►
From YouTube: 2021-01-27 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
E
My
headset
died
today.
So
sorry,
it's
a
computer
audio.
E
E
All
right
hey
this
is
this
is
the
first
time.
Yes,
we've
had
the
all
three
of
us
since
in
the
new
year.
A
A
A
Do
I
understand
correctly?
Okay
and
1.2
is
just
like
a
stability
guarantee.
This
is
stable
and
probably
will
not
change
in
backward
incompatible
way
so
and
do
understand
it
correctly
that
we
as
an
agent,
have
two
different,
separate
public
services.
One
is
telemetry
and
another
is
like
programmatic
api
that
we
have
and
probably
environment
variables.
As
part
of
that.
E
B
A
A
A
Yeah
but
in
specification,
tigran
has
a
proposal
about
like
telemetry
schema.
A
What
I'm
trying
to
say
is
that,
even
if
we
adopt
the
notion
of
telemetry
schema
for
me,
that
doesn't
mean
that
we
cannot
change
the
value
of
attributes
or
span
a
that
still
can't
be
changed,
because
well,
we
fix
bugs
well
whatever
what
can
be
fixed
only
is
like
if
there
is
a
an
attribute
with
such
a
name.
That
means
this.
E
B
B
C
A
If
user
uses
a
java
agent
that
we
will
export
also
telemetry
produced
by
third-party
library,
which
instrument
themselves,
which
means
that
we
cannot
guarantee
the
value
of
the
value
of
attributes
and
and
what
not
at
all.
Well,
how
can
we
distinguish?
These
paths
are
produced
by
agent
and
we
guarantee
their
stability,
but
these
pants
are
produced
by
third-party
library
and
we
don't
guarantee
anything
about
that.
That
doesn't
make
sense
to
me
like.
E
A
All
of
a
sudden
we
have
spring
boot
application
which,
which
you
use,
is
open.
Telemetry,
monitoring,
sprints
loose
at
last,
integrate
us
into
the
core
and
they
change
something.
A
E
A
E
A
So
my
opinion
that
I
am
going
to
express
in
specification
as
well
is
that
the
only
that
the
only
thing
that
we
can
claim
to
be
stable,
in
my
opinion,
is
the
meaning
and
type
of
attributes.
A
A
E
I
I'm
I'm
all
good
with
that.
We
tackled.
E
So
I
mean
it
sounds
like
primarily
the
the
pushback
has
come,
or
this
is
coming
from
light
step.
Then
so
it'll
be
good
to
see
what
ted
thinks
is
there.
I
don't
know
well
he's
been
driving
this
discussion
for
brought
it
up.
He's
brought
it
up
a
few
times
in
the
past.
A
E
Which-
and
I
think
I
recall
him
mentioning
it
like
way
back
as
well,
but
I
I
I
think
the
I
like
the
I
I
have
a
feeling
as
one
of
those
things
is
gonna
get
so
bogged
down
and
and
that
it's
gonna
just
get
pushed
to
like
some
eventual
2.0
release
where
that
they'll.
A
A
But
I
will
just
in
case
ask
ask
again
for
explicit:
do
we
understand
that
correctly,
just
in
case
but
yeah?
It
seems
in
the
theme
that
you're
right
so
which
the
only
thing
that
we
need
for
java
agent,
1.0
environment
variables.
C
E
We
haven't
spent
enough
time
and
I'm
sure
we're
going
to
want
to
as
we
spend
more
time.
I'm
sure
we're
going
to
want
to
make
lots
of
changes.
A
E
Probably
the
best
the
I
think
I
was
thinking
the
instrumentation
api
and
the
java
agent
api.
Okay,
those
are
those
are
a
mess.
I
mean
not
a
mess,
but
there's
a
lot
of,
I
think,
there's
a
lot
of
work
to
be
done.
There
still.
B
A
B
A
Still,
I
think
that
typed
instrumentation
instrumentation
module
that
system
is
well
good
enough.
For
now
I
mean
if
my
recent
pull
request
documenting
all
those
extension
points
or
extension
possibilities
is
relevant.
Then,
what's
big
things
missing
in
instrumentation,
around
type
implementations
and
instrumentation
modules.
