►
From YouTube: 2021-06-10 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
Awesome,
yes,
zoom
has
gotten
really
finicky
with
my
camera
and
mike
in
the
past
week.
I
don't
know,
what's
going
on
with
it.
B
B
A
Hello,
welcome,
hey!
You
need
to
talk
to
your
boss
about
conflict
conflicting
with
our
weekly
meeting.
C
C
B
B
Aws
she's
a
part
of
the
outreachy
program,
which
I
don't
know
who
runs
that.
D
It's
a
cncf
sponsored
initiative
where
they
offer
interns
that
have
to
basically
submit
to
it
and
then
can
contribute
to
open
source
projects.
It's
actually
a
really
great
program.
We've
worked
with
some
awesome
people
in
jaeger
through
the
program
and
there's
a
couple
of
folks
on
the
jager
maintainer
team
that
actually
mentor
these
interns
and
get
them
productive
and
they're
great
contributors.
It's
basically
a
six-month
internship.
D
They
generally
look
for
underprivileged
or
underrepresented
communities
in
open
source
and
try
to
get
diversity
and
different
opinions
involved.
Instead
of
all
of
us
white
males
that
tend
to
dominate
these
discussions
and
projects,
and
it
is
pretty
awesome,
it's
a
great
program.
It's
nice
cncf
sponsors
it,
along
with
you,
know
some
other
companies
like
red
hat
that
are
pretty
involved.
That's
pretty
cool
cool
program
for
sure,
so
awesome
and
I'll
I'll
paste,
the
url
if
you're
interested-
and
it
gives
you
more
details
in
the
outreachy.org.
D
B
Hopefully,
hopefully
our
cheater
will
show
up
to
present
what
her
plan
is.
D
Yeah-
and
I
think,
there's
a
way
like
if
you
want
more
folks
getting
involved
in
this
project
and
specific
to
like
appeal
to
that.
I
don't
know
how
it
works.
I
can
ask
you're,
probably
familiar
with
jurassi
in
a
hotel
and
yeager
he's
one
of
the
mentors
and
does
a
lot
of
work
with
them.
So
if
you're
interested
definitely
ping
him
on
slack
or
whatnot,.
B
B
Mainly,
we
ran
into
giant
bumps
trying
to
do
various
and
sundry
things
on
windows
that
just
didn't
work
at
all,
so
our
development
experience
is
not
fantastic
on
windows,
at
least
for
a
couple
small
cases.
It
looks
like.
B
D
B
Will
say,
switching
from
the
original
wsl
wsl-2
is
not
painless.
However,
really,
oh,
they
have
a
simple.
B
D
That's
why
I
joined
there
hasn't
been
much
movement,
but
I've
gotten
some
progress.
That's
gonna
start
next
week.
Basically,
I
know
that
there's
also
some
interest
in
other
folks
participating
in
it
outside
of
our
organization.
D
So
I
was
approached
yesterday
on
forget
what
language
it
was
exactly
one
of
the
newer
java
languages,
not
kotlin,
which
we
use
a
lot
at
our
company.
What's
the
other
java
based
modern
language
thing,
people
use,
I
forget,
yeah
scala,
one
of
the
scala
people
was
kind
of
interested
in
participating,
and
I
said
well
we're
doing
it
with
you
know
log
for
j
log
back.
Is
there
like
some
difference
that
happens
with
scala
or
you
just
use
those
same
libraries?
I'm
not
sure
I
don't
know
the
answer.
I'm
not
a
scala
well-versed.
A
Yeah,
we
don't
have
much
scholar,
expertise
except
when
josh
joins
actually,
which
is
sometimes
josh
soreth.
If
you
have
any
scala
questions
ping
him.
What
was
his
last
name?
Sorry
stretch
us,
let's
see
if
you
joined,
could
have
been
the
person
that
was.
D
B
A
So
the
what
are
you
starting
planning
to
start
with.
C
D
D
And
that's
outlined
in
the
open
issue:
that's
there
that
kind
of
breaks
it
out
into
a
few
different
chunks.
I
forget
the
issue
number,
but
I
can
take
a
look.
It
should
be
like
21
days
old
or
something
like
that.
At
this
point
you
can
just
do
jake
owl.
I
opened
it.
A
B
B
B
Because
almost
I
mean
yeah
because
everybo
almost
everything
supports
or
wants
to
be
async
with
logging,
because
you
want
you
don't
want
to
be
doing
all
that
work
on
the
on
the
application
path.
