►
From YouTube: Gateway API GAMMA Meeting for 20230523
Description
Gateway API GAMMA Meeting for 20230523
A
All
right,
hello,
everyone
and
welcome
to
the
May
23rd
reading
of
the
Gateway
API
gamma
initiative.
As
always,
contact
rules
are
in
effect,
so
please
treat
each
other
respectfully
and
that
I
will
share
my
screen
in
a
moment
and
we
have
a
few
agenda
items
to
discuss.
So
we
can
get
into
that.
A
A
All
right,
I'm
going
to
assume
everybody
can
see
this
and
yeah,
so
we
as
always
have
an
open
Agenda.
So
if
there's
a
topic
that
you'd
like
to
discuss,
please
feel
free
to
add
meeting
notes
and
also,
if
you
can
attach
your
name
to
the
list
of
attendees,
that
is
always
appreciated.
So
you
can
keep
track
of
who
was
able
to
attend
these
videos,
particularly
as
we
have
two
different
time
slots.
So,
let's
go
to
understand
and
be
able
to
each.
A
We
spent
a
bit
of
time
covering
it
the
past
two
meetings,
but
basically
the
kind
of
path
forward
for
defining
the
behavior
between
gateways
and
East-West
HTTP
routes
and
like
when
to
layer,
consumer
producer
rules,
things
like
that
were
intending
I
I'm,
proposing
that
we
remove
it
from
the
milestone
for
gamma-ish.
Here,
it's
ready
for
implementation
track,
I
have
been
working
with
Flynn
and
we've
done
a
heck
of
a
lot
of
research
and
iteration
and
basically
found
the.
A
There
is
a
lot
of
things
that
are
difficult
or
tricky
about
it
for
numerous
reasons,
be
those
technical
implementation,
details
or
ux
or
cloud
provider
constraints
or
migrating
from
service
Discovery
to
mesh.
So
we're
going
to
try
to
advance
this
Gap
as
basically
like
informational
we've
done
that
in
the
past,
before
with
steps
that
start
out
as
like
informational,
professional,
to
kind
of
like
share
this
context
publicly
and
hopefully
set
the
stage
for
future
iteration
on
this
work.
A
But
trying
to
at
this
point
not
be
prescriptive
about
defining
explicit
behavior
in
order
to
basically
allow
message
to
start
implementing
gamma
and
have
room
to
kind
of
figure
out
what
makes
sense
and
what
works,
and
hopefully,
feedback
from
actual
users
will
be
a
part
of
that.
So
I,
don't
think
I
have
the
ability
to
Milestone
but
I
think
I
see
Shane
in
here:
oh
NES,
ROM
too
yeah,
so
that
is
kind
of
like
disposition
for
that
and
tldr
of
most
recent
developments.
B
We
ended
up
talking
at
the
last
gamma
meeting
on
some
detail
about
this,
about
Cloud
load
balancers
and
all
that
I
ended
up
talking,
I
think
a
little
bit
with
Rob
after
that,
even
and
yeah
I've
basically
been
convinced
that
it
does
make
sense
to
recommend
routing
to
endpoints
as
kind
of
the
default
behavior
for
a
Gateway
API
aware,
Ingress
controller
when
working
with
gamma
I'm
not
gonna.
B
It
would
take
too
long
to
go
through
and
summarize
all
of
why,
but
but
there's
a
fair
amount
of
wordsmithing
on
that
one
to
go
through
and
kind
of
clean
that
up
and
try
to
render
it
something
coherent.
So
we'd
like
to
get
that
out
of
the
critical
path
for
making
everything
ready
to
go
ahead.
A
Cool,
that's
that's
good
to
hear,
and
hopefully
that
can
be
a
second
pass,
as
we've
done
with
prior
depths
of
merging
an
initial,
traditional
or
informational
and
then
coming
back
with
another
forward
for
us
to
kind
of
updates
and
ads
prescriptive
recommendations
to
it.
C
No
sorry
I'm,
just
getting
caught
up
at
everything
I
heard,
sounds
spot
on
so
yeah.
