►
Description
Don’t miss out! Join us at our upcoming hybrid event: KubeCon + CloudNativeCon North America 2022 from October 24-28 in Detroit (and online!). Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Contour Ingress Intro and Deep Dive - Nick Young, Orlin Vasilev & Nigel Brown, VMware
Contour, a CNCF incubating project, is a high performance ingress and load balancer solution for Kubernetes. Contour offers a richer feature set than some common alternatives while maintaining a lightweight profile. At its core, Contour is providing a control plane for the Envoy edge and service proxy. This session will show you how to leverage Contour and Envoy for Kubernetes workloads in a multi-tenant environment as well as include a demo of recent Contour features. We will also focus on the project roadmap including enhanced support for Gateway API, the Contour Operator for enhanced lifecycle management, distributed tracing support, and much more.
A
Thank
you
for
attending
our
session,
so
late
in
friday,
just
before
we
go
or
go
for
for
fun.
So
that's
the
contour,
intro
and
deep
dive,
so
we're
gonna
we're
gonna
go
through
this
one.
It's
like
we're
the
three
spirits
of
contour
or
christmas,
I'm
the
past!
That's
the
present
and
the
future
over
there.
So
I'm
going
to
cover
what
happened
through
the
last
year
in
terms
of
community.
A
So
does
the
team,
as
you
can
see,
it's
primarily
built
from
vmware
folks,
but
we
have
a
few
from
bloomy
and
erickson
on
stage
we
have
nick,
who
is
a
technology
lead,
and
then
we
have
nigel,
the
current
community
manager
and
I'm
orlean
the
previous
community
manager.
I
said
the
past
so
with
that
said,
that's
from
the
last
may,
until
this
may,
so
we
have
230
contributors
in
total.
A
So
what
we
are
proud
of
is
the
average
average
hours
of
response
for
pr
so
from
open
to
merge.
We
tend
to
be
around
12
hours,
which
is
we
find
super
cool
and
we're
kind
of
responsive
community,
so
the
new
community
managers.
That's
me
and
nigel
thanks
to
jonas,
who
was
the
previous
previous
previous
community
manager.
What
else
we
did?
A
We
started
a
technical
docs
writing
group
wow,
which
is
an
effort
to
try
to
combine
the
the
community
to
support
the
technical
writing
and
to
to
be
able
to
include
more
diverse
folks
into
the
community
in
terms
not
to
be
only
technical
and
code,
these
contributions,
but
technical
and
writing.
Okay.
I
said
that
a
couple
of
times
also,
we
implemented
new
levels
of
community
rows,
so
we
have
now
reviewers
contributors
which
allows
the
community
to
be
more
easily
accessible.
A
C
D
A
A
No
I'm
not,
but
but
thank
you,
everyone
who
spent
their
time
and
and
and
made
that
project
super
nice
and
for
the
last
one
it's
again,
maybe
not
very
easy
to
read
it's
more
for
a
reference,
how
many
organizations
are
actually
willing
to
spend
their
time
and
their
employees
time
for
our
project
you
can
see.
The
list
is
super
huge
yeah.
There
are
a
few
on
top,
which
is
vmware,
ericsson
and
redhead.
Those
are
the
three
major
playing
in
the
in
the
field
right
now
with
that.
Thank
you
very
much.
B
Alright,
okay,
so
everyone
can
hear
me:
fine,
yeah,
sweet,
okay,
so
a
couple
of
things,
I'm
gonna
sort
of
give
a
bit
of
an
update
about
our
road
map
and
then
do
a
see
if
I
can,
if
the
demo
got
smile
on
me,
and
I
can
smash
out
a
quick
demo
for
you
of
what
what
our
current
gateway
setup
looks
like
so
yeah.
The
first
thing,
speaking
of
gateway
that
we
need
to
talk
about
is
on
the
gateway.
B
Okay,
so
who
here
has
seen
the
announcement
about
the
onboard
gateway?