If
you
can
so
anyway,
I
think
that's
that's
a
a
challenge
for
defining
the
right
interfaces
in
the
sdk
to
implement
as
a
part
of
the
pipeline,
and
I
I
I'm
not
an
expert
I
just
like
as
soon
as
anyone
starts
talking
about
this.
B
If
the
I
mean,
if
you
read
the
logging,
the
current
logging
sdk
otep,
which
I
don't
think
has
been
turned
into
a
spec
yet
but
there's
an
otep
that
tigran
wrote
it.
Doesn't
it
kind
of
waves,
its
hands
and
just
says
this
is
something
that
has
to
be
done,
but
obviously
doesn't
give
any
guidance
about
how
it
that
should
work
from
an
api
perspective
right.
B
B
A
B
D
If,
if
the
context
is
being
propagated,
then
you
have
to
inject
it
and
if
it's
not
in
the
log
and
if
it's
not,
then
you
don't
obviously
have
to
add
anything
in
the
blog.
A
Exactly
cool
anything
else
on
logging.
D
A
I
don't
know
I'll
pick
one
depending
on
my
mood:
the
data
model
and
otlp
and
exporter.
Work
will
probably
land
in
the
java
sdk
repo
and
then
like
the
log
back
and
log
for
j
appenders
will
probably
land
in
the
instrumentation
repo
yeah,
but
yeah.
We
all
we
all
are
in
both
places,
so
you
won't
miss
us
sounds
good.
D
D
Now
that's
my
15
year
old
cat.
That's
really
wants
my
attention.
Currently,
I
think
he
wants
some
of
that
wine.
You've
got
there.
D
A
All
right,
the
next
agent
extensions
nikita
we've
got
two
pr's
here.
A
Do
you
wanna,
walk
us
through
anything.
A
I
hope
cool
so
for
everybody
nikita
already
added
the
extensions
supporting
a
single
jar,
so
we
can
load
extension
sdk
extensions
now
that
people
write
their
own
samplers
spam
processors,
propaganda.
A
Yeah,
your
own
auto
instrumentation
and
there's
a
nice
example.
A
Extent
here
that
shows
off
writing
a
bunch
of
those
things
and
tests
too.
I
wanted
to
ask
nikita.
You
mentioned
that
the
tests
aren't
running
on
ci.
Is
that
something
that
would
be
possible.
C
C
A
A
A
Yeah,
I
thought
this
was
pretty
cool
because
I
didn't
know
how
that
was
going
to
work
with.
We
have
all
these
different
extension
jars.
That
are
all
we
want
to
be
isolated,
not
among
themselves
so
that
if
one
extension
uses
embeds,
something
that
doesn't
conflict
with
other
extensions,
but
we
still
need
one
class
loader
that
we
do
everything
in
that
then
pulls
these
in.
C
Well
that
in
fact
it's
quite
so
my
original
idea
was
easy.
I
just
write
a
class
loader
which
takes
like
a
bunch
of
delegates
and
delegates
to
them
like
in
order.
The
the
only
problem
problem
is
that
one
of
our
code
paths
is
actually
uses.
Get
resources
which
returns
enumeration
and
merge.
Several
enumeration
in
java
is
its
total
pain
in
the
butt.
C
C
C
C
A
So
the
so
there's
the
agent
class
loader
and
that's
how
you
jump
you
jump
over
into
the
tooling
via
the
agent
class
loader.
And
then
you
use
the
extension.
C
A
But
then
we
we
don't
use
it
directly
anymore.
We
only
use
extension.
Yes,
yes,
yes
got
it
nice,
so
we're
definitely
we're
planning
to
for
that
to
be
in
the
release,
get
these
merged
and
because
this
is
one
of
our
most
popular
issues,
extensions
we've
got,
I
don't
know
if
there's
a
way
to
run.
Reports
on
this
is
a
lot
of
thumbs
up
for
us
on
an
issue.
So
it's
one
of
our
most
popular
the
other.
Just
is
health
checks.
A
That's
probably
our
other,
the
one
that
I've
noticed.
Other
top
issue
really
should
look.
I
should
just
google
it
I'm
sure,
there's
a
way
to
search
sort
issues.
A
C
A
B
I
was
typing
I
was
just
I
just
put
in
the
agenda
item.
Just
you
know
tell
everyone
to
get.
We
need
to
get
1.3.0
upgraded,
get
that
stuff
released,
so
nikita's
work
will
actually
be
used
by
people.
A
So
yeah,
I
think
our
plan
we
discussed
with
honorag
earlier
this
week
was
to
release
target
this
for
early
next
week.
A
B
The
button
the
java
button
was
super
fast
and
seamless.
