►
Description
[SIG-Network] Ingress NGINX Bi-Weekly Meeting for 20220929
A
Hello:
everyone,
it's
James,
strong,
it's
September
29th.
This
is
the
sub
project
meeting
for
Sig
networking,
Ingress
nginx,
which
means
as
cncf
project,
and
it
is
subject
to
the
code
of
conduct,
which
means
be
awesome
to
each
other.
If
you
have
any
issues,
please
report
those
to
myself
or
the
Sig
networking
leads.
Thank
you,
and
with
that
we've
got
a
couple.
People
I
haven't
seen
their
names.
If
you'd
like
to
introduce
yourselves,
tell
you
your
interest
in
the
project.
A
How
are
you
using
Ingress
or
what
you'd
like
to
see
and
help
with
in
the
community
that'd
be
that'd,
be
great,
no
pressure.
You
can
just
ignore
me.
Yes,
just
for
information.
In
fact,
we
have
a
problem
in
France
when
you
say
the
name,
we
just
put
the
name.
So
that's
why
I
just
put
the
order,
but
it
is
so.
I
have
joined
your
meetings
since
the
beginning,
but.
A
All
right,
that's
fun.
Well,
I
did
not
do
any
of
the
pulling
up
issues
like
our
good
friend
Noah
used
to
do
for
us,
so
we
can
I
want
to
make
sure
that
we
do
the
triaging,
because
we're
really
bad
at
not
doing
triaging.
A
We
need
to
spend
the
15
minutes
doing
triaging,
and
then
we
can
go
back
to
our
conversation,
topics
of
the
1
4
release
and
looking
at
the
stabilization
projects,
and
hopefully,
by
that
time,
Gentile
will
be
able
to
join
us
and
talk
about
the
open
Rusty
upgrade.
Does
this
sound
good
to
everybody
and
crowd?
That
means
you
have
to
pay
attention
yeah.
A
A
Well,
before
we
jump
into
all
of
these
I
know
long,
you
had
a
couple
that
you
were
looking
at
that
and
Verona
did
you
say
you
had
an
issue
that
you
wanted
us
to
look
at
or
you
just
here
to
hang
out
I'm
sorry
I
didn't
catch
that
basically.
B
Basically,
I
think
if
we
want
to
triage,
then
it
will
really
help
if
you
triage
the
issues
that
are
marked
as
bug
or
help
wanted.
For
example,
there
is
two
grpc
stuff
that
only
Ricardo
or
Jinta
somebody
who
knows
grpc
can
look
at
and
then
there
is
a
couple
that
are
monochized
bug
again
only
Ricardo.
A
And
look
at
that:
okay
publisher
poll
list,
those
together
there's
any
grpc
issues
but
okay.
This
looks
like
a
support
question
as
a
possible
to
Target
multiple
kit,
Services
behind
a
single
server
name
and
location.
B
A
Okay,
can
you
and
you
drop
the
needs
triage
in
priority,
then,
since
you're
supporting
them
on
the
slack.
B
Again,
this
is
an
install
thing,
they're,
not
giving
enough
information.
You
said
they
manually
messed
with
the
maps
and
stuff
because
they
had
incident
so
just
to
make
sure
that
the
incident
was
caused
by
a
manual
intervention.
They
experimented
and
once
again
they
manually
tried
to
mess
with
it
yeah.
Then
they
started
having
issues
and
I
suggested
they
do
a
upgrade,
so
it
will
reconcile,
but
instead
of
getting
information
about
the
exact
reconciled
or
upgrade
process
or
reinstall
process,
we
just
sent
one
message.
B
A
B
C
B
So
they
had
an
incident
and
they
suspected
that
somebody
messed
with
a
Helm
release
manually.
So.
B
Reproduce
to
verify
that
somebody
did
do
that
they
went
ahead
and
did
a
second
messing
around
of
manual
deletion,
yeah.
A
I,
don't
care
okay
until
they
manually
deleted
it
and
we're
just
waiting
to
hear
back
on
if
the
fresh
install
fixes
the
issue
yeah,
okay,
I'll
get
another
support.
One
I
think
I
read
through
this
one:
the
k3s
and
soul.
A
B
Basically,
you
see
if
the
user
Ubuntu
server
that
uses
mdns
a
system
to
resolved
and
if
that
gets
pushed
out
into
4
DNS
the
the
name
server
there
is
going
to
be
loopback,
so
k3s
will
I
mean
no,
nothing
would
work
with
blue
back
DNS
server
right.
