►
From YouTube: 2022-05-17 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
Yeah
yeah,
that's
interesting.
We
I
used
to
live
in
an
apartment
and
we
had
sort
of
like
these
two
door.
It
was
like
two,
like
it's
called
french
doors,
yeah
side
by
side
and
they
had
a
frame.
It
was
like,
probably
a
building
from
the
19th
century
sometime
and
we
bought
two
bookshelves
to
put
on
either
side
of
the
door
and
we
built
the
we
built
them.
The
ikea
billy,
six
foot
tall
bookshelves.
A
C
A
B
A
Okay,
so
there's
also,
I
live
in
this
neighborhood
where
they're
I
live
in,
like
kind
of
a
ranch
house
that
was
built
in
the
mid
50s,
it
was
just
like:
hey
you
fought
in
world
war
ii.
Come
live
in
a
ranch
house
near
you
know,
yeah
in
denver.
You
can
have
a
ranch
house
for
like
not
that
much
money
and
then
that
people
are
just
knocking
them
down
and
building
these
like
enormous
monstrosities
of
houses,
which
is
good
for
them.
A
But
there's
one
there's
one
that,
like
you
know
on
any
block,
there's
one
or
two
under
under
underway
and
then
one
of
them
across
the
street
from
a
park.
They
just
poured
like
the
concrete
basement
foundation
and
it
really
does
not
look
level
like
just
just
like
on
the
face
of
it.
If
you
stare
at
it.
A
You're
like
that
doesn't
something
seems
wrong
here,
but
I
kind
of
want
it's
totally
fenced
in,
but
I
want
to
like
get
in
there
with
a
level
and
see
if
it's
like
a
matter
of
perception
or
if
it's
like
they
actually
messed
up,
because
that's
that's
a
big!
That's
a
big
boo-boo!
If
you
pour
the
foundation,
it's
not
level,
I
I
would
imagine,
but.
C
A
B
E
E
E
B
B
E
E
Weight
problems
all
right,
that's
definitely
one
way
to
solve.
It
could
also
just
you
know,
like
burn
the
whole
place
down,
while
I'm
at
it
like
that'll,
also
solve
it,
while
we're
on
solutions
that
will
technically
work,
but.
B
Spec
sig
definitely
not
brought
to
you
by
tito's
corn
vodka
because
they
don't
know
that
we
exist,
but
I'm
sure
it's
fun
for
the
whole
family.
Friends,
family
extended
family
all
right.
There
was
not
a
whole
lot,
probably
to
mention
from
the
spec
sitting,
so
we
should
probably
be
able
to
get
to
this
pretty
quickly
and
on
the
agenda.
E
B
B
B
A
E
B
I
think
that's
true,
but
I
think
to
to
actually
reach
stability
and
say
that
you
implement
the
semantic
inventions
you
must
comply.
I
guess
with
this
and
just
looking
at
the
comment
which
has
been
helpfully
edited,
I
feel
like
this
makes
kind
of
sense
required
will
be
like
a
must.
Conditional
will
be
like
must.
If
yeah
edition
recommended
will
be
a
should
and
optional
will
be
a
may,
so
it
is
still
kind
of
more
or
less
using
rfc
blah
blah
blah.
B
Then
we
kind
of
quickly
got
into
metrics
and
there
was
some
update
of
the
exponential
histogram.
There
was
a
lot
of
words
said,
but
nothing
actually
linked
to
it,
and
I
kept
hearing
some
talk
about
2
to
the
55
versus
2
to
the
64..
There
was
something
that
was
going
to
leave
you
between
those
two
goal
posts
and
one
that
would
put
you
in
the
255
range
was
sub
optimal.
For
obvious
reasons,
I
should
have
listening
a
little
bit
more
closely
on
that
one.
B
Oh,
I
just
want
to
read
my
ass.
We
will
keep
this
quick,
but
there's
like
the
there
are.
There
are
things
that
actually
have
units
when
you
look
through
like
the
metric
specs
and
then
there's
these
things
that
have
like
these
pseudo
units
and
then
there's
you'll
often
just
see
like
one
as
like
a
unit,
and
it
has
always
been
a
little
bit
mysterious
to
me
what
any
of
this
actually
means,
but.