Yesterday,
very
unusually,
I
think
the
performance.
B
B
B
So
the
issue
was
so
we
added
in
1
1.2.
We
added
this
new
feature
with
that.
We
generate
public
api,
diffs
and
stable
modules
as
a
part
of
the
build
process
and
honorable
had
made
an
api
change,
but
had
not
run
the
full
build
locally.
So
those
api
diffs
changes
did
not
get
into
his
pr.
B
A
A
Oh,
we
wanted
to
ask
particularly
folks
other
folks
from
other
vendors,
nettie,
instrumentation
and
tyler
you
anybody
who
has.
I
don't
see
anybody
here
from
new
relic
today,
but
do
you
have
as
many
problems
with
neti
instrumentation
as
we
seem
to
in
our
repo?
For
example,
we
just
had
another
many.
A
Somebody
reported
this
we're
capturing
these.
It
looked
like
there
was
an
error
here.
We
would
capture
something
that
looked
like
an
error
on
connect.
That's
right!
This
is
really
tiny,
but
it
wasn't
really.
It
was
just
our
instrumentation
was
thinking
that
they
can
that
something
hadn't
gone
right,
but
it
really
had
gone
fine.
F
There's
so
I
remember
when
I
first
wrote
the
neti
instrumentation,
there
was
some
special
stuff
that
I
had
to
add
to
capture
when
a
client
request
was
being
made
and
the
connect
failed,
because
the
initial
handlers
didn't
capture
that
properly,
and
so
that's
where
this
like
connect
span
might
be
coming
from,
is.
Is
that
step?
I'm
not
sure.
A
C
By
the
way,
let
me
interject
for
a
second,
I
hear
a
lot
from
our
clients
that
who
want
to
know
who
want
to
get
information
about
all
connection.
Create
creation,
not
just
failed
ones
they,
so
they
want
to
like
they
want
to
know
if
new
unexpected
connections
are
created
instead
of
using
pooled
ones.
C
C
I
didn't
get
it.
Why.
F
C
You're
not
yes,
true.
That's
that's
exactly
like
counter
argument
that
we
may
need
low
or
low
level
for
to
provide
this.
This
kind
of
information,
one
one
more
piece
of
information
that
some
clients
are
interested
in
are,
for
example,
the
the
at
least
events,
if
not
spans,
of
dns
dns,
resolvement
and
tls
handshakes.
C
Yes,
that
totally
may
be
the
right
approach,
so
I'm
not
currently
saying
anything
about
how
to
represent
that
information.
I'm
just
saying
that
there
are.
There
are
interests
for
that
information,
so
yeah
either
metrics
or
an
attribute.
I
don't
know
events
or
metrics.
Yes,
yes,.
C
F
Yes,
the
other
thing
that
we're
probably
missing
is
nettie,
I
believe,
is
a
fairly
commonly
shaded
library,
and
so
you
know,
just
by
virtue
of
it
being
shaded
we're
not
capturing
a
lot
of
those.
A
I
noticed
a
couple
of
new
folks
on
today.
If
you
have
any
any
questions
or
didn't
want
to
share
anything,
I
would
love
to
know
who
you
are
and
what
you're
interested.
C
G
Joining
to
see
what's
up,
I
was
going
to
file
an
issue
talking
about
supporting
container
correlation
out
of
the
box,
so
I'll
give
that
file
shortly.
G
Like
like
resource
like
the
hotel,
yeah.
A
G
To
the
resource
out
of
the
box,
so
I'll
file
that
issue
talking
about
why
we
want
it.
A
H
Hey
I'm
ryan.
This
is
my
first
time
at
the
the
java
sig.
I
think
it
was
at
a
time
that
didn't
work
for
me
before,
but
happy
to
join.
Now
I've
been
kind
of
lurking
in
the
general
spec
sig
keeping
an
eye
on.
What's
going
on,
I
forgot
to
add
I'm
also
from
app
dynamics.
I
work
with
jason
and
yeah
between
us.
H
We've
got
like
a
fair
amount
of
java
agent
experience,
so
hopefully
we
can
fill
in
some
of
the
gaps
along
the
way
and
yeah
one
other
interest
that
has
come
up
from
our
side.
I
apologize
if
this
isn't
the
right
form
to
bring
it
up,
but
we've
got
some
kind
of
larger
customers
and
we're.
We
have
some
interest
in
more
of,
like,
I
guess,
a
higher
level
control
plane
for
things
like
instrumentation
or
configuring
sampling.
