►
From YouTube: Gateway API GAMMA Bi-Weekly Meeting for 20221115
Description
Gateway API GAMMA Bi-Weekly Meeting for 20221115
A
Hello:
everyone
welcome
to
the
November
15th
instance
of
the
Gateway
API
gamma
meeting.
As
a
reminder,
this
meeting
is
governed
by
the
kubernetes
code
of
conduct
and
you
know
that
basically
boils
down
to
be
nice
to
one
another,
so
nobody
gets
mad
at
you.
A
The
this
meeting
has
an
open,
Agenda
and
I
have
pasted
the
very
scary,
looking
Google
Docs
link
to
the
to
the
chat
here
and
please
feel
free
to
add
an
agenda
item
and
add
your
name
as
well,
so
that
we
can
have
a
record
of
your
attendance
as
I
just
messed
up.
Rob
writing
his
name
on
there.
That
was
funny
anyway.
A
Looking
at
our
agenda,
we've
got
a
couple
of
items
to
talk
through,
so
the
first
one
is
to
recap
the
last
meeting
for
those
who
might
not
be
who
who
for
those
who
might
not
have
been
able
to
attend
English
is
hard.
Last
meeting
we
chatted
a
bit
about
kubecon
had
some
general
feedback
from
get
14
26.
I'm
gonna
do
some
copy
paste
here.
A
B
I
haven't
heard
much
other
than
that.
A
lot
of
people
wanted
to
take
a
look
at
it.
So
I
don't
know
if
that
means
they
read
it
and
agreed
with
it.
I
didn't
feel
the
need
to
comment,
or
they
haven't
got
to
it.
I
expected
it
to
be
more
controversial
than
it's
been
so
far,
so
I
kind
of
suspect
people
haven't
looked
at
it
but
feel
free
to
correct
me
from
wrong
I.
C
Owe
you
and
apologize?
Oh
you,
an
apology.
English
is
hard
Keith,
because
I
definitely
did
not
get
a
chance
to
look
at
it
yet,
but
I
will
commit
to
doing
that
this
week.
A
All
right,
so,
yes,
please
take
a
look
at
that
at
that
PR.
We
had
a
pretty
involved
discussion
around
multi-cluster
mesh
last
week
and
there's
some
pretty
good
notes.
Here
from
that
I
think
the
if
I
had
to
summarize.
Essentially
we
were
trying
to
decide.
Okay,
do
we
want
operations
in
gamma?
A
What
for
things
like
multi-cluster
mesh,
the
semantics
of
adding
a
cluster
to
a
nest
resource?
If
that's
desirable,
Kuma
has
some.
You
know,
Shane
was
saying,
and
others
were
saying
that
Kuma
has
some
things
that
they're
working
on
where
that
construct
might
be
useful
to
have
in
gamma,
and
there
was
a
several
different
viewpoints
about
whether
or
not
into
what
degree
operation
should
be
in
gamma.
A
I.
Think
that
we
left
that
conversation
at
a
let's
do
some
more
research
into
this
I
realize
I'm,
not
sharing
my
screen.
Apologies
for
that
happens
when
I'm
putting
my
screen
the
entire
time.
Can
everybody
see
this
now?
Finally,
yeah
all.
D
Right,
maybe
maybe
doing
just
the
window,
could
you
make
the
like?
Could
you
un-maximize
the
window
that
way
if
it's
not
taken
up
your
hold
I?
Think
generally.
A
All
right,
my
desktop
situation
changes
like
each
week
depending
on
where
I
am
so
apologies
for
the
for
the
confusion,
all
right
yeah.
So
everyone
should
Heckle
me
if
I
don't
manage
this
okay,
so
Flynn
is
committed
in
public
public
heckling.
A
If
he
does
not
get
a,
he
does
not
review
John's
PR
here
it
has
been,
it
has
been
stated,
but
yeah
we
had
a
great
conversation
about
multi-cluster
mesh
I
think
this
is
probably
something
that
we
might
be
I
don't
know
about,
might
be
delaying
a
bit
or
if
there's
more
strong
opinions
multi-cluster,
then
we
can
chat
about
it.
More
Shane
I
will
surface
this
discussion
that
Shane
posted
and
try
to
pull
it
up
here.