B
I
think
in
general,
if
you
have
like
a
fractional
unit
and
you
are
receiving
telemetry
from
multiple
sources,
it's
like
you
kind
of
those
will
help
you
combine
stuff
and
aggregate
things
coming
from
more
multiple
sources
and
in
theory
this
one
unit
should
help
you
do
that
as
well,
but
apparently
it's
kind
of
is
often
ambiguous
and
when
and
in
the
ambiguous
cases,
it's
not
helpful
for
for
these
reasons,
but
I
think
somebody
was
asking
about
having
like
percent.
B
Percent-
and
I
think
it
was
kind
of
brought
up
in
the
in
the
context
of
one
being
kind
of
ambiguous-
is
this
100
of
something
or
is
this.
B
B
B
F
F
Not
me
directly,
but
I
have
some
teammates
that
are
trying
to
tune
in
and
participate
in
what
basically
front-end
instrumentation
will
look
like
yep,
I'm
in
the
same
boat.
What
about.
B
F
B
E
A
E
E
B
B
I
have
been
kind
of
interested
because
I've
dabbled
in
the
js
sig
a
little
bit
and
I
think
they're
yeah,
like
in
open
telemetry,
at
least
the
the
js
work
has
been
kept
abstract
enough,
that
it
can
be
used
in
the
browser
or
kind
of
back
end
for
node
type
stuff.
B
B
The
use
cases
tend
to
be
a
little
bit
different
and
like
code
sizes
in
browser
are
very
important,
being
like
very,
very
lightweight
in
the
browser
is
way
more
important
than
than
a
back-end
app,
and
I
know
at
some
of
the
various
apm
providers.
It's
like
their
browser
framework
is
very
different
than
kind
of
the.
The
data
emitted
from
a
browser
is
very
different
than
the
data
emitted
from
a
backend
application.
B
F
I
was
going
to
say
there's
this
whole
philosophical
debate
about
where
do
you
start
a
trace
in
activity
in
a
browser
in
a
in
a
web
app
and
like
yeah,
which
is
why
there
was
the
whole
talk
about
a
rom.
A
fourth
rum
signal
that
like
well.
Maybe
we
can
just
do
semantic
conventions
on
our
three
signals,
and
so
yes,
if
you're
interested
tune
in
to
the
hotel,
js
sig,
and
I
think
there
might
be
a
specific
sig
about
front-end
instrumentation.
B
E
Definitely
gonna
have
to
like
go
catch
up
and
understand
why
why
we're
obsessed
with
this
and
or
mail,
all
of
you
bottles
of
tea?
G
It
has
absolutely
nothing
to
do
with
with
ariel
and
the
hat
I
believe,
hey
speaking
of
which
brought
you
by
adidas
athletic
wear.
F
Drinking
vessel
of
choice
of
rubyists
everywhere.
F
Know
I
will
bring
up.
I
saw
the
the
otop
grpc
exporter.
Has
a
branch
open
kind
of
whippy?
I
started
looking
at
it
this
morning
just
to
keep
up
since
I
have
cares
about
that
thing.
I
will
try
to
contribute
to
a
thing
that
I
care
about
yeah.
I
don't
have
any
comments
here
for
this
venue,
but
it's
on
my
radar.
F
B
Yeah,
I
think
robert
has
laid
out
the
scaffolding
and
would
be
happy
for
anybody
to.
F
B
F
One
of
his
comments,
francis,
had
made
a
comment
about
hey,
is
this
that
right
and
he
goes?
I
look
forward
to
somebody
who
cares
when
they
pick
this
up.
Resolving
this
issue
is
my
summary
on
that
and
so
I'd.
I
hope
to
be
one
of
those
people.
E
Yeah,
I
think
I
think
the
pr
needs
some
things.
It's
definitely
very
whippy,
as
you
mentioned,
I
don't
think
it
has
any
like
error,
handling
or
retries,
or
anything.
Some
of
that
is,
would
definitely
keep
it
from
being
what
I
would
call
production
ready.
So
I
think
that's
probably
one
area
that
could
probably
use
some
love
is
working
on
that
kind
of
logic.
E
F
Stupid
hardening
of
network
coms
yeah
yeah,
so
I
don't
have
anything
more
useful
to
say
other
than
that
one's
on
my
radar
awesome.
B
About
the
extraction
of
the
instrumentation
registry.
A
A
I
also
made
a
doc
that
just
attempts
to
enumerate
the
different
combinations
of
gems
that
could
exist.
