►
From YouTube: CNCF Serverless Working Group 03-25-2021
Description
CNCF Serverless Working Group 03-25-2021
A
A
A
B
C
D
E
A
F
A
I
get
it
there,
we
go.
Okay,
thanks,
lance
howdy,
hello,
hello,.
A
Yeah
yeah,
that's
what
I
think
just
hey!
It's
interesting.
I
got
pinged
from
the
from
the
person
who
manages
our
recordings
and
I
guess
somebody
stayed
on
the
call,
and
so
she
started
looking
at
the
recording
at
starting
at
1
pm
eastern.
I'm
not
quite
sure
why
her
time
was
shifted
for
european
time,
since
I
think
she's
in
california,
but
she
couldn't
understand
why
there
was
no
one
on
the
call
she
thought
we
may
be
canceled
this
week
or
last
week,
so
that
was
interesting
anyway.
Klaus
you
there!
A
A
G
I
just
wanted
yeah
there's
a
topic
on
cubecardiu
but,
as
I
will
be
presenting
for
the
whole
group
next
week,
I'll
submit
the
I'll
send
it
to
you,
the
presentation
with
like
my
nuts.
So
if
you
can
give
me
feedback
but
like
I'm
a
bit
late
stuff,
but
I
will
get
it
ready
for
next
week.
So
you
can
all
review
because,
like
obviously
just
representing
this
whole
group,
so
I
don't
want
to
say
anything
that
you
would
feel
uncomfortable.
C
A
That's
good.
We
usually
do
tend
to
try
to
share
the
presentations
beforehand.
That
way
in
case
there's
something
obviously
wrong
that
we
need
to
fix,
but
yeah,
usually
it's
okay,
but
yeah.
Thank
you
for
sharing
that
all
right
anything
else
from
a
community
perspective,
that's
not
on
the
agenda
all
right!
In
that
case,
we
have
sdk
calls
scheduled
for
this
week.
Let
me
just
double
check.
A
A
We
don't
even
have
to
talk
about
it
here,
but
consider
yourself
nagged
if
you're
planning
on
doing
something
that
we
did
plan
on
doing
something
starting
next
week
for
interop
testing.
I
know
myself
I'm
still
way
behind,
but
other
people,
please
keep
in
mind
that
that
was
our
original
goal.
So
try
to
do
what
you
can
for
next
week.
A
E
Oh
just
real,
quick,
we
released
version
0.6
of
the
specification
and,
alongside
all
the
sdks,
the
vs
code
plug-in
and
updated
the
website
as
well.
So
that
was
like
all
we
did
last
week
and
then
now
I
guess,
preparing
for
next
release
or
starting
to
think
what
what's
going
to
be
included
and
that's
all.
A
All
right,
in
that
case,
let's
move
forward.
So
it's
interesting.
If
I
remember
correctly,
I
think
jennifer
may
have
a
question
on
the
conventions,
one
and
since
she's,
not
on
the
call.
Yet
let
let's
skip
that
one
for
a
little
bit
later,
just
in
case
she
joins
I'd
like
to
see
if
she
had
any
feedback
or
something
you
know,
I
didn't
notice
anything
in
the
issue.
I
just
want
to
give
her
a
chance,
so
let's
hold
off
on
that
one
for
a
minute.
A
If
that's
okay
with
folks
and
let's
look
at
this
next
one
okay,
so
this
one
is
still
in
draft
form.
Let
me
go
ahead
and
hide
the
comments,
but
I
did
want
to
bring
it
to
people's
attention
to
make
sure
you
see
it
so
this
this
person
is
suggesting
a
new
extension
to
basically
add
another
field
called
classification
where
people
can
then
indicate,
I
guess,
the
the
type
of
data
or
whether
it's
private
versus
public
or
something
like
that
and
the
example
they
give
here
is
that
it's
labeled
as
public.
A
Obviously
you
know
take
your
time
to
read
it
and
stuff.
It
is
just
an
extension.
So
the
bar
is
fairly
low.