H
Things
like
that
from
what
I've
seen
thus
far,
the
only
ways
to
kind
of
control,
some
of
that
are
more
like
file
based
or
system
properties,
and
so,
if
there's
enough
interest,
I
guess
we'd
like
to
kind
of
help
explore
the
possibility
of
you
know.
Specking
out,
you
know
how
these
agents
and
sdks
specifically
java,
but
also
you
know
if
it
works
out
to
kind
of
you
know.
Other
sdks
might
also
want
to
a
similar
type
of
control
plane.
You
know
what
it
might
take
to
get
that
into
place.
A
Cool
so
like
the
agent
starts
up
and
contacts
the
control
plane
and
gets
like
a
current
configuration,
desired
configuration
from
the
control
plane
and
applies
them.
H
Yeah
yeah,
something
just
to
be
able
to
kind
of
centrally
dictate
like
any
instrument.
Instrumentation
library
with
these
resource
attributes
should
be
applying
this
instrumentation
just
a
mechanism
for
applying
custom
instrumentation,
like
maybe
vendor
specific
such
that
they
don't
have
to
go
to
every
individual
agent
and
load
a
specific
library
or
set
certain
properties.
They
can
kind
of
handle
that
at
kind
of
a
higher
level
of
abstraction.
A
Cool
and
if
I
do
you
mind,
if
I
ask,
are
you
because
I
know
you
have
your
own
java
agent?
Are
you
having
do
you
have
some
customers
now
using
the
open,
telemetry
java
agent
and
you
have
a
collector,
an
exporter
from
the
collector.
H
Yeah
so
we're,
I
guess,
hoping
to
move
toward
leveraging
the
commodity
agents
just
so
that
you
know,
customers
can
kind
of
you
know
in
an
open,
telemetry
ecosystem.
They
won't
have
to.
You
know
wrestle
with
deploying
hours
on
some
and
you
know.
Maybe
some
teams
already
have
the
open
source
one
available.
So
I
guess
it
kind
of
like
our
back
end
for
open
telemetry
we're
designing
to
just
accept
the
otlp
format.
H
H
Of
accommodate
that,
and
so
we're
trying
to
kind
of
find
this
middle
ground
of
like
functionality
that
was
in
our
existing
agent
that
we
know
will
want
you
know
what
would
seems
like
it
would
be
worthwhile
in
the
open
source
agent
as
well,
and
especially
for
larger
customers.
We
know
that
you
know
dealing
with
these
big
deployments
of
you
know
hundreds
or
thousands
of
agents.
If
the
configuration
is
on
a
you
know,
per
container
basis
based
on
environment
stuff,
it
can
be
a
bit
of
a
pain
point.
A
Yeah
anyone
from
nikita
john,
have
you
all
heard
on
the
splunk
side
interest
in
central
configuration.
B
H
H
Has
but
we
could
I
you
know,
I
know
the
language
is
capable
of
doing
those
kind
of
runtime
re-transformations,
whether
we
want
that
or
whether
the
community
kind
of
wants
that
in
the
open
source.
One
is,
I
guess,
tbd,
but
that
is
also
an
option
for
like
applying
custom
instrumentation
as
well.
A
Yeah,
so
we
have
shied
away
from
like
true
like
dynamic,
attach
where
you
know
your
app
is
fully
running
and
then
we
attach
just
because
it
is
more
complicated
and
like
some
of
the
instrumentations
like
did
we
already
like
if
you
already
created
your
grpc
connection,
we
normally
instrument,
you
know,
we
add
the
interceptor
when
you
create
it.
So
we've
already
missed
that,
so
we've
missed
that
opportunity.
To
so
adds
some
constraints.
I
guess
on
how
the
instrumentation
would
have
to
work.
H
Yeah,
I
I
understand
from
like
a
framework
perspective,
they're,
probably
plenty
of
use
cases
like
that.
I
mean
we've.
You
know,
we've
seen
use
cases
even
just
like
kind
of
pojo
instrumentation
like
the
customer.
Just
has
some
method
they're
interested
in
and
it's
very
convenient
to
be
able
to
just
go
and
while
the
application's
running
say,
hey,
I'm
actually
interested
in
that
method.
Could
you
create
a
span
there
or
even
to
go
further?
You
know
for
that
method,
I'm
actually
interested
in
a
certain
argument.
H
A
I
see
so
when
you're
talking
about
applying
custom,
instrumentation
you're
kind
of
talking
about
those
like
configuration
based
instrumentation
of
customers,
saying
I
want
to
instrument
this
method
and
capture
this
parameter
there
right
right,
cool.
Yes,
I'm
actually
very
familiar
with
that
in
glow
root.
