►
From YouTube: 2021-11-30 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
A
B
A
A
A
A
B
Should
we
go
ahead
and
kick
off
a
recap
and
and
get
on
with
things.
B
B
B
Still
open
waiting
for
waiting
for
approval.
B
I
don't
know
a
whole
can
of
worms
was
opened
by
this
fairly
simple
idea
of
adding.
B
B
So
I
think,
like
it
was
decided
that
service
name
is
kind
of
like
the
thing
that
that
helps
you
identify
like
an
application
in
a
distributed
system,
but
I
think
the
people
from
like
the
browser
side
and
just
other,
like
various
weird
clients,
feel
that
that
is
like
not
really
a
good
name
for
for
some
of
these
things.
B
So,
for
example,
for
yeah
for
for
browser,
app
is
kind
of
like
the
preferred
name,
and
these
would
be
kind
of
the
preferred
attributes
that
would
go
with
it.
B
And
then
device
would
be
for
like
an
in
iot
or
or
mobile.
B
So,
there's
just
a
lot
of
bike
shedding
about
how
all
these
things
should
line
up
properly.
I
guess
and
ultimately,
whether
or
not
like
how
baked
and
stable
the
resource
semantic
conventions
are-
and
I
think
I
don't
know
during
that
whole
discussion.
There
were
some
some
interesting
weird
things
I
think,
like
the.
B
The
the
sdk,
the
sdk
resource
spec,
is
like
stable
but,
like
maybe
the
semantic
conventions
are
not
so
it
does
seem
like
there's,
maybe
a
door.
The
doors
may
be
open
to
change
the
conventions,
but
there's
definitely
some
discussions
about
how
inconvenient
that
could
be
for
users
and
and
other
problems.
So
so
yeah.
I
don't
know.
If
I'm
doing
this
conversation,
all
I
want
to
justice
other
than
there
are
some
complications.
I
don't
really
know
what
what
directions
or
outcomes
there
there
will
be,
but
there
are
definitely
some
discussions.
A
So
interesting,
and
that
makes
sense
coming
into
open
telemetry
from
a
non-open
optometry
compliant
specification.
A
B
Yeah
I've
seen
the
component
terminology
out
there,
and
that
is
generic
enough
where
it
seems
like
it
could
work
for
most
things,
but
I
think
yeah,
I
think,
that's
a
problem.
It's
like
you
get
kind
of
like
a
generic
name
that
doesn't
really
apply
to
anything.
The
more
specific
you
get
the
more
it
starts
not
applying
to
other
things,
but
we'll
see
where
this
goes.
It
would
be
nice
to
just
know
a
name
of
an
attribute
that
you
can
point
to
to
say.
B
This
is
the
thing
that
that
generated
this
telemetry,
but
it
seems
like
it's
a
whole
bucket.
B
So
metrics
there
is
a
project
and
some
burn-down
stuff
happening.
I
think
the
biggest
thing
was
there's
this
pr
to
mark
the
sdk
as
mixed,
slash,
stable,
which
is
interesting.
But
if
you
go
through
it
some
parts,
the
the
full
spec
is
mixed,
but
there
are
like
parts
within
it
that
are
stable
and
then
you'll
find
parts
within
it
that
are
still
kind
of
experimental.
B
B
So
yeah,
that's
that
I
think
the
goal
is
to
kind
of
shore
up
what
is
kind
of
agreed
upon
before
people
disappear
for
the
holidays
and
then
kind
of
figure
out
what
needs
to
be
worked
on
after
this
act.
B
B
There
is
a
draft
pr,
at
least
in
go
for
the
proposed
exponential
histogram.
So
it's
the
aggregator
component
as
well
as
export
support.
E
No
not
for
me,
I've
been
just
like
for
reference.
My
last
week,
I've
just
been
moving.
I
just
finished
like
settling
into
a
move
on
the
weekend,
so
I'm
like
poking
my
head
up
this
week,
trying
to
figure
out
where
everything
is
kind
of
like
a
prairie
dog
looking
around,
but
no,
I
don't
think
I
think
the
last
time
I
went
to
it.
