►
From YouTube: 2021-04-15 Governance Committee private 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
D
Yesterday,
with
with
using
open,
telemetry
javascript.
D
C
Is
that
a
documentation,
buff
or
a
how
it
happens?
Buff
is.
D
A
B
A
Yeah
it
takes
some
time
to
you,
know
get
I
mean
if
you
have
all
the
npm
packages
and
everything
well
documented.
I
mean
it's
kind
of
a
scattered
experience.
Right
now
I
mean
when
we
added
the
x-ray,
you
know
propagators
and
the
resource
detectors.
Again.
It
was
just
something
that
it's
still
work
in
progress,
yep,
there's.
C
A
A
Something
that
I'd
love
to
see
more
folks
in
the
dc
take
up
on.
C
Okay,
separate
maintainership
for
contrib
repos.
I
suspect
that
something
we
can
push
if
we
need
to.
E
Yeah,
so
just
the
the
background
there
as
part
of
figuring
out
our
story
around
instrumentation
there's
just
the
central
issue,
which
is
we
don't
feel
like.
We
currently
have
capacity
to
maintain
instrumentation
to
a
large
degree,
that's
problematic,
because
for
one
you
know,
one
of
the
main
purposes
of
this
project
is
to
provide
high
quality
instrumentation
to
people.
So
it's
problematic.
If
we
can't
do
that
at
the
same
time,
it's
definitely
the
case
that
we
don't
appear
to
have
the
sig
staffed
in
such
a
way
that
right
that.
E
Yeah
so
part
of
it
is:
we
need
to
actually
have
a
game
plan
together
for
like
how
we
are
going
to
approach
instrumentation.
This
is
something
we're
actively
working
on
right
now,
I'd
like
to
devote
more
time
to
it,
myself
and
part
of
that
is:
how
are
we
going
to
manage
this
ecosystem?
E
There
really
is
a
tension
here
between
just
if
we
just
to
make
it
clear.
The
problem
is,
if
we
say,
hey
people
just
go,
write
your
own
instrumentation
for
the
time
being,
like
we're
just
going
to
say
like
no,
it's
better,
you
go
maintain
it
over
there.
Then
it's
least
it's
clear
who's.
Maintaining
this
thing,
rather
than
saying
yes,
come
in,
we'll
add
your
thing
to
the
garbage.
E
We
call
contrib
and
then
we'll
make
you
like
an
owner
code
owner
of
that
and
then,
like
everyone's
gonna,
be
confused
like
just
go,
maintain
it
over
there.
If
we
do
that,
then
it's
not
clear
how
we
install
those
things
for
people,
because
open
telemetry
really
really
benefits
from
that
instrumentation
being
discoverable
and
automatically
installed
for
people
to
the
degree
to
which
we
can
do
that.
E
That's
like
a
huge
difficulty
in
getting
started,
because
if
you
fail
to
install
a
piece
of
instrumentation,
it
breaks
context,
propagation,
end
users
who
don't
have
a
lot
of
information
or
understanding
of
what
they're
supposed
to
be
getting
it.
Just
like
the
more
we
can
install
that
stuff
for
them,
the
better.
E
So
if
we
say
instrumentation
should
be
written
somewhere
else
which
we
have
to
no
matter
what
there's
always
going
to
be
instrumentation
written
somewhere
else.
We
have
to
solve
that
problem.
So
that's
actually
like
something
that
requires
like
design
and
thought
and
effort
to
be
put
into
it
like
we
have
a
registry.
So
we
think
we
can
improve
this
registry
to.
You
know,
maybe
actually
even
being
an
api
that
these
things
like,
like
we
use
like
as
part
of
open
telemetry,
to
identify
what's
available
in
a
operational
sense.
E
That's
like
a
heavy
load,
we're
going
to
put
on
a
service
if
everyone
using
open
telemetry
is
pinging
some
kind
of
api
when
they're
trying
to
install
open
telemetry,
and
then
we
need
to
also
solve
the
like
security
aspect
of
this
to
some
degree,
which
is
like
open
source
in
21st
century
is
a
supply
chain
attack,
and
this
is
like
I,
I
just
there's
a
general
feeling
like
like.