A
multi-cluster
mesh.
A
Yeah
except
I'll
spell
it
correctly,
but
yes,
if
you've
got
opinions,
we
should
probably
maybe
move
the
conversation
there.
A
Yeah
I
still
got
some
more
to
think
about
in
that
conversation,
so
they
either
put
in
the
agenda
or
the
or
the
discussion.
If
you've
got
things
you
want
to
to
say
about
that.
A
Okay,
well,
in
that
case,
Rob
you've
got
the
next
agenda
item
yeah.
E
So
I'm
really
just
trying
to
figure
out
what,
if
anything,
from
gamma
we
should
be,
including
in
the
vo.60
API
review,
so
I'll
give
a
little
bit
of
context
here.
The
way
we
do
releases
in
Gateway
API
is
based
on
our
Milestones,
so
we
we
are
nearly
closed
out
on
the
vo60
Milestone.
It's
taken
some
time,
but
we're
about
ready
to
go,
but
before
we
release
anything,
we
need
approval
from
a
higher
level.
So
in
that
case
it's
Sig
Network
API
reviewers.
E
We've
scheduled
that
review
to
start
on
Friday
we're
hoping
to
and
again
take
this
with
a
grain
of
salt.
But
I
I
am
optimistic
that
we
might
even
get
a
release
candidate
out
next
week,
we'll
see
we'll
see.
But
with
that
said,
I
don't
really
know.
This
is
our
first
release
that
gamma
has
existed
and
has
been
doing
some
work
and
I'm
not
really
sure
what
part
of
that
we
want
to
include
in
our
release.
If
anything,
we
can
just
say
you
know,
there's
this.
E
This
release
doesn't
include
anything
related
to
gamma,
that's
coming
still
in
progress.
We
don't
have
anything
to
release
yet
or
we
could
release
some
kind
of
you
know
experimental
something
and
say
hey
this.
This
initial
concept
is
something
that
we
consider
experimental
and
want
to
get
reviewed
in
this
session,
or
we
can.
We
can
go
for
a
third
option
of
we
don't
think
we
have
anything
to
release,
but
while
API
reviewers
are
looking
at
things,
why
don't
they
look
at
this
too?
E
E
You
know
we
have
been
saying
this
for
a
while
and
I
hope
it
will
be
true.
One
day
we
have
been
trying
to
push
for
faster
release.
Cadence,
and
you
know
what
one
of
the
things
we've
been
really
pushing
for
in
the
next
release
is
to
have
a
very
small
release.
So,
for
example,
things
like
redirects
rewrites,
maybe
TLS
route,
that
our
existing
features
just
graduating
them
on
to
another
level,
so
from
experimental
to
standards.
So
that
would
be
a
pretty
tiny
release.
C
E
D
My
my
feedback
there
is
I
would
I
would
really
like
to
see
at
least
the
IPR
reviewers
have
a
look
over
the
HTTP
route
binding
to
service
part
of
that
thing.
Even
if
they
don't
go
deep
into
the
details,
I
would
like
to
get
some
early
feedback
from
them
on.
If
they
think
that's
a
good
idea.
I'll
be
honest.
I
am
worried
that
they
will
be
like
you
know,
that's
terrible!
We
don't
do
it.
D
So
if
they're
gonna
say
that
it
would
be
good
to
know
right
now,
so
that
we
don't
spend
a
lot
of
Cycles
doing
it.
You
know
like
so
yeah
I'd,
rather
I'd,
rather
find
out
sooner
rather
than
later.
If
it's
going
to
be
a
big.
No,
if
it's
a
qualified,
yes
keep
look,
keep
working
on
this
and
think
we're
fine,
actually.
C
If
we
should
go
so
far
as
to
say
hey,
we
could
make
that
bit
of
it
experimental,
but
I
have
not
gone
through
this
process
before
so
I
don't
know
if
it
makes
more
sense
just
to
ask
for
generalized
feedback
or
to
say
hey,
we
can
make
this
experimental
and
then
see
what
they
say.
A
But
for
me
personally,
I
think
there's
still
too
much
ambiguity
in
that
initial
Gap
I.
Don't
know
that
we're
going
to
clear
that
up
by
the
current
timeline
of
when
API
review
is
going
to
be
seeing
this
so
I'm
in
favor
of
option.