A
However,
what's
interesting
is
they
do
mention
specifications
not
because
I
necessarily
adhering
or
they're,
not
because
necessarily
implementing
that
spec
per
se,
but
rather
they
just
sort
of
reference
it
as
saying
this
is
an
example
of
where
you
might
find
the
data
to
stick
in
there,
and
some
of
those
specs
are
private.
That's
maybe
a
little
bit
of
a
concern
for
us,
but
anyway
you
mean
isil.
A
I
mean
iso
is
usually,
unfortunately
behind
paywalls
yeah.
Maybe
it
was
the
s1,
then
okay,
I
honestly
can't
remember,
but
anyway
I
don't
know,
take
a
look
at
it,
since
it's
still
in
draft,
we
don't
have
to
vote
around
today
or
anything,
but
I
just
want
to
draw
people's
attention
to
it.
A
Yep,
it
does
seem
kind
of
weird
doing
the
state
of
affairs
to
have
something
like
that
be
behind
the
firewall
but
c
levy
all
right
since
the
owner
gave
her
the
owner's
name
yeah
since
matt
is
not
removable.
Jackson,
that's
not
on
the
caller.
I'm
not
sure
we
can
do
a
whole
lot
here,
other
than
just
bring
it
to
your
attention.
But
jim
did
you
get
your
hand
up.
I
Yeah
I
mean
I
I
commented
on
this
because
I'm
not
sure
what
to
make
of
it.
Without
seeing
that
spec
I
mean
we,
we
certainly
deal
with
data
privacy
and
having
to
annotate
data
and
all
that
sort
of
thing,
but
I
I
I
wasn't
clear
what
anybody
would
do
with
this
information
that
was
really
when,
where
it
wasn't
sort
of
resonating
for
me,
did
he
answer
you?
H
So
but
that
raises
a
good
point,
and
that
is
if
we
need.
So
if
a
cncf
working
group
needs
an
iso
standard
to
do
its
work.
G
So,
even
if
it
talks
about
either,
probably
because
it's
from
my
uk
from
what
I've
seen
it
can
apply
to
several
other
norms
like
security
numbers
like
the
izoton
ceo
one
is
like
a
security
nom.
H
Yeah
yeah,
but
my
my
question
is
it's
a
different
one?
It
is
like
two:
can
we
get
to
the
either
or
pdf
yeah
exactly
it's
like
it's
like.
So
what
is
what
is
the?
What
is
the
process?
So
iso
is
obviously
capital
s
standards.
So
I
think
that's
it's
without
question
that
we
we
should
be
able
to
go
reference.
An
iso
standard.
H
It's
just
that
for
us
to
judge
whether
it
is
you
know
our
specifications
make
sense
in
the
context
of
those
I
suspects,
we
would
have
to
go
and
take
a
look
at
those
iso
specs.
So
what
is
this?
Does
the
cncf
help?
H
A
I
I'd
say
I
don't
know
the
answer
that
question.
I
know
they
have.
They
have
help
for
going
through
the
the
past
process.
For
some
of
these
folks,
you
know
it's
the
fast
path
process,
but
I
don't
know
about
getting
access
to
things
behind
the
firewall
that
I
just
don't
know
so
so
I
think
that's
a
tntf
question.
K
Yeah,
well,
isn't
that
why
we
use
the
nrc
3339
and
for
a
time
stamp
instead
of
an
isis
standard.
H
G
H
H
H
The
cncf
bodies
need
to
go
and
and
acquire
this
through
cncf
so
that
we
can
go
in
and
and
share
it
first
rule.
What
I
saw
is
that
you
don't
talk
about
iso.
Exactly
that's
great
mark.
A
So
I'll
I'll
I'll
pay
some
comments
into
the
into
the
pr
in
this
space.
But
to
be
honest,
I
kind
of
interpreted
this
as
sort
of
an
example
of
some
folks
that
may
want
to
actually
use
it,
but
then
it
does
relate
to
the
to
the
next
question.
I
was
going
to
ask
them
at
some
point,
which
was
don't
you
need
to
give
some
guidance,
what
values
to
put
in
there?
I
understand
he
says
it's
maybe
proprietary
and
he's
not
proprietary.
