►
From YouTube: Gateway API GAMMA Meeting for 20230207
Description
Gateway API GAMMA Meeting for 20230207
A
Hello,
everybody
Welcome
to
the
February
7th
occurrence
of
the
Gateway
API
gamma
meeting
as
a
reminder.
As
always,
this
meeting
is
covered
by
the
kubernetes
code
of
conduct,
so
please
be
respectful.
Everybody
as
we
as
we
discuss
here,
reminder
that
this
is
this
meeting
has
an
open
Agenda.
So
that
means,
regardless
of
what
you
want
to
talk
about
now,
you
can
put
it
here
on
our
agenda
and
we
will
and
we'll
talk
about
it.
A
We
also
ask
that
you
go
ahead
and
add
your
name
to
the
attendees
list
here
in
that
agenda.
Doc,
I
put
the
link
in
the
zoom
chat
here.
This
helps
us
to
know
just
who
showed
up
and
be
able
to
track
the
you
know
how
much
people
care
about
this
kind
of
stuff,
so
thankful
for
everybody
who
is
here
I'll
go
ahead
and
get
started
with
a
recap.
A
What
we
covered
last
meeting
during
our
8
A.M
Pacific
meeting
last
week,
we
talked
about
the
remaining
areas
of
focus
before
we
get
to
the
Gateway
API
070
release.
Api
review
did
some
Milestone
pruning
I.
Don't
spend
too
much
time
here
because
we're
going
to
do
some
of
the
same
things
today,
but
get
some
good
administrative
work
here
to
move
us
forward
in
that
milestone.
A
We
got
a
couple
of
big
PR's
merged
that
expanded
the
scope
of
our
original
get
a
bit
so
adding
support
for
Consumer
routes
and
other
x
route
types
like
ECP
route
and
TLS
route
so
got
those
in
really
excited
about
that
talked
about
conformance
test
difficulties
that
exists.
You
know
within
Gateway,
API
implementations.
A
The
big
one
here
was
assumptions
that
your
your
load,
balancer
or
your
gateway
is
already
created
instead
of
or
I,
can't
remember,
which
way
which
direction
it
went
but
yeah,
basically,
one
percent
as
a
pattern
for
a
for
for
those
implementations
that
expect
a
user
to
manually
provision
the
resources-
and
there
is
some
desire
to
kind
of
create
a
pattern
for
this
within
the
components
tests
and
within
the
API
in
general
and
I.
A
Think
I
saw
Shane,
create
a
issue
or
PR
on
some
of
this
and
there's
some
new
discussion
in
that
which
reform
that
was
so
good
to
see
that
that
made
its
way
up
into
the
general
meeting.
A
There's
a
good
bit
of
discussion
about
proxy
configuration
in
in
mesh
and
looking
for
a
place
to
put
some
of
these
things
like
H2
settings,
window
sizes,
Etc
istio
in
particular,
had
interest
in
doing
this.
It
didn't
really
seem
to
be
a
good
place
for
it,
and
this
was
kind
of
moving
like
the
last
point.
A
This
is
that's
what
led
into
our
discussion
on
policy
attachment
and
how
several
folks
felt
like
it
was
a
bit
burdensome
for
some
of
the
more
common
needs
in
mesh
world
and
we
bubbled
that
conversation
up
to
the
Gateway
API
meeting.
Now
the
main
meeting
yesterday
actually
and
I've
got
an
action
item.
A
Actually
I
got
a
tab
with
the
discussion.
The
new
discussion
all
like
filled
out.
We
need
to
like
finish
writing
it
in
in
Hixson,
in
order
to
do
that
and
just
spur
a
conversation
on
policy
attachment
and
take
another
look
at
it,
and
Flynn
actually
also
has
an
action
item
for
timeout,
specifically
because
there
is
a
good
bit
of
interest
in
in
doing
some
standardization
around
timeouts.
A
So
that
was
the
the
last
meetings.
Anybody
have
any
questions
or
comments
about
our
recap.