No
okay,
cool,
excellent,
a
few
people
great
so
envoy
gateway
is
envago,
is
a
new
initiative
from
four
organizations,
vmware
tetrad,
ambassador
and
fidelity.
So
the
vmware
is
obviously
me
and
the
other
vmware
controller
maintainers
we're
going
to
build
a
gateway,
api
implementation
that
uses
envoy
as
go
as
the
data
plane.
But
this
one
is
a
branded
envoy
sub
project.
B
So
this
is
a
different
thing
to
contour,
that's
built
under
the
envoy
banner,
and
so
like
one
of
the
things
I
was
worried
about
with
this
announcement
is
that
people
would
be
worried.
You
know,
like
that's
the
same
sort
of
thing
as
contra
is,
and
so
one
of
the
things
I
was,
I
thought
that
people
might
be
worried
about.
B
Is
you
know,
what's
going
to
happen
to
contour,
so
yeah
I've
got
that
qr
link
is
the
link
to
the
blog
post
that
I
wrote
that
sort
of
explains
where
we're
at,
and
so,
if
you
want
any
details
about
what's
going
on
that,
you
definitely
should
grab
that
qr.
Think
code,
thing
and
sort
of
these
are
some
of
the
questions
I
anticipated
people
asking
like
you
know,
will
contour
stop
being
released,
is
contour
going
to
be
absorbed
into
envoy
gateway,
and
so
here's
our
sorry?
B
Does
anyone
want
me
to
go
back,
so
you
can
take
a
picture
of
that?
No
cool,
yeah,
okay,
good.
So
this
is
our
commitment
to
our
users.
The
short
answer
is
no
contour's
not
going
anywhere.
Contra
is
staying
where
it
is.
We
will
continue
feature
development
on
contour
we've
got.
You
know
I'll
talk
more
about
the
roadmap
in
a
minute,
but
we're
not
going
anywhere,
and
we
want
to
give
you
this
commitment
like
no
one's
going
to
be
left
behind,
no
matter
what
happens
with
envoy
gateway.
B
As
long
as
people
are
using
contour,
you
know
we
will
help,
maintain
and
support
it
for
those
people
who
do
want
to
move
to
the
gateway
api
and
on
the
gateway
we'll
provide
tooling
to
help
you
convert
http
proxy
resources
to
gateway
api
resources.
Actually,
so
the
bit
I
missed
at
the
start
of
this
is
the
headline
here:
is
that
we're
going
to
be
moving
the
gateway
api
development,
though
resources
over
to
envoy
gateway,
so
we're
going
to
sort
of
round
out
the
gateway
api
support?
B
That's
in
contour
that
I'll
demo
for
you
in
a
little
bit
and
then
we're
going
to
sort
of
move
that
effort
over
to
onboard
gateway,
but
the
rest
of
contour's
development
will
continue
as
per
normal,
okay,
so
yeah
we
want.
We
want
to
make
it
so
that
you
can
move
to
gateway
api
gateway.
Api
is
the
future
of
ingress
and
kubernetes.
B
Definitely,
and
if
any
of
what
I'm
saying
right
now
does
change
we'll
give
you
we'll
we'll
tell
you
shout
as
loud
as
we
can
on
every
every
channel
we
have
to
shout
at
you
at
and
make
make
sure
we
give
you
any
notice
if
any
of
these
changes,
but
the
the
key
part
is
that
I
want
you
to
know
is
contra
is
not
going
anywhere.
We
pledge
to
you
when
we
all
versions
after
119
would
always
have
at
least
one
year
of
support
and
that
that
has
not
changed.
B
So
you
know
nothing.
Nothing
will
change
that
unless
we
tell
you
otherwise,
in
which
case
we
still
got
to
give
you
a
year's
notice,
because
that's
what
we
promised
you
okay,
so
I
just
wanted
to
set
everyone's
mind
at
ease,
who
might
have
heard
the
invoke
gateway
news
that
you
know
there's
nothing
to
panic
about
we're
not
going
anywhere
contra's
staying
around.