So
probably
something
like
that,
so
we
just
are
great
for
information.
A
Yeah,
okay,
I'm,
not
gonna,
look
at
the
oh.
This
is
a
feature
request,
schedule,
name,
field,
support.
B
They
got
into
this
so
waiting
again,
I
think
we're
waiting
for
information
for
trying
to
be
to
be
able
to
reproduce
something
and
but
yeah
it
could
be
a
large
size
or
some
something.
It's
just
locked
up.
It's
just
wedge.
The
process
itself
is
West,
so
I
asked
if
the
web
server
was
working.
I
asked
if
there
was
any
logs
that
the
web
server
did
process
a
a
HTTP
request
from
the
same
part
or
from
outside,
but
I'm
still
waiting
for
information.
A
A
C
Yeah
the
healthy
on
on
the
on
the
control
plane.
It
is,
if
I
record
correctly,
it
is
from
engine
X
and
not
from
from
the
go.
The
go
modeling.
C
A
C
I
got
I
got
a
problem.
Is
this
running
on
kind
or
something
like
that
or
Kita?
Yes,
no,
that's
Cube
spray.
Are
there
any
solid
yeah,
so
I
face
it?
Something
similar
yesterday
running
the
end-to-end
tests
and-
and
it
was
some
cctl
needs,
I'm,
not
sure
if
something
changed,
if
this
is
like
a
new
version
or
not
I,
don't
think
so.
C
Yeah,
so
maybe
much
probably,
this
is
something
on
the
up
on
the
operational
system
and
not
on
on
engine
X
like
this
calls
to
the
liveliness
and
the
Readiness
they
are
made
by
cubelet
right,
they
are
not
made
by
by
us
by
any
Saints
and
and
in
China
x
should
be
closing
this
connection.
So
there
are
two
possibilities.
One
is
the
program
that
deals
with
this
health
check.
I
I,
don't
remember
where
this
is,
but
this
slash,
health
in
Gen
X
is
like
blocking
connections
or
it
needs
some.
B
No
see
the
people
who
are
using
thousands
of
Ingress
resources,
they
haven't
reported
running
out
of
sockets
in
the
pod,
so
it
is,
and
these
sockets
and
connections
not
closing.
My
first
suspicion
is
where
to
get
information
if
they
are
customizing
their
operating
system,
a
PCP
stack
windows
and
buffers
and
stuff.
C
Will
definitely
reflect
in
the
spot
and
yeah
yeah
I
I
agree
with
you.
Long
I
just
want
to
make
sure
that
that
there
is
no
program
in
our
side
getting
stuck
on
that
like
if
this
house
check
is
not
it's
not
blocking
right
if,
where
these
health
check
runs
because
I
don't
remember
where
it
runs.
C
B
B
A
B
Basically
he's
saying
that
he'll
open
a
TCP
board,
but
he
wants
Ingress
functionality.
Like
you
know,
then
we
should
he'll
have
10
parts
and
then
he
wants
the
connection
to
go
to
the
cluster
IP
type
service.
That
is
going
to
be
that
you
would
have
created
and
all
that
so
it
to
like
contradictory
to
the
whole
concept
of
opening
a
TCP
Port
that
directly
routes
to
the
back
end.
A
B
No,
that
is
a
confusing
point,
so
we
need
to
know
istio
well,
but
he's
his
his
requirement
from
English
controller
is
very
simple.
He
wants
English
Engineers
controller
to
open
PCP
port,
but
he
wants
to
route
that
connection,
so
the
back
end
can
that
when
the
port
is
open,
he
wants
the
traffic
to
go
to
the
service
and
the
end
points
that
istio
and
all
that
is
involved.
A
B
A
A
A
A
C
A
There's
a
lot
to
make
sure
it
is
an
interesting
problem
and
it
is
an
interesting
feature,
but
we'll
have
to
have
someone
test.
It
I'll
take
this
one
as
a
help
on
it
and
it
is
a
feature
request.
A
So
we
can
just
do
what
we
always
what
we've
been
doing.
Go.
C
B
Is
that
is
that
for
HTTP
and
https
traffic,
even
if
assuming
that
the
HTO
Gateway
is
bypassed,
really
is
using
the
HTTP
and
https
endpoint
from
English
nginx
controller?
But
would
it
be
able?