A
It
may
be
domain-specific
and
he's
just
sort
of
defining
a
field
where
this
common
information
can
go,
but
then
it's
up
to
each
domain
to
find
what
those
things
are.
But
it
still
would
be
nice
to
point
to
non-proprietary
samples
that
people
can
look
at
and
say:
hey,
that's
cool,
I'm
gonna
copy
what
they
did
instead
of
reinvent
their
own.
D
I
I-
and
I
think
this
is
what
I
was
really
questioning-
I
certainly
in
our
world.
We
can't
get
away
with
saying
this.
Complex
object
is
of
a
particular
class
or
sensitivity
yeah.
It's
actually
all
the
data
items
inside
that
object,
so
I
I
was
still
bemused
as
to
how
you
could
apply.
I
assume
this
would
apply
to
the
entire
data
payload
of
the
of
the
event
yeah
now.
A
Okay,
so
if
you
guys
can-
and
if
you
have
time,
go
ahead
and
add
your
own
comments
to
there
jim,
I
know
you
already
did
and
I'll
try
to
poke
them
behind
the
scenes
to
see.
If
you
can
answer
the
questions,
but
you
know
remy
clemens
anybody
else.
If
you
have
additional
questions,
please
add
them
to
the
pr
and
I'll
try
to
add
some
as
well
based
upon
some
of
the
feedback
on
the
call
here
but,
like
I
said
it's
just
in
draft
form,
so
we
still
have
time
to
hash
through
it.
A
L
D
A
D
A
That
that's
fine
all
right,
let's
see
slinky
is
not
on
the
call.
Oh
yes,
slinky
is
on
he's
joining
right
now
hold
on.
A
A
Oh
there
we
go
slinky
you
there
now
slinky
yeah,
all
right,
just
you
joined
just
in
time.
So
we're
just
talking
about
this
pr
in
here.
I
was
wondering
if
you,
because
I
don't
think
the
author's
on
the
call.
Let
me
just
double
check
drew
I
don't
see
drew
so
did
you
want
to
talk
to
this
one?
Is
it?
Do
you
think
it's
aside
from
this
type
right
here?
Do
you
think
it's
ready
to
go?
Do
you
want
more
time
to
review
it
since
you
kind
of
own
the
sql
spec
as
it
right
now.
M
Well,
can
you
can
you
pick
me
there
so
I'll
give
a
look:
yeah,
okay,
yeah,
because
it's
some
something
seems
just.
A
A
Go
look
yeah
yeah,
like
I
said
I
think
when
I
did
my.
N
A
D
Yeah
again
cc
me
in
the
vr:
oh
okay,
cc,
slinky
done.
A
M
Yeah
so
for
who
doesn't
know
antoiler,
so
I
basically
went
ahead
and
started
sketching
the
grammar
for
antero,
which
is
a
pretty
famous
tool
for
like
one
of
the
most
famous
tools
for
a
password
generation.
M
In
the
process
of
doing
this,
I
found
some
minor
issues
with
the
spec
that
I
fixed
I
before
going
emergencies.
I
I
want
to
have
one
of
my
colleagues
to
look
at
it
because
he
is
more
expert
than
me.
M
Okay,
yeah
there
were
some
things.
Some
say
some
leftovers
in
the
spec
that
I
forgot
like
this:
the
yeah,
the
expression.
Maybe
an
f,
was
missing
the
the
like
operation
axis
operation
and
in
operation.
Then
some
I
fixed
some
naming
consistencies
in
the
eb
and
f
rules,
but
yeah
stuff,
like
that.
I
didn't
add
that
nor
removed
anything
new
in
this
pack.
A
Anybody
have
any
experience
with
this.
M
No,
no,
I
want
to
wait
for
another
week
because
I
want
one
two
lionel
highs
on
it:
okay,.
D
A
Let's
go
back
to,
I
don't
see
jennifer,
yet
yeah.
She
usually
would
join
by
now,
if
she's
going
to
join.
So
let's
go
ahead
and
do
this
anyway.
Did
anybody
have
any
comments
on
this
one?
A
This
is
the
one
where
grant
was
trying
to
get
us
to
use
the
commit.
The
conventional
commits
guidelines
for
tagging
issues
in
pr's.
