►
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).
B
A
Yes,
yes,
so
I
threw
four
more
options
for
a
total
of
five
options
to
discuss,
so
I
could
live
with
this.
I
wasn't
loving
it.
I
don't
necessarily
love
any
of
these
alternatives,
but
I
wanted
to
get
your
thoughts
well,
let's
see.
C
A
I
set
my
slack
enter
to
not
be
submitted
yet.
A
A
Thank
you.
We
have
an
important
instrument
or
api
topic,
so
the
background
on
this
is
the
time
extractor.
A
A
A
A
I
guess
in
the
start
time
also
yeah
right
yeah.
You
would
use
that
both
in
both
to
get
the
timing
it
kinda
would
be
nice
if,
like
extract
start
time
could
return.
The
cut
could
update
the
context,
because
then
it
could
stick
the
timer
in
there.
But
then
what
do
you
do?
Have
you?
I
mean?
I
guess
you
could
yeah.
A
So
let's
look
at
the
other
options
and
then
see
what
you
all
think
so
trying
so
I
was
just
thinking
of
like
we
have
this
one
use
case.
So
I
was
trying
to
think
of
you
know
a
more
specific.
A
We
can
do
better
by
just
solving
that
specific
use
case.
So
this
idea
puts
two
new
methods
on
the
instrumenter,
no
time
extractor
and
you
tell
it
okay.
I
want
to
essentially
mark
this
as
potential
start
time.
A
It
puts
the
current
time
into
the
context,
returns
that
and
then
you
call
start
and
end
later
on.
With
that
context,
nice
things
is
user,
doesn't
need
to
deal
with
clock
timers,
it's
all
handled
by
the
instrumenter.
A
Other
good
thing
is
you:
it
doesn't
expose
the
ability
to
change
the
end
time,
which
is
nice
since
metrics.
Don't
support
that
anyways
the
cons.
Of
course.
These
are
very
ugly.
I
could
not
think
of
better
names
and
it
you
know
we
have
this
nice
clean
instrument
or
api
contract
that
this
totally
ruins.
A
So
let's
move
on-
and
these
are
a
no
there's-
no
order-
they
don't
get
better
or
worse,
they're
just
spaghetti.
I
was
throwing
on
the
wall,
so
this
one
is
add
a
single
new
method
on
instrumentor
start
and
end,
which
takes
both
the
start
time
and
end
time.
A
A
So
we
would
have
a
public
timer
api
and
the
context
key
user
would
start
the
timer
whenever
they
want
pass
it
past.
The
timer
object
and
then
put
it
in
the
context
before
they
call
start
and
then
end,
and
we
don't
need
the
combined.
In
that
case,.
A
Because
yeah,
so
with
this,
we
can
still
hide
basically
we're
still
in
control
of
the
clock
and
we
can
hide
customizing
the
end
time.
It's
definitely
more
magical
and
I
didn't
really
it
would
sort
of
require
checking
the
context
for
that.
Timer
object,
even
in
the
normal
case,
which
nothing.
B
And
yeah,
so
that
you
have
some
kind
of
timer
class
which
shouldn't
really
need
a
huge
api.
It
would
be
just
like
one
method.
Just
this
mark
start
and
that's
it.
B
There's
also
a
sixth
option,
which
is
making
this
feature
internal
for
now
and
it's
you
know
just
hiding
all
the
time
stuff
yeah
and
using
it
another
in
our
instrumentation.
So.
C
A
So
one
question
about
that,
so
would
say
you
could
add
links
to
spans
late.
Would
you
at
the
beginning
of
because
we
also
use
it
in
the
case
where
we
may
not
create
the
span,
but
could
we
call
start
span
and
just
if
we
never
want
it?
Just
never
call
end.
A
Do
we
have
any
any
like
gc
check?
I
know
we
have
for
like
scopes
that
were
never
closed.
A
A
A
Possibly
if
we
did
create
a
span.
B
Yeah
yeah,
even
if
we
created
them,
I
think
it
would
be
reasonable
to
have
a
configuration
flag
that
disables
creation
of
every
poll
like
the
equation
of
stand
for
every
vocal,
because
they
can
happen
fairly
often
like
one
each
one.
Every
second
or.
B
A
Yeah,
but
we
don't
know
if
we're
gonna
get
a
record
back
until
pull
in
it's
like.