You
know,
of
course,
maybe
two
years
in
the
future
like
envoy
gateway,
is
a
completely
awesome
solution
and
there's
no
reason
for
you
to
use
contour.
B
Then
then,
maybe
we'll
have
a
chat,
but
there's
a
lot
of
things
that
have
to
happen
for
something
like
that
to
happen.
So
there
is
no.
I
still
strongly
recommend
if
you're
interested
in
contour,
it
is
100
fine
to
pick
up
and
get
using
right
now,
and
any
change
that
you
need
to
do
in
the
future
will
hold
your
hand
every
step
of
the
way.
Now
I
think
I
am
going
to
just
stop
for
a
sec
to
say.
B
Okay,
cool,
okay,
no
way!
So
if
you
do
have
any
other
questions
about
it,
I'm
young
nick
on
twitter,
young
nick
on
github,
please
feel
free
to
hit
me
up
and
ask
okay
and
those
those
of
you
playing
at
home
same
goes
for
you.
If
you
have
any
questions
about
any
of
this,
read
the
blog
post
and
hit
me
up
about
it
if
need
be.
Okay,
okay!
B
B
B
So
you
can
configure
the
oauth
filter
that
sits
inside
envoy
so
that
you'll
be
able
to
when
that's
done,
you'll
be
able
to
use
a
http
proxy
resource
and
request
oauth
for
any
anything
that
you
are
protecting
like
that,
you
are
exposing
via
contour
with
a
pretty
simple
setup.
The
oil
filter
is
a
little
tricky
and
there's
a
lot
to
figure
out
with
it,
and
it's
still
alpha.
So
there's
quite
a
lot
of
work.
There
left
to
do
it's
taking
a
lot
longer
than
I
thought
it
would.
To
be
honest.
B
To
the
extent
that
I
know
that
the
issue
number
is
number
399
in
the
contour
repo
because
it
has
been
around
for
so
long
and
I've
had
to
reference
it
so
many
times.
That
is
what
that
one
is
held
up
on
a
redesign
of
our
extension
service
crd,
which
is
linked
to
the
http
proxy
one,
so
yeah.
These
are
all
of
these.
B
Things
are
coming,
so
that
is,
that
is,
for
those
of
you
who
you
know
quite
a
bit
about
contour,
already
we're
kind
of
putting
the
deep
dive
a
little
bit
before
the
intro,
so
sorry,
but
now
next
up,
I
would
like
to
run
you
through
a
quick
demo
now.
So
it's
this
way.
Okay,
that's
good!
Is
the
text
big
enough?
Everyone
can
see
that
yep
cool
cool
okay.
So
for
those
of
you
who
haven't
seen
the
gateway
api
before
the
gateway
api
involves
a
set
of
resources.
B
That
basically
will
do
in
this
case,
I'm
going
to
show
you
what
the
gateway
api
setup
that
replicates
what
an
ingress
would
look
like
in
kubernetes
looks
like
okay,
so
that
needs
first,
you
need
a
gateway
class
which
is
just
like
an
ingress
class
in
you
know,
with
ingress,
it's
just
basically
just
a
bucket
to
hold
gateways.
Now,
there's
a
possibility
that
you
can
use
that
to
configure
like
special
parameters
and
stuff.
B
B
And
so
yeah
I've
got
the
I've
got
our
gateway,
automatic
gateway,
provisioner
already
up
and
running
what
that
does.
Is
it
watches
the
the
cluster
for
gateway,
class
gateway
resources
and
provisions
a
whole,
a
contour
and
envoy
to
match
your
gateway
request,
so
the
gateway
is
the
key
resource
in
the
gateway
api.
B
It's
like
the
it's
the
way
that
you
define
how
the
traffic
should
get
into
your
cluster,
so
in
in
a
lot
of
it's
a
little
bit
like
in
some
ways
the
bits
of
the
service
of
type
load
balancer
that
lets
you
expose
certain
ports
and
stuff
like
that.
But
it's
also
where
you
attach
the
routes
to
which
I'll
get
to
in
a
minute.