I
guess.
A
A
A
Now
I
guess
tom
suggested
what
about
making
it
a
structure,
so
we
can
have
not
just
the
time
but
an
alternative,
and
then
we
got
a
whole
bunch
of
questions
about
the
alternative.
So
after
thinking
about
it,
my
proposal
was,
if
we're
going
to
keep
the
alternative
in
there.
A
I
think
I
did
include
text
in
there
already
that
talks
about
how
the
subscriptions
or
what
happens
to
subscriptions
after
that
the
the
expiration
date
is
out
of
scope
for
us.
I
got
a
little
more
specific
here
saying:
subscriptions,
do
not
carry
over
to
the
new
service,
because
there's
no
guarantee
that
a
new
service
actually
provides
the
same
semantics.
A
In
fact,
the
alternative
it
doesn't.
The
alternative
could
be
a
completely
different
service.
It's
just
what
this
person
who
owns
this
service,
it's
kind
of
recommending
that
they
consider
it's
not
meant
to
be
a
one-to-one
replacement.
At
least
this
specification
knows
places
no
guarantee
that
it's
a
one-for-one
replacement.
A
Therefore,
I
propose
that
we
don't
do
any
transitioning
or
transferring
of
subscriptions
at
all
and
also
must
be
a
url
to
the
alternative
service.
It
can
be
in
another
discovery
endpoint,
which
means
there's
no
guarantee
that
the
client
can't
access
it.
H
I
D
A
A
Life
would
be
so
much
easier
if
everybody
had
that
attitude.
You
know.
Okay,
anybody
else
want
to
chime
in
that
sounds
good
to
me.
Okay,
thank
you.
Remy
lance,
you
came
up
with.
L
K
Yeah
I
agree.
I
I
think
that
having
an
alternative
and
then
your
recommendations
around
it
are
good.
I
you
know
I
if,
if
it
is
expanded
like
that,
then
I
might
suggest
changing
the
term
removal
to
deprecated,
because
you're
specifying
a
time.
B
I
Gem,
I
think,
did
we
go
over
this.
I
think
that
did
we
want
to,
I
believe.
Originally
we
had
deprecation,
but
then
there
was
a
confusion
about
what
about
removal.
G
I
G
A
This
smells
like
wordsmithing
by
committee,
but
it's
all
good.
So,
let's
start
with
one
at
a
time.
First
of
all,
the
top
level
element.
I
have
no
problem
with
changing
it
from
remove
removal
to
something
else.
My
question
is:
is
some
variation
of
the
word
deprecate
problematic
time?
I
wasn't
sure
whether
you,
whether
you
think
that's
too
confusing,
because
it's
not
like
you're
saying
the
word
deprecate-
doesn't
mean
or
means
it's
going
to
be
deprecated
at
this
date,
as
opposed
to
it,
is
currently
deprecated
right.
A
I
I
think
that's
the
marker
yeah
and
then,
if
it's
typical
from
that
you
infer
or
we
imply,
you
know
that
it's
gonna
be
that
it
is
deprecated
and
then
the
the
date
after
that
is
informative,
because
you're
really
saying
on
a
on
this
particular
day,
this
thing
is
gonna
disappear
altogether
and
if
there's
a
replacement
for
it,
here's
the
reference
to
it
and
you've
got
until
that
time
to.
Actually
you
know
make
any
changes
that
you
need
to.
I
So
maybe
it
should
say
some.
You
know
you
could
change
the
language
to
say
it's
supported
until
or
something
like
that.
If
the
word
time
is,
is
not
clear
enough,
I
mean
for
me.
If
it's
in
the
context
of
removal,
then
I
sort
of
the
way
I
tend
to
read
stuff
is
that
that
would
mean
it's
the
removal
time,
but
that's
just
the
way.
My
my
brain
works.
I
guess
okay.
A
I
Jim
sorry,
yes.
N
Yeah
just
wanted
to
add
a
so
add
an
experience
here,
so
we
actually
had
this
at
some
other
discovery
both
and
we
had
the
deprecation
date,
which
is
the
time
when
the
deprecation
will
become
active
and
that
might
be
in
the
future.