Three
that
Rob
said
and
echoing
what
Nick
kind
of
just
just
said,
I
would
love
to
get
Sig
Network
to
take
a
look
at
this
and
say
no.
This
is
awful
here's.
A
Why
and
let
us
pivot
or
say:
okay,
just
be
careful,
be
careful
about
ABC
and
keep
that
in
mind
I
also,
while
we
have
their
eyes
on
this
I
love
for
them
to
maybe,
if
we
can
maybe
ask
them
about
the
group
question,
so
that's
come
up
in
the
past.
A
You
know
we
I
think
there's
probably
a
similar
discussion
that
Network
said:
Network
reviewers
were
thinking
about
with
reference
Grant
right
now,
and
so
it
might
be
appropriate
to
to
kind
of
bundle.
Those
and
say:
hey:
we've
got
this
gamut
thing.
Do
we
keep
this
in
the
same
group
as
Russia,
Gateway
API?
A
Does
the
fact
that
now
measures
in
this
mean
that
the
group
just
get
get
some
feedback
on
that
early
and
at
the
very?
If
we
don't
get
an
answer,
that's
totally
fine.
We
can
at
least
I
think
start
the
conversation
get
some
people
thinking
about
it.
F
A
D
We
do
that
so
this
is
talking
about.
If
we
add
new
resources
as
part
of
gamma,
do
we
add
them
in.
F
D
Don't
know
yeah
yeah,
so
so
that's
like
it's
not
it's
not!
So
it's
not
so
much
like.
Are
we
gonna
add
another
HTTP
route
in
a
different
Gateway
I
think
that
would
be
a
terrible
idea
personally,
but
I
think
it's
more
in
a
different
API
Group,
it's
more!
If
we're,
if
we
do
add
extra
resources
as
part
of
gamma,
do
we
add
them
in
the
gateway
to
our
networking.kubernetes.io
API
Group,
or
do
we
make
a
mesh.networking
for
kubernetes.
F
My
question
is
really
to
have
as
many
resources
shared
between
Gateway
and
and
Gamma
as
possible
and
and
not
diverge,
but
if
there
is
a
resource
that
is
only
specific
to
mention
can
never
be
applied
to
gateways,
and
it
may
be
so
far.
I
haven't
found
any
case
where
resource
or
configuration
will
priority
to
measures
so.
E
D
It
was
just
that
we
we
it
sort
of
came
up
when
we
were
saying:
oh,
maybe
we
might
look
at
making
another
resource
and
one
of
us
said
hey
what
API
Group
should
they
feed
we're
all
black,
so
yeah
yeah,
that's
right
that!
That's
that
that's
the
that's
the
sort
of
the
the
thing
that
I
would
want
to
ask
them
as
well
is
like
hey.
C
If
I
recall
correctly,
there
was
a
big
part
of
the
discussion
around.
Oh,
maybe
we
should
add
this
was.
Maybe
we
should
try
to
avoid
adding
more
series.
D
I
mean
I
think
that
that
is
in
general,
a
good
goal,
but
yeah
well.
I
I
have
also
found
that
I
mean
the
the
back-end
properties
thing.
Slash
breaking
capabilities
thing
has
sort
of
also
indicated
that
sometimes
adding
more
smaller
cids
puts
you
in
a
much
better
design
space.
So
that's
trade-off.
Yeah.
F
E
F
D
Yeah
so
I
think
the
other.
The
other
question
is
yeah,
the
the
just
that
so
just
the
general
okay.
How
bad
an
idea
is
it
or
how
good
an
idea
is
to
attach
that
HTTP
routes
directly
as
with
the
services,
apparently,
if
that
would
that's
the
sort
of
the
question
that
I
would
want
to
be
honest
if
I
had
been
a
coupon
I
would
have.
I
would
have
asked
him
that
in
person,
but
but
I
was
not
so
I
would
like
to
find
out
what
he
has
to
say.
A
A
So
look
forward
to
to
hearing
what
they
say,
but
I
think
that
one
of
the
takeaways
here
is
that
by
the
time
070
runs
around
we
might
do
we
do
we
think
that
that's,
if
you
know
we're
going
to
be
promoting
resources
to
and
make
070
small,
does
that
make
sense
for
the
first
gamma
release.