I
don't
know
if
all
of
those
like
that
matrix
represents,
like
actual
things
that
are
possible
to
install,
but
I
think
so
so
maybe
I'll
try
to
get
robert
to
comment
on
that
doc
to
just
localize
all
the
concerns
in
one
place,
but
also
maybe
he'll
comment
on
the
pr,
but
it
seemed
reasonable
to
me.
A
A
B
B
B
B
On
the
question
about
the
resource
detectors,
do
they
use
the
registry?
I
I
know
for
sure
the
instrumentation
does
I'm
not.
G
D
B
D
D
D
But
you
know
what's
kind
of
quirky
about
this.
One,
though,
is
like
that
that
gcp
one
requires
you
to
put
the
gcp
gem
in.
I
don't
think
that
it's
a
runtime
detectable
one.
I
could
be
wrong
about
that,
but
but
I
think
in
general,
like
our
resource
detectors,
are
a
little
anemic
and
I
think
we
had
like
one
contribution
that
somebody
wanted
to
make
which
was
like
a
rails
resource
detector,
but
it
never
got
merged.
D
B
I
think
maybe
they
will
in
time
I
know,
there's
quite
a
few
in
the
js
repo
just
from
from
what
I've
seen.
D
D
F
D
Work
like
yeah.
I
think
it
would
work
the
same
way
that
you
would,
with
the
instrumentation,
only
it's
a
little
more
quirky,
because
it's
like,
if
you
have
like,
like
the
gcp
detector,
we
looked
at
the
implementation,
it's
relying
on
the
gcp
api
gem
and
then
the
aws
sdk
one
would
I'm
assuming
that
that
would
do
something
similar.
I
don't
know
like.
F
A
F
Fraught
with
peril,
like
the
implementation
of
how
does
it
detect,
would
probably
have
to
be
sort
of
battle
hardened
so
that
it
doesn't
take
down
your
process
looking
up
like
stuff
for
us
to
auto,
run
it
like
auto
instrumentation,.
F
If
you
want
to
know
what
things
about
your
infrastructure
make
it
an
infrastructure
concern,
not
a
an
application
level
concern
yeah
and
all
of
it
is
always
a
little
tricky,
but
not
everybody
runs
in
coop.
So.
B
I
think
there
is
definitely
a
lot
of
future
possibilities
for
making
these
resource
detectors
a
little
bit
more
automatic.
But
I
think
the
question
was:
do
they
depend
on
the
registry
and
do
we
need
to
worry
about
it?
For
ariel's
registry
gem,
pr
yeah.
D
A
So
don't
forget,
we
do
use
the
gcp
resource
detectors
here
at
shopify
just
select,
so
you
could
delete
it.
No.
B
Okay,
so
I
think
robert
had
mentioned
in
slack
that
that
he
would
review
this
at
some
point.
So
I
feel,
like
I'm
happy
to
review
this,
and
I
think
this
is
the
right
approach
and,
as
I
was
saying,
the
worst
thing
that
will
happen
is
a
warning
for
a
user
running
kind
of
like
a
mismatched
like
an
older
sdk
with.
B
A
What
is
on
my
mind?
Okay,
yaniv
d
opened
up
a
pr
related
to
like
context
propagation
in
sqs,
and
I
mentioned
it
in
slack.
It's
1248..
If
other
people
have
more
important
stuff.
We
can
talk
about
that.
A
I
think
eric
was
the
last
person
that
paid
attention
to
this
and
eric
is
like
not
currently
sleeping
and
is
on
leave
right
now.
So
I
don't.
A
I
haven't
been
party
to
any
of
these
discussions
yet
and
I
think
we
thought
there
was
some
discussion
in
one
of
these
cigs
a
while
back
but
like
what's
the
right
way
to
do
context
propagation
for
sqs.
I
don't
know
if
people
have
thoughts,
I
I
was
I
was
like
we
gotta
read
this
pr.
It's
been
open
for
a
little
bit,
but
I
don't
you
know
the
code.
A
Is
the
code
is
not
going
to
be
that
complicated,
but
like
what
the
philosophy
around
context
propagation
is
is
going
to
be
more
important,
so
I
was
wondering
if
anyone
cared
and
if
they
care
what
their
thoughts
are.
A
A
Well
so
yeah,
I
don't
I
not
the
mechanism
of
context
propagation,
I
don't
think
is
really
at
issue,
so
maybe
I
misspoke,
I
think
it's
just
like
you
know.