E
D
Chain
attacks,
my
hope
is
that
the
tc
will
take
up
incident
management
for
the
supply
chain
attack
that
just
impacted
us.
A
Yep
yep
I
mean
I
I
was
talking
to
liz.
I
was
talking
to
some
of
the
maintainers
this
morning
and
again
you
know
they
were
also
asking
about.
You
know
we
should
define
some
process
for
triaging.
A
You
know
when
there
are
incident
reports
of
this
sort
and-
and
you
know
some
kind
of
a
process
because
there's
no
process
right
now.
Right
I
mean:
do
we
all
assemble
in
a
triage
room?
Do
we
you
know
what
decisions
do
we
need
to
take
and
how
do
we
handle
p0
vulnerabilities.
A
Yep
yep,
but
I
mean
I
don't
know
ted.
If
you
heard
about
the
curl
sorry,
the
code
code,
cov.
D
D
D
So
right,
the
question
is:
just
you
know,
double
checking.
Do
we
have
any
secrets
that
could
have
been
exfiltrated?
Do
we
need
to
rotate
them
number
two
like
do
we
need
to
you
know
you
know:
do
we
do
we
need
to
rebuild
any
artifacts
right
that
might
have
been
tainted.
E
Yeah,
so
that's
that's
the
thing
and
here's
how
I
think
the
blue
check
approach
to
instrumentation
doesn't
fully
solve.
The
problem
which
is
instrumentation,
is
a
requirement
which
means,
if
someone
needs
kafka,
js
plug-in
like
if
they
need
kafka,
js,
instrumentation
and
job
supplies,
kafka,
js,
instrumentation
for
open
telemetry.
E
It
doesn't
matter
if
there's
a
blue
check
next
to
the
name
or
not
like
that's
just
what
the
end
user
is
is
stuck
with.
So
in
a
world
where
we
have
like
multiple
copies
of
instrumentation,
it's
like
helpful
to
know.
Oh,
this
is
the
trusted
one,
but
but
this
is
like
gonna
be
like
long-term.
Managing
our
instrumentation
ecosystem
is
like
the
big,
the
the
big
effort
in
this
project,
so
anyways.
C
Sorry
can
we
not
leverage
an
existing
repository,
sorry
registry
of
some
sort
or
several
of
them
in
order
to
do
this,
and
instead
of
talking
about
our
own
registry,
because
that
always
makes
me
nervous.
E
Yeah,
so
so
the
first
thing
that
needs
to
be
investigated-
and
I
was
talking
about
this-
with
some
aws
people,
jana
and
crew.
Yesterday,
we
did
by
the
way
at
the
end
of
the
lambda
call,
which
is
like
first
thing
that
needs
to
be
investigated
is
this
is
a
problem
that
gets
like
language,
specific,
so
kind
of
like
step.
One
is
actually
like
what
are
the
needs
and
requirements
in
every
language.
E
A
That's
great
exactly
exactly,
and
that's
precisely
the
point
so
I
mean
sarah.
I
think
that
you
know
we
we
absolutely
as
a
project
should
be
doing
an
evaluation
of
existing
tools.
You
know
that
are
just
used
like,
for
example,
github
is
a
container
registry,
but
that's
only
for
you
know,
I
mean
very
specific
usage.
E
Yeah
and
specifically
like
requirements
gathering,
we
just
need
to
understand
for
each
language
what
are
the
requirements
here
so
that
whatever
tools
we're
picking,
we
know
they're
gonna
work,
so
so
that
that's
effort,
you
know
that
that
needs
someone
to
do
do
research.
This.
C
This
may
actually
be
in
in
the
vein
of
the
same
question
that
I
got
asked
last
week
or
a
week
and
a
half
ago,
which
was
anarang
from
aws
mentioned
last
week
that
he
and
ted
were
chatting
about
the
costs
of
about
possibly
hosting
vendor
java
agent
distros
in
the
hotel
repos.
C
What
is
this
you
know?
What
is
this?
What's
the
back
story,
what
does
this?