Oh.
A
I
don't
see
John
in
here.
Oh
maybe
I
do
I.
D
A
It
well
yes:
oh
okay,
cool
all
right
I
just
wanted
to
highlight
that
your
PR
for
marking,
gamma
gets
implementable,
has
merged,
so
I
think
that's
really
exciting,
and
it's
kind
of
like
a
big
step
forward
and
worth
celebrating
in
this
group,
we're
interesting
closer
to
actually
having
gamma
included
in
a
formal
release
and
I
hope
that,
with
closing
out
that
Milestone
that
we
should
be
on
track
to
be
able
to
do
that
for
the
upcoming
zero
eight
zero
release.
A
B
D
C
C
Yeah
I
think
I
think
we've
discussed
this
before,
but
I
am
I,
am
fully
supportive
and
would
love
to
have
a
whole
section
devoted
to
Gamma
or
more
or
whatever.
But
yes,
please,
that's
that's
my
the
only
thing,
I'll
say
yeah
I
would
love
to
see
that.
E
A
A
I'll
just
go:
ask
an
issue
open
tracking
that
but
yeah
so
Keith
well.
E
Let
me
see
if
I'm
assigned
us
to
to
one
of
these
things
and
no
yeah
I,
don't
I,
don't
see,
I,
don't
see
one
so
but
I.
Remember
it
being
like
an
action
item
on
the
maybe
I
created
it
but
didn't
assign
it.
That
may
have
been
what
it
was
anyway.
My
other
comment
is
kind
of
about
gamma
and.
B
E
I,
actually
don't
think
it
exists
anymore.
I
did
it
I've
been
talking,
I've
been
looking
while
I'm
talking
and
I
actually
I
see
one
about
version,
clarification
in
the
docs,
but
I
don't
see
any
on
making
gamma
more
prominent
in
the
docs.
So
if
you're
looking
to
help
with
it
would
be
great
feel
free
to
create
that
issue.
If
you.
E
Appreciate
it
but
yeah
this
is
kind
of
I'm
gonna
not
go
into
my
agenda
topic
quite
yet,
but
this
I
couldn't
hear
see
who
it
was
in
Zoom,
but
when
somebody
said
can
we
make
it
more
get
more
prominent
on
the
website,
instead
of
just
an
initiative
kind
of
got
me
to
thinking
you
know,
gamma
was
when
we
started
this
was
it
was
just
an
initiative
to
see
if
we
could
make
this
work,
I
would
think
Gateway,
API
I
think
you
know.
E
We've
got
around
three
implementations
at
this
point
and
we
see
that
increasing.
It
does
feel
more
like
this
is
no
longer
an
initiative,
but
rather
a
for
a
full
part
of
Gateway
API
I.
E
Don't
know
Rob
Shane
if
you
feel
the
same
way
here,
so
maybe
the
Claire,
maybe
the
doc
clarification
is
less
of
make
a
gamma
more
prominent
and
versus
like
make
it
first
class
within
the
Gateway
API
I'm,
not
saying
it
wasn't
before,
but
I'm
just
kind
of
thinking
about
where
gamma
fits
you're
kind
of
like
this
Sub
sub
project
within
Sig
Network,
but
I'm
interested
in
in
thoughts
about
the
right
I'm
losing
the
word
here,
but
like
I,
guess,
threat,
Machinery
or
or
orchestration
around
how
how
we
talk
about
Gamma,
moving
forward.
C
Yeah
I
agree
agree
with
what
you're
saying
I
mean
I.
Think,
for
example,
I.
You
know
one
of
the
things
we
have
on
Gateway
API
is
a
good
set
of
guides.
It
seems
to
make
sense
to
kind
of
interweave.
C
Maybe
some
gamma
related
guides
in
there
too
again
I'm
not
really
sure
how
we
structure
it,
but
I
feel
like
it.
It
deserves
prominent
placement
throughout
the
other
thing.