Do
we
want
it
if
you
shove
a
message
into
sqs
like
when
someone
picks
it
off
the
queue
like?
Do
you
want
that
to
be
a
continuation
of
the
span?
Do
you
want
that
to
be
a
link
span
like
what
what's
the
approach
there,
but
I
only
took
a
cursory
look
so.
F
Having
this
come
up
in
in
our
so
speaking
as
a
honeycomb
person,
vendory
type,
our
instrumentation
has
generally
left
this
up
to
the
instrumenting
end
user.
There
were
ways
to
to
get
the
current
contexts.
Propagation
data
like
what
we
would
put
in
the
header
for
an
http
call.
F
You
can
get
as
a
value
and
you
could
choose
to
put
that
in
your
message,
payload
and
then
in
your
job
or
whatever,
that's
taking
a
message
off,
choose
to
go
and
extract
it
out
of
the
payload,
so
it's
very
much
opt
in
and
then
opting
in
you're
choosing
what
form
it
comes
in.
Does
it
continue
to?
Is
it
a
span?
That's
part
of
the
participating.
F
It
a
link:
is
it
an
event
because
message:
cues
are
so
like?
How
are
you
using
the
message?
It
depends
on
the
shape
you
want.
The
instrumentation
of
that
handoff
to
take
depends
on
like
how
you're
using
a
message
so
not
having
read
this
code.
My
initial
recommendation
would
be
make
this
configure
not
configurable.
D
F
F
Yeah
that
that
would
be
like,
if
there's
an
api
in
the
instrumentation,
that
would
say
like
what
to
do,
and
maybe
it
defaults
to
make
a
span
that
participates
in
the
trace
and
people
who
don't
want
that
you
can
do
a
global
nope.
All
my
messages
are
events
and
do
that
or
something
at
the
enqueue
or
the
dq
choose
so
that
you
have
a
default
and
then
you
can
override
in
particular
cases
where,
generally
you
want
them
to
be
spans,
but
in
some
cases
you
want
them
to
be.
F
Events
is,
is
my
hand
wavy
high
level.
This
is
what
we've
done
or
generically
message
queues,
because
there's
not
a
good
convention
for
us
to
decide
ahead
of
time.
F
So
this
is
another
one
I'll
put
on
my
radar
to
read
more
closely.
Since
I
have
thoughts,
I
will
try
and
share
them
so.
D
A
D
Because
in
the
use
case
that
they
had,
they
had
java,
you
know
node.js
clients
that
were
pulling
messages
from
sqs,
and
then
this
one
is
where
now
ruby's
receiving
messages
from
sqs,
and
if
we
go
back
to
did
they
reference
the
link
to
the
original
pr
that
had
that
very
long
conversation?
D
Okay,
the
only
thing
that
I
remember
contributing
was
similar
to
what
robert
said
is
that
all
of
our
other
asynchronous
instrumentations
take
an
attribute.
That's
like
how
would
you
like
to
link
how?
How
would
you
like
to
continue
the
trace
right
link,
none
or
or
whatever
the
hell?
It
was
spam
or
events.
I.
F
D
They
picked
it
up.
There
was,
they
were
using
the
same
key
value
pairs,
I'm
putting
in
the
message.
One
of
the
problems
that
I
brought
up
was
the
fact
that
there's
limited
space
in
what
you
can
put
in
a
message
header,
so
another
thing
that's
not
implemented,
is
multi-type
or
or
a
multiple
context,
propagation
support.
You
only
have
like
one
of
the
only
ones
yeah
one
propagator
in
there-
that's
like
you,
can't
do
a
mix
of
x-ray
with
trace
pairing
with
ot
trace
with
whatever
for
interoperability.
D
So
I
don't
know,
I
don't
know
how
we
landed
on
that,
but
we
had
to
have
like
we
have
to.
We
have
a
fixed
width
on
perfect
size
of
what
we
can
put
in
the
headers
at
all
for
what
can
go
in
there,
and
I
think
that
headers
was
the
was
the
mechanism
that
we
were
using
to
like.
Do
the
transport,
because
that's
how
x-ray
does
it.
E
F
D
They're
like
because
it's
basically
like
designed
after,
like
the
postal
service,
it's
like
you,
put
things
in
an
envelope
and
then
you
put
a
stamp
on
it
and
a
return
address
and
a
sending
address.
