►
From YouTube: 2021-04-29 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
A
B
B
How's
your
week
been
nacho
my
week.
A
Yeah,
no
not
much,
not
just
working
on
on
some
some
issues
waiting
yeah.
I
think
I
I
added
some
feedback
in
one
of
the
pr's
and
also
waiting
on
on
the
network,
instrumentation
vr,
so
yeah.
He
asked
for
more
reviews,
so
james
sherlock
already
had
some
review.
It
would
be
great
if
some
other
review
can
appear
there,
because
it
looks
like
as
it's
a
big
pr
and
it
has
some
very
low
level.
Swift
stuff.
I
think
series
was
not
very
comfortable.
Just
approving
that.
B
B
B
Fun,
I
guess
I
don't
know
if
if
james
is
gonna
show
up,
should
we
wait
a
couple
more
minutes.
A
A
A
Yeah
so
first
first
topic
is
reviewing
last
meeting
actions
just
to
know
if.
A
So
let
them
dlp
exporters
are
done
at
least
decided
and.
C
Yeah
so
I
spoke
with
the
james
one-on-one,
I
chatted
with
him
and
he.
A
C
He
actually
did
not.
You
know,
spend
much
time
on
this
as
well,
because
apparently
they're
using
the
collector
now
so
they're,
not
they
don't
need
the
okay
yeah,
so
so
yeah.
I
think
I'd
like
to
do
more
research
on
how
to
do
a
remote
write
so.
A
A
Okay,
that's
from
the
meetings
for
the
beta
release
plan.
Still,
I
think
same
things
missing.
A
Yeah
also,
I
I
started
with
a
blog
post.
I
was
checking
how
other
blog
posts
had
been
posted.
They
are
basically
using
the
same
structure
for
all
the
open,
telemetry
announcements
more
or
less
so
yeah.
I
I
I
as
network
instrumentations
must
be
integrated
or
merged
with
the
rest
of
the
code.
I
I
was
not
gonna
rush
with
that,
but
yeah.
Let's
I
I
will.
I
have
more
or
less
the
points
written
but
not
finished,
so
I
I
will
let
you
I
will
send
you
a
copy
bryce.
A
Okay
about
the
network
status
information,
we
we
talked
about,
keeping
that
like
like
it
was
the
code
is
commented,
but
it's
in
the
official
code.
So
if
someday
someone
wants
just
to
take
on
the
instrumentation
part,
all
is
done
just
that
missing
and
also
the
trace
propagation
between
processes
is
something
I
I
did
in
the
last
week
and
I
have
added
a
point
just
to
talk
about
it
and
and
explain
how
it's
done,
because
it's,
I
think,
it's
very
powerful,
but
it
was
very
basic
to
implement.
A
I
even
added
that
to
a
pr
that
I
had
with
a
trace
the
3s
provider
sdk,
because
it
was
very,
very,
very
simple,
in
fact
so
yeah
that
from
that
last
meeting
actions,
so
the
next
point
is
price
cross
application
time
stamp
disparities,
yeah.
B
Yeah,
so
this
is
an
issue
that
I
was
running
into
you
know
doing
so,
generating
some
data
with
elastics
back
in
we've
got
some
like
services
set
up.
That
kind
of
you
know
generate
data
and
stuff
like
that
here.
Let
me
share
my
screen
here.
Oh
oh,
could
you
uncheer
nacho.
A
B
There
you
go
cool,
so
not
anything
too
exciting
to
look
at
here,
but
so
running
into
this
issue,
where
you
know
we
generate
a
span
on
the
mobile
device
and
it
sends
data
to
the
back
end,
and
you
know,
there's
an
issue
with
the
timing
between
the
two,
the
two
clocks
right
on
on
the
mobile
device
and
on
the
back
end
device
and
okay.
B
Going
to
line
up
quite
right-
and
I
was
wondering
I
I
haven't-
bothered
digging
around
in
the
in
the
open,
telemetry
spec,
but
is
there
you
do
you
know
off
the
top
of
the
top
of
your
head?
If
there's
any
sort
of
you
know
spec
for
compensating
for
this,
I
know
that
good.