Would
it
be
really
possible
for
TCP
traffic
he's
already
saying
that
PCB
traffic
non
non-tls
works
and
that
is
to
assume
that
HTTP
and
https
traffic
within
history
integration
will
definitely
work
but
TCP
traffic.
So
example,
if
you
open
a
MySQL
or
something
like
that,
and.
C
Then
no
I
see
what
I'm
understanding
here
is
that
the
but
I
didn't
find
out
how
to
do
the
similar
TCP
traffic.
Okay,
oh
yeah,
I
missed
that
line,
so
it
works
with
HTTP
and
they
want
to
use
Ingress
to
Route
TCP
endpoints
right
yeah.
C
C
The
English
controller
would
know
base
it
on
this
DS
and
the
Sni.
You
expose
the
service
with
a
name,
and
you
say
that
that's
a
virtual
host
with
a
name
and
then
we
call
that
service
using
TCP.
You
should
say
that
you
are
using
TLS
and
because
of
the
header
of
the
Sni
saying
Hey
I
want
these
present
me.
The
certificate
of
this
thing
Ingress
knows
that
you
are
asking
to
be
routed
to
a
TCP
service
behind
the
TLs
with
that
name,
and
it
can
do
the
routing,
so
you
wouldn't
need.
C
So
the
thing
is
that
if
people
stopped
exposing
Services
via
TCP
and
UDP
on
Ingress,
without
using
TLS,
you
could
just
rely
on
all
of
the
layer,
7
stuff
and
not
on
that
config
map,
exposing
like
doing
what
from
two
from
ports
right,
I,
don't
know
if
this
is
the
case
and
and
honestly,
my
point
here
right
now
is
that
this
seems
like
an
interesting
case.
I
just
don't
have
time
to
test
it,
but
Sim
something
valid
at
least
to
test
and
see.
If
we
can
support
that
or
not
right,
yeah.
C
C
If,
if
you
want
to
take
a
look
on
that
one,
because
I
think
I
think
I
mean
I
think
this
is
something
that
can
be
used
in.
Let's
say
if
you
have
Sni
TLs,
but
I
want
to
understand
a
bit
more
instead
of
just
saying
hey,
we
don't
support
that
like
we
don't
support,
because
this
is
like
a
constraint
for
us.
This
is
a
constraint
for
engine
X
or
we
just
don't
have
time
or
our
architecture
doesn't
support
that
right
or
we
can
support,
but
we
need
to
implement
something
else.
C
A
C
C
C
A
Well
or
don't
no
I've
yeah
I've
seen
that
before,
where
it
doesn't
actually
remove
it.
Okay,
I'm
good,
all
right,
that's
time
for
the
the
triage
issue,
so
let's
go
ahead
and.
A
Go
ahead
and
talk
about
where
we're
at
with
things
so
the
the
3618,
that's
the!
Is
that
the
119
one
or
is
that
the
no?
That's
the
that's
the
nginx
upgrade
for
the
open,
Rusty
stuff,
so
we'll
so
graveyard
a
a
fogging
for
saying
he's,
gonna
be
here
and
then
not
no
kidding.
C
A
C
Yeah,
it
may
be
unstable
because
we
are
doing,
we
are
just
doing
compiling
using
the
latest
beats,
but
we
want
to
know
how
this
is
going
and
how
this
looks
like
for
you.
A
So
this
might
be
a
question
for
further
conversation,
because
I've
been
really
digging
into
our
nginx
build
for
the
last
week
and
a
half
and
I
don't
understand.
The
open
Rusty
component
is
that
the
lure
Rusty
core
that
needs
to
be
upgraded
because
there's
like
there's
22
dependencies
and
over
half
of
them
or
lower
code
dependencies.
C
A
C
A
This
is
why
patches
still
do
happen.
C
C
Yeah,
so
we
we
only
do
this
way
because
we
need
to
be
compliant
with
open
rest
right
or
open
the
rest
version.
So
I
wanted
more
to
do
that
thing
as
an
experiment
and
and
say
hey.
We
have
this
other
version
that
may
or
may
not
have
problem
may
or
may
not
be
stable,
and
we
want
you
folks
to
tell
us
yeah
look
at
this
man.
This.
