►
From YouTube: 2021-10-01 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
So
what
does
that
mean
like
in
the
city
like
heavy
heavy
winds,
flooding.
D
A
A
C
Oh
yeah,
that's
a
good,
so
you're
so
you're
in
america.
I'm
sure
it's
still
the
30th
for
you,
yeah.
A
A
Yes,
we
have
very
confusing
meetings.
We
have
three
meetings
a
week
at
different
times
because
of
honorable
in
japan,
nikita
and
matej
in
paul
and
dystonia
and
myself
in
the
us,
so
we've
agreed
to
to
at
least
make
it
less
con
slightly
less
confusing.
We've
marked
them
as
all
utc
time.
C
The
other
java
meetings,
but
they're
at
like
3am
or
something
for
me,
so
it's
good
that
this
is
available.
Yeah.
D
C
C
D
C
A
C
I
don't
have
anything
super
specific
to
talk
about
today.
I
just
kind
of
kind
of
wanted
to
look
and
just
figure
out
what
you,
what
you
guys
are
talking
about
and
just
get
more
yeah,
because
I'm
going
to
be
working
on
the
java
stuff
a
little
bit
so
just
getting
familiar
with
it
all.
A
Right
so
yeah,
so
usually
in
this
meeting,
if
there's
nothing
at
least
the
the
baseline
for
this
meeting
is
reviewing
stuff
that
we
talked
about
in
the
morning
meeting
with
honorag
and
then,
if
there's
sometimes
like
last
week,
we
had
a
couple
of
folks
joined
to
talk
about
reactor
instrumentation
and
sometimes.
D
A
D
A
E
A
Awesome
so
yeah,
so
I'm
going
to
start
with
going
through
kind
of
topics
we
talked
about
this
morning.
C
Far
just
the
sdk,
we
don't
use
the
instrumentation
internally,
so
we,
but
we
do
use
the
sdk
so,
but
I'm
still
interested
in
hearing
about
the
instrumentation.
Definitely
yeah.
A
There's
there's
some
tricky
stuff.
That's
actually
decently
has
some
decent
documentation,
where's
the
muzzle
contributing
muzzle
so
yeah
so
honorary.
This
was
nikita.
A
Talking
about
now
that
we're
auto
now
that
we're
moving
things
to
that
interface
and
so
users
sort
of
then
have
to
use
muzzle
to
generate
those
classes
like
they
can't
just
manually
write
a
get
helper
class
names.
A
A
It's
the
only
benefit
of
sort
of
separating
it
out
in
the
way
that
nikita
has
been
doing
is
to
make
it
more
clear
to
end
users
which
parts
that
they
use
explicitly
versus
which
parts
get
auto-generated,
which
still
seems
like
a
win,
but
we
could
leave
those
other
parts
that
gray
area
into
a
separate
module,
maybe
that
you
could
still,
if
you
wanted
to
manually,
pull
that
in
implement
the
interface
implement,
get
helper
class
names.
B
D
A
A
So
he
seemed
to
think
that
he
was
going
to
think
on
that
some
more,
but
it
sounded
good
and
or
he
was
just
going
to
revert
everything
back
if
if
there
was
no
benefit
to,
we
couldn't
hide
them
and
make
them
internal
anyway.
A
Could
just
move
it
all
back,
though
I
do
think
there's
some.
I
mean
what
I
do
think
there's
some.
Obviously
more
modules
is
less
good,
but
it
does.
It
could
be
more
clear
for
users.
A
And
then
some
discussion
around
how
you
know
the
path
to
api
stability-
and
this
is
james
and
sean.
This
is
the
instrumentation
api,
so
we
have
three
apis
extension
api,
which
is
for
extending
the
java
agent
instrumentation
api,
which
is
for
general
library
instrumentation,
which
you
might
be
interested
in
using
this
once
it's
stable
or
taking
a
look
at
it.
It's
the
new
instrumenter
apis.
A
It's
sort
of
a
convenience
makes
writing
instrumentation
a
little
bit
more
convenient
and
wraps
around
both
tracing
and
metrics,
so
that
it
you
start
and
end,
and
it
will
generate
the
span
and
the
metrics
for
you.
Since
typically,
they
share
a
lot
of
the
same
attributes.
C
Yeah,
that's
funny.
We
actually
did
something
similar
to
that
internally
in
our
own
kind
of
instrumentation
library,
so
that
that's
yeah,
that's
cool.
It's
definitely
very
interesting
to
me
that
instrumentation
api
yeah
definitely
something
we
could
probably
use
once
it's
stable.