So
first
thing
I'll
apply
the
gateway
class.
B
Okay,
no
problems
there.
Now,
let's
show
you
the
show
you
the
gateway,
pretty
simple
gateway
here.
This
is
just
saying:
hey,
create
a
gateway
in
namespace
version.
Contour,
the
gateway
class
name
is
the
one
that
I
just
created,
the
contour
gateway
class
and
it
has
a
http
listener.
I'm
not
going
to
worry
about
all
the
shenanigans
involved
in
tls
at
this
point,
it's
a
little
too
hard
for
a
simple
demo
like
this.
B
So
you'll
see
that
there's
one
of
the
different
things
to
ingress
is
that
the
route,
the
details
of
like
what
host
name,
what
you
know,
what
path
and
all
that
sort
of
stuff?
That's
not
in
this
object.
All
this
object
has
is
like
port
and
the
protocol,
and
what's
allowed
to
attach
to
it
right
so
I'll,
apply
that
one.
B
Okay,
so
you
can
see,
we've
started
con
two
contour
pods
and
an
envoy
demon
set
has
started
up
as
well.
So
when
I
created
that
gateway
object,
these
were
not
present
before
I
don't
know
if
you
noticed
before
when
I
did
kgetpod
here,
it
was
only
the
gateway
provision
of
running
when
I
create
the
gateway
object.
All
these
the
the
gateway
provisioner
magically,
goes
off
and
creates
the
the
contour
and
envoy
pods
for
you.
So
you
don't
need
to
go.
B
You
go
and
figure
out
how
to
install
like
contour
itself
and
envoy
the
gateway.
Provisioner
knows
how
to
do
that
for
you
and
does
that
in
response
to
you
creating
a
gateway?
Okay,
so
that
one
is
ready.
The
contour
pods
are
ready.
We
are
good
to
go
so
in
order
to
do
that,
hang
on
I'm
going
to
switch
to
the
default
namespace,
so
no
pods
in
the
default
namespace.
B
No,
that
is
not
the
command
I
intended
to
type
okay,
so
we've
got
just
a
really
simple,
simple
deployment
here
you
can
see.
There's
a
single
deployment
single
service
like
this
is
a
very
basic
web
service.
It's
just
going
to
show
that
I
can
get
some
html
out
of
it.
Basically,
okay,
so
we've
got
a
service
we
can
now.
Normally
you
could
just
create
an
ingress
object
to
expose
this
service.
Contour
would
pick
that
up
do
its
magic
and
then
you'd
be
able
to
see
everything
as
per
normal.
B
Is
we're
going
to
create
a
http
route
object?
Now
this
hp
route
object
is
sort
of
the
http
specific
bits
that
you
can
attach
to
the
gateway.
Now,
each
type
of
traffic
that
you
might
want
to
send
through
a
gateway
has
a
separate
type
of
route.
So
there's
http
route,
there's
a
tls
route
for
things
where
you
don't
want
to
terminate
your
tls
traffic.
You
want
to
just
forward
the
tls
traffic
through
there's
a
tcp
route
and
a
udp
route.
B
Now
contour
only
supports
the
http
and
tcp
route
because
we're
a
layer,
7,
ingress,
controller,
okay,
so
key
parts
about
this
one.
You
can
see
the
parent
ref
section
that
is
saying,
hey
my
parent.
I
want
you.
I
want
to
bind
to
the
contour
gateway
in
the
project,
contour
namespace,
that's
the
line
I
just
created
and
I'm
going
to
listen
on
the
local.projectcontour.io
hostname
and
just
sir
send
everything
to
the
quad
service
that
I
just
installed
right,
very
basic
config.
B
A
E
E
B
There
this
is
the
sort
of
normal
describe
you'll
see,
but
the
key
part
is
the
status.
You
can
see
that
the
status
contour
has
updated
this
status,
it's
a
valid
http
route.
It's
been
accepted.
