►
From YouTube: CNCF Serverless WG Meeting - 2019-09-19
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
A
A
A
A
We
really
didn't
have
that
many
questions
and
definitely
I,
don't
think
we
had
any
hard
questions
or
it
seemed
like
anybody
was
questioning
our
existence
or
why
were
even
going
for
this
I
think
the
next
step
in
the
process
is
for
us
to
formally
make
the
request
through
a
proposal
file
in
the
TOC
s,
get
a
repo
which
Mark
and
I
are
working
on
offline,
which
hopefully
be
done
either
today
or
tomorrow
and
I.
Believe.
A
The
next
step
in
COCs
perspective
is
to
actually
start
the
formal
vote
which
I'm
not
exactly
sure
when
that
starts,
but
hopefully
you'll
be
very
soon.
I
know
that
Chris
anatec
is
very
eager
for
us
to
not
only
get
the
1.0
but
to
also
have
this
next
status
for
us
to
go
to
incubator.
So
we
can
do
a
double
announcement,
both
at
the
same
time
and
get
even
more
press
and
an
excitement
around
our
stuff.
So
that's
kind
of
what
happened
there.
Mark
can't
leave
anything
that
I'm
forgetting
no.
C
I
I
thought
there
went
very
well
thank
you
for
putting
together
all
that
all
those
slides
and
doing
the
presentation
gives
a
great
job.
The
I'd
say
the
the
one
area
that
or
the
one
question
that
came
up-
that
we
didn't
have
an
answer
to
was
a
pro
the
buff
spec
that
we
have
explicitly
excluded
for
now.
But,
aside
from
that,
I
don't
know
that
there
was
any
any
questions
that
didn't
make
sense
and
that
we
can
answer
yep.
A
All
right
any
questions
from
the
community
on
the
meeting
or
proposal
that
I
put
forward
all
right,
cool,
moving
on
then
koukin,
nothing
worth
mentioning
there
I,
don't
think
anybody's
made
any
changes
to
the
outline
for
our
sessions.
Obviously,
sometimes
that's
not
an
issue.
However.
I
did
want
to
mention
that
the
service
practitioners
summit
CFP
is
now
live.
I,
don't
know
why
a
note
wasn't
sent
out.
A
I
am
I
expected
one
will
be
sent
out
soon,
but
if
you
actually
go
to
the
website
for
the
I
think
it's
called
day,
0
co-located
events
or
something
like
that,
you
will
see
that
listed
there
and
here
is
a
link
to
the
CFP.
If
you
guys
do
want
to
submit
one,
it
closes
October
4th.
So
we
don't
have
a
whole
lot
of
time.
If
you
guys
want
do
something,
if
someone
thinks
that
there
should
be
something
there
from
our
group,
you
know
please
speak
up.
We
got
obviously
put
something
together.
A
We
want,
but
we
obviously
have
our
sessions
that
coupon
itself
so
I
don't
feel
a
little
bit
awkward
about
repeating
things,
but
you
know
it's
been
done
before
so
think
about.
If
you
guys
want
to
do
something,
we
can
always
do
that.
Alright.
So
with
that
Before
we
jump
into
PRS.
Are
there
any
other
sort
of
community
related
topics?
You
want
to
bring
up
alright
look
forward,
then.
A
So
there
was
an
outstanding
issue
about
open
by
Evan
about
how
to
handle
extensions
in
binary
format,
in
particular
when
the
extension
has
its
own
serialization,
meaning,
for
example,
in
the
ACP
case.
It
won't
be
prefixed
with
a
see,
II
and
I
believe
on
last
week's
call.
I,
don't
know
where
we
looked
at
remember
where
real
families
call
but
I'd
leave
either
through
that
call
or
offline
groovers
kind
of
agreed
on
a
Khurana
proposal
that,
basically,
let
me
see
if
I
have
this
here.
Basically
there's
a
couple
things
one.
A
We
do
allow
for
a
secondary
extension,
so
they
may
have
a
secondary
extension,
but
they
still
must
also
have
the
Cee
prefixed
ones.
If
that's
what
the
transport
costs
for
okay.
So
that
way,
we
still
have
the
consistency
of
a
single
civilization
for
extensions
and
it's
in
a
location,
but
the
data
can't
be
replicated
someplace
else.
However,
if
upon
receiving
those
that
data
in
two
different
locations,
if
that
data
differs,
then
the
extension
must
explain
what
to
do.