I
don't
know
what
kind
of
state
it's
in
at
the
moment.
Is
it
experimental.
A
Yeah
it's
alpha
currently,
and
so.
We've
have
this
slow
march
here
we're
at
73
out
of
99
instrumentations,
converted
to
the
new
instrumenter
api.
A
So
if
you
look
at
the
we've
got,
if
you
look
at
the
instrumentation
and
let's
take
like.
A
Grpc
so
we've.
D
A
So
it's
so
our
goal
is
to
have
parity
there
and
then
the
testing
module
apply
basically
to
both
to
make
sure
that
they're
generating
the
same
telemetry,
whether
using
the
top
agent
or
the
library
instrumentation
yeah,
okay.
Well,
that
makes
sense
a
lot.
C
A
Don't
have
like
we're
going
to
see
apache,
we
do
have
now
a
4-3
library
instrumentation,
but
a
lot
of
things
don't
have
library
instrumentations
yet,
but
that
is
a
goal
and
if
you
look
at
this
is
the
list
of
a
current
library
instrumentation
that
we
have
so
far
so
definitely
if
you're
interested
in
library
instrumentation.
That
would
be.
You
know
if
you
are
using
say
cassandra.
C
No,
this
is
definitely
something
interesting
to
us
and
yeah.
We've
done
a
few
instrumentations
ourselves,
but
if
we
could
utilize
this
like
be
super
useful.
Actually,
could
you
scroll
down
a
little
bit
yeah,
there's
a
spring
a
couple
spring
ones:
oh
yeah,
so
we've
got
we've
yeah
spring.
Those
spring
ones
are
particularly
interesting
to
us,
so
yeah
I'll
definitely
have
a
closer
look
at
that.
I
wish
I've
had
a
close
look
at
that
earlier
on,
so
I
could
yeah
no.
D
Spring
is
awkward
because
there's
also
sleuth,
and
so
it's
like
there's
two
types
of
live
instrumentation
for
spring,
one
that
spring
owns
one
that
sort
of
an
intern
developed
here.
I
would
probably
recommend
looking
at
sleuth
rather
than
here
for
spring,
because
that's
probably
going
to
be
more
maintained
going
forward.
They
have
an
hotel
based
thing,
also.
A
A
D
D
A
B
D
D
D
C
Yeah
yeah
sure.
D
E
D
A
E
A
E
A
Yeah
good
to
put
a
face:
yeah
yeah
yeah.
Well,
this
is
sort
of
the
the
plan
for
the
this
auto
translation
like
here,
we're
proposing
to
replace
this
by
the
generic
db
name
and
this
by
the
generic
db
name.
So.
C
Yeah,
that's
a
really
cool,
think
something
something
I
actually
called
out.
One
of
the
issues
on
the
specification
like
breaking
dangers.
Getting
translated
between
semantic
convention
versions
is
like
super
useful
to
make
it
stable.
So.
E
A
A
D
A
So
then
we
just
talked
about
the
review
process:
how
to
then
review
the
api
changes,
and
so
I
suggested
it
that
nikita
could
open
up
issue
with
like
breakout
packages,
and
you
know
wanted
to
start
looking
through
them
and
checking
them
off.
And
at
least
you
know,
phase
one
review
phase
two
review
something
like
that,
but
then
also.
A
Because
I
think
I
I
was
niki
was
asking
about:
can
we
make
declare
this
stable,
but
I
was
kind
of
pushing
back
like
that.
I
don't
see,
there's
a
whole
lot
of
benefit
to
declaring
this
stable
until
we
can
declare
this
stable.
Also,
since
that,
doesn't
we
can't
really
say,
extensions
are
stable
until
all
three
of
these
are
stable.
A
And
then
I
was
going
to
ask
john
or
you,
you
went
through
it,
what
kind
of
tips
from
the
like
how
what
worked
didn't
work
with
the
sdk
one,
zero
review
cycle.
D
A
Did
you
try
to
split
it
up
or
did
everybody
just
kind
of
look
through.
D
C
D
A
Hey
just
what
do
you
call
beard,
my
soul?
I
did
my
communion.
B
A
D
D
A
D
A
You,
people's
patch
release,
builds
won't
work
because
they
can't
deploy
that'll.
A
B
C
D
A
Yeah,
I
want
to
tell
slack:
don't
call
me
I'll
call
you,
I
just
need
to
mute
it
and
go
in.
Actually,
I
think
yeah.
A
Oh
this,
no,
but
he's
going.
It
will
be
a
pr
at
some
point.
C
B
D
True,
it
does
appear
to
not
be
utility.