I've
been
meaning
to
say,
is
we've
gotten
more
than
just
clearance,
but
actually
encouragement
to
have
our
Gateway,
API
and
I
think
by
extension,
gamma
documentation
of
some
sort
on
kubernetes
I
o,
probably
as
a
reference
to
you
know
our
side
I
mean
we
probably
we
could
put
as
much
as
we
want
on
kubernetes.
C
I
o
but
I,
think
it's
it's
harder
to
maintain
that
because
none
of
us
are
approvers,
but
at
the
very
least
we
should
have
a
good
overview
of
probably
the
full
scope,
including
gamma,
and
if
we
could
have
good
entry
Pages
for
both
both
sides
of
this
you
know
like
Gateway,
API
is
Broad
I,
don't
think
we
mentioned
anything
about
Gamma
on
our
main
page.
Yet
that
should
be
fixed.
C
Many
things
about
our
main
page
should
be
fixed
and
and
then
from
there
I
think
we
really
need
some
like
good
overview
landing
pages
for
both
the
Ingress
side
and
for
the
gamma
side
and
then
some
merged
content
throughout
I
know
that's
that's
kind
of
hand
wavy,
but
all
of
that's
to
say.
Yes,
we
need
huge
improvements
there
and
I
appreciate
everyone.
Who's
already
contributed,
but
yeah
I
think
this
is
a
big
project
and
any
any
incremental
steps
to
get
us
a
bit
closer
be
very
welcome.
A
Awesome,
that's
exciting.
Yeah
I
think
that
that'll
be
great
and,
like
I,
think
maybe
like
linking
this
to
the
like
08
over
lease
and
like
as
part
of
moving
formally
into
the
experimental
Channel
like
that
feels
like
the
right
impetus
to
like
Elevate
this
on
the
website
and
yeah
make
it
formerly
part
of
Gateway
API,
so
yeah
I
think
that's
great
cool
I
have
just
two
other
topics
here
on
just
like
issues
that
or
issues
in
DRS
that
I
notice,
whereas
stuff
in
progress.
A
It's
all
worth
highlighting
to
this
group
I,
don't
know
how
much
there
is
to
discuss,
but
first
off.
If
you
want
to
talk
yeah
talk
about
some
of
the
performance
work
that
you've
been
doing.
I
saw
one
thing
open
for
adding
mesh
consumers
for
gamma.
F
Yeah,
it's
just
one
test
case.
It's
getting
into
every
consumer
risk
producer
semantics
of
gamma,
but
I
think
it
would
be
good
too
yeah,
like
you,
wrote,
have
other
implementations
test
it.
That's
passing
in
Kuma,
I,
don't
know
John
if
what
the
status
of
istio
is
but
yeah.
D
Yeah,
it
failed
in
Easter,
but
it
was
a
bug
on
our
side,
so
I
slapped
together,
fix
which
I
haven't
merged
yet,
but
I
think
the
test
is
is
right.
After
Maya,
after
some
of
the
fixes
you
made
so
I'm
I'm
good
with
merging
this.
A
A
All
right,
yeah,
that's
exciting
and
looking
forward
to
seeing
this
merge
shortly.
It
sounds
like
and
then
John
you
opened
an
issue
for
dropping
the
provision
to
Omit
back-end
rep.
Do
you
want
to
expand
that
a
little
bit
more.
D
Yeah
this
started
out
with
I
think
it
was
a
conversation
with
Rob
that
it's
kind
of
weird
that
we
say
that
implementation
should
allow
omitted
backend
rough.
This
really
feels
like
something
that
should
either
be.
They
must
do
it
or
they
must
not
do
it,
and
so
I
wanted
to
clean
that
up
and
there's
kind
of
two
options.
One
is
not
do
it
or
do
it
I
think
personally,
it
makes
sense
to
not
do
it.
So
that's
what
I
submitted
in
this
PR.
D
The
reason
is
which
I
listed
here
is
that
it's
inconsistent
with
gateways.
You
can't
you
do
this
with
gateways,
so
it
may
be
a
bit
odd
that
you
know
you
want
to
have
the
same
type
of
route
for
both
and
suddenly
the
back
end
rough,
which
is
seemingly
unrelated
to
the
type,
is
different.