I
shared
my
updates
from
there.
I
do
intend
to
go
to
it
today.
So
hopefully
I
can
share
something
interesting
next
week.
E
B
E
Repo
yeah,
I
think
that
makes
sense
just
for
like
reference,
I
really
need
to
create
a
release
for
pretty
much
everything
this
week.
So
I'm
hoping
to
do
that
soon.
I'm
gonna
try
to
just
release
as
much
stuff
as
possible.
It
might
end
up
just
being
a
big
bang
release
of
just
like,
because
I
think
most
everything's
changed
at
this
point.
I
need
one
more
pr
to
go
into
the
api
for
deprecating
the
rack
indicator
so
that
we
can
eventually
remove
it
from
the
api.
E
A
The
q
polar
is
failing,
but
that
looked
good
to
me
when
I
reviewed
it,
I'm
not
sure,
maybe
there's
something
I
missed
in
the
review.
It
was
a
pretty
small
change
that
I
think
was
pretty
relevant
for
this
user.
All
he
was
doing.
He
was
essentially
doing
the.
A
I
guess
I
should
have
let
the
test
run,
what
sidekick
offers
which
is
like
you
can
disable
the
polling,
because
it's
incredibly
noisy
and
spammy.
He
did
like
a
metoo
of
that
for
q,
polar
it
was
pretty
uncontroversial,
but
me
I
will
why
don't
I
take
this
I'll
look
after
the
meeting
and
see
why
the
tests
are
failing.
I
thought,
but
that
would
be
the
only
one
the
others
are.
A
I
think,
could
use
yeah
or
just
not
need
review
or
need
feedback
implemented
or
whatever
is
that?
Are
you
is
that?
Okay,
if
you
yeah.
A
Okay,
well,
the
timing
works
out
good
then,
but
I'll
try
to.
I
don't
know
where
the
can.
We
push
up
changes
on
branches
of
kind
of
contributes
or
is
that
all
right
I'll
see
what's
wrong
with
it.
A
E
If
you
get
the
sense
that
it's
taking
longer
like
it's,
not
it's
not
an
issue
to
do
an
individual
release
for
that
gem
right
like
after
it's
done
it's
just
the
the
big
bang,
release
gets
really
messy
and
that
one
will
be
kind
of
painful,
but
doing
the
individual
one
will
be
pretty
easy.
Like
that's.
That's.
A
Worse
well,
either
way
I
could
have
sworn
it
was
green
when
I
reviewed
it,
so
I
don't
love
given
that
I
just
yelled
at
someone
last
week
for
proving
stuff
when
tests
were
failing.
I
don't
love
that.
I
just
did
that
karma,
but
yeah
I'll.
Look
at
that,
and
hopefully
we
can
sneak
it
in,
if
not
I'm
sure,
a
small
one's
fine
yeah,
the
others
are
in
various
states
of
disarray.
A
A
I
think
it's
pretty
close,
they're
missing
it's
doing
propagation
across
these
queues,
sqs
and
then
sms
doesn't
really
have
an
offer
opportunity.
So
sqs
can
do
in
inject
and
extract
sns.
You
don't
really
that's
just
like
sending
text
messages,
so
you're
not
really
going
to
extract
propagation
from
there,
but
they
are
missing
extraction
at
the
moment
and
it's
a
little
bit
confusing.
A
How?
Basically
one
issue
is
like
when
you
extract
context.
The
messages
come
in
about
the
messages
can
come
either
as
a
single.
The
you
can
receive.
One
message
in
which
case
extract
in
context
is
pretty
transparent
and
and
the
thing
that
you're
the
span
represents
processing
of
all
the
messages.
A
A
You
know
trace
and
they
can
be
from
various
things,
so
you
can't
necessarily
extract
a
child
context
for
that,
like
enveloping
operation
for
each
message
processed
the
operation
representing
each
message
process,
you
could
do
that,
but
they
don't
have
that
in
this
instrumentation.