D
D
D
D
A
And
so
what's
so
you're
saying
in
this
there's.
D
And
the
as
with
any
library,
the
assumption
is
that
if
you're
always
on
the
latest
versions,
it's
going
to
work
so
1.6.
It
is
okay
to
film
1.38
if
it
works
1.41,
because
palm
changes
are
all
collected
together
like
if
a
newer
version
of
your
library
requires
in
your
dependency,
that's
not
a
breaking
change,
because
you're
updating
your
dependencies.
D
D
D
D
A
D
D
Nope,
maybe
it
should
be
so
that's
another
option,
so
you
always
have
to
because
I
think
even
this
grpc
midi
shaded,
maybe
it's
okay.
Now
without
the
grpc
portable,
like
I
think
the
nadi
shader
does
have
a
transitive
pc
core
there's
this
set
of
three
grpc
core
grpc
protobuf
and
grpc
neti
and
grpc
protobuf
is
independent,
like
grpc.
Id
definitely
doesn't
depend
on
that,
so
you
would
have
this
bad
situation
anyways.
As
long
as
I
was
there
now
that
the
depends
on
your
fpc
protocol
is
gone.
D
Yeah
I
mean
jpsc
has
three
things:
it's
a
wire
format,
that's
its
core
thing,
so
you
can.
Any
code
can
generate
that
wireframe,
it's
same
as
protobuf,
and
then
there's
also
the
generated
stubs
that
the
portal
compiler
can
generate
for
trpc
stumps
and
then
there's
the
runtime
library
to
use
those
steps.
D
B
D
B
D
A
channel
I
wish
we
didn't,
we
didn't
have
to
that's
unfortunate.
We
did
that
because
it
was
probably
not
required
because
of
that
it
would
be
awkward.
I
mean
I'm
still
thinking
like.
Maybe
another
approach
to
this
is
just
if
the
user
excluded
grpc
detect
the
class
not
found
and
switch
to
this
implementation
instead,
that
might
be
possible.
A
Still
feeling
good
about
by
exposing
bite,
buddy.
D
D
A
Apis
moving
things
that
are
only
used
by
internal
instrumentations
can
move
to
the
bootstrap
module.
That
seems
easy.
A
So
I
mentioned
that
I'm
pretty
comfortable,
given
that
the
sdk
has
gone
down
this
path
already
with
a
lot
of
internal
packages,
but
I
think
nikita
was
still
thinking
he
might
try
something
via
reflection
or
something
to
avoid
it
being
in
an
internal
package.
A
Let's
see
what
that
looks
like
I'm,
not
convinced,
that's,
given
that
we've
kind
of
already
gone
down
that
path,
I'm
not
sure
this.
If
there's
any
benefit
to
us,
not
using
internal
packages.
I
think
nikita
is
just
worried
because
of
like
gradle
issues
where
people
say
hey,
you
broke
what
the
this
api
and
it's
some
internal
class.
I
guess
those
issues,
yeah
yeah.
D
C
A
Cool
yeah,
I
mean
we
can
definitely
open
another
issue.
Another
one
of
these
x
out
of
99
issues
of
converting
tests
from
groovy.
D
I
don't
know
if
nikita
would
allow
us
to
have
such
an
issue,
because
it
I
mean,
of
course
it's
it's
better
if
it
was
java
that
doesn't
mean
that
there's
a
huge
motivation
to
do
a
migration
like
things
that
actually
affect
the
business
logic
downstream.
So
it's
still
just
our
repo.
So
I
think
it's
good
to
do
in
our
spare
time,
but
it's
not
something
to
prioritize.
B
D
A
Yeah
I'll
try
to
get
in
in
the
next
couple
of
days
the
the
prs
we
had
a
little
bit
of
a
back
log
because
we
didn't
want
to
cause
more
conflict
on.
A
A
This
went
through
a
few
revisions,
but
this
is
what
we
ended
up
with
computing
and
sedif
null,
although
this
one,
we
think
we
can
probably
get
rid
of
it.
A
Would
think
that
the
places
calling
this
today
don't
really
need
this
behavior
could
probably
just
call
set,
and
if
they
really
did
need
sedif
null,
they
could
call
get
and
set,
because
there's
nothing
atomic
about
this
because
they're,
not
initially
it
was
returning
the
field,
but
that
wasn't
being
used
anywhere.
A
D
B
A
Yeah
this
one
felt
like
compute
of
absent
in
map.
