
inside,
like
in
kind
of
well
the
moving
open
sensors
to
open,
telemetry,
well,
libraries
that
are
using
using
them
but
yeah.
It
won't
really
be
like
involved
anymore
with
the
I
guess,
the
zigzag
anything.
Basically
anything
that's
not
related
to
the
project,
because
the
goal
is
to
basically
finish
the
project
as
soon
as
possible,
and
then
so
I
can
move
on
to
my
current.
The
new
teams
projects.
H
H
Yeah,
yes,
so
just
wanted
to
come
and
yeah
say
bye
and
yeah,
and
thank
you
for
helping
me
on
board,
even
though
it's
something
long.
F
Yeah,
I
think
we
are
making
the
new
team
somewhere
on
the
east
coast.
If
I
I
talked
to
another
guy,
he
contacted
me.
The
other
guy
contacted
me
to
ask
how
he's
open
television
if
we
are
a
healthy
thing
and
stuff,
and
I
made
a
good
advertising
for
the
project
and
most
likely
the
team
will
be
in
on
the
east
coast.
If
I
heard
correctly.
H
Yes,
I
think
basically
yeah.
That
was
another
thing
I
wanted
to.
Let
you
know
is
yes,
someone
will
be
replacing
me
for
this
role
and
open
telemetry,
but
it'll
probably
be
someone
from
the
us.
G
A
G
A
Well,
thanks
for
thanks
for
stopping
by
and
letting
us
know,
zoey,
it
was
great
having
you
here
while
it
lasted
and
best
of
luck
with
with
spanner,
I
don't
know
what
that
is.
F
A
A
It
does
sound
familiar
now,
but
at
first
I
was
thinking
it
was
a
span
I
was
making.
I
was
trying.
H
E
H
A
G
F
B
So
speaking
of
24
hours,
talk
to
carlos
today
he's
going
to
cut
a
release
of
0.9.0
of
the
api
and
sdk
tomorrow
morning
is
sometime
his
time
so
we'll
get
on.
I
know
released
tomorrow,
so
I
don't
think
we
should
put.
F
Right
now
the
context
there
personally
yet
or
do
you
think
we
should
put
it
right
now?
Well,.
B
B
Oh
about
whether
we
want
to
try
to
release
that
tomorrow
or
not-
I
don't
know
bogdan,
do
you
want
to
have
more
time
to
look
at
that
or
do
we
want
to
merge
it
and
move
forward,
since
this
sdk
only
changes,
I'm.
G
That
brainstorm,
like
it's
related
to
this
interface,
I
think
this
interface
makes
that
even
better,
so
I.
F
A
F
A
Oh,
you
didn't
see
the
last.
Oh
man,
the
last
the
last
snap
I
updated
the
snapshot
just
like
was
it
a
week
ago,
but
it
was
a
ton
yeah
anyway,
honorable
reviewed
it
and
he
was
sweating
after
because
it
was
you
know,
yeah.
A
Are
making
some
massive
changes
but
good
changes?
Yeah
the
like
when
I've
been
updating
the
snapshots,
I
can
see
like
all
the
the
ripple
effects
and
they've
been
good.
A
Cool,
I
think
this
I
had
put
this
on
the
agenda
for
well.
I
can
give
people
an
update
on
the.
I
did
go
over
the
weekly
digest,
so
the
this
was
a
splunk
change
for
being
able
to
supply
default
configuration
for
the
java
agent
and
it
ended
up
spawning
some
really
good
improvements
in
the
config
and
so
now
config
is
just
a
properties
bucket,
instead
of
having
like
all
these,
like
instrumentation
specific,
like
kafka,
config
properties
in
the
config
object.
A
Now
it's
just
a
properties
bucket
and
all
the
instrumentations
read
that
at
startup
cache
that
value
and
use
that
we
now
support
otlp
span
like
so
you
can
disable
essentially
metrics,
which
is
how
the
light
step
is:
go
working
around
the
metric
export
log,
spamming
problem.
Currently
we
had
got
debugging
docs,
oh
yeah,.
A
F
Thanks
everyone,
by
the
way,
I
owe
you
some
couple
of
prs
more
for
the
metrics
data.
I
will
probably
send
them
tomorrow,
see
you.
A
G
A
G
There
yeah
so
this
sorry,
this
spam
is
a
different
spam,
but
so
in
my
playground,
since
it
everything's
running
in
docker,
I
have
to
override
the
endpoint
for
otlp
and
I
found
the
otlp
endpoint
variable
to
been
removed
and
I
switch
it
to
the
spanning
point
and
then
the
metrics,
because
I
didn't
do
that.
The
metrics
expert
just
always
has
a
connection
refused
in
my
playground.
Oh.