A
In
that
particular
case
and
barring
them
about
saying
anything,
the
assumption
is,
the
receiver
will
just
pick
up
the
the
C
version
of
stuff.
Obviously,
if
it's
an
unknown
extension,
they
won't
even
know
about
the
other
stuff,
even
worried
about
a
conflict,
but
in
the
tracing
case,
I
did
add
some
logic
there.
That
says
both
should
be
passed
on
to
the
application,
but
make
it
clear
that
they
are
two
different
properties.
A
One
is
more
of
a
transport
level
one
and
one
is
more
of
a
clown
event
level
one
and
it
represents
what
was
originally
provided
when
the
clown
event
was
first
created.
Okay,
so
the
actual
changes
that
I
made
basically
represented
what
I
just
described
most
of
the
changes
and
the
documents
are
talked
about,
how
things
can't
have
a
secondary,
a
mapping
serialization
or
whatever
you
want
to
call
it,
but
then
they
also
say
they
must
also
include
the
previously
defined
primary
mapping.
A
Oops
I'll
fix
that
typo
later
yes,
so
here
basically
talks
about
how
the
state
the
CD
version
one
is
the
one
that's
supposed
to
be
used
as
the
cloud
event
attributes.
When
you
create
this,
you
know
content
object
as
a
receiver,
but
the
secondary
one
may
be
picked
up
and
offer
it
up
to
the
receiving
application
as
additional
metadata.
Okay
I
mean
she
be
transport.
Does
the
same
thing,
let's
jump
down
to
this
back.
That
might
be
the
only
other
change
worth
mentioning.
A
So
it
makes
it
clear,
cuz
I
think
you
might
have
been
Alan.
Who
was
asking
about
this?
That
all
extensions
must
use
the
same
must
use
our
type
system
attributes.
They
cannot
just
define
some
random
type,
I
think
that's
what
we
were
implying
before,
but
we
weren't
as
clear
as
we
should
have
been
just-
and
this
just
makes
that
clear
this
here,
just
talks
about
how
they
can't
have
that
secondary
or
extensions
could
have
a
secondary
serialization,
nothing
new
there,
everyone
whatever
to
talk
about
there.
A
I
think
that's
basically
it
Oh
down
here
Kathy
added
this
text
a
long
time
ago,
when
we're
talking
about
extensions
and
I
was
going
through
there.
I
noticed
that
we
have.
The
word
should
here
and
talks
about
have,
and
then
provider
should
also
add
data
someplace
else.
The
reason
I
changed
this
from
it
should
to
a
would
then
remove
the
normative
nough
Stu.
It
is
basically
because
this
is
an
example
and
having
normative
text
ending
example
is
not
appropriate
because
we're
not
actually
mandating
people
actually
do
stuff.
A
A
So
I
made
a
change
to
to
those
documents
so
that
every
time
or
they
started
the
first
time,
I
reference,
a
secondary
civilization
I
changed
that
secretary
civilization
wording
to
be
a
pointer
to
the
extension
inspection,
I'm,
sorry
extension
section
in
the
main
spec,
so
that
people
understand
lean
by
secondary
serialization,
no
Nurit
of
changes
in
the
in
the
commits
that
just
went
in,
and
it
was
strictly
synced
tactical
just
to
give
people
a
pointer
to
reference
this.
This
notion
of
secondary,
but
thank
you
mark
for
those
comments
that
was
good.
A
Now
Before
we
jump
to
evans
PR,
which
is
other
PR,
which
is
not
required
to
one
point,
though,
what
I'd
like
to
do
is
jump
down
to
I
said
on
a
formal
item,
for
it
all
I,
guess
kind
of
do
what
I'd
like
to
do
is
see.
If
people
think
we
are
ready
to
go
to
1.0
release
candidate,
one
is
there.
Anybody
have
any
concerns,
because
that
has
always
been
our
plan
and
once
we
resolve
all
1.0
PRS.
But
if
you
have
any
concerns
about
that
or
lingering
doubts,
please
speak
now.
A
A
I
would,
according
to
our
schedule,
we
were
supposed
to,
or
we
are
supposed
to
start
a
two
week
review
period.
For
one
point,
though,
that
does
not
mean
we
can't
make
changes
even
large
changes
if
we
want
to-
and
you
know
as
we
make
those
changes
we
may
want
to
reset
the
the
timeline
if
we
so
choose.