A
Let's
see,
if
I
can
log
in
here,
if
it
will
remember
me,
do
you
remember
me
yeah,
you
can
add
new
instrumentation
on
the
fly
and
you
can
tell
it
you
know
what
to
capture
and
what
you
know
kind
of
your
template,
what
args
and
then,
when
you
save
it,
you
can
apply
it.
You
can
apply.
Let's
see
if
we
make
a
change
over
here,
I
will
break
break
the
demo
yeah
and
then
you
can
apply
without
restarting,
which
then
gives
you
this
friendly
warning.
E
A
A
A
I
think
something
like
this
is
way
cooler
more
useful,
but
that
pure,
like
most
at
least
our
customers
and
again
it
kind
of
depends,
I'm
sure
run
it
all
the
time
right.
They
install
it
and
run
the
java
agent
all
the
time
and
are
continuously
monitoring.
A
So
I
think
we
weren't
quite
sure
if
there
was
enough
like
what
the
but
I
was
very
interested
to
see-
and
I
would
love
if
there
was
somebody
from
new
relic
here
to
talk
about
sort
of
the
customer
demand
side.
Of
that
you
know
what
use
cases
that
was
addressing,
I
don't
know,
jason
or
ryan.
If
you
do,
does
the
app
dynamics
do
dynamic
attached?
Do
you
have
customers
doing
that?
You
do
supply.
H
It
it
is
supported
yeah,
I
I
wouldn't
say
I'm
not
too
regularly
communicating
with
customers,
but
I
I
guess
from
like
conversations
with
support.
I
I
don't
see
it
come
up
too
often,
I
think
it's
pretty
common
to
just
kind
of
go
with
the
dash
java
agent
just
start
up
the
jvm
with
an
agent
attached,
but
it
is
pretty
common
to
have
use
cases
where,
once
the
jvm
is
running
and
is
instrumented
that
we
have
these
kind
of
dynamic,
instrumentation
changes.
H
A
Oh
yeah
yeah,
these,
the
dynamic
instrumentation
saved
me
on
a
bunch
of
production
issues
where
you're
you
end
up
in
some
bad
state
over
some
time,
and
then
you
can,
you
know
dynamically,
add
some
stuff
and
capture
stuff
without
having
to
restart
and
lose
that
state.
C
A
I
think
we
should
support
really
what's
the
use
case,
yeah.
E
A
If
you
attach
in
the
like,
I
would,
I
would
be
for
supporting
attaching
in
the
first
line
of
maine
right
like
okay,
if
we
could
restrict
it
to
that,
or
you
know
just
do
that-
I,
I
would
be
good
with
that,
and
we
do
I
mean
it.
It
does
actually
work.
We
had.
G
A
We
had
yeah,
we
had
somebody
in
our
repo
wanting
to
do
that.
I
don't
know
where
they
got
even
came
up
with
this,
but
we
had
removed
the
agent
class
manifest
attribute
in
our
wrapped
in
our
distribution
of
the
hotel
java
agent,
and
so
of
course
they
couldn't
that
didn't
work
anymore.
A
So
yeah
I
was
like.
I
don't
think
we
want
to
do
that
and
then
I
was
like
okay.
Well,
I
guess
for
well.
I
must
have
put
some
comment.
Oh
yes,.
C
A
Cool
yeah,
I
think
underog
was
previously
also
in
favor
of
that
thought
that
that
would
be
potentially
even
interesting
like
we
could
do
some
spring
boot
configuration
at
some
point,
although
that
would
be
after
the
first
line
of
maine.
So
that
goes
against
my.
F
Yeah
in
my
mind,
anything
after
the
first
line
of
maine
you're,
risking
losing
the
opportunity
for
adding
fields
to
classes
that
are
are
necessary
because
falling
back
to
the
the
map
based
field
injection
is
a
huge
performance
penalty.
H
A
Back
to
jason's
static
right.
C
E
A
Happy
all
right:
we've
got
three
minutes
left
until
our
five
minute
cut
off.
B
A
Worse
because
we
hide,
we
do
hide
everything.
Almost
everything
gets
hidden
in
that
jar
under
this
inst
folder
and
all
the
class
files
are
renamed
to
class
data,
so
that
there's
actually
not
a
lot
that
gets
pulled
in
there.
A
H
C
A
A
B
A
All
right:
well,
we
hit
five
till
and
I
think
we
had
a
great
discussion.
I'm
really
excited
to
have
a
new
vendor,
joining
ryan
and
jason.
Thanks
for
giving
us
updates
and
hope
to
see
you
back.