A
Yeah,
so
this
it's
this
issue
here
we
need
to
update
them
on
the
readme
yeah
yeah.
G
A
B
No,
there
should
be
a
common
one
still
here,
this
giovanni's
pr
preserved
it:
okay,
okay,
so
so
the
way
it
works
is
it'll.
Look
for
the
specific
one
and
it'll
fall
back.
G
B
Yeah,
you
can
see
exporter.
Exporter
got
shoved
into
all
of
the
names.
G
A
B
G
G
B
G
A
Okay,
so
yeah,
so
so
powell
had
added
these
the
debugging
docs,
and
then
I
had
just
updated
it
because
I
was
following
up.
I
was
gonna,
give
updates
from
last
week
on
these
these
things,
so
I
was
retesting
that
and
realized
that
you
can
john
was
right.
B
A
Right
right,
yeah
yeah,
so
we
also
now
have
pavel
did
a
really
nice
job
of
taking
the
discussion
we
had
last
week
in
the
sig
and
turning
that
into
a
high-level
muzzle.
You
know
what
the
heck
is
muzzle
and
this
will
give
us
a
good
place
to
start
adding
more
muzzle
dock.
G
A
We
are
updated
to
zero
nine
zero
snapshot.
As
of
what,
like,
I
don't
know,
five
four
or
five
days
ago,
so
at
least
the
oh
yeah.
This
was
the
problem.
Sorry
jason.
This
was
the
problem
with
the
jfrog
snapshots,
going
going
going
bye-bye.
A
D
G
D
G
A
A
And
yeah,
so
hopefully,
when
zero
nine
is
released,
we
won't
have
we'll
see,
is
won't
have
too
much
more
to
update,
but
we
will
see
only
nikita
said
only
three
decorators
left,
so
that
long
effort
is
almost
done.
So
that
will
be
a
deletion
party.
After
that
we
have
a
new
approver
matias.
A
A
A
That
was
a
good
comment.
I
think
mateos
and
you
on
our
were
chatting
about
this,
of
how
oh
for
the
kafka
test
of
splitting
out
and
only
updating
the
config
at
the
beginning,
because
eventually
we'll
want
to.
We
want
to
run
the
test
with
the
real
java
agent
and
only
load
the
config
once
and.
H
A
H
A
G
Was
there
any
other
response?
I
forget
what
the
statement
was.
A
If
we
are,
we
should
get
the
artifact
name
correct
if
we
can,
but
now
after
this
morning's
discussion,
I'm
wondering,
should
we
not
push
these
at
all
to
maven
central
to
avoid
sort
of
them
ending
up
in
both
places
or
do
we
think
that
we.
H
A
A
G
A
A
G
A
G
A
Okay
well
at
least
yeah,
we'll
we'll
try
to
hold
off
as
long
as
possible,
then
to
give
time
to
figure
out
the
whole
sleuth
plan,
but
it
sounds
like
they're
they're
pushing
yeah.
I
was
surprised,
they're,
the
the
dates
that
they're
talking
about
they're
they're
pushing
hard
on
that.
So
that's
great.
I
think
that
made
me
optimistic.
A
What
do
you
think
about
that?
The
ga
I
shared
with
you
on
slack
the
ga
schedule
from
the
maintainers
meeting
john
I
mean,
is
that.
B
Yeah,
I
have
no
idea.
My
guess
is
that
the
like,
I
think,
that,
what's
his
name
from
lightstep
top
ceo
guy,
oh.
B
So
I
yeah-
I
don't
I
mean
it's
if
they're
not
gonna,
if
we're
gonna
keep
on
tweaking
the
apis,
it's
gonna
take
a
long
a
lot
longer
to
get
to
ga.
Then.
G
B
B
Yeah
I
mean
my
guess:
is
that
we're
not
announcing
metrics
ga
at
kubecon
by
we
I
mean
I
have
no
idea
who
that
is
the
power's.
The
bee,
like
I
can't
imagine
like
we're
not
even
done
with
the
trace
spec
freight
spec
isn't
frozen
it's
october
now
right
we're
already
like
a
week
off
the
schedule,
at
least,
if
two
not
two.
B
H
B
Mean
it's
so
similar
that
I
don't
think
it's
going
to
be
a
huge
deal.
No,
I'm
just
concerned
that
they,
like
the
whole,
like
spain
context,
and
I
mean
I'm
sure
that
as
soon
as
people
actually
start
implementing
baggage,
there's
going
to
be
all
sorts
of
things
that
change
speaking
of
baggage
yeah.
I
forgot.
B
Yeah
I'm
actually
starting.
I
I
like
to
bogdan
today,
I'm
gonna
take
over
the
propagator.