And
then
we
have
a
removal
date
which
is
from
this
date
on
this
actually
has
been
removed.
N
So
that's
actually
a
window
of
time
you
have,
which
is
the
deprecation
phase
right.
A
I
I
think
two
is
more
informative.
You
know,
because
you're
really
saying
you're
just
advertising
the
date
on
which
the
deprecation
started,
but
I
mean
it's
really
the
removal
time
that
becomes
important.
You
know
for
anybody
consuming.
So
I
think
I
don't
you
know
I
wouldn't
push
back
if
you
wanted
to
add
that,
but
I
don't
think
it's
it.
It's
it's
candy
yeah!
It's
not
really
required
from
a
sort
of
processing
perspective.
A
In
my
opinion,
I
agree
so
I
think
what
I'm
hearing
is
possibly
look
at
adding.
A
I
No,
I
I
don't
think
so,
because
you
may
be
turning
this
capability
off
altogether
yeah,
so
I
there
may
well
not
be
an
alternative
to
it.
A
Right,
okay,
I'm
just
trying
to
get
all
the
comments
that
were
in
the
slack,
I'm
sorry,
it's
like
webex
chat,
make
sure
I'm
getting
them
all.
A
A
A
P
Maybe
I
missed
an
instant
story,
hello,
everyone
and
it's
the
first
time
for
me,
remy
invited
me
and
what
I
understand
it's
when
something's
deprecated,
we,
we
added
a
new
field,
called
the
art
to
transform
what
we
what
we
are
using
before
with
this
one.
This
alternative
is
that
correct.
A
I
wouldn't
say
it's
transforming
it
as
much
as
the
owner
of
this
service
is
saying
it's
going
away
and
here's
an
alternative
you
can
consider.
It
is
not
necessarily
a
direct
replacement,
it
could
be,
but
it
also
may
not
be
it's
just
something
that
the
user
may
want
to
consider
moving
to
if
they
want
to
try
something
in
its
place.
P
P
Yes,
maybe
so
it's
just
a
singing
as
I
I'm
not
an
expert
in
a
cloud
event,
I'm
just
new,
but
what
I
hear
what
I
see
here
is
when
we
remove
something,
maybe
it's
to
to
split
it
in
several
other
things,
and
maybe
it
could
be
useful
to
to
have
the
link
to
those
new
alternatives.
A
I
A
G
A
Yeah
I
was,
I
was
actually
wondering
something
similar,
because
obviously
a
singleton
is
conceptually
easier,
but
if
people
have
real
use
cases
where,
as
you
just
said,
remy
one
service
gets
split
into
many
microservices,
I
can
definitely
see
the
need
for
more
than
one.
A
A
A
N
A
Okay,
clemens,
I
believe
you
missed
the
call
last
week,
so
I
just
want
to
draw
your
attention
to
the
fact
that
we
did
talk
about
it
last
week
and
people
seem
to
be
leaning
more
towards
your
option
b.
So
do
you
want
to
create
a
pr
around
that
option.
H
A
Yes,
all
right,
okay!
Well,
thank
you
for
agreeing
to
do
that.
One
okay
of
these
four
down
here.
Does
anybody
have
any
updates
they
would
like
to
bring
to
our
attention?
I
don't
think
anybody's
made
any
changes
yet,
but
since
we
have
extra
time,
I'm
asking
okay
not
hearing
any,
does
anybody
have
any
other
business
they'd
like
to
bring
up
on
the
call,
because
we're
at
the
end
of
the
agenda.
A
Not
hearing
it,
okay,
we'll
do
scott's
favorite
part,
then
last
call
for
roll
call,
so
daniel
you're
there.
Yes,
I
am
excellent
and
how
hello
hello,
granty.
A
O
O
A
So
there
you
go
all
right,
any
other
topics
for
the
sdk
call
or
any
topics.
Sdk
call
all
right
cancelled
and
that's
it
then
we're
done.
Thank
you,
everybody
and
don't
forget.
If
you're
doing
the
interrupt
you
got
till
next
week,
then
we're
going
to
start
testing.
Hopefully
all
right
have
a
good
day.
Everybody
all
right.