►
From YouTube: 2022-01-04 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).
C
A
Solidarity
for
the
not
mourning
people.
A
A
Which
okay,
yeah
yeah
yeah,
I'm
very
interested
in
both
of
the
alternatives,
as
opposed
to
this
I
mean
I
don't
have
anything
necessarily
against
this,
but
the
either
doing
your
your
proposal
or
this
one.
E
A
Be
really
nice
also
from
a
performance
perspective.
A
I
remember
battling
with
that,
trying
to
pair
down
some
benchmarks
and
being
annoyed
with
the
rappers,
sometimes
they're
proxied,
sometimes
there's
multiple
rappers
and,
like
you
have
to
do
these
checks
in
each
layer.
So
if
we
could
not
instrument
those
that
would
be
cool.
D
And
the
data
that
way
kind
of
makes
sense-
I
mean
it
makes
sense
more
than
ignoring
chosen
wrappers
because
there's
like
a
infinite
number
of
sql
databases
and
sql
database
drivers
and
wrappers
well,
you
can
have
as
many
as
you
want
right.
I
was
even
worried
that
technically
you
can
generate
the
statement:
implementation
using
java
proxy
right
so
yeah,
but.
C
D
Yeah
yeah,
this
really
makes
sense
now
that
I
think
of
it.
Although
the
complete
list
of
all
statement
connection
and
everything
in
the
implementations
is
well.
B
A
And
luckily
we
have
the
the
starter
list
here
and
datadog
they
did
have
several.
There
were
several
follow-up.
A
I
remember
seeing
this
one
when
they
did
it,
because
I
was
interested
in
it
from
way
back,
but
that
they
had.
There
were
several
followers
follow-ups
like
support.
This
jdbc
driver
support.
This
jdbc
driver.
A
A
So,
as
opposed
to
the
because
I
was
thinking
the
the
proxies,
I
mean,
are
there
that
many
I
mean
if
we
could
match.
A
At
least
prepared
statements
if
they're
cashing
those
what
did
david.
A
A
D
Our
yeah
that
there's
a
context
or
slash
virtual
field
for
connections
and
statements
too
so
we'd
be
adding
fields
to
all
those
rappers,
probably.
D
Fields
probably-
but
I
mean
the
richard
mentioned
something
in
the
pr
description,
so
I
I
just
remembered
about
it
and
yeah
anyway,
it's
kind
of
interesting
how
they
went
around
it
if,
if
they
went
around
it
at
all.
A
A
Well,
I'm
I'm.
Certainly
in
favor
of
a
known
I
mean
going
down
a
known
route.
They've
had
this
for
a
year,
so
have
a
good
degree
of
confidence.
D
A
So
it'd
probably
be
better
for
one
of
us
to
merge
so
I'll
I'll
reply
on
this,
and
I
can
probably
attempt
that
merging
stuff
over.
D
A
Why
why
is
that?
What,
because
that
it
feels
to
me
like
there's
less,
but
I
guess
I
mean
yeah.
It
could
be
a
long
list,
but
at
least
it's
not
broken
like
I
mean
the
problem,
if
the
new
jdbc
driver
that
we're
not
that's
not
in
our
list,
it's
not
going
to
capture
anything
versus
if
we
do
the
inverse
and
ignore
proxy
specific
proxies,
the
fallback
is
better.
D
Yeah-
and
I
guess
that
you
could
probably
use
proxies
like
completely
on
demand
if,
for
example,
if
you
want
to,
I
don't
know,
add
some
some
custom,
metrics
or
logging
around
your
connections,
then
you
could
use
java
proxy
to
do
that
in
your
in
your
application,
like
without
the
additional
library.
A
True,
okay,
so
maybe
but
it's
not
the
case
then
not
necessarily
generated.
I
mean
I
mean
it's
so
easy
to
generate
your
own
proxy
in
intellij,
oh
yeah.
Let's
do.
A
But
still,
even
in
that
case,
because
the
the
whole
downside
right,
I
think
I,
if
I
understood
this
correctly,
the
the
whole
problem
here
was
just
that
the
wrong.