A
Yeah,
I
think
that
the
solution
for
this
is
modifying
the
clock.
You
have
in
open
telemetry.
There
is
a
in
the
provider
sdk.
There
is
a
way
to
modify
the
clock,
so
I
think
it
should
be
modified
there
using
something
like
an
ntp
when
you
create
it,
I
have
not
done
it,
but
I
think
it
should
be.
That
way
I
mean
not
adding
ntp
or
something
like
that
into
the
spec
itself.
A
B
Yeah
well
so
the
the
way
that
at
least
elastic
has
kind
of
solved.
This
problem
is
it'll
normalize.
The
data
coming
in
from
like,
like
another
service,
like
let's
say
like
like
a
web
browser,
has
an
agent
in
it
or
something,
and
it's
sending
data
to
the
backend
it'll
normalize,
the
the
span
times
based
off
of
the
server
side
span
times,
because
those
are
generally
going
to
be
more
accurate
than
some.
B
You
know,
users,
you
know,
watch
your
phone
or
laptop
or
whatever,
and
then
and
then
basically
time
stamp
everything
based
off
of
the
received
time
that
the
server
got
the
initial
request
from
rather
than
from
the
agent
or
rather
than
from
like
the
mobile
side
of
things.
Okay
and
yeah,
I'm
one
I'm
wondering
if
like
it
might
be
better
to
run
everything
in
offsets.
Like
I
don't
know,
yeah
I
mean
I
think.
A
You
could
you
could
change
the
clock
if
you
want,
but
you
could
also
I
mean,
send
attack
with
your
local
time
or
something
like
that.
I
don't
know.
A
Work
if
the
back
end
just
modifies
times
after,
because
the
rest
of
the
values
should
be
valid,
I
mean
just
the
start
and
end
time
should
be
modified,
but
I
think
it
could
also
be
done
changing
in
the
clock.
B
Yeah,
I
I
just
figured
since
this
is
I
mean
this
is
going
to
be
a
oh,
my
god,
I'm
not
sure
anymore
I'll
be
good.
I
feel
like
this
is
going
to
be
a
pretty
common
problem
with
any
sort
of
cross.
B
You
know
system,
but
it
might
not
be
as
bad
between
servers
but,
like
any
anything,
that's
sending
data
across
different
systems
that
are
depending
on
different
blocks,
they're
going
to
have
this
problem,
so
I
would
think
that
it
should
be
in
the
open,
telemetry
spec
on
how
to
deal
with
this
I'll
I'll,
bring
it
up
at
the
maintainers
meeting
and
see
if
they
have
anything
to
say
about
it.
A
B
Yeah,
I
I
wonder
how
like
how
well
that'll
actually
solve
any
problems
I
feel
like
at
the
end
of
the
day,
if
you
have
two
different
clocks
that
you're
depending
on
then
they're
always
going
to
be
they're
always
going
to
be,
you
know,
misaligned,
you
know
you
can't
ever
get
them
exactly.
A
Yeah,
that's
true
yeah
because
I
mean
probably
your
phone
is
synchronizing
every
day
yeah,
but
maybe
maybe
they
are
not
desynchronizing
but
just
have
a
constant
option.
Yeah.
B
A
That's
why
you
might
solve
it
just
by
adding
one
constant
time
I
don't
know,
but
probably
the
phone
should
be
so
be
very
precise.
With
time
I
mean
they
connect
with
the
gps
antennas
also,
and
they
must
have
a
very
specific
time
but
yeah
it's
right.
It's
it's
true
that
when
you
have
short
spans
you,
you
probably
will
have
a
millisecond
difference
with
the
server
and
if
you
have
like
one
millisecond
expanse
or
even
shortest
it
will,
it
will
show
like
crap
always
but
yeah.
A
There
is
nothing
in
the
spec,
especially
for
for
that
stuff,
except
maybe
changing
the
clock
is
the
only
thing
I
can
think
of
the
spec
that
could
work
out
of
the
box.
B
Yeah
but
yeah
we
yeah,
I
mean
yeah.
I
I
can
definitely
figure
out
a
solution
in
our
system
over
here.