Is
that
a
good
Target
for
us
to
get
this
HTTP
route?
Readiness
Milestone
knocked
out.
F
A
In
in
my
mind,
I
would
Envision
070
it's
like
the
the
first
release,
where
mesh
is
supported
with
semantics
I
that
that's
the
best
way
to
be
a
phrase.
It
because
there's
not
really
a
lever
for
like
we've,
got
conformance
levels
and
we've
got
like
the
channels
and
and
we've
got
Alpha
Beta
And
GA,
but
there's
not
really
a
a
lever
for
here's,
a
new
semantic
for
how
this
might
work.
So
that's
probably
how
I
would
I
would
phrase
that,
and
so
it
sounds
like
070
is
going
to
be.
A
The
should
be
our
goal.
Look
should
be
our
Target
for
getting
this
Milestone
closed
out.
Any
agreements
disagreements.
F
A
Yes
and
I'm
not
saying
touch
for
clarity,
I'm,
not
saying
to
make
it
beta
at
zero.
Seven,
oh
I'm,
saying
that
okay
implementations,
mesh
implementations,
you
just
feel
comfortable
getting
the
crds
in
070
to
be
able
to
import
and
use
and
Rob
Do.
You
have
a
we've
been
talking
about
more
frequent
releases,
but
we
have
a
kind
of
a
current
ETA
on
when
we're
planning
for
070
I.
E
Mean
my
own
personal
goal
is
q1,
but
I've
been
wildly
wrong
before,
but
certainly
the
first
half
of
next
year.
E
No
I
mean
it
every
time,
there's
something
that
that
bogs
us
down
I.
You
know
it
could
be
gamma
this
this
next
time
that
but
but
we'll
see,
I
I
do
want
to
make
sure
we
have
meaningful
releases
and
and
and
also
maintain
a
release
Cadence.
So
it's
a
it's
a
tough
balance,
but.
E
I
I
think
I
think
at
this
point.
I
I
do
really
want
to
see
something
implementable
for
gamma
and
070
I.
Think
that's
a
good
goal
and
and
I
think
what
you're,
how
you're
phrasing?
That
of
something
that
is
implementable
in
an
experimental
state
is
probably
the
best
starting
point,
and
then
you
know
similar
to
what
we
have
with
other
apis.
We
say
that
something
needs
to
be
implemented
by
you
know
several
implementations
that
are
passing
conformance
tests
before
we
graduate
something
to
to
standard
or
beta,
so
I
I.
Think
070.
E
I
think
that
that's
a
great
call
and
definitely
let
me
know
how
I
can
help.
A
All
right
sounds
any
other
questions
on
our
comments
on
this
topic
before
you
move
on.
A
All
right,
I
have
the
the
last
topic
on
there.
The
Thanksgiving
holiday
is
this
next
week
right,
that's
next
week,
I
think
so.
Yes,
Thanksgiving
is
next
week.
There
are
going
to
probably
be
probably
be
several
people
out.
I
myself
will
be
out
all
of
next
week.
What
do
we
want
to
do
with
the
meeting
for
next
Tuesday?
Do
we
want
to
keep
it
cancel
it?
What
do
we
do?
We
have
any
preferences,
yeah.
D
Any
preferences
next
week's
meeting
is
a
morning
one,
so
I
can't
make
it
so
as
one
as
probably
the
main
person
who
is
definitely
not
doing
anything,
Thanksgiving
related,
I
I,
would
say,
probably
there's
a
good
chance.
You
should
cancel
it
because
everyone
who's
going
to
be
coming
is
probably
going
to
be
Thanksgiving.
A
I'm
happy
yeah
I'm
happy
to
cancel.
It
speak
now
forever
hold
your
peace.
B
B
A
Good
I
will
add
a.
A
A
All
right
awesome!
Well,
that
is
all
of
our
agenda
items.
For
today
we
have
a
while
left
in
the
meeting,
but
as
I
mentioned
at
the
top
of
the
at
the
beginning
of
the
meeting,
I
have
to
run
here
in
about
five
minutes.
Okay,
any
last
topics
of
conversation
or
anything
like
that.
I.
