►
From YouTube: 2021-03-24 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
B
C
Yeah
and
since
I
start
to
look
at
this
stuff,
I
I
kind
of
go
enjoy
the
meeting,
so
I
don't
forget
you
know
because
it's
like
oh,
I
have
two
minutes
more
and
then
I
decided
to.
C
B
B
C
Yeah
but
but
but
it's
good,
we
we
wait
a
little
bit
more.
Let's
wait
until
five.
D
B
C
C
C
Is
this
beginning
of
the
meeting?
If
people
I
don't
know
if
people
see
the
recordings,
but
it
should
be
very
kind
of
boring
and
open
just
us
here,
waiting.
C
Yeah
yeah,
I
I
think
I
went
back
to
some
stuff
like
in
the
beginning.
We
had
no
talk
about
a
lot
of
things
about
the
profiler.
I
think
I
went
back
to
some
of
those
some
time
ago,
but
then
I
we
should
check
these
statistics
on
those
videos.
You
know.
C
Okay,
I
just
have
the
the
I
would
like
to
check
the
pr's
and
be
sure
that
there
are
nothing
kind
of
blocking
any
progress
from
those,
and
that's
the
only
thing
in
my
mind
and
last
week
I'm
trying
to
have
the
guy
that
implemented
the
event
pipe
for
the
collector,
the
open,
telemetry
collector
to
do
a
demo
for
us,
and
I
think
this
is
something
that
we
we
want
to
keep
an
eye
on
and
build
on
top
of
that,
as
we
mentioned
many,
I
mentioned
at
least
many
times
before
so
yeah.
C
And
I
I
want
to
be
sure
that
my
screen
sharing
is
working
because
sometimes
yeah
it
is.
B
C
Okay,
this
was
some
initial
discussion.
E
Yeah
on
this
one,
there
is
one
thread
that
was
a
little
bit
lengthy,
that
I
think
we
wanted
to
bring
discussion
to
this,
the
sig
group,
but
with
people
being
gone,
we
might
want
to
hold
off
on
that
discussion.
C
Okay,
do
you
remember
which
thread
is
that
so
so.
E
The
the
main
thing
I
so
I
went
in
and
started
reviewing
it
in
terms
of
just
like
implementation
details,
but
it's
pretty
breaking
for
data
dogs,
so
we
might
have
to
consider
well.
We
have
to
talk
about
the
merits
of
making
the
changes
and
then
having
data
dog
be
out
of
sync
or
trying
to
abstract,
like.
E
B
B
C
Let
me
if
we
had
an
additional
data
structure.
That
is
also
a
problem
and
just
for
me
to
understand
the
internal
problem
kind
of
if
we
keep
the
old
structure
and
have
one
on
top.
Is
this
still
a
problem?
I'm
not
saying
should
for
us
to
do
that.
I'm
just
kind
of
trying
to
understand
the
the
scope
of
the
impact.
E
Yeah
so
at
least
so
the
trace
id
as
it's
represented
in
our
internal
apis
and
external
as
well.
People
are
using
nuget
package
it.
The
trace
id
is
represented
as
a
as
a
ulong,
so
even
just
changing
that
structure
is
a
breaking
change.
E
What
your
one
idea
that
you
know
could
work
is
adding
sort
of
an
upper
trace
id
just
basically
adding
a
new
year,
long
fields
that
we
keep
internally
and
could
add
to
the
internal
interfaces,
while
keeping
just
one
sort
of
view-long
trace
id
in
the
public
apis,
at
least
for
us,
since
we
we
need
to
keep
those
the
same
for
now
and
until
we
do
some
sort
of
planned
version,
change
that
could
work,
it'll,
probably
just
be
duplicating
quite
a
bit
of
code,
but
especially
if
it's
only
going
to
be
used
for
hotel.
E
E
B
That
as
well
as
well-
let's,
let's
say
someone
has
an
order
out
of
instrumentation
package
and
down
is
the
new
one
and
he
creates
this
bank
context
using
the
new
constructor
and
the
old
dll
is
loaded
in
run
time.
It
will
break.
C
But
but
but
then
the
old
one
is
just
going
to
see
the
lower
part
right.
E
Got
it
in
this
case
in
this
case
that
would
still
work
because
well
it
just
depends
on
sort
of
how
the
assemblies
work
out
and
for
the
way
it's
currently
constructed
is
where
that
main
datadog
trace
package,
the
one
that
carries
all
those
definitions,
that
one
is
the
one
that
they
compile
against
so
they'll
definitely
be
using
the
newer
one
and
the
older
automatic
instrumentation.