But
if
this
feels
like
something,
that's
you
know
going
to
be
a
global
problem,
so
it
probably
should
be
defined
how
to
deal
with
this
in
open,
torment.
A
Okay,
next.
C
B
B
Yeah,
it
seems
it
seems
like
that's
the
case
right,
I
don't
have
a
totally.
I
don't
have
a
full
knowledge
of
how
like
the
trace
exporter
works,
but
I
think
it
just
uses
the
timestamp
from
the
you
know
from
the
system.
Clock
and
you
know,
assigns
it
that,
and
so
you
know
you
get.
The
back
end
has
its
own
clock
and
then
you
know
iphone,
has
its
own
clock
and
then
they're
not
going
to
necessarily
line
up.
C
A
A
A
Is
the
yeah
automatic
context,
preparation
to
child
process
yeah?
This
was
more
or
less
how
I
said
last
week:
it's
what
happens
if
you
have
a
process
that
is,
that
is
instrumented,
but
you
launch
it
as
a
child
of
another
process.
So
this
allows
you
to
use
the
instrumentations
of
the
child
process
as
as
part
of
your
instrumentation
of
the
parent
process.
A
So
basically,
what
I
added
was
an
environment
context,
propagator
that
it's
based
on
the
on
the
standard,
w3
c
trace
context,
propagator
that
the
only
things
that
does
is
taking
the
trace
parent
and
thresh
state
from
the
headers
that
it
will
create,
or
the
dictionary
created
from
the
b3c
and
move
it
to
environment
variables.
A
With
this
format,
it's
this
will
be
the
environment
variables.
So
what
it
does
is
that
if
you,
when
you
create,
let
me
show
here
yeah.
The
idea
is
that
when
the
tracer
starts,
it
checks
the
in
the
environment,
context
propagator.
A
If
it
has
in
the
processing
for
environment,
it
checks,
if
it
has
the
context
and
the
the
trace
context
and
the
trace
state,
and
if
it
does,
it
uses
it
when
it
creates
a
spam
in
the
spam
builder,
when
it
gets
the
parent
context,
what
it
does
is
if
there
is
no
parent
context,
but
it
has
the
the
environment
context,
it
will
assign
it
as
the
as
the
parent
context.
A
Just
like
that.
So
if
you
are
going
to
launch
a
process,
you
can
use
the
environment,
you
can
use
the
you
can
inject
the
environment
that
will
create
here.
Decay
in
the
carrier.
Add
it
to
the
environment
of
the
of
your
next
process,
and
it
will
automatically
will
take
it
as
a
parent.
B
A
Yeah
that
that
that
was
very
very
easy.
I
mean
it
was
just
starting
this
line
here
and
just
reading
on
the
when
it
initialize
and
you
can.
You
can
do
that
so
yeah
just
watch
to
to
show
this
this
small
changes.
So
now
there
is
a
in
the
shared
state.
There
is
a
launch
environment
context
that
will
probably
be
nil,
always
except
if
there
is
a
if
there
is
an
environment
set
so
yeah
that
that
will
will
work
for
really
any
process.
B
Yeah,
so
I
opened
up
a
spec
proposal
for
network
connectivity
for
monitoring.
You
know
what
what
the
network
state
for
the
mobile
device
is
so
it'll
capture
things
like
or
well
it'll
support
things
for,
like
the
carrier
name,
the
network
connection
type,
you
know,
2g
3g,
wi-fi
and
then
the
various
like
mobile
codes
for
like
country
and
and
that
sort
of
thing.
B
I
just
wanted
to
yeah,
let
you
guys
know
so
that
you
could,
if
you
want
to
chime
in
on
that
on
that.
B
B
Oh
yeah,
so
the
combination
of
the
mobile
network
code
and
the
mobile
country
code
can
they're
slightly
different
things.
So,
if
you
can
buy
you
find
them
that
there.
So
actually,
if
you
yeah
that
link,
there
shows
a
good
like
breakdown
of
everything,
but
combining
those
two
numbers
can
allow
someone
to
identify
exactly
like
what
what
cell
carrier
plus
like
what
like
a
a
cell
like
system
that
that
their
connection
is
using.