I'm
gonna
try
to
implement
the
baggage
propagator
at
least
from
w3c-
hopefully
tomorrow,
hopefully
by
end
of
tomorrow,
I'll
get
a
in
for
that
it's
kind
of
a.
I
don't
know
if
anybody's
look
at
the
baggage
spec
in
w3c,
it's
kind
of
weird,
so.
H
B
Bear
keys
in
the
baggage
as
well
anyway,
there's
some
there's
some
oddness
in
there.
B
B
B
Yeah
I
mean
I
mean
not
that
we
couldn't
break
it
if
we
needed
to,
but
I
think
the
idea
is
at
that
point:
we've
got
a
stable
api,
an
implementation,
that's
a
release
candidate
which
for
java,
I
think
we're.
I
mean
api
tweaks
here
and
there,
like
we've,
got
a
stable
like
this
tracing
implementation
works
like
so
a
few
tweaks
here
and
there
on
the
exact
structure
of
the
api
is
not
going
to
slow
down
the
implementation
except
just
to
adjust
for
whatever
that
is
so
metrics
is
a
whole.
Other
story
like
I.
B
D
Everybody's
going
to
have
their
pet
in
their
pet
instrumentation
too,
and
they're
going
to
have
piles
of
metrics
for
every
instrumentation.
Is
my
prediction.
A
So
for
tracing
I
mean
so
there's
still.
I
was
just
thinking
of
the
stuff
that
we
wanted
to
get
in
around
like
the
java
8,
some
java
8
simplifications
in
the
api.
B
A
Spam
context
was
the
big
span.
Contract
is
a
big
one.
You
saw
bogdan
gave
thumbs
up
to
the
spec
issue.
G
B
B
A
G
A
Yeah.
Thank
you,
though,
for
on
rug
for
doing
that
stuff.
Like
the
I
know
that
that's
fan
context
was,
I
know
you
were
very
hesitant
to
go
down
that
road,
but
thanks
for
persisting.
B
Yeah,
it's
interesting
to
me.
This
is
just
my
impression
that
I
think
maybe
bogdan
has
let
go
of
open
census
and
now
he's
thinking
a
little
bit
more.
I
don't
know
whether
it's
splunk
oriented
or
just
generally
like
not
worrying,
about
open
census.
So
much
and
it's
like
all
bets
are
off,
throw
crap
out
get
rid
of
it.
Destroy,
delete,
delete,
delete,
I'm
like
it's
all
good.
A
G
F
B
A
B
B
A
Oh
yeah,
that
was
onorak.
What
are
your
feelings
about?
The
x-ray
propagator
spec
and
revert
discussion.
A
G
A
A
No,
the
the
new
this.
G
E
A
That
was
opened
and
merged.
Yesterday.
B
I
mean
I
think
I
created
that
issue,
the
issue
that
kicked
this
off
originally
like
six
months
ago,
yeah.
I
think-
and
I
thought
that
everyone
had
agreed,
that
it
was
okay
to
have
vendor
propagators.
B
G
G
A
A
H
A
G
G
I
wonder
if
even
just
renaming
it
makes
people
feel
a
bit
better.
So
yeah
like
it's
not
meant
to
be
x-ray
specific
in
the
first
place,.
A
Yeah
yeah
and
same
with
the
ot
tracer,
the
the
the
light
step.
A
H
B
G
A
Anyway,
I'm
fine
with
the
stand
of
if
it's
in
open
telemetry,
you
know
we'll
fix
it,
but
we're
not
going
to
prioritize
that
work
right,
we'll
fix
it
eventually,
probably
when
somebody,
you
know,
submits
a
pr
with
some
good
tests.
B
Yeah,
I
think
that's
why
that's
why
carlos
is
a
little
hesitant.
It's
just
like
you
know
what
what
are
maintainers
on
the
hook
for
and
if
we're
on
the
hook
for
like.
If
so,
if,
let's
just
say
hypothetically
and
there
was
a
security
vulnerability
in
the
propagation
format
and
it
was
hosted
in
open
telemetry,
would
the
maintainers
be
on
be
on
the
hook
for
security
fixes
to
that
when
it
wasn't
something
that
they
necessarily
had
expertise
or
really
owned?
That's
probably
why
carlos
is
pulling
the
pulling
the
the
stop.
The
train
lever.
G
A
G
B
G
A
Yeah
yeah
we'll
see
what
comes
up.
I'm
gonna,
you
know
I
I
would
prefer
the
propagator
stay.
I'd
prefer
that
the
the
resource
detectors
stay
again
because
those
benefit
you
know
there's
tons
of
people
on
aws.
Why
wouldn't
we
want?
You
know
at
least
the
cloud
provider
resource
detectors
seems
like
a
win
for
open
telemetry,
yeah.
B
A
A
B
B
Good
to
see
y'all
I'll
see
you
next
week.