That
means
that
it's
now
been
configured
okay,
so
I'm
just
going
to
because
I'm
running
on
kind
on
a
local
machine.
I
did
not
want
to
be
relying
on
the
conference
wifi,
so
I'm
just
going
to
do
a
quick
port
forward
to
the
to
the
onboard
pod
that
was
spun
up.
B
B
Okay,
if
I
repeat
that
same
command,
nothing,
okay,
that's
because
we've
taken
out
the
ingress
thing
so
yeah,
that
is
the
system
working
like
I
said,
there's
not
there's
not
a
lot
of
there
there
in
this,
but
the
whole
point
of
the
http
route
object
is
to
have
a
great
place
to
put
all
of
the
config
that
currently,
if
you're,
using
engine
ingress
engine
x
or
our
http
proxy
object
lives
in
a
bunch
of
other
fields.
This
is
a
standard
way
to
do
that
right.
The
whole
point
of
this
is
it's
standard.
B
B
D
D
So
community
is
integral
to
any
open
source
project.
Open
source
doesn't
really
work.
If
we
don't
have
it
and
we
are
very
sensitive
to
the
needs
of
what
our
folks
are
asking
for,
and
we
want
you
all
to
be
as
much
a
part
of
it
as
you
would
like
to
be
so
wherever
you
are
in
your
contour
journey,
orlan
mentioned
it
earlier.
D
D
So
where
can
you
find
us
all
of
the
information
about
our
community
is
on
the
website,
so
project
contour,
dot,
io,
slash
community.
It
has
all
the
information
about
when
we
meet
where
we
meet,
but
I'm
going
to
go
through
some
of
it
for
you
right
now.
We
have
community
meetings
every
tuesday
in
alternating
time
zones.
I
don't
know
if
you
noticed
nick
has
a
bit
of
an
australian
accent,
so
we
have
our
australia
friendly
time
zones
on
the
first
and
third
tuesdays
of
the
month,
and
that
happens
at
9
30
a.m.
D
Eastern
and
our
north
america
friendly
time
zones
happen
on
the
second
and
fourth
tuesdays
at
1
pm
eastern
and
then
we
are
starting
a
brand
new
google
group
for
handling
email.
I
can
assure
you
if
you
join,
you
won't
get
blasted
with
a
bunch
of
emails,
because
I
don't
particularly
like
writing
emails.
D
I
will
just
give
you
the
essential
updates,
so
we'll
probably
average
between
one
and
two
emails
a
week,
but
this
is
where
you'll
find
requests
for
comments.
You'll
find
what
are
what
kinds
of
things
we're
working
on
any
topics
that
we
need
to
bring
up
in
our
community
meetings.
Anything
to
keep
you
updated
will
be
happening
at
this
google
group
and
we'll
start
getting
that
up
online
next
week.
D
If
you
have
a
chance
to
check
out
the
talk
that
nancy,
lancaster
and
karen
chu
did
or
john
mcbride
about
what
happens
when
you
don't
have
a
large
community,
especially
of
maintainers
around
projects
as
well
as
the
keynote
this
morning
from
who
was
it
this
morning?
Lee?
Yes,
thank
you
from
lee
with
that.
We
want
to
open
it
up
to
questions.
There
are
three
of
us
here.
There
are
names
and
handles
if
you
would
like
to
direct
any
questions
to
us
in
particular.
D
B
The
the
simplest
one
is
that
contour
is
a
thing
that
you
can
run
right
now,
envoy
is
not
on
the
gateway
is
not
envoy
gateway.
Is
the
intent
to
create
a
project,
not
a
project
as
yet
yeah.
We
have
zero
code.
We
are
still
in
discussions
about
exactly
how
we're
going
to
do
that.
But
contour
is
usable
today,
any
stuff
you
do
with
a
gateway
api
today
with
contour
will
you'll
be
able
to
literally
pick
it
up
and
use
it
with
onboard
gateway.