Does
this
sound
familiar
ted
first,
yeah.
E
So
I
think
it's
familiar
to
both
me
and
alawita.
I
believe,
assuming
this
isn't.
Some
other
thing
we
are
specifically
talking
about.
Lambda
is
the
area
where
this
has
come
up,
where
we
want
to
start
publishing
artifacts
for
people
to
to
run
open
telemetry
on
lambda
right.
E
So
there's
you
know
the
thing
that's
most
easy
for
people
is
to
have
these
lambda
layers
be
published
in
every
region
that
have
open
telemetry
stuff,
and
then
you
hit
this
question
of
like
okay,
so
the
open
telemetry
that
you
get,
which
is
like
a
collector
combined
with
like
some
python,
open,
telemetry
or
whatever,
what's
actually
installed
there
and
it's
especially
tricky
with
lambda,
because
the
actual
image
size
matters.
E
The
kitchen
sink
collector
in
there
is
actually
problematic
because
the
the
what
the
user
pays
for
is
like
total
image
size,
and
that
thing
is
large.
So
we're
like
looking
at
stripped-down
ones
and
the
initial
thing
was
like.
Well,
you
know
amazon
can
make
made
like
a
stripped-down
one
yeah,
but
then
this
led
to
also
naturally,
all
the
vendors
being
like
well.
But
I
want
my
thing
in
there
and
then
I'm.
C
A
Yeah,
but
I
mean
sarah,
the
specifically
what
you
know
again,
my
guidance
has
been
on
on
our
aws
side
is
to
actually
make
sure
that
hotel
has
a
lambda
layer
available
for
any.
You
know
a
generic
one
is
a
generic
enough
with
otlp
support
that
actually
is
serves
as
a
baseline
and
then
anywhere
anyone
any
vendor
or
anyone
else.
You
know
who
has
a
component
that
they
want
to
add
to
a
initial
configuration,
can
can
easily
do
so
right.
A
So
exactly
as
ted
said,
you
know,
can
we
actually
support
otlp
as
a
as
in
baseline,
which
means
that
having
the
otlp
exporter
and
the
and
and
just
that
layer
being
available.
C
A
Also,
you
know,
we
plan
to
add,
build
your
own
documentation
and
you
know
script
so
that
that's
easy
to
build
out,
but
providing
at
least
a
baseline
on
the
project
which
is
maintained
by
us
and,
at
the
same
time,
yeah.
E
E
Right
yeah,
we
made
some
sorry
yeah.
We
made
some
progress
on
this
in
the
last
landed
meeting
yesterday
as
far
as
like
kind
of
really
getting
this
to
gel
up,
which
is
to
provide
basically
like
upstream,
a
cicd
process
that
publishes
all
these
lambda
layers
for
a
super
stripped
down.
Super
basic
version
of
this
that
just
exports
otlp
and
just
contains
the
amount
of
like
x-ray
stuff
that
you
actually
need
to
run
this
stuff
on
lambda,
because
lambda's
got
some
restrictions
right
now
around
like
what
headers
it
uses,
anyways
and
then
downstream.
E
C
C
But
you
know
what.
A
It
does
serve,
as
is
in
reference
implementation
that
actually
can
be
then
extended
for
any
other.
You
know
serverless,
you
know
implementation,
so
the
idea
is
not.
You
know
that
it's
lambda
specific,
it's
just
that
it's
a
reference
implementation
and
then
that
can
be
actually
made
available
for
any
any
provider.
Can.
C
We
potentially
publish
it
then
as
a
minimum
artifact.
You
know
reference
implementation
kind
of
thing,
as
opposed
to
a
lambda
artifact
yeah
yeah.
A
Be
very
specific,
unfortunately,
on
the
implementation,
when
you
are
providing
an
initial,
you
know
reference
right
and
then
we
can
extend
it.
E
And
the
secret
here
is
that
what
we
want
to
offer
to
the
community
is
a
a
cloneable
cicd
platform
yep.
Yes,.
A
E
C
F
My
my
my
concern
here
is
not
about
why
data
splunk
is
not
there.
