►
From YouTube: 2021-06-16 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
That
was
pretty
funny
yeah.
I
remember
I
remember
doing
the
butt
I'm
being
responsible
for
a
lot
of
those
myself.
C
B
B
B
B
B
I
see
that
is
a
good
point
yeah,
so
the
stack
trace
was
I'm
not
really.
B
There
was
no
transform
in
that
stack
trace,
but
I
did
look
in
here
so
the
typical
the
one
above
it
is.
C
B
B
It's
currently
loading
it,
so
it
can't.
You
know
it
can't,
has
no
like
fallback.
C
B
B
This
problem
yeah,
so
this
is
slightly
different
where
it
was
not
loading
fork,
joint
pool
when
it
happened
to
access
fork,
drain
pool
the
first
time
so
under
transform.
If
you
load
any
of
our
code,
touches
a
class
and
causes
it
to
get
lost.
C
B
Loaded,
the
class
but
didn't
use
it.
I
know
it,
can
you
load
it
and
use
it,
but
it
won't
transform
it
transform,
won't
anything
you
load
under
transform,
won't
get
transformed
itself.
B
B
B
B
B
B
I
had
free
initialize.
C
B
Ourselves
and
it
figured
that
yeah
yeah,
so
this
was
something
that
I
had
in
glory.
I
liked
this
quote
a
lot.
B
And
this
was
specifically,
I
had
it
because
of
different
problem,
but
it
basically
is
the
same
thing:
initializing
other
classes,
inside
of
transform
in
prior
to
parallel.
B
To
deadlocks,
but
apparently
it
can
lead
to
lots
of
other
problems,
and
so
I
had
this.
I.
C
C
C
B
Java
dot
see,
I
didn't
preload
java
classes
only
this
one
because
of
some
random
problem
I
had
seen
but
yeah.
Clearly
it's
just
luck.
I
mean.
C
C
C
B
You
know
what
I
do.
A
lot
is
in
essentially
in
on
open
telemetry
in
your
pre-main,
for
I
I
I'll
put
right
at
the
beginning
of
pre-main.
I
add
a
class
file
transformer
that
just
looks
for
a
specific
class
that
gets
loaded
and
dumps
stack
there.
So
I
can
see
where
the
first
time
it
gets
loaded
is,
and
it
would
be
yeah
a
hash
set.
C
B
C
B
You
want
to
talk
about
spring.
B
B
C
B
B
B
I
thought
there
was
for
a
test,
but
I
know
I'm
just
not
finding
it.
Oh,
this
must
be
it
yeah,
because
this
has
the
rabbit,
mq
spring
integration
and
rabbit
test,
so
we've
got
http,
then
we've
got
we're
producing,
and
so
this
is
the
spring
integration
internal.
It's
an
internal
span
called
produce
spring
integration.
C
A
B
B
This
is
the
producer,
but
this
this
is
not
our
norm.
This
is
not
the
message.
B
A
A
A
I
mean
they
are
not.
If
that
comment
above
is
correct,
then
it's
understandable
why
they
are
actually
not
ch,
parent
and
child
because,
like
stick
trace
wise,
they
are
separated.
B
B
A
B
Oh,
I
see
why
and
the
kind
of
the
point
is
with
the
spring
integration
is.
We
can
capture
the
consumer
here,
but
at
this
point
we
don't
necessarily
have
the
this
information
anymore.
A
A
So
if
we
don't
have
rabbit
and
queue
integrations,
for
example,
then
sprint
sprint,
integration
still
can
guarantee
message,
propagation,
another
context,
propagation.
B
A
That
that
that
I
think
depends,
I
mean
how
exactly
underlying
messaging
system
and
integration
like
work
if
they
that
may
be,
or
maybe
not
need
it.
If
underlying
system
just
reads
a
message
and
pass
that
for
the
process
in
the
same
thread,
we
don't
need
anything
from
from
sprint
integration
right.
If
that's
like
multi-threaded
hopping,
then
yes,
we
can.
We
should
use
that
for
context
propagation
as
well.
Yes,.
B
A
B
Yes,
I
see,
and
so
maybe
this
maybe
we
can't
change
this.
If
this
is
the
our
existing
mean,
maybe
it
is
actually
a
process.
We
are
processing
the
rabbit
and
q
message
we're
just
dumping
spring.
Integration
is
just
processing
it
and
dumping
it
into.
A
C
A
A
A
A
B
So,
okay,
the
receiving
side,
makes
sense
to
me
that
that
makes
sense
to
me
on
the
sending
side.
B
B
B
Nice
with
matthias
on
thursday,
just
to
get
I
mean.
A
My
I'm
I'm
pretty
sure
that
this
spring
integration
instrumentation
currently
is
far
from
ideal,
but
my
our
point
is
to
get
that
in
to
give
that
to
the
client
and
get
actual
feedback
yeah.
So
that's
why
I
am
very
eager
to
get
that
merch
and
release
in
whatever
shape
info
form.
This
is
right
now
to
get
actual
feedback.
A
B
A
B
B
B
But
you're
welcome
to
merge
it
with
just
your
approval.
Is
that
good
for
you.