A
E
C
Yeah,
I
had
a
quick
question:
it's
not
too
much
trouble
because
I
I
I'm
just
like
kind
of
just
trying
to
become
familiar
with
the
instrumentation
like
repo,
and
I
was
just
kind
of
wondering
what
is
the
the
pattern
of
if,
if
you're
not
using
the
java
agent,
what
is
the
pattern
of
consuming
those
instrumentation
libraries
like?
Is
it?
Is
there
a
general
like
kind
of
thing
or
is
it
like
or
like
best
practice,
or
is
it
just
kind
of
up
to
the
user
to
consume
those
libraries?
C
A
Some
of
them
don't
have
documentation,
so
the
goal
is
so
this
is
user
code.
The
goal
is
for
the
library
instrumentation
to
have
something
like
this
kind
of
a
class
like
okay,
http,
tracing
or
grpc
tracing,
and
you
construct
a
new
one
of
those
passing
in
your
injecting
your
open
telemetry
instance
or
the
global.
A
If
you
prefer,
you
build
it
and
then
it
has
methods
to
get
you
back,
something
that
is
like
an
interceptor
for
that
library
or
a
wrapped
client
for
example,
and
so
you
can
say,
like
you
have
a
kafka
producer,
you
could
say
give
me
a
new
give
me
a
new
tracing
producer.
C
A
C
D
D
C
Probably
more
leaning
towards
library
instrumentation,
I
think
the
java
agent,
it's
kind
of
in
our
circumstances,
can
be
a
little
bit
difficult
to
to
run
in
the
different
kind
of
deployment
environments,
but
I
think
the
instrumentation
from
what
I
could
see
like
instrumentation
library
is
a
little
bit
more
flexible.
In
that
way,
you
can
kind
of
configure
it
how
you'd
like
but
yeah,
that's
kind
of
the.
E
We've
historically
had
issues
trying
to
install
agents
into
our
cloud
products
where
the
agents
is
taking
up
most
the
cpu
trying
to
collect
all
the
telemetry
from
all
the
different
packages.
So
there's
a
bit
of
history
of
untrust
with
installing
new
telemetry
agents.
E
A
Yeah,
I
definitely
recommend
the
if
you
are
finding
that
you're
writing
instrument-
library
instrumentation
for
things
that
your
for
libraries
you're
using
to
you
know,
contribute
those
and
will
help
iron
out.
You
know
it's
definitely
not
a
seam,
you
know
it's,
it's
it's
some
work.
This
is
a
good
example
of
a
a
user
who
one
was
doing
library
instrumentation
for
kafka,
and
so
you
know
it.
It
went
through.
A
You
know
a
lot
of
review
and
but
you
know
it
turns
into
something
really
solid
and
following
all
the
conventions
and
will
be
supported
ongoing
with
new
versions.
We
have
a
lot
of
tests
and
a
lot
of
things
that
pick
up
auto
detect
when
new
releases
come
out
and
we
fix
them,
and
so
you
get
a
lot
of
sort
of
maintenance
once
once
that's
landed.
C
Yeah,
that's
super
useful
for
us
actually
sean.
Your
team
is
working
on
some
kafka
stuff.
Now
currently
aren't
you
so
it
could
be.
Even
that
could
be
useful,
so
especially
the
cement
convention.
Stuff
is
really
important
to
us
as
well.
So
if
just
get,
if
we
just
get
this
mandate
convention
for
free,
that's
a
win
huge
win.
C
A
Cool
right,
well,
I
know
you
made
honorable
day
because
he's
always
looking
for
more.
D
C
Yeah
I
mean
we
currently
have
our
own
kind
of
like
instrumentation
library
that
uses
like
around
like
engineers
around
atlassian
would
consume
and
that
it
kind
of
configures
the
sdk
for
them
and
provides
kind
of
instrumentation
libraries.
But
I'm
always
looking
for
better
ways
to
to
do
it,
and
I
think
this
could
utilizing
this.
Could
this
kind
of
work
is.
A
Yeah
and
it
would,
it
would
help
also
to
get
your
feedback
on.
You
know
to
use
it
in
anger
as
they
say,
and
let
us
know
what
works
and
you
know
what
what
could
be
better.
C
C
Yeah
cool
and,
like
I
appreciate
the
the
help
we
are
kind
of
new
to
this
space,
so
there
might
be
a
few
silly
questions
here
and
there,
but
bear
with
us.
I
guess
yeah.
D
A
Yep
we
have
a
standing
meeting
here
and
then
yeah.
The
tuesday
meeting
also
you're
welcome
to
drop
by
and
chat
anytime
anytime.
A
Cool
well
great,
to
great
to
see
you
and
there
was.
A
Yeah,
all
right
so
until
next
time,
all
right,
catch
ya,
bye,.