A
B
Working
at
new
relic,
they
we
had
some
customers
who
are
very
interested
in
this
sort
of
information
coming
out
of
their
mobile
device.
So.
B
B
And
then
the
other
ones
are
more
convenient,
like
the
iso
code,
like
gives
you
a
just
a
general
country,
information
which
is
good
and
then,
of
course,
the
carrier
itself
is
human-readable
carrier
information.
A
Yeah
yeah
I
yeah
also,
maybe
I
was
thinking
more
about
the
statues
of
the
network,
also
as
a
as
a
possibility
to
add
us.
I
was
thinking
when
I
read
about
network
conditions.
I
was
expecting
more
about
if
it
was
what
kind
of,
if
it
had
good
connection
but
connection
that
stuff,
oh
yeah
or
unconnected,
or
something
like
that.
That
was
I
what
I
thought
about
conditions
but
yeah.
B
Yeah,
so
I
guess
I
didn't
I
didn't
mention
in
this,
but
like
I
would.
I
would
think
that
yeah,
maybe
under
network
effective
type,
that
might
not
be
the
best
name
for
it,
but
under
that
it
would
say
if
it
was
disconnected
or
not.
B
Okay,
but
yeah.
I
don't
know
if
there's
a
way
to
access
like
the
actual
signal
strength.
That
would
be
interesting
to
see.
A
I
I
I
don't
really
know
that
that's
an
area
I
have
not
played
much.
I
know
they
are
doing
something,
some
with
that
in
the
data
with
the
ram
instrumentation,
but
with
real
user
monitoring.
They
are
using
something
about
the
network
and
that,
but
I
have
not
digged
much
deep
in
that
area
yeah,
but
I
think
it's
it's
great,
because
these
are
important
information
for
mobiles.
A
Definitely
great
okay,.
C
B
Oh
yeah,
I
mean
you,
can
you
can
detect
if
the
there's,
no
internet
connection,
sure
yeah
and
that's
that
was
part
of
the
intention?
I
didn't.
I
didn't
specify
that
in
the
actual
spec,
though
so
that
I
should
probably
update
it.
A
Yeah,
the
only
thing
with
that
is
that
apple
asks
you
not
to
check
for
your
network
connectivity
when
implementing
network
services.
A
Interesting
yeah
this
you
have
an
api
to
know
if
you,
if
it's,
if
you
have
connection
or
not
before
trying
something
but
apple
actively,
says
not
to
use
that
api
and
configure
your
network
request
to
wait
until
you
have
connection
and
things
like
that,
instead
of
actively
checking
that
status,
maybe
because
of
battery
consumption
or
something
like
that,
maybe
they
have
to
switch
on
some
some
silicon
just
to
save
the
status.
A
But
but
yeah
I
I
mean
it's
useful
information,
but
the
connectivity
thing
is
something
that
I
remember
having
read
apple,
asking
people
not
to
to
use
their
own
framework.
B
Yeah
yeah,
so
I've
written
up
some
instrumentation
for
this
and
the
intention
is
it
checks
when
the
network
connection
is
happening,
so
maybe
the
hope
is
any
sort
of
antennas
are
already
all
turned
on.
So
maybe
it
won't
be
very
expensive
but
yeah.
If
that's
a
problem,
that's
not.
A
Good
yeah,
okay,
so.
A
Okay,
then
we
can
finish
here
all
right.
Okay,
let's
see
if
next
week
we
can
have
a
a
better
release,
or
this
with
everything
just
ready
to
to
create
the
release.
Yeah
sounds.
B
A
Yeah,
I
hope
so
I
mean
it
it's
like
I.
I
have
that
as
a
as
a
I
want
to
reach
that
point
to
have
something
publicly
released.
We
need
more
people,
definitely
using
this
stuff,
because
we
are
finding
some
issues
that
should
be
shouldn't
appear
and
releasing
with
those
parents.
We
will
lose
yeah.
The
credibility
I
mean
the
project
if
you
are
releasing
source
code
or
code
that
has
not
been
well
tested.
A
B
Thanks,
that's
good
good,
seeing
y'all.
B
Have
a
good
thursday,
friday
and
weekend.