My
my
big
concern
is
why
aws
lambda
versus
azure,
serverless
or
or
gcp
function.
So
this
is
where
the
question
we
need
to
answer
more
than
I
think
I
think
for
for
the
for
the
other
one.
As
long
as
we
support
otlp,
we
are
done.
We
can
say
that
hey
otlp
is
the
writing
to
design
thing,
but
why
aws
lambda
versus
azure.
C
A
I
think
I
think
bogdan,
it's
just
a
question
of
adding
those,
so
the
idea
was
that
again,
you
know
with
the
reference
example
available.
Initially
then
azure
functions
or
you
know,
google
serverless
implementations
can
be
available.
So
I
mean
again,
I
don't,
I
think
it's
just
timing.
Also
yeah.
F
But
but
but
I
think
that's
where
the
questions
to
sarah
came
from,
I
don't
I
I'm
not,
I'm
I'm
pretty
confident
that
they
did
not
ask
about
insight.
They
more
likely
ask
about
why
azure
function
is
not
there
so
so,
which.
C
C
A
A
Already
reached
out,
I
mean
I've
already
spoken
with
punier's
team
at
google
and-
and
you
know
they
were
totally
going
to
I
mean
they
have
been
involved
in
the
lambda
discussions.
You
know
which.
C
A
C
A
C
Yeah
totally
exactly
totally
understand
that,
and
and
no
I'm
not
suggesting
that
I'll
lisa
should
pony
up
engineers
to
go
make
it.
You
know,
make
a
reference:
implementation
for
for
microsoft,
functions
or
azure
functions,
but
yeah
you
know
just.
I
want
to
make
sure
that
that
there
is
not
a
first
among
equals
perception.
E
E
A
like
minimal,
open,
telemetry
thing
either:
that's
like
the
collector,
with
all
the
things
or
minimum
whatever
it
is
like,
like
some
non-vendor
specific,
open,
telemetry
thing,
and
but
that
thing
is
then
clonable
for
people
to
adjust.
However,
it
is
that
they
need
and
so
open
up
upstream
we
we
provide
like
the
reference
architecture,
for
how
one
publishes
this
stuff
or
whatever,
maybe
maybe.
A
F
It
would
be,
it
would
be
appreciated
from
from
the
competition
of
aw
yeah.
A
I
agree,
but
again
we
had
initially
proposed
serverless,
you
know
as
a
group,
but
I
think
because
it
was
specific
to
the
initial.
You
know:
implementation
discussion,
yep
totally.
E
D
And
I
think,
with
regard
to
other
kinds
of
instrumentation,
I
anticipate
less
of
a
problem
right
because,
when
it
comes
to,
like
you
know,
a
framework
like
express
right
like
either
it
is
baked
into
the
open,
telemetry,
contrib,
repo
or
the
framework
right
decides
to
just
make
the
hotel
apis
into
their
into
their
code.
Right
like.
I,
don't
think
that
I
don't
anticipate
to
have
a
need,
for
you
know
a
bunch
of
third
parties
that
are
not
the
library
framework
and
they're,
not
in
hotel
contrib.
D
Right
like
I,
don't
anticipate
that
that's
going
to
be
a
very
common
case
like
it's
just
this
particular
case
where,
where
you
know,
obviously
aws
is
not
going
to
be
modifying
every
single
lambda
function,
that's
run
and
where
it
can't.
Obviously,
you
know
easily
go
into
contrib
repo,
where
we
have
this
kind
of
template
pattern.
That
needs
to
be
done.
B
C
B
D
Right,
like
you
know,
I
applaud
lightstep,
and
I
applaud
everyone
else.
Who's
done
like
these
just
distros
or
launchers,
right
like
you've,
very
clearly
brand.
You
know
this
is
the
proprietary
bit
and
this
is
where
the
upstream
bits
are
and
we
support
either
right.
The
thing
we
don't
want
to
see
is
right
like
where
people
are
like.
You
know
you
are
on
your
own
if
you're
not
using
our
launcher
right,
because
that
that
sucks,
right
and
and.
E
A
A
No,
no,
I
mean
the
and-
and
you
know
the
objective
exactly
was
especially
for
the
lambda.