B
That's
the
whole
point
of
mumble
gateway,
but
even
if
you,
as
I
said
before,
even
if
you
want
to
learn
about
how
to
do
some
of
our
more
advanced
features,
like
you
know,
weighted
load,
balancing
rate
limiting
external
auth
and
a
whole
bunch
of
the
other
cool
stuff
that
we've
got
built
into
the
http
proxy.
Like
I
said,
we
are
going
to
provide
a
tool
to
help
convert
that
config
across
so
yeah.
B
There's
we're
going
to
make
that
process
as
smooth
as
possible,
but,
like
look
to
be
real,
I
really
don't
see
we're
hoping
to
have
sort
of
our
first
prototype
cut
for
envoy
gateway
before
cubecon
america
later
this
year,
but
that's
like
the
first
prototype
and
there's
a
lot
of
work
to
do
to
get
our
gateway
to
parity
with
the
with
contour
yeah.
I'd
be
pretty
I'd,
be
pretty
happy
if
we
were
like
at
parity
with
contour
by
kubecon
north
america
2023,
which
yeah
it's
still
like
nearly
two
years
away.
E
B
Okay,
sure,
can
you
review
the
question,
yeah
sure
so
yeah
so
we're
so
you
so
you
asked
about
tls
support
and
automatic
tls
certificate
selection
yeah.
So
obviously
we
support
tls
right.
Like
that's
a
no-brainer
table
stakes
feature
everyone
should
do
that
the
and
we
support
the
usual
ways
of
doing
this
with
having
you
know
a
secret
with
your
tls
certificates
in
it.
Now
I'm
not.
Can
you
maybe
explain
to
me
a
little
bit
more
about
what
you
mean
by
like
selecting
multiple
certificates.
B
B
Right,
okay,
cool,
so
so
yeah
look!
So
if
I
were
to
summarize
that
since
you
didn't
have
the
mic,
what
you'd
like
to
be
able
to
do
is
to
have
like
have
to
make
it
easier
for
developers
to
be
able
to
expose
their
their
services
on
their
domain
on
in
individual
domains
safely
without
having
to
like
revision,
their
own
search.
You
want
to
be
able
to
have
a
central
platform
team
provision,
asserts
and
and
then
have
the
things
so
yeah
right
right
now.
Like
I
mean
we
don't
do
that
for
you.
B
We
certainly
would
anticipate
that
you
use
something
like
that
manager
to
help
with
that.
So
what
so?
The
way
that
I
would
suggest
that
you
do
something
like
that
is
depending
on
exactly
how
magic
you
want
to
make
it.
What
I
would
suggest
looking
at
is
we
have.
We
do
have
another
cid
called
a
tls
certificate
delegation
that
lets
you
drop
a
tls
certificate
into
a
secret
in
one
name
space
and
then
refer
to
that
secret
from
another
namespace
right.
B
So
if
you
know
allen's
got
allen's
running,
here's
a
super
awesome
blog
service
in
namespace
in
all
in
namespace,
and
I'm
the
platform
owner
and
I'm
like
well,
I
want
to
make
I
want
to
make
that
available
on
like
orland.com
and
yeah,
and
so
he
wants
there
to
be
tls
for
that.
But
I
want
to
make
issue
the
keepers.
I
can
issue
the
key
pairs
in
a
secret
in
the
infra
name
space
and
create
a
tls
certificate
delegation
that
sits
next
to
my
certificates.
B
In
my
name,
space
that
says,
hey
orlan's
allowed
to
use.
You
know
to
use
this
tls
certificate
from
his.
You
know
all
the
namespace
okay,
and
so
that's
what
I
would
recommend
doing
for
this
you
can.
You
can
have
a
single
tls
certificate
delegation
that
covers
like
all
secrets
in
a
namespace,
so
you
could
have
a
single
namespace
that
held
all
of
the
secrets
for
all
the
dynamically
created
things
in
terms
of
actually
dynamically
creating
them.
That
is
not
the
business
that
we
want
to
be
in
with
contour.
B
You
know,
other
people
are
doing
that
doing
it
better.
Creating
and
managing
certificates
is
a
hard
problem.