D
A
Back
on
track
all
right,
so
one
thing
we
could
do
is
try
to
look
at
the
spec
and
the
semantic
conventions
and
see
if
anything
has
changed.
That
might
give
us
guidance
on
this,
because
I
feel
like
the
one
one
thing
cueing
like
one
language
cueing
in
another
language.
Picking
it
up
is
like
a
real
use
case,
so
we
might
want
to
look
at
that
and,
like
I,
don't
really
want
to
try
to
push
anything
through
but
like
someone
might
want
to
push
something
through
in
the
spec
to
make
that
happen.
A
But
then
I
don't
know,
I
guess
I
can
go,
read
the
old
pr
and
this
pr
and
try
to
just
do
a
response,
and
at
least
be
like
thank
you
for
thank
you
for
doing
this.
We
appreciate
it
because
we
do
and
then
figure
out
what
to
all
what
all
to
do.
So
I
I
don't
know
that
I
really
said
anything
there,
but
I'm
happy
to
look
at
this
pr
and
give
a
response
along
the
lines
of
you
should
give
the
user
an
option
to
have
a
particular
propagation
or
particular.
A
D
D
D
The
kafka,
the
not
rescue,
what
is
it
hey
worth?
It's
a
active
job.
B
And
is
that
the
link
child
or
none
option.
F
B
B
I
guess
tbv
for
for
homework
for
all
of
us
to
actually
figure
out,
if
that's
true
or
not,
but
I'm
not
seeing
those
three
options:
anyways
at
a
glance
and
yeah.
I
guess
the
the
wrap
up
on
this
is
to
investigate
the
spec
and
see
if
there's
any
guidance
there,
I'm.
B
I
would
be
pleasantly
surprised
if
there
is
anything
comprehensible
over
there
as
to
what's
new
in
the
situation.
Yeah
there's
a.
B
B
I
don't
know
if
any
other
languages
have
have
sqs
instrumentation,
but
it
might
be
worth
a
slight
comparison,
but
if
there's
a
reasonable
way,
if,
if
the
ruby,
sig
is
kind
of
gone
with
this
lynx
child
or
none
propagation
style
for
message
queues,
and
if
this
is
something
that's
reasonable
to
implement,
it
might
not
be
a
bad
suggestion
and
resolution
to
all
this.
A
No,
this
is
this.
This
amounts
to
no
more
than
rambling,
and
I
think
what
you
said
is
correct.
There
is,
I
just
wanted
to
drop
a
link.
I
don't
think
this
really
solves
our
problems
at
all,
but
this
is
like
a
thing
that
exists
like
nerds
are
trying
to
describe
event
data
in
a
common.
G
A
B
All
right
probably
worth
a
look.
E
If
you're
curious
about
why
there's
been
build
failures
with
j
ruby
in
the
past
24
hours,
they
released
a
bad
version
and
I
think
it
got
cached,
and
so
I
figured
it
out
now.
I
just
got
to
like
maybe
bust
the
cash,
so
at
least
that
mystery
was
just
solved
moments
ago,
but
I
think
that's
the
only
thing
that
was
actually
on
my
mind
I
mean
I
had
prs,
people
could
review
or
approve,
but,
like
you
know,
matters
so.
B
Yeah,
I
think
definitely
prs-
could
do
some
reviews
anything
else
worth
talking
about,
while
we're
all
here.
F
Meeting
well,
if
only
we
had
all
gone
to
spain,
we
could
we're
going
to
be
hanging
out
at
the
at
the
hotel.
Oh.
F
F
D
A
A
A
It's
in
detroit
in
october,
that's
all.
I
know.
E
Actually,
maybe
I
might
go
to
that
detroit
is
home,
so
I
might
oh
I'm
out
of
10..
That's
interesting.
A
24
to
28.
yeah,
you
should
all
also
why
to
talk.
B
E
Gearing
up
to
go
carrying
up
to
go
this
weekend
actually
for
weddings.
Otherwise,
I
also
don't
usually
go
back
very
often.
B
D
F
F
F
Argo
khan
do
we
have
any
instrumentation
for
argo,
look
at
bay
area
september.
D
F
C
E
For
everyone
who
was
here
at
the
very
beginning,
I
think
I
got
all
the
ants
in
the
kitchen
I
think
they're
dead.
So
I'm
personally
very
happy
about
that
murder.