A
E
Could
set
a
default,
I
mean
we
could
set
basically
modify
the
existing
constructors
to
null
basically
set
the
higher
trace
id
as
zero,
but
that
that
would
actually
work.
E
I
think
we'd
probably
take
some
poc
or
exploration,
but
that
seems
possible.
C
Yeah
and
I
I
also
want
to
consider
the
other
possibility
that
you
said
about
when
you
eventually
do
a
version
bump,
because
then
we
we
could
perhaps
do
this
switch
on
open
telemetry.
C
We
you
don't
pick
up
this
part
and
then
when
there
is
a
version
bump
on
your
side,
then
you
pull
because
then
it's
a
breaking
change.
C
Just
just
for
me
to
to
have
an
idea
is
this
something
that
you
think
is
work
about?
You.
C
Of
course,
then
we
have
to
batch
those
breaking
changes,
but
we
could
do
some
of
like
this
one.
You
don't
pull
this
one
for
some
time,
but
then,
when
you
do
your
version
bump,
then
you
pull
this
one
and
we
are
exactly
on
the
same
page.
This
is
also
workable.
C
C
Yeah-
and
this
puts
a
little
bit
a
burden
on-
let's
say
the
the
people
doing
the
the
upstream
pool,
but
that's
something
that
we
can
handle
for
some
time.
You
know
so
I
I
think,
as
I
said
I
on
one
hand,
I
don't
like
to
keep
blocking
the
the
pr,
so
we
can
make
progress,
but
I
think
we
need
to
kind
of
come
to
an
agreement,
at
least
for
the
short
term.
I
think
long
term.
C
We
eventually
need
the
tracer
to
kind
of
be
on
its
own
thing
and
eventually
we
need
to
support
activity
so,
but
but
this
is
much
more
long-term
and
on
the
short
term,
I
just
want
to
kind
of
be
able
to
kind
of
be
sure
that
they
we
don't
stall.
You
know
so,
do
you
want
to
take
some
time
and
get
back
to
us
via
slack?
Whatever
about
this
alternative
of
perhaps
do
the
breaking
change.
We
handle
the
merge.
C
C
All
right
so
I'm
gonna
be
so
for
now,
let's
hold
off
a
little
bit
on
merging
that
one.
Just
we
we
have
confirmation,
but
I
I'll
try
to
note
this
on
this
pr.
So
we
are
aware
that
kind
of,
because,
from
a
perspective,
kind
of
open
source
projects,
I
don't
like
when
we
have
dprs
kind
of
going
for
a
long
time,
and
there
is
no
updates.
You
know
we
have
to
eventually
make
a
decision
on
top
of
that.
C
I
think
hotel
started
the
last
that
we
mentioned.
I
think
robert
had
mentioned
that
we
we
had
a
way
forward
left
meeting,
but
I'm
not
sure
if.
E
Yes,
oh
yeah,
so
this
was
actually.
This
is
the
thought
I
was
thinking
about,
because
I
suggested
just
making
the
api
compatible
by
keeping
the
api
and
just
adding
new
one
and
just
making
it
backwards
compatible.
So
we
could
consume
it,
but
then
we're
just
thinking
long
term.
E
Is
that
the
right
solution,
and
so
this
is
where
robert
proposed
a
couple
of
options.
Can
you
scroll
up
a
little
bit.
E
Yeah,
so
this
was
that
this
is
what
I
was
thinking
of.
I
was
just
number
one:
just
making
it
backwards
compatible
two
is
leaving
the
backward
compatibility
just
burden
on
on
datadog,
and
that
was
also
you
know
a
fine
option
and
then
three
was
trying
to
use
this
as
a
a
jumping
point
to
better
describe
the
api
that
hotel
wants
moving
forward,
and
then
that
would
you
know
likely
introduce
some
sort
of
you
know
forking
of
some
of
the
code
from
what
we
currently
have.
C
I
see
so
so
in
the
end,
it's
it's
kind
of
a
similar
situation,
but
the
idea
here
that
we
want
is
really
preserving
the
backwards
compatibility
right
so
yeah.
So
I
think
we
need
to
to
decide
if
we
are
going
to
do
kind
of.
C
In
similar
way
that
we
discussed
it
for
the
trace
id
and
then
and
then
this
kind
of
thing
should
be
batched.
So
when
you
eventually
do
the
version
bump,
then
you
can
get
this
and
this,
but
on
the
other
hand,
I'm
starting
to
get
afraid
that
we're
going
to
have
a
lot
of
those
you
know
yeah.
It
does
seem
like
that.
C
Yeah,
okay,
yeah.