A
D
Identified,
as
I
don't
know,
postgres,
but
calcite
instead
right
right.
D
You
still
got
the
query,
but
the
b
system-
db
name,
I
think-
is
also
extracted
from
the
connection
string
and
I
don't
know,
maybe
that's
beauty,
tanks.
A
Well,
I
think
I
was
like
in
theory
I
like
the
idea
of
the
skipping
proxies
because
of
the
slightly
better
fallback,
but
the
the
precedence
I
mean
it's,
it's
awfully
nice
to
have
something.
That's
proven
that
we
can
merge
over.
A
Thought
of
we've
been
doing
the
practice
right
of
not
merging
doc
changes
until
the
release,
but
it's
kind
of
pain,
it's
kind
of
a
pain
and
especially
now
that
we're
we've
decided
to
go
with
the
website,
docs
being
hosted
separately
and
they
can
then
pull
in.
A
You
know
that
can
be
up
to
date,
I'm
kind
of
leaning
towards
just
going
stopping
that
practice.
D
I'm
saying
we've
we've
had
a
similar
discussion
and
similar
problem
in
the
splunk
distro
repository
and
at
the
end
we
went
with
the
main
branch
contains
the
current
development
version
dox2
and
if
you
want
dogs
for
particular
version,
just
check
out
the
tag
or
splunk
official
box.
So
if
we
you
know,
but
much
probably
the
website
thoughts
would
be
kind
of
like
the
official
box
right.
A
A
The
key
is
not
here,
so,
let's
skip
that
oh
yeah
bogdan,
had
oh,
you
had
replied.
I
missed
your
response.
A
A
D
E
D
And
I
will
like
read
through
this
class
and
how
it's
used,
but
right
now
I
think
that
the
current
approach
also
has
the
advantage
of
using
alternative
concepts
like,
for
example,
if
you
use
micrometer
distribution,
somehow
it
gets
translated
to
open
telemetry
histogram
instead
of
you
know,
you're
just
getting
a
bunch
of
cages
if
you
use
like
the
raw
micrometer
data
that
it
provides.
E
D
C
D
You
can
you
effectively
get
a
separate
gauge
for
each
bucket,
okay,
with
tax,
saying
that
it's,
for
example,
like
you
know,
less
than
50
percent,
or
something
like
that.
A
E
Oh
geez,
that
being
said,
I
mean
I
don't
know
how
much
how
well
specified
prometheus
ii
hotel
spec
is,
but
of
course,
if
there
was
some
conversion,
we
would
be
able
to
probably
apply
it
to
the
micrometer
electric
producer
if
you
wanted
to
so
maybe
we
shouldn't
think
too
much
into
those
aspects,
because
any
sort
of
converting
gauges
histogram
would
apply
for
printing
this
data
anyways,
regardless
of
micrometer,
that's
supposed
to
be
one
thing:
that
hotel
supports
first
class.
As
far
as
I
know,
a
good
export
from
prometheus
and
total.
D
D
Also
one
one
more
minor
reason
for
not
using
metric
producer
is.
D
E
Yeah,
as
far
as
the
histograms
go,
this
issue
configuring
histograms
the
api
like
asking
users
to
duplicate
that
into
their
sdk
configuration
when
setting
up
the
bridge.
That
doesn't
seem
like
the
end
of
the
world,
because
I
think
that
should
we
could
even
log
like.
Maybe
we
can
have
a
log
statement,
here's
the
code
to
your
view
and
then
people
can
just
add
that
that's
not
the
end
of
the
world.
I
think
something's
going
to
be
hard
to
use
so.
D
D
Yeah-
and
I
mean
I
I'm
not
sure
why
would
anyone
use
it
because
when
I
tried
it,
I
completely
did
not
understand
what
the
data
was.
E
A
Cool
the
login.
E
E
E
E
D
Oh,
so
all
the
of
the
auto
codes
that
you
know
they
were
using
instrumentation
api,
it's
in
the
tracing
project.
Right
now,
is
it
yes,
this.