B
E
Oh
yeah,
I
think
that
was
part
of
that
maven
artifacts
issue
that
never
got
finished.
They
were
gonna,
encourage
them,
artifacts,
merging.
B
B
Agent
installer
would
then
be
moved
out.
That
was
the
plan
because
that's
not
part
of
an
api.
B
B
A
E
The
ones
I
think
that
we're
missing
are
the.
I
think
that
all
the
global
ones
are,
but
I
don't
think
we've
documented
the
the
instrumentation
specifically
okay.
A
B
A
B
A
A
Okay,
does
htk
auto
configuration
module
documents,
its
environment
variables.
A
B
A
B
B
B
A
E
Yeah,
that
was
a
great
thank
you
for
pushing
all
of
that
into
the
sdk.
B
A
A
Can
we
not
repeat
that
date
with
the
totally
unclear
source,
but
but
but
taking
into
account
these
code,
which,
essentially
more
or
less
just
reviewing
our
configuration
story
and
documenting
that
and
the
sdk
is
planning
to
have
like
a
release
candidate
next
week.
A
So
I
I
would
say
that
we
well
that
we
don't
need
any
definite
deadline,
but
we
have
to
create
a
one,
zero
milestone
and
make
sure
that
it
has
necessary
tasks
and
just
well
just
do
it.
It
should
be.
B
A
B
I
think,
like
from
what
I
understand
this
date
came
from
marketers
and
there's
going
to
be
lots
of
different
companies
doing
a
lot
of
marketing
on
the
date.
That's
my
impression
yeah,
and
so
I
just
like,
and
so
that's
why
the
whole.
The
idea
is
like
yeah,
something
something
one
point,
or
at
least
it
doesn't
have
to
be
the
agent
and
so
we're
already
discussing
like
the
agent
is
going
to
come
later.
It's
fine
we're
sort
of
wondering
when
we
would
talk
about
it,
but
it's
not
as
important
powerful
business
just
because.
A
B
B
B
A
A
E
We
did
yeah,
we
had
snow
today.
I
don't
think
it's
related,
but.
E
A
E
All
right
shall
we
roll
into
100
triage
then.
D
B
A
A
Some
kind
of
gega
when
we
have
at
least
like
the
top
tier
instrumentations,
already
programmatic
interfaces
already
metrics
already.
So
we
are
ga
like
we
still
have
some
ga
notion,
but
for
1.0.
What
I
pro
is
to
make
it
is
to
make
a
milestone
and
just
assign
those
tasks
that
we
want
for
1.0,
which
is
like
fix
our
environment
variables
and
configuration,
and
that's
it.
E
B
A
C
E
Yeah,
I
think
this
issue
is
scoped
down
now
to
the
remaining
parts
which
were.
B
E
Because
it's
a
little
different
than
right,
normal
libraries
and
dependencies
since
we
bundle
there's
different
licenses,
there's
different
license
restrictions
or
things
that
apply
to
bundling.
E
The
cncf
help
desk.
That
was
how
like
with
the
github
actions
stuff
like
and
we've
gotten,
the
only
the
technical
committee
can
open
cncf
help
desk
issues,
though,
but
I
think
that's
generally
built.
A
A
B
And
we
suspect
we
have
to
embed
these
licenses
into
the
jars
somehow,
whether
there's
a
directory
structure,
that's
expected
or
file
format,
we
don't
know
and
our
expectation
could
be
wrong.
Maybe
nothing
is
required
that
I
doubt,
but
there
has
to
be
some
way
that
complies
with
their
legal
department.
Basically,.
E
And
I
didn't
see
any
like
they
have
some
guidance,
but
there
was
not.
A
E
So
I
mean
I
guess
we
could
start
just
by
asking
kind
of
a
general
generally,
if
there's
any
specific
guidance
that
they
already
have,
that
they
could
point
us
to
for
projects
that
embed
binary
distributions
of
third-party
code.
B
We
could
also
help
this
test
by
maybe
also
explaining
what
we
already
do
like.
I
know
the
shadow
plugin
does
make
an
attempt
to
do
something
like
it
seems
to
copy
palm
files
or
something
like
that
in
the
pump
files
into
a
license,
and
maybe
that's
enough,
maybe
not,
and
so
we
do
do
something
right
now.