All
they're
instrumenting
is
the
broad
message,
the
broad
batch
process,
so
that
was
my
feedback
was
like
if
we
want
to
go
down
the
path
of
and
and
so
if
they
want
to
go
down
the
path
of
doing
one,
there
needs
to
be
extraction.
A
A
Really
really
hacky
approach,
which
I
would
hate,
which
is
like
kind
of
just
porting
over
what
datadog
did,
which
is
like.
You
only
extract
context
when
it's
not
a
batch,
but
in
practice
nobody
uses
sqs,
like
that
everybody
batches
their
messages
in
sqs,
because
it's
more
performant
and
if
you
want
to
so
then
it
becomes
hard.
You
can
only
do
spam
links
if
you're
going
to
attempt
to
extract
a
batch
of
context
for
that
one
broad
operation,
you
can
only
just
say
like
okay.
A
These
are
all
maybe
spam
links,
but
you
couldn't
say:
there's
a
child
parent
relationship
and
then,
if
they
wanted
to
do
child
parent,
they
would
have
to
add
individual
patching
for
the
spans
that
represent
just
the
single
messages
of
the
batch
being
processed
so,
which
is
how
node
open
telemetry
node.
Does
it
currently?
A
So
I
was
like
this
is
kind
of
weird
and
there's
no
great
specification
here.
There's
no
specification
for
like
how
this
should
be
represented,
so
I
reached
out
to
anthony
mirabella
and
elita
and
anarag
from
aws
and
was
like
hey
like
these
will
end
up
being
your
users,
who
probably
you'll,
be
the
one
who
has
to
handle
the
support
tickets
on
this
stuff.
A
Like
do
you
guys
have
any
feedback
anthony
had
some
interesting
feedback
that
I
think
is
relevant,
which
is
that
the
group,
the
the
span
that
represents
pro
the
entire
batch
you
assign
as
span
links
and
then
the
spans
that
represent
the
individual
messages
you
represent
as
childs
of
the
you
know,
whatever
the
extracted
context
was
so
ascent,
which
is
interesting,
might
get
represented
weird
in
uis,
but,
like
maybe
the
semantically
correct
way-
and
I
was
like
oh
that's
interesting
and
then
anarag
said:
why
do
we
even
do
that's
propagation?
A
Doesn't
make
sense
for
sqs
at
all?
We
shouldn't
be
doing
it.
It's
just
an
rpc
aws
sdk
is
just
an
rpc
framework
and
he
then
he
was
like.
Actually
I'm
not
sure.
Let
me
look
into
it.
So
that's
kind
of
where
it's
like
left
is
there's
this
weird
open-ended
issue
with
extraction.
A
That's
kind
of
goofy
and
I
think
there
needs
to
be
some
resolution
to
just
like
having
some
way
of
doing
it,
but
I
don't
necessarily
want
to
like
end
up
implementing
or
supporting
something
that
just
ends
up
being
this
like
totally
off
the
rails,
no
pun
intended
way
of
doing
propagation
for
aws,
sdk
and
then
like
aws
down.
A
The
line
is
like
that's
not
how
we
want
to
do
it
so,
okay,
that
was
my
that's
now
you're
all
up
to
date
on
what
I
did
on
friday,
which
is
looking
into
this.
So
I
thought
it
was
interesting,
but
it's
like
I
don't
really
know
the
right
answer,
but
there's
definitely
an
unresolved
part
of
this
instrumentation.
In
my
mind,.
B
Yeah
thanks
thanks
for
that
summary.
Just
one
question,
maybe
more
as
ss
is
currently
structured
context.
Propagation
aside,
do
we
end
up
with
a
span
for
the
batch
and
then
a
span
for
each
individual
message.
A
A
That's
what
so
node
has
this
really
interesting
helper
utility?
That's
called
like
it's
just
called
like
a
pub
sub
like
spam.
I
forget
what
it's
called,
but
it's
essentially
it's
just
a
utility
where
you
pass
in
a
where
they
they
do
it
per
message.