D
Got
one
really
short
one
that
document
I
poked
you
all
about
last
time,
he's
still
is
still
open,
I
think
I'm,
gonna
break
I,
think
I
said
in
the
main
gateway.
D
Api
chat,
I'm,
going
to
break
that
up
into
like
the
the
main
update
to
to
sort
of
make
a
change
about
talking
about
meta
resources
versus
policy
attachment,
and
that
sort
of
thing
I
think
that
the
big
things
that
have
come
out
of
that
disturbation,
though,
are
that
I,
don't
think
it's
viable
for
policy
attachment
to
have
much
to
do
with
gamma
resources,
although
I
think
it
should
eventually,
we
need
to
solve
the
status
communication
problem
which
I'm
pretty
worried
about
so
I'm
like
that
is
all
still
on
my
list.
D
You
know
it's
all
still
on
most
off
to
do,
but
so
I
just
wanted
you
all
to
know.
I
haven't
forgotten
about
it.
It's
still
on
there.
D
I'm
just
got
to
get
to
it
and
I
want
to
be
cautious
about
how
I
do
it
I'm
really
worried
about
the
status
thing
that
for
policy
attachment
it
runs
a
very
big
risk
of
like
fan
out
API
server
load
where
one
a
small,
a
change
to
one
resource
can
end
up
with
changes
to
like
30
resources,
and
so
I
want
to
be
really
deliberate
in
how
we
design
that
status
to
to
try
and
avoid
that
so
I
want
to
kind
of
I
want
to
try
and
talk.
D
Maybe
I
may
actually
try
and
tell
sometime
I
know:
Tim
has
had
some
problems.
They've
had
some
problems
with
that
before
and
in
a
few
different
places
in
kubernetes,
so
I'd
like
to
sort
of
try
and
talk
to
some
people
about
fan
out
problems
and
stuff.
D
So
I
might
have
been
happy
on
Machinery
about
that
and
if
there's
any
examples
of
good
and
bad
ways
to
handle
that
problem
yeah,
so
just
wanted
to
know
just
wanted
to
sort
of
note
that
everyone
that
yeah
for
everyone
yeah
that's
definitely
still
happening
and
yeah
I
haven't
forgotten.
F
This
topic
did
we
ever
discussed
using
kubernetes
events
to
propagate
some
of
this
state
or
transitions
because
I
agreed
you
know,
status
will
become
pretty
messy
with
with
everything
that
is
there
and.
D
Events
might
be
viable
yeah.
That's
definitely
something
to
consider
I
guess,
because
they're
I
tend
to
not
think
about
them
too
much
because
they're
they're
so
ephemeral,
you
I
know
you
know
I've
sort
of
worked
in
many
clusters
where
I
worked
in
a
couple
of
classes
where
we
separated
out
the
events
stream
to
a
separate
LCD
and
then
didn't
notice
that
it
filled
up
because
no
one
ever
looked
at
events.
So
there
were
no
events
for
like
nine
months
or
something
like
that.
So
yeah
I
mean
like
I,
think
yeah.
C
F
But
it's
it's
it's
it's
I
mean
even
if
they're,
not
reliables,
you
know
it's
useful
to
Define
at
least
what
events
should
be
generated
and
have
kind
of
a
mixed
solution
where
you
kind
of
really
relate
some
of
the
pressure
on
status
by
by
having
some
stuff
propagated
to
events,
especially
since
that
is
very
ephemeral
by
Nigeria.
Attaching
attached,
you
know
and
then
just
keep
I
don't
know
it.
I
found
events
very
useful
in
many
ways.
D
Yeah
I
I
hadn't
really
thought
about
them
because
the
stuff
I
thought
about,
but
I
I
will
make
sure
I
add
them
to
my
mental
list
thanks
custom.
That's
a
good
point.
A
All
right,
if,
if
that's
everything,
then
we'll
we'll
go
ahead
and
call
it
I
appreciate
everybody
coming
and
being
a
part
of
this.
Just
to
reiterate
tell
your
friends:
next
meeting
is
canceled,
so
the
the
next
time
we
meet
here
will
be
in
December
at
this
same
time,
slot,
if
you
celebrate
Thanksgiving,
have
a
Happy
Thanksgiving
week,
since
we
won't
be
meeting
and
yeah
looking
forward
to
chatting
with
everybody.