C
B
B
B
I
can
live
with
that:
okay,
let's
jdbc
library,
instrumentation.
B
So
this
is
seems
interesting,
like
a
cool
approach.
B
B
B
This
so
there's
definitely
gaps
right.
Like
I
mean
we
would
not
use
this
library
instrumentation
in
our
java
agent
instrumentation,
because
there's
too
many
jdbc
drivers
that
have
their
own
sort
of
public
methods
that
people
call
casting
it
to
their
oracle
driver
or
my
sql
driver,
but
I
don't
know
if
I
mean
as
far
as
library
instrumentation
goes,
I
mean:
do
we
really
want
to
have
a
different
one
for
every
driver
out
there
I
mean.
C
B
Oh,
they
use,
they
have
the
p6
spy
tracing
event,
listener
thing.
B
B
Wrapper
at
which
point
I
mean
maybe
given
that
brave
and
don
alexi,
why
don
propose
p6
spy.
B
C
B
Yeah,
so
it's
definitely
easy
to.
B
Oh,
I
see
with
the
with
the
my
sequel,
you
just
specify.
C
B
For
tracing
yeah,
that's
cool.
C
B
It's
basically
like
this,
where
I
mean
you
know,
we
would
have
jdbc
hotel,
for
example,
and
then
everything
following
it
is
the
real
jdbc
url.
C
A
B
Directly
for
it.
C
B
C
You
could
hand
those
prepared
statements
and
all
that
stuff
properly,
but
not
yeah
like
I
know
that
mysql
you
don't
have
to
worry
about
prepared
and
whatnot.
It's
all
just
given
to
you
in
a
way,
that's
usable,
I
wonder
if
the
jdbc
instrumentation
can
get
pretty
complicated.
If
you
deal
with
all
those
weird.
B
C
B
C
Okay,
in
that
case,
yeah,
if
this
approach
is
easier
for
us
to
maintain
than
driver
specific
ones,
then
we'd
want
to
go
with
it.
On
the
flip
side,
if
each
driver
has
a
reasonable
interceptor
mechanism,
it
might
actually
be
less
code
to
have
three
different
instrumentations.
That's
the
trade-off,
possibly.
B
Oracle
sql
server,
just
just
the
popular
yeah
yeah
h2,
h,
sql
dpd.
C
C
B
B
C
B
C
B
Let's
see
if
I
can
remember
so,
we
can't
return
parent
span
directly
here,
because
then.
C
B
C
C
C
B
I
was
not
I
I
don't
think
I
had
been
convinced
that
it
would
work.
I
don't
remember.
C
C
C
B
Same
depression,
if
we
put
a
dummy
span
in
here
that
did
suppress
everything,
basically,
including
and
but
then
spans
to
get
started
downstream,
would
they
would
get
the
parent
so
we'd
have
to
stick
the
current
parent?
Basically,
the
span
context
in
there,
which
would
be
fine,
then
down?
Oh.
B
C
C
C
I
think
it's
possible
and
we
can
even
change
our
mind
like
if
we
want
to
model
it
differently.
That
opens
that
up
as
long
as
they
should
start.
I
think
that's
not
possible
where
it
becomes
harder,
I
mean
it
might
we
might
just
make,
should
start
to
know
I
mean
so
I
guess
it's
always
some
way
to
back
out
of
that.
But
ideally
we
can
just
do
everything.
B
C
B
Agree,
I
agree,
that's
better,
I
think
you
know
I
feel
like
we
may
get
to
a
point
sometime
in
the
future,
where
we
want
to
optimize
some
things,
but
that
could
be
you
know
at
the
agent
level
optimizations
that
are
hidden,
but
you
know
I
agree
that
we
should
start
with
the
simplest.
B
B
So
releasing
let's,
I
will
plan
on
hitting
the
button
tomorrow.
If
nobody
hits
the
button
today,.
C
B
C
B
End
users
could
write
distress
and
even
extensions
like
could
potentially.
C
B
C
Spi's,
what
was
the
question
like
my
simple
question
is
just
do
we
need
some
special
way
to
highlight
but
change
the
testing
common
where
ok,
http
isn't
returned,
because
I
think
that
tricked
you
off.
C
B
Yeah,
okay,
so
we'll
call
this
extension
plus
distro
author
section
yeah.
Okay,
so
we
need
to
add.
C
B
B
B
C
B
Know
I
won't.
I
need
to
run
that
script
that
scrapes
and
pulls
together
the
the
this
list
of
people
in
case.
There's
anybody
else
in
the
last
couple.
So.
A
B
But
I
will
I'll
update
that,
so
maybe
you
wait
and
I
guess
I'll
hit
the
plan.
I
don't
want
to
hit
the
button.
I
will
hit
the
button
tomorrow.
B
B
B
No
cool,
I'm
hoping
that
maybe
ago
or
laurie,
somebody
who
understands
project
reactor
better
than
me
we'll
look
at
this.
B
Yes,
after
my
after
my
recent
total
failure
at
the
reactor
yeah,
there
was
some.
C
B
B
C
B
Okay,
all
right,
but
if
it
fails
worse
than
that,
I'm
I'm
I'm
I'm
phoning
at
phoning.
Phoning
y'all
on
slack.