C
A
C
A
Well,
I
mean
actually
it's
a
much
better
branding
than
salute.
A
The
logging,
I
think
it
feels
clear
in
my
head
now
at
least
honorary.
I
think
I
will
go
ahead
and
do
what
you
had
proposed
of
moving
our
modules,
making
them
internal,
basically
hiding
the
so-called
logging
api
and
sda.
A
A
A
E
C
A
I
think
his
his
main
interest
right
now
is
the
the
version.
The
version.
A
Yes,
he
doesn't
like
the
I
mean,
not
that
any
of
us
love
it,
but
the
the
v216
being
confusing
of
like
well.
What
if
I'm
on
v217
is?
Do
I
use
this
this.
D
E
A
Right
as
we
and
even
then
we
would
still
be,
I
mean
if
we
couldn't
then
break
anything
like
if
the
library
itself
had
breaking
changes
in
v3.
We
wouldn't
necessarily
without
versioning
our
package
name.
E
E
That
can
sometimes
still
cause
problems
because,
of
course,
libraries
don't
always
get
their
semiconductor
right,
but
it's
often
still
possible
for
us
to
work
around
it
with
some
classic
attacks.
Anyways
like
using
reflection
to
determine
what
version
we're
on
and
taking
different
branches
and
lazy
questions
and
whatnot.
So
it's
still
possible
to
support
multiple
versions
like
it's,
probably
more
likely
we'll
have
to
do
weirdness
like
that.
If
we
do
this
versus
right
now,
we
tend
to
not
need
that
weirdness,
because
we
would
allow
having
both
the
class
path
at
the
same
time.
E
C
E
C
A
E
A
B
A
So
I
think,
were
you
doing
this
matteo
sure
was
laurie.
D
D
A
Yeah,
so
there
was
this:
we
had
a
discussion
about
generalizing
that
span
server
span,
naming
into
like
a
hd,
http
router
and
starting
to
capture
that
and
starting
to
pass
that
around
consistently
of
the
instrumenters.
D
E
E
A
E
I
did
change
the
response.
Type
response
from
call
like
the
call
type
in
the
response
handler
might
be
a
rooting
call
similar,
like
the
request
handler,
doesn't
have
the
route
on
the
call,
as
we
have
here,
this
arc
dot
root
stuff
yeah,
it's
a
bit
weird,
like
the
call.
The
type
of
the
call
is
different,
depending
on
what
the
pipeline
phase
this
thing
called
call.
E
A
E
A
To
yeah
it
was,
it
was
an
optimization
so
that
we
didn't
need
to
use.
There
were,
oh,
the
main
reason
yeah.
The
main
reason
it
takes
a
supplier
is
so
we
don't
have
to
do
some
expensive
operations
in
some
of
the
instrumentation
in
particular
spring
web
mvc
does
a
fairly
complex
calculation.
A
Yeah
yeah
servlet
instrumentation
also
is
a
little
has
some
weird
stuff
because
it
has
to
look
it
up
in
the
like
the
filter,
mappings
to
remap,
where
the
url
would
have
mapped
to
or
something
some
weird
stuff
like
that.
Okay.
E
A
A
E
A
This
one
I
did
I
looked,
we
can,
I
didn't,
merge
it
in
case
anybody
wanted
to
take
a
look
at
it,
but
feel
free
not
to
emerge
it
tomorrow
and
that's
the
end
of
the
instrument.
That's
the
end
of
the
bass,
tracers.
A
E
A
And
bogdan
already,
yeah
bogdan
already
looked
over
stuff,
so
he
he
got
his
chance
to
complain.
So
that's.
C
A
E
A
Oh
this
one,
I
I'll
take
a
look
at
this
because
I
had
asked
this
user
to
try
to
write
a
test
and
they
did
and
had
some
issues.
So
I'll.
Take
a
look
at
this
in
the
next
day
or
so.
E
D
Cool,
oh
yeah,
on
a
somewhat
related
note:
should
we
rename
all
those
you
know,
let's
bring
something
tracing
into
spring,
something
instrumentation.