►
From YouTube: 2020-11-18 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
A
B
First
thing
I
wanted
to
mention
was
just
that:
the
the
specification
freeze
for
tracing
baggage
and
context,
so
I
know
we've
been
talking
about
ga
or
rc
or
whatever.
We
want
to
call
it
for
a
while,
but
this
is
the
first
step
towards
actually
making
it
happen.
So
we
don't
have
too
many
things
left
to
do
for
this,
but
we
don't
have
a
lot
of
time
left
to
do
it
either.
So
I
just
thought
everybody
should
be
aware
of
that.
B
One
thing
I
did
notice
this
morning
was
that
our
baggage
api
was
still
called
correlation
context,
so
I
went
ahead
and
I
made
a
rename
pr
for
that.
B
B
One
thing
that
I'd
like
to
talk
about
that
was
brought
up
on
the
github
action
resource
detector.
Pr
is
where
should
these
resource
detectors
live
right?
Now
we
have
aws
and
gcp
resource
detectors
in
the
main
repo.
B
B
But
I
figured
we
should
talk
about
it
if
anyone
has
any
reason
that
they
should
live
in
the
main
repo.
I
guess
now
is
the
time
to
speak
up.
B
Yeah
and
I
think,
as
we've
added
things,
it's
it's
become
pretty
clear.
Most
things
should
be
in
contrib
anyways,
unless
they're
actually
core
shared
functionality,
which
these
are
not
so.
B
B
B
B
A
So
I
mean
one
thing:
I
think
that
we're
talking
to
have
before
ga
was
the
automation
tests,
some
integration,
but
I
think
it
was
probably
very
deep
somewhere.
B
Is
integration
testing?
Is
there
anyone
here
that
would
like
to
try
to
work
on
integration
testing?
I
think
it's
something
that
that
we
should
have.
I
know
that
it
comes
up
in
the
maintainers
meeting
every
now
and
then,
but
we
haven't
talked
about
it
in
a
while
there,
either.
A
B
Yeah
the
cncf
would
pay
for
that.
I
will
bring
that
up
with
the
governance
committee
and
I'll
try
to
find
out
how
we,
how
do
we
get
paid
accounts.
B
D
What's
your,
I
don't
know,
target
requirements
for
browser
stack
testing
because
we
do
have
something
for
hotel
right
now.
I
just
don't
know
if
it's
sufficient.
B
What
do
you
mean?
We
have
something
for
hotel
right
now,
who
is
we?
We
is
splunk.
Splunk
is
testing
open,
telemetry
internally
with
browser
stack?
Yes,
okay,
I
mean
essentially
right
now
we
run
all
of
our
browser
tests
in
a
headless
chrome
browser,
and
I
think
the
idea
is
just
we
would
like
to
test
them
on.
B
I
think
at
a
minimum
we
should
run
all
of
our
unit
tests
in
like
chrome,
firefox,
safari,
internet
explorer
and
edge.
D
Yeah,
we
do
actually
use
it
for
integration
tests
rather
than
unit
tests,
but
and
by
the
way
browser
stack
is
free
for
open
source,
so
auto
should
qualify
for
that.
B
Okay,
you
mentioned
a
paid
account.
Was
there
anything
not
on
that
free
tier,
that
your
features
that
you
think
we
would
need
to
use.
A
I
mean
I
was
always
using
the
paint
account
because
it
wasn't
open
source,
but
if
they
can
get
us
like
a
no
fully
working
account,
that's
fine
yeah,
okay,
but
someone
from
the
governors.
Anyone
needs
to
you
know
apply
for
that
here,
to
proof
it's
open
source
and
and
so
on,
yeah.
A
It's
basically,
you
know
connected
with
the
number
of
the
minutes
and
the
tests
that
can
be
run
in
parallel,
because
if
you
want
to
run
this
on
many
different
browsers,
we
need
you
know
they
should
be
running
quite
fast,
so
they
can
be
run
against
every
pull
request
and
this
will
consume
much
resources
than
just
you
know.