But
what
this
is
meant
to
do
is
to
basically
put
a
nagging
pressure
on
everybody
to
review
the
spec,
because,
hopefully,
the
end
of
the
two-week
review
period.
A
A
A
C
C
G
A
E
A
A
E
A
Interesting
that
might
be
useful
yeah,
but
it
sounds
like
it.
Okay,
so
either
way,
though
I
don't
think
it's
appropriate
to
vote
on
this
one,
and
there
are
no
questions
worth
that
people
have
so
there's
nothing
to
discuss.
I
think
we
just
need
to
wait
for
Evan
to
come
back
with
another
revision
based
by
508
and
possibly
your
comment
there
Scott
so
I
think
we
can
move
forward
and
you
disagree
with
that
or
move
on
to
the
next
topic.
A
E
A
Was
just
an
issue:
he
opened
up
this
PR
as
a
result
of
that
issue,
which
is
modifying
508.
We
can't
vote
on
this
one
right
now
because
he
just
opened
it
like
an
hour
or
so
ago.
I
believe.
However,
he
wasn't
as
a
planning
on
this
being
normative
changes.
I
think
this.
These
were
just
clarifications.
That's
why
I
marked
it
as
nice
to
have
them
for
one
point
overnight
requirement.
Why?
A
No,
but
please,
when
you
get
a
chance,
take
a
look
at
this
and
put
any
comments
you
have
in
the
PR
itself,
so
I
just
wanted
to
bring
it
to
your
guys
attention,
but
since
it
is
related
to
the
extension
stuff,
we
should
probably
give
it
a
very
careful
review.
So
if
that
is
kind
of
a
tricky
area,
all
right
any
comments
on
that.
A
A
A
A
A
It's
just
a
nagging
reminder
that
we
have
these
as
things
someone
at
one
point
in
time,
but
you
really
nice
to
have
for
one
point
though,
but
I
don't
think
anybody's
taking
the
action
I
actually
make
it
a
reality,
and
that
means
an
open,
a
PR
or
suggested
to
close
it.
So
please,
when
you
get
a
chance,
somebody
take
a
look
at
one
of
those
all
right.
Now
it
gets
all
three
of
those
you
know.
For
example,
Clemens
may
be
the
webhook
one
might
be
of
interest
to
you,
since
that
was
your
spec.
A
Yes,
yep
I,
don't
know
who
what
the
oh
I
just
key.
One
I
opened
the
issue,
but
I
was
willing
to
get
someone
mentioned
it
during
a
call.
You
know
lose
track
of
it.
So
maybe,
if
you're
interested
in
the
petition
key
extension,
you
can
look
at
maybe
doing
an
example
there.
So
anyway
take
a
look
at
those
three.
If
you
want
to
jump
on
an
omen
of
those
that'd
be
really
nice
anyway,
we
can
close
these
things
out
these.
A
These
try
for
v1
issues
all
right
and
with
that
we
actually
might
be
done
with
the
agenda.
Okay.
Now,
while
I
asked
Thank
You
Ronnie,
so
any
other
topic
you
want
to
bring
up
before
I.
Do
the
final
roll
call
and
let
you
guys
all
have
a
full
35
minutes
back
of
your
day.
That
is
quite
amazing.
Amazing.
B
G
E
A
In
that
case,
one
final
note
to
forget
to
mention
I'm
going
to
have
vacation
for
the
next
two
weeks.
I'll
send
an
official
note
to
the
list
asking
for
leave
of
absence.
However,
mark
will
be
running.
The
call
on
October
3rd
I
believe
is
the
date.
So
in
two
weeks
time
next
week,
depending
on
who
is
actually
able
to
make
it
the
order
of
potential
run
people
running
it
will
either
be
Clemens,
Hines
or
ginger,
depending
on
whether
they're
meeting
or
whatever.
It
is
right
before
this
call
runs
late.
H
A
A
C
A
So
that
means
for
the
German
folks
or
anybody
else
who
may
have
a
reason
not
to
make
the
call
that
week.
If
you
have
any
issues
with
the
speck
go
into
1.0,
obviously
raise
them
offline
in
some
fashion,
preferably
in
PR
form.
So
if
you
disapprove
the
PR,
but
then
yes,
the
vote
starts
and
have
a
week
after
that
to
to
vote.
So
you
know
you
can't
make
the
call.
You
still
get
to
vote
all
right
anything
else
before
that
Clemens
gets
to
his
soccer
match.