A
A
Just
one
take
another
look
at
our
Milestone
here,
so
this
I'd
like
to
end
this
progress
about
two
weeks
ago,
we
had
I
think
like
six
or
seven
different
issues
in
this
Milestone
we're
down
to
three,
which
is
very
exciting.
A
The
top
one
here
is
defining
how
gateways
shooter
should
not
interact
with
gamma
router
configuration.
There
is
a
DOT
here
that
was
created
several
months
ago.
It's
gone
a
little
stale,
but
Glenn
was
assigned
to
it
last
week.
He
volunteered
to
help
Mike
out
in
tackling
this,
making
sure
that
the
story
is
clear
for
the
070
release,
so
feeling
good
that
there
is
some
some
ownership
on
that
particular
issue.
A
My
accountability.
You've
had
a
chance
to
talk
with
Flynn
about
this
at
all
in
the
past
week.
No,
we
have
not
all
right.
That's
that's
fine,
so
yeah
we'll
look
we'll
check
again
next
week
for
an
update
from
Flynn
and
Mike
on
this
one.
A
This
next
item
is
performance
levels
defining
how
they
should
be
applied
to
gamma
and
I
actually
had
a
question
on
this
this
afternoon.
The
last
kind
of
consensus
that
we
gained
from
the
main
gateway
API
meeting
is
that
conformance
performance
tests
aren't
required
well
at
in
the
old
framework.
Anyway,
the
the
status
quo
performance
tests
aren't
required
until
a
promotion
to
Beta.
A
So
we
have
some
time
to
decide
exactly
how
we
want
that
to
be
represented
with
for
gamma
and
the
as
far
as
performance
levels
and
support.
There
have
been
a
lot.
There's
been
a
lot
of
talk
about
conformance
profiles
actually
and
that's
kind
of
where
we
feel
that
gamma
conformance
test.
What
makes
the
most
sense
the
I
I'm
pretty
sure
this
is
just.
This
is
still
kind
of
the
state
of
things
I
I.
My
question
here
this
afternoon
was:
does
this
new?
A
That's
this
new
issue
that
was
created
cover?
What
we
need
to
you
know
what
needs
to
be
done
for
the
070
release.
At
the
very
least,
we
need
to
have
a
plan,
that's
written
and
documented
as
part
of
our
draft
spec,
that
talks
about
how
we
plan
to
do
conformance
tests.
A
B
A
If
this
new
issue
1686
is
sufficient,
my
POV
it
is
1686,
is
probably
sufficient
here,
based
on
our
previous
conversations
but
I'm
hoping
to
hear
others
opinions
here.
C
So
my
interpretation
of
this
is
that
1686
is
a
super
set
of
what
would
be
required
for
1488.
C
the
Define,
how
conformance
levels
release
channels
should
be
applied
to
gamma
my
expectation
for
the
deliverable,
and
that
is
basically
like
a
one
paragraph
added
PR
to
the
existing
gap
of
like
we
should
use
conformance
profiles.
We
will
use
the
same
conformance
levels,
experimental
or
sorry,
core,
extended
implementation,
specific
as
main
gateway
API,
and
we
will
use
the
experimental
Channel
with
no
immediate
plans
to
move
anything
into
standard.
Basically,
a
summary
of
like
what
I
just
said,
PR
into
the
gap.
C
So
I
think
we
have
pretty
close
to
consensus
on
that
being
what
we
feel
yeah
it's
the
path
we
want
to
take.
A
D
Go
ahead,
Rob
yeah!
Just
so
this!
This
testing
plan
for
mesh
implementations,
I
know
when
we
did
our
first
conformance
test
plan
for
Gateway
API
as
a
whole.
It
was
a
gap.
Is
that
kind
of
what
we're
talking
about
here
as
well
or
would
it
be
some
other
format,
some
other
level
of
detail?
Any
anyone
thought
about
what
this
looks
like.
C
When
we
first
discussed,
this
I
had
recently
seen
the
Sig
multi-cluster
testing
plan
that
they're
working
on
for
MCS
API
for
their
conformance
tests.
So
could
it
be
a
gap
sure
probably,
but
that
was
not
necessarily
required,
but
it
if
we
want
to
make
it
required.