Standard
accounts
that
it's
free.
I
guess
yeah
I'll,
look
into
it
and
see
what
we
can
get.
B
I
guess
like
testing
and
releasing
is
something
that
I
think
needs
to
be
solidified
generally,
you
know
browser
testing
and
a
more
automated
release
and
a
stricter
release
schedule
and
things
like
that.
B
I
was
I've
been
working
on
the
github
action
for
the
testing
and
canary
release
this
morning.
I
know
we're
supposed
to
be
moving
off
of
circle
ci
because
the
cncf
wants
to
stop
paying
for
it.
So
I've
been
working
on
that
stuff
all
morning,
but
adding
browser
stack,
I
think,
is
a
good.
B
B
Sort
of
open-ended-
I
think
this
is
the
eks
resource
detector
reviews
wanted.
B
E
Yeah,
that
was
me,
and
actually
it's
more.
I'm
gonna
start
for
like
more
of
a
meta
question.
If
I
have
a
question
about
the
pr,
is
it
better
to
discuss
it
in
the
pr
via
text,
or
is
it
okay
to
discuss.
B
E
Okay,
well,
I
just
wanted
to
point
out
that
we
thought
the
pr
is
ready
to
go,
but
then
there's
some
concern
about
what
happens
if
there's,
for
example,
in
the
content,
encoding,
header,
gzip,
comma
identity,
if
that's
considered,
compressed
or
uncompressed,
that's
something
that
carlo
and
I
we've
been
pairing
on
this
issue
and
we've
been
working
on
it
at
new
relic,
and
we
didn't
really
know
the
answer.
To
that.
E
We
don't
know
if,
like
gzip,
comma
identity
means
that
you
know
a
proxy
got
the
request
and
then
passed
it
on
uncompressed
or,
if
it
just
didn't,
add
any
additional
compression.
If
that's
what
identity
means
like
a
second
identity
in
the
content
encoding
list-
and
he
was
running
some
tests
to
try
to
figure
out
what
that
heading
means
and
so
we're
not
we're
there
we're
just
kind
of
like
unsure
about
our
like
how
our
code
works.
E
We're
we're
just
testing
for
a
simple
identity,
and
that's
like
we
assume
that
that's
the
only
way
that
content
could
be
uncompressed
in
any
other
content.
Encoding
header
like
if
it's
gzip,
comma
identity.
We
would
consider
that
compressed,
and
we
just
don't
know.
I
don't
know
if
anybody
if,
if
anyone
here
knows
for
sure,
but
that's
something
that
we're
going
to
be
researching
for
that
pr
before
we
can
consider
it
like
ready
to
go.
B
Yeah,
as
far
as
I
know,
if
you
have
multiple
entries
that,
like
it
could
theoretically
be
compressed
twice,
so
you
know,
I
don't
know
why
you
would
do
that.
But
if
you
have
gzip
comma
identity,
I
would
expect
that
to
be
to
be
gzipped
essentially
to
be
functionally
equivalent
to
just
gzip.
E
Okay,
that's
what
I
thought
too,
and
that's
that
was
our
concern
is
like
carlo
thought.
Maybe
that
gzipcom
identity
means
that
gzip
was
applied
by
one
part
of
you
know
one
part
of
the
request,
then
maybe
a
proxy
got
it
and
passed
it
on
and
like
uncompressed
it,
and
that
would
I
don't
know
so.
E
That
was
something
that
he
thought
that
that
might
mean,
but
that
was
my
thinking
too,
is
that
it
was
just
like
a
list
of
encodings
that
had
been
applied
and
if
you
undo
an
encoding,
you
just
remove
it
from
the
header.
You
don't
list
out
like.
E
If
so,
if,
if
there's
some
kind
of
proxy,
that's
like
decompressing
the
request,
it
would
just
remove
the
gzip
anyway
yeah
thanks
for
the
feedback,
I
think
I'll
I'll
put
what
we
just
discussed
in
the
pr
and
just
to
kind
of
like
document
our
discussion
about
that
and
try
to
move
that
forward.