B
C
B
C
There
is,
but
some
say
like
I
guess:
theoretically,
we
should
always
create
the
receive
span
unless
the
user
doesn't
want
them
or
something
yeah
like
I
don't
know.
If
we
need
a
well,
I
mean,
I
guess
it's
an
open
question,
but
I
guess
you're
saying
risk
only
create
receive
spams
if
there's
messages,
but
this.
C
A
So
yeah,
that's
a
good
that
so
maybe
we
should
talk
about
the
neti
connect
example,
which.
B
Yeah
and
in
case
of
nettie
it's
it's
sort
of
complicated
as
usual,
because
the
connection
phase
happens
before
the
http
request.
So
you
can't
add
an
event
on
the
spam
that
will
not
be
started
because
the
prerequisite
has
failed
and.
B
B
Not
possible,
I
think
it's
not
possible,
because
it's
steady
that
that's
how
we're
doing
things
in
react
already
where
there
is
a
sort
of
level
abstraction
over
there.
So
the
the
http
client
span
is
the
parent
of
the
connect
science
fan,
but
in
virality
it's
not
that
easy.
A
This
is
making
me
like
the
idea
of
making
it
internal
for
now
and
give
a
few
months
to
see
sort.
A
C
A
I
think
just
noise
noise
to
value
ratio
well
historically,
because
that's
how
it
was
when
we
brought
the
code
base
over
from
datadog.
A
B
I
want
I
mean
we
think
value
aside
and
the
connect
spans
shouldn't
really
generate
that
much
noise
pretty
much.
Every
http
client
uses
some
kind
of
movement
underneath,
so
they
should
have
been
like
a
couple
of
times
and
that's
it.
The
number
should
be
significantly
smaller
than
the
number
of
http
goals.
C
C
C
B
A
B
I
followed
the
the
behavior
of
neti
connect
in
that
one.
B
It
always
gets
false
as
the
default
in
both
properties
that
it
needs,
so
it
only
creates
a
response.
A
A
C
B
A
Oh
so
back
to
the
honorary,
do
you
think
the
ssl
time
is
worth
having
on
by
default.
C
A
Matisse
you
had
commented
in
the
slack
about
this
yeah,
so
the
only
thing
that
the
thing
that
I
wanted
to
get
your
thoughts
on
was
whether
we
should
do
this
or
not
adding
the
exception
message
as
the
status
code
description.
B
A
And
someone
was
asking
in
slack
about
contrib
repo
release
because
they
want
the
jmx
metrics
for
the
collector.
So
I
will.
A
C
C
A
Do
you
want
to
take
care
of
the
snapshot
or
show
I
push
a
separate?
I
guess.
C
A
I'm
not
going
to
worry
about
anything
else,
making
it
for
the
release.
Thank
you,
matthias
for
volunteering
to
take
a
slightly
closer
look
than
scroll
scroll,
scroll
girl.
I
don't
see
any
my
my
definition
of
a
minimal
review
is:
there
was
no
malware
injected
into
this
code.
C
B
A
A
Whether
it
should
be
more
community
decentralized
or
if
you
know,
if
isn't
that
the
whole
point
of
github
in
general
of
like,
if
you
don't
need
a
centralized,
just
people
can
publish
their
own
library.
C
A
All
right
anything
else
that
we
should
chat
about
while
we're
here.
A
B
Yeah,
so
I
have
one
small
thing:
I'm
gonna
go
on
a
vacation
starting
next
week
and
I
will
be
absent
for
a
month
so
see
you
all
back
in
june
nice,
which
probably
means
that
some
of
the
instrumentation
api
things
will
be
left
until
june.
But
that's
fine.
I
guess
I'll
do
what
I
can
this
week
and
then
yeah.
B
I
don't
really
have
like
one
huge
trip:
I'm
planning
a
couple
of
several
smaller
ones,
so
at
least
one
trip
to
the
baltic
sea
several
trips
to
the
mountains,
which
are
a
bit
closer,
so
I
can
go
there
every
other
day,
yeah
so-
and
I
wouldn't
just
I
just
want
to
rest
in
general.
More
than
you
know,
sightsee
and
travel
the
world.
A
So
you're
gone
mate.
Friday
is
your
last
day
around
yeah
all
right
cool.
Well
then,
I
will
see
you
thursday,
on
friday.
A
All
right
well
have
a
good
tuesday
and
oh,
let's
see
how
the
release
is
going.