That's
fine,
too,
I
just
think
that
this
would
probably
be
more
extensive
than
like
the
one
paragraph
on.
How
do
we
view
it
with
the
levels,
whereas
this
should
have
like
we
want
to
test
these
specific
things.
We
plan
to
use
this
tool.
D
A
Okay,
so
yeah
I
agree
that
that
context
of
money,
I,
agree
1686,
is
probably
as
a
super
set
of
the
other,
the
other
issue,
and
that
the
deliverable
for
1686
should
have
like
actual
some
implementation
detail
on
the
the
specific
scenarios
that
we're
going
to
test
as
part
of
the
testing
plan.
So
what
is
the
action
plan
here?
A
Do
we
feel
like
we
should
move
this
I
mean
I,
guess
they're,
both
part
of
the
Milestone,
so
they
should
probably
both
stay
here
or
we
can
close
1488
like
document
what
Mike
just
said
as
far
as
and
then
just
like
put
that
as
a
footnote
here,
I'm,
not
I,
don't
know
any
any
opinions.
C
On
that
I
think
we
just
keep
them
separate
because
they'll
be
easier
to
close
the
smaller
one.
First
like
we
can
just
do
one
small
PR,
Target,
it
close
it
and
then
it's
complete,
rather
than
leaving
that
unblocked
on
the
potentially
more
extensive
one
pair.
B
D
I'm
fine,
you
know
I
think
this
is
sufficient
detail
that
we
want
to
have
that
I'm
fine
with
a
brand
new.
You
know
I
I'm,
not
saying
it
has
to
be
a
gap,
but
the
right
now
Gap
is
the
way
that
we
really
the
only
mechanism
we
have
to
formally
approve
docs
of
this
size.
So
it
can
start
in
any
format,
but
it
probably
needs
to
end
up
in
in
gap
form
as
my
guess
but
Shane.
It
looks
like
you've
got
an
idea.
D
Yeah
no
worries
yeah,
it's
it's
Mike
added
a
good
Link
in
Zoom
chat
here,
which
we
should
add
to
the
agenda
or
to
an
issue
or
somewhere,
but
it's
the
MCS
test
plan,
which
I
think
is
a
good
Baseline
for
the
kind
of
detail
we
could
include
in
ours
I.
You
know
right
now.
D
Basically,
what
I
was
saying
is
the
only
real
mechanism
we
have
in
Gateway
API
to
approve
docs
of
this
scale
is
going
through
the
get
process,
so
that's
probably
the
best
fit,
but
I
know
it's
not
really
a
perfect
fit.
It's
not
really
something!
That's
going
to
graduate
from
experimental
to
stay.
You
know
it's.
C
I
mean
Gap
seems
good
to
me
like
it's.
We
want
a
formal
doc
that
will
be
reviewed
by
Sig
Network
API
reviewers
yep
seems
like
a
fine
fit
for
it.
Yeah.
B
A
I
I'm
already
working
on
the
conformance
test
framework
for
this
kind
of
in
spare
time
so
I've
had
some
thoughts
here.
I
can
take
it
if
nobody
else
is
I'm.
E
I'm,
not
gonna
I'm,
not
gonna.
Personally
stand
up
to
take
this
one,
but
I
have
somebody
who's.
Not
here.
I
would
like
to
ask
because
they
might
be
interested
so
just
if
you'd
give
it
a
day,
I'll
ping
them.
B
A
E
I
love
that
we're
talking
about
conformance
tests.
It
fills
me
with
good
feelings
like
that.
We're
at
that
point
with
camera,
I'm
not
I
mean
I'm,
not
making
a
joke
like
just
good
job,
guys.
I
know
it's
been
a
lot.
There's
a
lot
of
there's
a
lot
of
people
interested
in
this
project.
It's
a
hard
one
to
navigate,
but
we're
at
the
point
where
we're
starting
to
talk
about
tests
and
that's
good.
A
Yeah
yeah
tests
tests
make
things
feel
real
and
yeah
happy
to
be
here
as
well.