So
I
think
right
now
we're
also
missing
sort
of
our
current
state,
which
I
guess
we
could
write
up
in
this.
B
E
Well,
anurag
was
so,
I
think
what
we
need.
I
mean
reach
out
to
cncf
help
desk
ask
if
they
have
any
specific
guidance
on
embedding
binaries.
A
E
And
then
there's
probably
because
I
know
the
let's
see
it
looks
like
I
already
docked
the
different
licenses
so
like
mit.
I
don't
think,
there's
well
and
I'm
not
sure.
I
know
the
one
thing
that
is
different
about
apache
is:
if
any
of
these
projects
have
a
notice
file,
then
we
have
to.
A
A
B
E
Okay
yeah,
this
one
is
not.
I
assume
we're
okay
with
this,
not
going
one
zero
or.
E
And
I'll
span
attributes,
so
this
was
when
I
was
concerned
about
telemetry
stability.
E
A
A
E
E
E
E
D
E
E
A
E
But
that's
that's
the
there.
That's
the
byte
code.
Now
honorable,
doesn't
one
zero
mean
byte
code,
no
byte
code
compatible
api
yeah.
So
you
you
would
not
add
a
new
method
to
an
interface,
a
non-default
method,
to
an
interface.
A
A
A
B
A
E
A
A
You
tr,
even
if
that's
what
you're
trusting
is
true
and
joey's
api
will
only
add
def
default
methods
to
interfaces.
That
does
that
mean
that
we
java
agent
are
okay,
because
if
we,
if
we
intercept
some
methods
of
interfaces,
but
not
all
of
them,
will
we
continue
to
work?
Can
we
guarantee
that
we
will
continue
to
work.
A
B
A
B
E
E
Okay,
I
kind
of
thought
this
was
a
okay
yeah,
maybe
you're
right.
Your
lemonade,
shaded
properties
are
not
docked.
E
Okay,
let's.
A
Or
you
can
just
assign
that
to
me
and
I
will
write
a
comment
and
close.
It.
E
D
D
D
A
E
E
E
E
Spring
boot
starters,
grizzly
logging
doesn't
record
exceptions
flushing
on
well
flushing
spans.
I
think
mateosh
was
gonna.
Do
this.
E
E
E
Document,
okay,
nice
to
have,
but
not
required
for
g,
not
required
for
one
zero
yep.
E
D
D
D
E
E
D
E
D
Got
it
trace
your
name.
E
Nested
client
instrumentation,
yes,
revision.
D
D
E
I
think
this
one's
closed
now
I'll,
stop
derailing,
mark
jetty.
E
E
E
This
is
gone
right.
This
is
all
different.
E
E
E
E
A
E
B
E
All
right
and
then
yeah,
if
you
all,
have
thoughts
for
how
we
can
release
party
for
1-0,
we
should
plan
something.
B
E
E
A
D
E
The
there's
a
lot
of
a
lot
of
activity.
I
don't
know
how
you
keep
up
with
two
repos
honorary.
I
can
barely
keep
up
with
one.
B
C
A
E
Yeah,
look
at
the
last
my
last
pr
was
probably
weeks
ago.
Well
that
doesn't
count
but
yeah
there
you
go
yeah.
I
have
so
same.
Yes,
I
like
doing
what
I
can
just
to
keep
up
with
the
pr
reviews
and
following
other
repos.
E
But
it's
good
thanks
for
bringing
nikita
all
of
your
team
on
board
yeah.
Definitely.
B
B
A
B
A
E
I
always
just
like
when
I
run
into
a
template,
I
typically
just
erase
it
and
start
the
scratch,
but
I
mean
that's
because
I
know
I'm
going
to
write
out.
You
know
like
so
at
least
then
it
would
be
intentional
on
somebody's
part
to
ignore
the
template.
A
B
A
But
then
then
I
I
see,
for
example,
specification
some
pull
request
which
just
has
have
emptied
the
empty
template.
Template
is
in
place,
but
just
that
is
not
filled,
which
is
even
even
more
horrible.
In
my
opinion,.
E
I
I
I've
been.
I
think
I
started
to
write
a
response
on
this.
I
was
looking
over
this
code,
so
I
will
do
that.
That's
the
and
that's
all.