A
C
Yeah
I
I
wanted
to
check
if
we
just
use
like
version
122,
we
still
need
all
of
those
patches
or
not
if
they
are
related
to
like
Gen
X
or
if
they
are
related
to
some
something
with
Lua
and
also,
as
you
see
as
you've
seen
that
we
have
22
dependencies,
the
majority
of
them
being
been
lower
as
well,
if
which
part
of
the
code
we
use
those
and
if
we
can
get
rid
of
them
and
just
keep
with
like
Lua,
recycle
and
some
other
car
stuff
and
and
start
removing
and
and
making
smaller
our
footprint
right.
C
So
I
I
think
that's
a
good
experience
actually
first
of
all
say:
hey.
We
have
this
Gen
X
version
122,
which
we
just
apply:
Lua,
resicor
and
other
stuff
and
you're
going
to
start
removing
things
from
this
compilation,
and
we
we
want
you
community
to
tell
us
if
this
is
making.
If
you
miss
something
here
or
not
right,
okay,.
A
A
C
No
I
was
I'm
tempted
even
in
this
new
compilation
in
this
new
build,
removing
like
mod
security
and
checking
like
if
mod
security
is
not
like
available.
Just
like
we
accept
the
configuration,
but
we
can
as
an
example
on
the
admission
web
hook,
say:
hey,
we
don't
have
much
security.
So
it's
like
we.
We
are
not
going
to
do
nothing
with
that.
Like.
B
C
B
What
I'm
hearing
from
what
Ricardo
is
saying-
and
this
is
kind
of
obvious
to
me-
also
my
opinion,
voice,
I'm,
saying
that
it
we,
if
we
cannot,
we
don't
have
a
choice
of
making
only
one
change,
because
I
think
there
is
it's
like
a.
They
are
like
a
mesh
Rusty,
plus
bass,
Plus
controller
plus.
So
they
have.
We
just.
A
C
Let's,
let's
Okay
I
I
got
I,
got
James
point
and
I
and
I
agree.
I
mean
in
this
case
I'm
I
have
this
anxiety
things.
It
is
anxiety
thing
of
like
I,
just
want
to
clean
up
everything
and
get
rid
of
everything,
and
then
everything
breaks.
C
A
C
B
Can
you
close
on
the
couple
of
beers
that
are
like
really
waiting
the
the
cube,
a
book
search
gen1,
which
one.
A
Oh,
the
new
web,
so
I'm,
okay,
so
I
know
which
one
you're
talking
about
long,
because
I
was
reading
it
yesterday
or
this
morning.
So
a
web
hook.
Surgeon
I
commented
on
it.
When
we
build
a
new
web
exertion,
it
doesn't
follow
the
nginx
tag.
So
it's
got
that
random
Sean
date,
I'm
fine
with
it
having
the
Sean
date,
because
we
can
change
those
images
faster
than
we
can
change
the
controller
image
tag.
B
Ricardo,
the
if
you
look
if
you
look
at
the
linter
golint
issue,
so
in
my
first
run
there
was
a
very
big
list
of
breaks
and
you
may,
and
you
suggested
I-
checked
what
what
they're
caused
by
right
the
litter.
But
then,
after
while
reading,
documentation,
there's
a
flag
there
to
check
by
release
or
tag
get
tag
and
when
I
check
for
the
latest
tag.
B
There
is
zero
lifting
errors
and
when
I
checked
for
one
three
zero
one,
three
zero,
there
are
six
linting
errors
and
all
of
them
are
related
to
the
to
the
difference
between
one
three
zero
and
one
three
one.
And
that
was
because
at
the
time
of
merging
that
the
that
PR
from
I
think
kamimos
I
don't
know
who
it
was
still
open.
B
B
C
Yeah
yeah,
so
one
one
of
the
things
that
I
wanted
to
fix
is
it's
like
past
things
right
so,
instead
of
passing
the
flags
checking
because
I
guess,
when
you
pass
the
flags,
it's
going
to
see
the
difference
between,
like
the
latest,
the
the
less
tax
right
so
between.
B
B
C
I
think
we
can
merge
the
the
the
the
CI
and
and
maybe
seeing
also
in
having
some
effort
to
correct
things
from
the
past
right
so
instead
of
like
just
looking
into
the
future,
if
we
have
something
from
the
past
as
an
example,
one
one
of
the
things
that
go
along
Sea
Island
takes
it's,
we
are
using
the
math
random,
which
is
kind
of
deprecated,
I,
guess
to
use
crypto
random
or
something
like
that
right.
So
we
can
we
we
can
look
into.
We
can
look
to
that
and
start
doing
those
fixes.
C
From
long
the
doubt
heard
on
yeah
this
one:
no
that's
not.