I
think
I
think
if
we
manage
to
get
because
the
next
thing
that
I
see
is
about
having
events,
but
events
is
an
addition,
is
basically,
let's
say,
tags
with
time
stamps
and
that
is
used
for
randomly
recording
exceptions,
for
instance,
that
I
don't
think
it's
a
problem,
because
this
is
a
addition.
E
Do
we
have
a
list
of
the
different
kind
of
big
important
work
items
like
the
events
and
the
span
choice
id
that
need
to
be
tackled.
C
We
don't
have
this
list,
but
we
can
put
one
together.
I
I
think
we
don't
have
this
list
ready
kind
of,
but
we
should
put
this
list
together
and
I
should
send
to
you.
E
Yeah
that
might
help
for
us
to
think
about
that
planning
that
yeah.
C
Okay,
I'll
try
to
come
up
with
this
list.
We
already
know
status,
trace
id
and
additional
events.
C
C
E
So
I'll
need
to
finish
looking
at
it.
I
started
writing
a
couple
notes,
but
yeah
it
adds
a
new
propagator
feel
to
the
tracer,
which
is
definitely
fine.
I
might
yeah
sorry
I'll
just
need
to
finish
reviewing
I
I
started,
but
I
haven't
finished
cool.
C
And
this
is
just
something
that
escaped
it
from
the
first
batch
of
renames.
So
I
think
we
are
good
here.
Okay,
so
we
have
those
trace
id
and
status.
That
are
the
big
ones
and
we
give
time-
and
I
think
uv
is
like
so
we
kind
of
can
make
progress
and
be
sure
that
those
tool
doesn't
get
blocked
forever.
A
C
I
see
dave
they've
joined
the
meeting,
just
just
so.
You
know
I'm
asking
pablo
to
come
next
week
to
do
a
demo
from
using
the
collector,
so
you
may
be
interested
in
in
seeing
that
you
know.
A
C
Yeah
yeah,
so
I
I'll
send
a
a
mail
to
you
if
I
have
a
confirmation
that
pablo
is
really
coming
next
week,
yeah,
okay,
so
I
I'll
send
my
mail
to
you
and
you're
looking
for
the
invitation
to
anyone.
If
you
are
okay
thanks
all
right,
that's
what
I
had
do
you
guys
want
to
bring
something
up
or
it's
time
for
zach
to
roll
their
credits.
E
I
have
nothing
else,
but
I
am
super
appreciative
of
the
work
you
guys
are
doing
and
I'll
you
know
try
my
best
to
make
sure
that
we
communicate.
If,
like
how-
and
you
know
if
things
are
blocked-
to
try
and
unblock
you,
because
I
really
appreciate
all
the
work
you're
putting
in.
B
Another
one
thing
from
robert:
I
think
that
he
was
the
one
that
picked
it
up.
There
are
a
lot
of
static
things
in
the
instrumentation
code
and,
like
there's
a
question,
if,
if
there's
any
way
that
the
methods
that
the
instrumentation
could
have
some
somehow
be
injected
with,
for
example,
tracer
and
not
use
the
singleton
one,
because
it's
really
hard
to
test
after
we
have
static
code
and
there
are
a
lot
of
new
things
without
the
instrumentation
namespace,
I
think,
but
that
there's
just
static
classes
or
aesthetic
matters.
E
B
F
F
B
Share
your
screams
give
me
a
second,
because
I
have
to
find
something:
okay,.
B
E
Yeah,
so
that's
one
example
that
one
we
could
again
do.
One
of
the
ideas
from
the
previous,
like
the
first
sort
of
span
decorations,
was
to
make
that
configurable
in
the
tracer
and
that's
that's
a
change.
We
can
definitely
do.
E
Okay,
we,
we
just
finished
a
release
earlier
this
week,
so
we
have
like
a.
I
don't
know
if
all
the
changes
so
far
have
been
pulled
up
in
that
upstream
pr,
but
there's
definitely
a
lot
there.
A
E
Right
now,
I'm
gonna
go
back
and
now
that
we've
kind
of
released
that
start
porting
that
list
of
changes
that
we
compiled
before
so
we
can
get
back
to
being
even
with
hotel.
C
Cool
cool,
okay,
very
cool
and,
by
the
way,
zach
thanks
very
much
for
really
participating
in
the
project.
They
were
being
instrumental
for
this
thing
to
really
happen.
Thank
you
very
much.
C
C
Live
cool.
We
we,
we
forgot,
the
the
the
conversation
that
should
happen
right
now,
so
people
have
to
watch
until
the
credits
are
completely
done.