F
B
I
mean
it
should
it
should
be
relatively
easy
to
test
too.
If
you,
just
you,
know,
put
an
express
server
behind
engine
x
and
then
send
a
compressed
request
to
nginx
engine
x
should
decompress
it,
and
then
you
can
just
look.
What
does
the
header
look
like
in
express
total.
B
Okay,
thank
you,
yeah
no
problem,
I
mean
I
don't
really
have
much
else
to
say
just
because
we
are
just
moving
along
with
the
trace
rc,
but
I
did
want
to
talk
about
what
comes
next.
I
know
that
the
specification
is
making
a
push
for
metrics
next
and
they
would
like
to
have
a
spec
freeze
on
metrics
near
the
end
of
january.
B
I
know
bart
and
matt.
You
two
have
probably
done
the
most
work
on
the
metrics
in
js
have
either
of
you
been
following,
like
the
current
state
of
metrics
and
where
we're
at
and
the
specification
or
is
it
I'm
just
trying
to
get
a
handle
on?
What
is
our
general
state
with
metrics.
A
I'm
not
sure
if
you
were
listening
to
the
train
market
d
yesterday,
but
there
is
the
biggest
problem.
Is
that
what
we
have
is
cumulative
with
regards
to
delta,
and
currently
we
cannot
implement
correctly
the
the
metrics,
which
is,
for
example,
the
sum
observer,
because
the
our
badger,
sorry,
our
batcha,
is
using
the
aggregation
and
by
default
we
are
already
aggregated
the
data
and
what
we
should
really
do.
A
We
should
not
aggregate
this
until
we
know
what
is
the
requested
from
the
configuration,
because
the
same
instrument
like,
for
example,
someone
abdul
sum
observer
might
be
requested
to
either
collect
deltas
or
to
collect
cumulative,
which
basically
means,
even
though
you
choose
certain
metrics,
you
should
wait
with
the
aggregation
based
on
some
configuration
and
this
way
basically
it
means
we
have
to
refactor
this,
and
we
also
have
to
rename
a
few
things
and
that's
it.
Yeah.
B
Okay,
so
I
I
am
aware
of
the
the
issue
between
having
a
pre-aggregated,
you
know,
delta
versus
cumulative.
B
A
B
B
Okay,
then
I
know:
okay,
I
was
looking
through
the
go
implementation
and
a
little
bit
of
the
spec,
and
I
don't
know
if
our
components
just
have
older
names
but,
for
instance,
there's
like
this
processor
component
in
the
go
implementation
that
I
think
does
something
similar
to
what
our
bachelor
does.
Is
that
just
a
name
difference
or.
A
So
you
can
basically
replace
it
easily
here.
B
C
Yeah
I
was
kind
of
looking
through
all
this
as
well
and
just
trying
to
come
up
with
like
a
a
list
of
what
was
different.
I
failed
in
this
task
so
far,
just
because
there
are
like
so
many
there,
there's
really
so
many
moving
pieces.
I
guess
to
the
metrics
api
and
a
lot
of
it.
You
kind
of
have
to
go
digging
in
the
go
implementation,
which
I
think
is
really
the
reference
implementation,
but
yeah.
There
were
some
things
that
definitely
some
a
lot
of
things.
It
looks
like
we
can
reuse.
C
There
are
a
handful
of
things
that
are
just
going
by
a
different
name
and
some
slightly
different
behavior,
and
I
do
think
we
need
to
kind
of
come
up
with
a
plan
for
for
how
to
align
our
our
metrics
sdk
with
what
is
currently
being
specced.
B
C
Yeah,
I
think
yeah
so
tldr.
I
don't
have
all
the
answers
to
this,
but
I
think
this
is
work
that
that
we
should
do,
and
it
would
be
nice
to
have
kind
of
multiple
people
looking
at
it.
I
do
think
bart
probably
has
the
most
in-depth
knowledge
about
what
we
currently
have
and
how
it
works.