C
A
A
A
B
B
So
regard
if
you're
saying
this,
then
then
it's
going
to
be
where
to
cover
all
the
errors.
If
you,
then
we
can't
use
flag
and
just
only
check
the
latest
tag,
so
I
think
you
have
to
fix
all
the
errors,
or
at
least
most
of
them
yeah
before.
C
I
mean
no,
we
can
we
can
merge,
we
can
you
can
open
up
your
to
the
CI
lint
right.
It's
it's
fine
and
we
can
add
that
already
to
have
this
working
in
in
the
future.
I
think
it's
it's
it's
fine,
but
also
I
I.
Don't
want
to
close
that
this
issue
instead
use
it
to
fix.
Also
what
we
find
on
the
past
and.
B
C
C
C
B
C
I
was
just
gonna
say
that
this
other
PR
that
James
sent
to
me
it's
something
that
I've
asked
actually
to
add
other
kubernetes
tests
into
Helm
shark
test,
because
I
have
removed
pod
security
policy,
and
now
this
is
gonna.
We
are
gonna,
have
more
tests
than
cold,
but
I
think
we
need
this
team.
B
But
Ricardo
so
on
the
same
thought
process.
You're
talking
this
this
we
are
talking
about
same
thought,
process
and
think
about
the
other
problem
right,
the
other
other
day
I
mentioned.
So,
if
you're
talking
about
linking
problems,
there
is
more
or
higher
priority
and
really
critical
security
problems
that
are
not
at
all
caught
from
image
Scan.
They
are
all
from
the
deprecated
or
insecure
versions
of
packages
that
are
including
in
go
mode.
So.
B
Westlake
reports
very
clearly
shows
there's
so
much
that
you
know
it
is
like
three
four
pages,
so
there's
so
much
in
go
mod
that
depend
about
must
be
catching
some,
but
not
or
there's
too
much
left
there.
Okay,
not
sure
the.
C
Thing
is
that
someone
I,
think
Zhong
and
he
is,
is
working
on
the
Google
check
right,
I've,
seen
that
on
on
the
channel
yeah
before
before,
assuming
that
something
that
that
we
have
a
lot
of
vulnerabilities
I
want
to
I
want
actually
to
see
the
result
on
that,
because
that's
the
that's
the
official
way
of
goal
right
now
of
go
Community
right
now
to
say
what
we
do
have
vulnerable
or
not
like
if
our
code
actually
touches
the
part,
the
vulnerable
part
of
the
code
or
not
right
right,
so
I,
I,
I,
I,
I
got
to
and
and
I
hear
you
along
on
that
and
I
and
I.
C
I
am
like
my
biggest
concern
right
now
with
all
of
these,
and
all
of
the
things
that
we
are
doing
is
security
as
well
right,
I,
I
just
wanted
to
make
sure
that
sorry
I
just
wanted
to
make
sure
that
we
are
not
changing
too
many
things
too
fast
on
this
part
of
vulnerability
check
and
we
end
up
having
more
more
false
positives
that
real
alerts,
and
then
we
are
just
in
this
mood
of
like
yeah,
whatever
Let's
ignore
it
right.
So,
let's,
let's
let's
work.
A
That'll,
be
the
other
nice
thing
too,
is
that
if,
if
things,
if
things
do
come
across
and
Global
says
that
our
code
doesn't
touch
it,
even
though
it's
there
that
does
I
hate,
saying
it,
but
being
the
question
of
why
we're
using
it
and
then
we
should
be
able
to
remove
it
right.
If
that's
the
idea
about
go,
vol
yeah.
B
Yeah
I
I
agree,
James,
that's
correct
way,
but
then,
if
we
had
to
close
this,
then
this
issue
that
you
created
with
the
go
vulnerable
use
case
it
has
to
we
have
to
like
you-
should
have
it
sooner
than
later.
Yeah
do
you
do?
Are
you
thinking
of
targeting
it
for
the
next
four
point,
X
release,
or
do
you
want
to
go
back
yeah?
It's.
A
C
So
if
we
have
some
real
some
some
arterio,
but
some
critical
concern
on
some
gold
model,
you
just
just
just
tell
me
just
send
those
to
me
and
and
and
and
I.
B
C
Can
I
can
I
can
take
care
of
at
least
like
the
the
mark,
critical
ones,
I.