D
It
also
limits
the
ability
to
have
like
right
now
we
have
Services
running,
but
we
talked
about
other
things
in
the
future.
It
kind
of
limits
those
because
it
expects
a.
D
F
Yeah
I
just
had
the
comment
that,
like
this,
makes
the
use
case
of
a
service
parent
ref
without
a
port,
pretty
negligible
I,
don't
know
if
it's
possible
to
just
say
like
apparent
ref.
That's
a
service
requires
support.
Is
that
something
we
can?
We
can
do.
E
D
A
That
feels
like
a
thing.
We
can
probably
take
a
look
at
and
figure
out
a
ux
that
makes
sense
for
that,
but
in
general
this
seems
reasonable
to
kind
of
remove
ambiguity
of
like
those
two
ways
to
do
this
and
yeah.
It
seems
to
make
sense
from
like
the
future.
Looking
thing
too
of
potentially
be
an
old
Target
apparent
refs
that
don't
have
back-end
endpoints
as
part
of
them
at
some
point.
So.
B
I
think
I
think
that
I
at
least
am
gonna,
have
to
think
about
this
one
and
kick
it
around
with
some
of
the
folks
over
in
buoyant.
B
Yeah,
it.
A
Doesn't
make
it
slightly
more
verbose,
verbose
for
kind
of
like
the
simple
case.
B
D
D
B
A
B
E
Noodling
this
for
for
a
bit
since
the
the
pr
came
out
and
I
think
this
is
a
thing
that
I
that
I
was
a
big
fan
of
when
we
were
initially
writing.
The
the
first
draft
of
the
of
the
gamma
spec
but
I
think
I.
Think
I
I
agree
that
it's
one
of
those
tricky
situations
where
yes,
like
like
Flynn,
said
eloquently
like
it
feels
like
it
really
helps
a
user
out
and
has
a
nice
ux
one
for
the
new
adopter,
but
over
time
it
actually
makes
life
more
difficult.
B
Not
fair
I
think
that
was
Mike
who
said
that
well
well,
while
we
were
talking,
oh.
E
Yeah
I
couldn't
quite
remember
who
it
was,
but
yeah
I
think
that
I
have
found
myself
like,
on
the
other
end
of
this
being
bitten
by
stuff
like
this
in
the
past,
where,
like
oh,
it's
super
easy
to
adopt
and
then
to
actually
mobilize
it
to
do
useful
things.
It's
it's
now,
this
extra
work
or
extrovertent
or
up
to
skates,
what's
actually
happening
and
so
like
having
a
be
explicit
in
your
back
end
and
sitting
100
of
traffic.
E
A
Cool
well
good
initial
discussion
and
yeah.
This
is
only
onto
this
for
anyone
who
maybe
wish
to
opine
or
dig
into
this
in
a
little
more
detail.
A
All
right,
so
that
brings
us
to
the
end
of
our
schedule
agenda
for
today.
E
C
Just
real
quickly,
I'll
mention
I,
think
I've
mentioned
this
in
slack
already,
but
since
we're
at
a
a
similar
time
zone
around
a
week
from
now
last
five
minutes
is
when
we're
having
another
Gateway
API
meeting
we're
trying
another
meeting
at
this
time
slot.
C
So
if
you
can
make
it
to
this
meeting,
you
can
hopefully
make
it
to
the
next
meeting,
we'd
love
to
get
more
discussion
and
definitely
more
people
from
time
zones
that
are
harder
to
work
with
our
other
meeting
time
so
yeah
next
week
at
you
know,
30
minutes
later
than
this
meeting
started.
E
I'll
also
add
that
next
week
is
also
the
near
this
time
slot
at
the
multi-cluster
meeting,
I
believe
and
we're
looking
to
I.
Think
next
week,
specifically
there's
you
know,
I've
got
an
item
on
the
agenda
about
the
kind
of
their
their
process
for
enhancements
and
hopefully
gonna
you
know
bring
some
feedback
about
potentially
bringing
that
lore.