A
Okay,
good!
Do
we
I
feel
like
I
asked
us
every
meeting,
Rob,
so
apologies,
but
do
we
have
a
I've
been
the
last
date
that
we
talked
about
was
first
quarter,
preferably
4070
so
into
first
quarter
that
looking
probable
right
now.
D
I
would
love
I
would
love
for
it
to
be
I.
Think
so.
The
the
thing
that
I'm
most
aware
of
that
seems
closest
to
get
across
the
line
to
em
to
include
an
070
is
path,
rewrites
and
no
all
of
rewrites
and
then
just
path,
redirects
and
right
now,
I
know
of
two
implementations
that
have
support
for
it,
and
usually
our
bar
for
graduating,
something
to
Beta
has
been
closer
to
three
or
more
implementations.
D
So
I,
you
know,
I
I'll
keep
on
doing
a
survey
to
see
if
anyone
else
has
support
for
it
or
or
has
it
soon,
but
if
not,
we
either
would
need
to
lower
our
bar
or
wait
or
find
something
else
to
include
an
070.,
not
sure
which
of
those
is
going
to
win
out,
but
that's
current
state
there
may
be.
There
may
be
some
other
things
on
the
roadmap
for
070.
That's
just
been
the
one
that
I'm
most
aware
of
is
seeming
like
the
easiest
lift,
because
it
is
so
close.
A
Okay,
so
what
I
got
from
that
is
hopefully
into
the
quarter.
So,
let's
try
to
to
close
on
these
as
soon
as
we
can
and
write
up
some
sort
of
okay,
cool
yeah.
D
Yeah
for
sure
I,
you
know
I
yeah,
that
that's
really
two
things.
It's
definitely
want
to
have
it
out
by
end
of
quarter
but
probably
need
it
would
be
a
lot
easier
if
one
of
the
implementations
that
hasn't
fit
in
support
for
it
could
do
that.
But
two
you
know
yeah,
it
I
think
it
could
be
meaningful
to
have
a
release.
That's
gamma
only
as
well.
Just
this
is
the
release
where
we
formalize
the
gamma
spec.
So
in
any
case,
I
think
having
that
timeline
in
mind
would
be
good
I.
D
Definitely
you
know
I
hate
to
say
Coop
driven
development,
but
q1
is
end
of
March
and
you
know
kubecon's
coming
around
the
corner,
just
mid-april
so
yeah.
A
Yeah
loud
and
clear,
okay,
so
for
next
next
meeting,
I
I
think
what
I'm
going
to
start
doing
that
I'm
doing
this
already,
but
I
think
we
should
be
this
Milestone
to
pop
up
every
meeting
and
make
sure
that
you've
got
ownership
for
these
things
and
make
sure
that
there
are
people
working
to
close
a
loop
on
it
so
that
we
can
sniff
out
problems
that
are
not
roadblocked
at
the
last
second
ooh
good
question
from
Mike
in
the
chat.
C
E
A
We
actually
opened
up
to
TCP
right
and
other
things
once
John
sponsored.
D
B
D
Think
I
think
I
I
think
I
get
the
question.
The
idea,
as
at
least
as
I
understand
it,
is
that
gamma
would
become
a
part
of
the
experimental
channel
in
Gateway
API.
So
it
would
be
a
subset
of
Gateway
API.
D
B
D
D
B
A
A
If
it
turns
out
that
that's
not
possible,
then
we
will
have
to
have
to
revisit
that,
and
you
know
for
what
it's
worth,
if,
if
any
implementations
are
playing
around
with
the
gamma
apis
and
find
things
find
that
things
aren't
usable
the
sooner
we
can
get
that
feedback,
the
better
before
Things
become,
you
know,
become
set
set,
I'll
say
any
other
questions.
A
All
right,
thank
you
all!
That's
not
the
right
one.
This
is
the
right
one.
Okay,
so
Rob
you've
got
a
question
about
emea,
friendly
meetings.