A
C
C
Yeah,
but
there
was
like
there
are
some
some
things
like
you
need
to
add
labels
and
other
things
and
they're.
Also
so
much.
C
A
C
A
That's
I
like
that.
Look!
Okay!
So
long
after
all
that
that
was
an
action
item
for
you.
So
I've
got
some
action
items
last
few
minutes.
How
are
we
doing
on
the
cpdp
split.
C
Great,
so
we
be.
C
We
are
getting
crazy,
no
I'm
just
kidding,
so
the
pr
is
getting
huge
and
and
and
I
have
some
some
things
that
I'm
just
leaving
as
a
to-do,
so
I
will
just
go
and
do
a
self
review
on
myself
to
see
things
that
I'm
missing
right
right
now.
I
am
writing
the
end-to-end
test,
I'm,
just
changing
the
end-to-end
test,
Suite
just
to
support
the
data
playing.
It
already
have
the
flag
and
it
already
deploys
the
data
plane.
C
For
some
reason:
it's
not
it's
not
connecting
the
data
plane
to
the
control
plane
right,
so
I
need
to
debug
that
but
other
than
that,
the!
If,
if
you
take
a
look
into
the
the
pr
you're
gonna
say,
that's
like
a
massive
huge
and
and
and
it's
it's
let's
go
in
slower
than
I
wanted,
but
but
it's
it's
going!
C
Yeah
yeah
yeah
yeah
yeah,
it's
there,
it's
a
take,
a
look.
There
is
like
one
from
me,
approve
it
but
EU
here
you
go,
you
pass
it
or
you
can
find
me
yeah.
There
is
something
else
yeah.
So
after
I
I
can
make
the
end-to-end
test
connect.
I
still
have
two
major
items
that
I
need
to
figure
out.
One
is
with
Ismail.
C
We
need
to
look
into
the
wire
communication
right,
so
retries
all
of
the
grpc
connection,
things
that
we
need
to
to
to
figure
out
like
the
retrial,
balancer
whatever
and
I,
need
to
make
sure
that
the
end-to-end
test
day
passes
and
if
something
doesn't
pass,
I
need
to
see
if
it's
like
a
fix
that
I
need
to
make
on
the
code
or
if
this
is
something
that
we
shouldn't
be
supporting
anymore
because
of
control,
plane
and
data
plane.
So
those
are
the
major
things
right
now.
C
B
C
A
Please
now
it's
it's
Halloween
season,
it's
not
Christmas,
awesome,
okay
and
then,
just
last
but
not
least,
I've
been
investigating
for
the
last
couple
of
weeks
week
and
a
half
on
how
to
use
apko
and
melange
to
rebuild
all
of
our
container
images
so
that
we
can
reduce
our
compilation
times.
I
was
able
to
it,
took
about
four
hours
to
produce
all
of
the
apks
for
all
of
our
dependencies.
But
then
it
takes
five
minutes
to
rebuild
our
containers
every
single
time,
because
we
don't
change
our
dependencies.
A
All
that
often
so
anytime,
we
need
to
do
a
complete.
Refresh
rebuild
still
takes
the
same
amount
of
time,
but
every
new
control,
Ingress
controller
version.
It
shouldn't
take
us
that
long,
because
we
don't
have
to
rebuild
the
nginx
image
every
single
time,
then
so
it's
gonna
be
pretty
awesome.
I.
Think
nice
once
I
get
that
working
for
those
I
can
start
working
on
the
other
images
that
we
have
and
then
we
won't
have
any
Docker
files,
and
we
can
programmatically
do
this.
A
A
Yeah
and
they'll
come
with
s-bombs
and
all
the
other
fun
Jazz
that
we
can
use
I
know
we
get
a
finalist
bomb
from
the
staging
to
production,
build
and
those
containers
get
signed.
We
also
don't
put
any
of
that
information
in
the
release,
notes,
and
that's
a
note
for
me
to
add
that
anyway.
So
that's
where
I'm
at
once
I
get
that
released.
I'll
do
probably
do
the
same
thing
that
we
did
with
the
V1
and
what
we're
going
to
do
with
cpdp,
just
put
out
a
Beta
release
and
then
eventually
move
to
those.
C
A
C
A
Yeah
all
right,
yeah
I,
will
drop
a
note
tomorrow
when
I
start
the
release.
If
people
want
to
follow
along
and
we'll
go
from
there.
C
C
Security
policy,
yeah
kubernetes
125,
doesn't
support
that
anymore
right.