That's
very
easy
to
screw
up
and
if
you
screw
up,
you
are
creating
huge
security
problems
for
yourself,
so
I
certainly
would
recommend.
If
you
want
to
do
dynamic
certificate
management,
don't
write
your
own
thing.
Use
cert
manager,
use
vault,
use
some
use,
something
that
someone
whose
job
it
is
to
do
nothing
but
build.
B
That
thing
has
built
for
you
so
yeah,
but
that
is
how
I
would
recommend
doing
that
to
make
it
a
bit
more
magic
have
have
it
set
up
so
that
the
certificates
get
get
provisioned
into
a
central
namespace
and
use
tls
certificate
delegation
to
move
across.
Now.
Of
course,
there
is.
There
is
one
other
option
here,
and
that
is
if
everybody's
domains
fit
under
the
same
tld.
B
You
know
if
everybody
is
at
example.com
and
you
can
give
them
subdomains
off
example.com,
then
you
can
just
issue
a
wildcard,
cert
and
use
to
tls
certificate
delegation
to
keep
that
in
a
safe
place
and
then
have
everybody
just
say.
My
tls
cert
is
the
the
wildcard
one
and
so
that
that
is
like
the
easiest
possible
way
to
do
it,
but
it
requires
you,
everybody
to
be
exposing
on.
You
know:
orland.example.com
yeah,
rather
than
all
on.com
yeah
thanks.
Does
that
yeah
cool
great
question?
Thank
you.
A
B
B
B
We've
got
an
example,
one
that
will
do
like
basic
auth
for
you,
but
lots
of
other
people
have
ones
that
you
know
will
do
full
oodc
or
you
know,
talk
to
your
a.d
or
do
a
bunch
of
other
stuff
like
that
right.
So
that's
what
we
have
support
for
today.
You
a
complaint
that
I
have
heard
that
I
think
is
completely
justified.
Is
that
there's
a
lot
of
philly?
D
B
I
would
say,
or
otherwise
but
yeah
so
yeah,
it's
like
there's
a
lot
of
fiddly
fiddly
stuff.
You've
got
to
do
there
to
actually
make
everything
line
up
and
work,
and
so
yeah
the
the
sort
of
the
effort
that
we're
doing
is
to
find
ways
to
make
that
a
little
easier.
Whether
that's
update
our
the
example.
B
So
you
tell
you
would
tell
your
http
route,
you
know
hey.
I
want
to
configure
my
oauth
flow
in
this
way
and
I,
if
who
here,
has
configured
off
in
the
past
yeah,
you
all
know
right.
There's
like
20
million
settings.
You've
got
to
have
for
that.
So
we're
going
to
have
to
wait
have
a
way
for
you
to
provide
those
20
million
settings
as
part
of
your
hp
proxy
or
to
your
contour
install,
so
that
then,
but
then,
if
once
we
do,
that
and-
and
I
you
know,
get
the
design
right.
B
D
And
if
you
want
to
file
an
issue
in
the
repo
and
tell
us
a
little
bit
more
about
some
of
the
specifics
of
what
you're
looking
for,
that
will
help
us
keep
track
of
it
and
be
in
communication,
and
that's
for
anyone.
If
you
have
any
ideas,
problems,
questions
file,
an
issue
against
the
repo
we'll
triage
it
and
get
back
to
you
about
how
we
can
incorporate
your
feedback
into
into
the
project.
Yeah.
B
I
guess
the
way
I
would
say
that
is
issues
are
not
just
for
blog
reports.
Yeah
you've
got
a
question
to
ask
us.
You
want
to
ask
us
for
something
like
the
worst.
That
we'll
say
is
yes,
sorry,
the
telescope,
like
that's
the
you
know,
the
worst
possible
answer.
I
can
give
you,
but
I
I
won't
say
that
very
often,
although
we
have
said
it
in
the
past
about
udp
support,
so
maybe
check
that
we
have
an
open
issue
about
that.
One
already.