A
B
B
I
haven't
seen
him
around
in
a
long
time
months,
so
I
don't
think
he's
gonna
be
coming
back
too
much
and
it
seems
like
mayor
is
moving
on
as
well.
He
reached
out
to
me
recently,
and
I
I
don't
think
he's
gonna-
have
time
to
continue
being
a
maintainer
too
much
longer
in
the
future.
B
He
obviously
already
doesn't
come
to
these
meetings
because
it
conflicts
with
something
else.
So
I
think,
as
far
as
that
goes
he's,
probably
not
going
to
be
able
to
help
too
much
with
the
metric
stuff.
B
Okay,
well,
I
guess
I
don't
really
have
anything
beyond
that.
I
was
just
trying
to
get
a
handle
on
what
is
our
current
state,
this
jaeger
propagator
point
who
added
this?
B
D
F
C
So
if
I
remember
the
exact
wording
from
the
spec,
it
says
that
b3
and
jager
should
be
available,
but
kind
of
distributed
as
extension
packages.
B
Yeah,
what
jacob's
talking
about
is
the
environment
variable
spec?
So
if
you
configure
the
environment
variable
to
to
use
the
jaeger
propagator,
then.
C
Yeah,
I
think
that's
probably
true.
My
interpretation
of
that
would
be
that
if
somebody
has
included
the
extension
package
and
they
set
the
environment
variable,
then
it
should
be
set,
but
I
mean.
B
That's
what
I
would
prefer
to
maybe
this
is
a
good
question
for
the
for
the
maintainers.
C
Meeting,
I
don't
know
what
others
have
done.
I
can
dig
up
the
wording
about
at
least
the
extension
package,
because
I
tried
to
change
that
at
one
point
where
I
felt
like
you
know,
at
least
for
b3
is
what
was
on
my
mind
that
it
would
be
pretty
harmless
to
distribute
b3
as
part
of
the
sdk,
but
I
I
got
some
pushback
when
I
tried
to
relax
that
wording.
C
B
B
And
then
yeah,
I
guess
like
if
you,
if
you
configure
your
environment
variable
you
also
theoretically,
have
control
over
your
dependencies.
You
should
be
able
to
just
add
it
as
a
dependency,
if
you
need
it,
maybe
jacob
instead
of
moving
it
into
the
core.
Maybe
if
jaeger
is
configured
but
not
installed
log,
a
warning.
D
I
guess
that's
that'll,
be
okay,
too
yeah.
I
think
that's.
D
Deployment,
so
so
tell
me
this:
in
that
case,
the
jaeger
propagator
package
currently
has
a
version
of
zero
tan.
D
D
Because
I
mean
it's
going
to
be
pure
dependency,
we
agree
on
that
or
it's
or
not,
really.
B
B
B
Right
so
I
think
possibly
I
think
npm
has
the
concept
of
optional
dependencies.
You
may
want
to
look
into
that,
okay,
but
then
well
in
any
case,
you
would
just
use
whatever
the
current
latest
version
is,
so
that
would
be
0.10.
D
B
We
may
have
to
use
the
create
require.
In
that
case,
are
you
familiar
with
that.
D
B
F
D
I
think
I'm
good
for
now.
Okay,
thank
you.
F
B
Problem,
I
guess
that's
the
end
of
the
agenda.
Has
anything
anyone
have
anything
else
that
they
would
like
to
talk
about.
B
Okay,
then
I
guess
I
will
is
everybody
going
to
be
around
next
week?
I
just
realized
it's
actually
the
week
of
thanksgiving.
Should
we
still
have
this
meeting
or
should
we
cancel
it
for
next
week
or
what
do
you
guys
think.
B
A
C
Yeah,
I
I'm
definitely
interested
in
coming
up
with
a
plan
and
and
helping
with
that,
to
whatever
extent
I
can
so,
if
that
is
gonna,
be
a
topic
on
wednesday,
I
I
can
show
for
an
hour
and
then
take
the
rest
of
the
day
off.