D
Yeah
so
I
hopefully
got
that
acronym
right,
but
basically
I
I
know
you've
been
rotating
between
Europe
friendly
time
zones
and
APAC
friendly
roughly
meeting
times
and
I
am
curious.
I
know
we
did
this
for
the
main
gateway,
API
meetings
and
it's
come
back
up
as
more
and
more
people
are
interested
in
attending
the
main
gateway
meetings
and
just.
B
D
Know
can't
make
the
current
time
fit
I'm
curious,
how
you've
seen
this
experiment
work.
You
know
I
know
when
we
were
originally
setting
up.
Gamma
was,
you
know,
let's
see
how
it
works
and
reevaluate
when
we
originally
did
this
with
Gateway
meetings,
we
had
a
pretty
significant
drop
off
in
the
Europe
timed
meeting,
but
it
seems
like
you
may
be,
having
the
opposite.
D
Okay
I
mean
that
that's
great
I
I
feel
like
we
may
inevitably
have
to
go
to
a
model
where
we're
kind
of
bouncing
back
and
forth
where
every
week
there's
one
meeting
that
is
in
each
time
zone
and
we
kind
of
alternate
whether
that's
Gamma
or
main
gateway
meeting
figure
out.
The
timing
figure
out,
but
just
just
trying
to
get
an
idea
of
how
that's
been
received
so
far,
and
it
seems
positively
so
cool.
The.
A
Yeah
one
more
thing
on
that
one
I
think
I
think
I
just
wanted
to
capture
Freeman
on
the
next
topic.
I
think
I
I
agree
with
Mac.
As
far
as
our
experience
I
think
I
take
that
with
a
little
bit
of
a
grain
of
salt,
because
I
know
of
several
big
mesh
folks
who
are
on
the
East
Coast
television
mesh
companies
that
are
East
Coast
space
solo
being
one
of
them.
A
Linker
D
has
several
folks
who
are
either
on
East,
Coast
us
or
are
in
Europe,
so
I,
don't
know,
I
mean
I,
know
that
folks
would
probably
love
to
attend
both.
So
you
might
get
more
mesh
Ingress
cross-pollination
if
there's
a
swap,
but
I
did
just
want
to
kind
of
point
that
out
that
there
is
some
mesh
specific
context
that
is
likely
leading
to
our
our
attendance
disparity.
There.
D
Yeah
that
that
makes
complete
sense
and-
and
you
know
it
is
unfortunate
that
our
current
time
is-
is
so
hard
for
even
Eastern
Time
us
to
to
attend.
So
we
can
experiment
a
bit
there.
D
D
D
If
there's,
you
know
just
start
thinking
about
what
might
be
valuable
to
to
have
at
kubecon
this
year,
but
but
first
off,
maybe
just
you
know
we
can
do
the
the
raise
hand
or
whatever,
if
you
plan
on
being
there,
that
could
be
an
easy
way
to
do:
Paul
of
sorts,
okay,
cool
and
yeah
anyone
else,
but
just
just
definitely,
if
you
have
ideas
for
what
we
can
do
in
in
person,
that
would
be
great
I.
Think.
D
Last
time
we
just
had
some
discussions
that
were
were
helpful.
We
completely
failed
to
find
a
way
to
make
it
a
hybrid
event,
I'd
love
to
figure
that
out
more
this
time,
because
clearly
there
are
people
on
this
call
that
are
not
going
to
be
able
to
make
it
so
yeah.
D
Cool
well,
that's
all
I
had
I
just
wanted
to
get
a
an
initial
feeler
out.
There
I
think
the
schedule
formally
is
released
tomorrow,
but
lots
of
good
Gateway
content
I
think
even
a
little
bit
more
than
we
had
in
Detroit.
So
should
be
a
good
event.
A
All
right,
yeah.
A
Like
I
said
my
working
on
getting
on
getting
all
the
t's
crossed
and
dotted
to
be
able
to
be
there,
so
I
hope
that
all
that
was
when
I
can
see
everybody
there.
A
That
is
the
end
of
our
agenda.
So
I
am
happy
to
give
folks
30
minutes
back.
If
there's
nothing
else
to
chat
about.