They
create
a
span
per
message
processed
instead
of
a
representing
the
group,
but
they
use
some
one
like
it's
like.
I
don't
necessarily
think
node
is
like
the
right
way
to
do
things.
A
A
Anyway,
I
think
I
think
it's
important
reflected
the
way
we
do
for
all
our
other
for
kafka
and
whatever,
where
you
get
trip.
Triplet
of
you
know,
span
link
or
none,
I
think
is
ultimately
the
right
way
to
offer
this
to
people.
But
it's
really
hard
to
give
like
that
optionality
here
and
then
also
like
support
all
the
different
things.
So
I
don't
know.
D
I
imagine
this
is
one
of
those.
It
depends
situations
because
how
how
the
service
that
is
using
the
queue
and
what?
What
do
the
messages
mean
and
what
does
it
mean
for
a
bunch
of
stuff
to
get
bashed
together,
and
do
you
want
the
individual
messages
in
a
batch
to
get
associated
with
the
trace
that
we
have
customers
who
are
like
I'd
like
to
see
that
from
the
thing
that
published
the
message,
I
would
like
the
processing
of
that
message
to
be
included
in
that
trace,
even
though
it
makes
for
weird
long
traces.
D
A
But
if
theory
and
practice
are
you
know
like,
and
certainly
my
ex
experience
previously
is
people
from
we
want
the
same
thing
with
sidekick,
even
though
that's
not
the
right
way
to
represent
psychic
or
whatever,
and
so
like
yeah
it'll
depend,
and
eventually
we
have
to
kind
of
give
that
optionality
to
folks,
but
it's
quite
hard
to
do
that
because
they
come
in
basically
because
it
comes
in
batches.
The
batches
make
it
almost
as
though
you
really
can't
represent
a
batch
as
anything
except
a
spam
link.
Well
you
mean
the
messages.
A
This
is
diagram
the
different
output,
the
different
mental
models
in
a
visual
like
excali
draw,
or
something
would
make
this
discussion
easier
and
also
make
communic
like
anthony
mirabella,
got
it
from
abs
because
he
just
gets
everything
but,
like
you
know,
to
the
implement
to
the
person
making
the
contribution,
I
think,
if
I
can't
explain
it
clearly
to
you
all
who
unfortunately
spend
way
too
much
time
in
this
field,
like
I
don't
think
they'll
understand,
so
I
think
maybe
I
need
to
diagram
the
different
options
and
that
would
help
okay.
I
like
pictures.
B
Cool
yeah.
No,
then
that
sounds
good.
That
would
be
super
helpful
and
thanks
for
I
don't
know,
kind
of
taking
us
on
and
and
reaching
out
to
the
folks
at
aws.
I
think
that's
definitely
the
right
move
and
yeah
kind
of
by
by
talking
through
this.
I
think
I
think
we'll
ultimately
arrive
at
at
a
pretty
good
solution,
but
definitely
some
discussions
seem
to
be
had.
B
Yeah,
there's
there's
no
reason
to
drag
this
meeting
on
if
there
isn't
anything
else
to
talk
about,
but
if
there
is,
we
should
cover
it.
Well,
while
we're
together.
A
A
quick
shout
out
ariel's
put
a
lot
of
time
and
thought
into
the
version
compatibility
pr.
A
I
won't
go
into
detail.
I
think
it's
a
novel
approach
but
worth
of
viewpoints.
So
if
people
are
interested
a
review
would
be
helpful.
I've
left
comments
on
there.
I
won't
go
into
it,
but
I
know
he's
keen
and
he's
on
vacation
right
now,
but
he
would
bring
that
up
if
I
was
if
he
was
here.
B
Cool
thanks
for
bringing
that
up.
I
think
he
brought
it
up
last
time
too
so
yeah
we
should
get
some
eyes
on
this.
B
Awesome,
cool
yeah.
Everyone
have
a
good
week.
I
guess
we'll
be
awaiting
awaiting
the
big
bang
release.
So
if
there's
anything
needed
reach
out
over
slack
fine.