B
If
that's
what
you
want,
but
you
know
yeah,
I'm
making
a
joke
there.
Please
just
I'd,
prefer
that
you
open
another
issue
and
I'll
point
you
to
the
one
that
we've
already
opened.
Issues
are
free.
I
don't
care
if
we,
if
we
burn
them
on
people,
asking
questions
that
we've
already
answered.
I
would
much
rather
have
you
ask
the
question
and
get
a
good
answer
than
for
you
to
be
too
worried
about
doing
that.
B
B
Even
if
we've
already
answered
the
question
you
opening
the
issue
tells
us
that
we
haven't
given
you
a
way
to
find
that
question
to
answer
that
question
after
yourself,
right
like
we
haven't,
we
need
to
update
our
documentation
or
something
like
that,
like
every
everything
that
you
do
for
us
helps
us
the
best
way
that
you
can
help
us
help.
You
is
to
tell
us
what
we're
not
doing
and
or
to
tell
us
what
we
are
doing.
Hey
positive
feedback.
Is
nice
too
right
like
yeah
yeah,
so
yeah
yeah
cool,
hang
on
wait
for.
B
B
The
one
thing
that
I
have
learned
after
nearly
three
years,
I'm
also
a
maintainer
on
the
gateway
api
and
the
one
thing
I
have
learned
after
nearly
three
years
of
being
maintained
on
the
gateway
api.
Is
writing
specs,
like
that
it's
really
bloody
hard.
Like
you
know,
it's
really
hard
to
write
a
very
clear,
very
well-defined
spec
that
that
people
can
actually
use
and
we
can
actually
implement,
and
so,
but
at
the
same
time,
we've
got
to
be
really
careful
that
we're
not
like
messing
it
up
by
rushing
okay.
B
So
I
really
want
so
external
auth
rate
limiting
circuit
breaking
tracing.
All
of
those
things
need
to
be
present
in
the
upstream
gateway
api.
They
are
all
on
the
list
of
things
to
do
in
the
upstream
gateway
api,
but
we
just
haven't
got
to
them
yet
so
yeah
yeah,
I
mean
yeah
if
nigel's
talking
about
finding
ways
to
contribute
gateway.
Api
desperately
needs
people
who
know
about
that
sort
of
stuff.
Who
need
it
to
to
tell
us
more
about
the
things
I
mean
giving
us.
B
B
E
B
So
yeah
the
the
beta,
the
beta
for
the
first
parts
of
gateway
api,
is
due
like
any
day
hoping
to
get
it
done
before
like
before
kubecon,
but
we
just
keep
finding
like
little
things
that
we've
got
to
fix
up.
I
hope
within
the
next
couple
of
weeks
you
know,
but
I
just
we
just
it's
like
a
hydrate.
You
know
you
just
keep.
We
we
think
we're
almost
there
and
we
get
down
to
the
last
thing.
We're
like
okay.
This
is
the
last
one
we're
going
to
resolve
this
pr.
B
Someone
really
is
a
really
helpful
contributor
helps
us
out
with
the
pr
and
they're
like
they're
like
oh.
I
just
found
this
question
about
this
edge
case
and
we're
like
oh
no
yeah.
Now
we've
got
to
go
back
and
design
some
more
so
and
that's
just
been
happening
heaps
and
heaps.
But
that's
that's
again.
We
don't
want
to
rush
this
api.
We
might
want
to
make
sure
that
it's
fully
specced
out
and
we
don't
leave
big
holes
there
and
so,
like
the
system,
is
working.
B
Sadly,
it
means
the
system
is
going
slowly,
but
yeah
so
and
like
real.
Soon,
now
is
the
only
sort
of
time
frame.
I
can
give
you,
after
that,
we're
going
to
look
to
move
some
of
the
the
thing
with
the
gateway
api
is.
We
can
move
the
objects
one
by
one,
so
we,
I
would
pre
anticipate
we're
going
to
be
having
pretty
regular
releases
for
like
the
next
year
of
the
gateway
api.