You
know
implementation,
it's
very
easy
for
aws
to
go
and
say
that
you
know
it's
very
aws
specific,
so
we'll
just
do
it
on
the
distro,
but
I
was
very
specifically
interested
in
hotel.
Actually,
having
a
you
know,
agnostic
layer
available
which
can
be
extended,
for
you
know
just
as
a
reference
which
can
be
then
used
by
any
anybody
to
extend
right.
So
if
you
have
to
maintain
that
we
are
committed
to
that
upstream.
E
And
we
do
need
I
mean,
I
know
we're
heads
down
just
trying
to
ship
right
now
and
that's
fine,
but
we
are
gonna
need
to
concretely
define
and
hopefully
legally
back
terms
what
distros
mean
because
well,
I
think
the
core
community
is
generally
going
to
be
like
focused
on
doing
the
right
thing.
It's
going
to
be
inevitable
that
two
things
happen:
someone's
gonna
ship,
a
client
thing,
they're
calling
open
telemetry
that
has
open
telemetry
api
extensions
right
like
yeah.
E
It's
open
telemetry,
but
with
this
better
api
with
more
features
so
like
use
this
thing
and
get
stuck
on
our
distro
quote-unquote
right,
that'll
be
one
form
of
embrace,
extend
and
then
the
other
thing
you'll
see
is
like
people
shipping,
some
kind
of
mystery
meat,
collector
like
where
they're
calling
an
open,
telemetry
collector.
But
it's
it's
just
like
a
bunch
of
proprietary
who
knows
what
and
and
we're
just
it'll,
be
better
to
figure
out
as
a
group
like
what?
E
What
do
we
want
these
things
to
be
defined
as
at
least
in
terms
of
terminology,
so
that
when
someone
does
ship
one
of
those
things
it's
open
source,
you
can
be
like
yeah,
you
can
ship
it,
but
you
just
can't
call
that
thing
open
telemetry,
like
you
have
to
call
that
something
else,
because
it's
either
mystery
meat
or
it's
like
a
proprietary
extension
of
something
that
that
we
think
you
shouldn't
extend
or
whatever.
Yes,.
A
A
We
just
don't
have
the
you
know,
focus
right
now
to
actually
kick
this.
C
E
C
A
You,
that's
all
folks,
folks
are
just
discussing
it,
but.
A
Sure,
yeah
and
just
one
more
thing
I
wanted
to
bring
up,
which
I
didn't
add
here
I
will
is
the
incubation.
You
know
update
status
where.
C
A
And
and
morgan-
and
I
have
been
following
up
with
elena
on
you-
know,
any
items
that
she's
requested
she's
also
been
on
the
dd
channel
on
slack
so
again.
Anyone
you
know,
please
feel
free
to
respond
to
her
whenever
she
has
questions,
but
I
think
they're
in
the
process
of
just
interviewing
and
chatting
with
different
customers
and
end
users
right
now,
yeah.
C
Cool
sarah,
have
you
heard
anything
else
I
have
not.
I
have
not
had
anything
come
my
way
that
wasn't
in
the
slack
chat
the
the
shared
slack
channel
or
previous
requests
that
have
been
handled
or
routed
to
the
shared
channel.
Of
course,
now
I'm
looking
for
where
my
cncf
connection
is
on
slack
there
we
go
no
little
red
dots,
brilliant.
E
One
final
thing
reminder:
we
have
a
user
study
out.
There
is
a
user
research
slack
channel,
matt
mccleary
has
done
a
great
job,
putting
together
a
user
study
form
and
we're
at
the
phase.
Now
we
should
be
shouting
that
out
and
spreading
it
around.
So
just
fyi.
If
you
haven't
checked
that
out,
please
tweet
it
out
or
and
share
it
within
your
organization,
otherwise
try
to
get
get
feedback.
Super
helpful
looks
like
we've
got,
maybe
15
responses
so
far.
So
that's
great.
A
Yeah
another
note
great
good
meeting.
Thanks
again,
everyone
have
a
good
day,
bye,
bye.