E
So
it's
not
like
this
long
cap
period
for
getting
things
into
the
and
the
spec
and
it's
another.
You
know
good
place
if
you've
got
that
feedback
about
MCS
API
or
the
multi-cluster
use
cases
as
it
relates
to
mesh.
That's
a
good,
a
really
good
venue
to
let
those
be
heard
so
for
free
to
come
to
that
as
as
well.
E
And
then
I've
got
to
follow
one
more
topic
that
from
the
Gateway
API
meeting
yesterday,
you
know
with
a
lot
of
stuff
kind
of
wrapping
up
for
080
I.
Think
actually,
the
only
action
item
like
that
we
have
before
080
for
for
gamma
is
the
website
changes
now
that
removing
the
mesh
Gateway
interactions
out
of
the
the
Milestone
laptop's
only
thing
left
so
be
thinking
about
what
you.
What
else
you
think
is
important
for
you
from
from
gamma
and
be
paying
attention
to
the
policy
attachment
discussion.
E
That's
been
happening
in
kind
of
Mainline
Ingress
capable
API
world
Rob's
got
a
really
great
discussion
on
GitHub
about
different
options
for
policy
attachment.
We
had
some
conversation
about
it
yesterday.
I
know
a
lot
of
people
here.
Care
about
policy,
so
it'll
be
great
to
get
your
feedback
in
those
conversations
so
that
we're
able
to
take
that
into
consideration.
E
As
that's
iterated
on
Broadway,
correct
me
if
I'm
wrong,
but
I'm
I'm,
pretty
sure
policy
is
one
of
those
things
that
we
are
wanting
to
figure
out
before
it
gets
to
stable
I
think
that's
kind
of
the
the
gist
of
it.
C
Yeah
we
want
to
have
a
an
approach
that
we're
confident
about.
It
is
my
perspective.
Before
we
go
ga
there's
a
discussion,
that's
I
think
it's
2012.
I,
don't
know.
C
I
I
can't
say
that
policy
attachment
will
graduate
to
GA
at
the
same
time
as
you
know,
HTTP
route,
because
I
think
there
are
different
levels
of
stability,
but
I
do
want
to
make
sure
that
we're
on
a
trajectory
that
we're
happy
with-
and
we
feel
has
a
path
towards
ga
I
have
my
own
biased
opinion
here,
but
I
think
there's,
there's
still
different
opinions
throughout
the
community,
so
I
want
to
make
sure
we
respect
all
of
those
can
get
to
a
point
where
we're
all
happy
with
the
trajectory
that
we're
on
whatever
that
is.
B
C
A
We
definitely
would
encourage
folks
to
check
out
some
of
the
all
the
attachment
stuff,
particularly
because,
after
we
kind
of
like
get
the
like
basic
functionality
of
routing
East-West
traffic
in
place,
one
of
the
use
cases
of
mesh
is
heavily
leveraging
policy
to
do
different
things.
So
it's
definitely
something
that
I
would
expect,
would
be
of
interest
to
folks
in
this
group
and
would
encourage
you'll
check
that
out
to
start
thinking
about.
A
If
some
of
the
proposed
approaches
are
going
to
be
workable
and
are
going
to
kind
of
make
sense
for
how
each
of
you
all
have
been
thinking
about
a
client
policy
in
your
various
implications
and
if
there's
gaps
or
things
that
kind
of
like
don't
make
sense
now,
would
probably
be
the
right
time
to
erase
some
of
that.
A
Well,
I!
Guess
with
that:
if
there
are
no
other
topics,
we'll
wrap
up
a
little
bit
early.
So
thank
you
all.
So
much
for
attendee
and
yeah
I'm
excited
to
see
Gamble
moving
ahead
to
an
actual
release
as
experimental
feature,
Gateway
API.
So
congratulations
and
thanks
everyone
teams
up
this
build
and
help
make
this
happen.
Cool
all
right,
take
care
y'all
and
have
a.