►
From YouTube: Cloud Foundry for Kubernetes SIG [May 2020]
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
B
B
A
A
A
B
B
B
B
A
B
D
E
A
D
C
E
D
D
D
D
D
A
B
D
B
A
Yeah
officially
welcome
to
this
week's
Cloud
Foundry,
uncoordinated
special
interest
school
calm.
I
try
to
follow
my
own
promises
and
kind
of
set
out
the
ball
a
little
bit
earlier
and
close
it
a
little
bit
earlier.
I
think
I
can
even
do
better
next
time,
but
for
this
time
at
least,
we
kind
of
gave
folks
a
little
bit
of
a
heads
up.
So
we
have
Zach,
Angela
and
and
folks
from
the
VMware
sites,
and
the
topic
for
for
this
time
is
talking
about
series
in
cloud
foundry
on
communities.
A
C
Don't
think
anyone
prepared
any
specific
presentation,
we
just
know
some
materials
been
going
out
and
Eric
sure
that
this
topic
was
gonna,
be
here
so
figured
we'd,
sweet,
hop
in
and
could
ask
questions
and
Angela
and
Connor
did
a
lot
of
the
sort
of
exploration.
So
they
may
have
some
some
drawings
here
in
there
they
can
pull
up
if
any
of
the
questions
sort
of
head
in
that
direction.
Otherwise
we
can
just
sort
of
chat,
and
if
we
need
some
more
visuals,
then
we
can.
We
can
come
back
with
more
later.
C
Yeah
I
can
maybe
do
a
quick
sort
of
high
level
of
at
least
some.
You
know
what
spreads
some
of
this
stuff.
You
know
we're,
obviously
we're
removing
cf2,
Cates
kind
of
had
a
number
of
ways
we
we
have
VRA.
We
have
this.
You
know
our
that
the
C
uppercase
effort
we
have.
We
have
the
cube,
CF
effort
and
they're
both
putting
and
moving
and
running
components
on
the
Cates
and
and
in
parallel
to
that
they're
ripping
out.
C
You
know
we're
finding
that
as
we
go,
it's
becoming
really
natural
to
start
wanting
to
either
consume
existing
CR
DS
in
the
ecosystem.
So
this
this
is
how
a
lot
of
tools
are
represented
to
us,
for
example,
Asda,
which
is
which
is
taking
over
the
the
routing
tier
essentially
and
as
we
build
out
new
functionality.
So
even
within
the
routing
org
they're.
C
Finding
that
you
know
it
would
be
nice
for
them
to
have
a
route,
see
Rd
that
look
like
a
mirror
of
essentially
the
CF
API,
and
allows
us
to
declare
that
state
Indic
indicates
and
essentially
then,
under
the
hood.
You
know
that
their
their
operator
can
turn
that
into
sto,
based
routing
and
theoretically
in
the
future,
maybe
that
that
could
create
a
point
for
plug
ability
of
different
types
of
routing
technology,
but
there's
other
things
too.
There's
lots
of
different
integrations.
C
You
know
we
look
at
some
of
the
work
that
serves
as
teams
in
in
vmware
are
doing
and
they
want
to
be
able
to.
They
want
to
know
about
different
parts
of
Cloud
Foundry.
They
want
to
know
about
orgs
and
spaces,
so
they
can
set
up
different
tenancy
constructs
in
their
integrations.
So
sometimes
it's
you
know
it's
not
always
core.
C
Let's
see
what
some
of
the
really
important
outcomes
that
people
are
doing,
we
know
there's
you
know
when
you
just
say:
extensibility,
there's
probably
gazillion
things
people
can
do,
but
can
we
maybe
sort
of
see
what
are
some
the
key
ones?
People
are
trying
to
do
now
and
then
does
that
help
us
understand
you
know:
is
there
a
more
deliberate
way
to
integrate
those
things
into
the
system?
Are
there
certain
CDs?
We
can
build
in
app
and
route
CDs
that
we
know
will
be
really
valuable
integration
points
for
people
and
will
simplify
our
implementation.
C
You
know:
do
we
need
to
think
about
orgs
and
spaces
in
various
different
ways,
for
integrators
to
be
able
to
vote
against
CF
and
have
the
opportunity
to
really
extend
things
and,
as
we
do
build
those
in
are
there
other
patterns
we
should.
We
should
follow
that
we
should
try
and
get
everyone
in
the
ecosystem
to
follow
so
for
voting
and
routing
charities
for
building
and
absurd
ease.
If
we're
building
in
you
know
whatever
other
service,
C
or
DS
or
something
alright
did
it.
C
B
Question
are
you
like,
as
you
mentioned,
service
series?
Are
you
thinking
about
like
looking
in
things
like
Service
Catalog
like
there
is
already
an
offer,
like
a
controller
based
always
be
platform
in
the
Kuwaitis
ecosystem,
yeah.
C
I
was
kind
of
just
listing
things
that
exist
in
in
the
API
that
just
sort
of
a
demonstration,
but
I
do
believe
that
that
some
of
the
teams
that
are
thinking
about
the
civics
experience
you
know
they're.
What
we're
finding
is
lots
of
folks
are
out
there
doing
sort
of
disparate
explorations
and
I.
Think
some
of
the
folks
responsible
for
the
service
experience.
Experience
have
been
looking
at.
You
know.
What
might
it
look
to
like
to
put
the
service
catalog
behind
the
CF
API
I?
C
B
Yeah
hi
hi
everyone-
this
is.
This
is
my
first
time
on
this
cool,
so
nice
to
see
some
some
familiar
faces
and
some
new
ones
mostly
my
face
well,
I
know
what
your
face
looks
like
so:
yeah
I
am
involved
in
some
of
the
services
work,
sort
of
out
of
the
London
office
or
out
of
London
at
the
moment,
and
yeah
Service
Catalog
is
something
that
we're
obviously
very
interested
in.
B
We've
got
a
few
deliberate
efforts
that
are
sort
of
kicking
off
at
the
moment
to
look
at
what
might
it
look
like
to
integrate
Service
Catalog
more
directly
into
CFO
gates,
but
we've
also
actually
just
been
using
it
sort
of
outside
of
Cloud
Foundry
for
some
other
purposes
that
we're
looking
at
at
the
moment.
So
we're
pretty
pretty
involved
with
it.
B
C
And
then
so
so
folks,
you
know,
have
read
the
docs
and
have
questions.
That's
great.
If
folks
are
going
to
read
the
docs,
you
know
I
can
sort
of
there's
a
couple
at
least
the
two
that
you
know:
I
worked
with
Angela
and
Connor
on
a
little
bits
are
the
exploring
series
in
the
CF
at
run-time
doc.
You
see
that
link
and
that
one's
really
sort
of
exploratory
we
put
down
some
of
the
things.
We
know
that
we're
aware
of
an
ecosystem
that
people
are
hoping
to
achieve.
C
It's
not
intended
to
be
a
proposal
of
like
hey.
We
should
do
this,
it's
just
it's
more
like
what
have
we
heard
right?
There
was
a
thread
that
that
came
up
and
you
know
it
was
like
hey.
We
want
to
declare
it
an
interface
for
creating
organs
and
spaces
right,
and
so
that
that
was
a
use
case,
and
then
you
know
I
shared.
C
So
that's
a
great
place
to
have
that
conversation
and
then
there's
a
follow
up,
one,
the
CF
app
runtime
Syria
proaches,
where
we
stepped
back
and
said:
okay,
given
the
goals
were
aware
of,
and
we
think
those
will
grow
as
people
you
know,
maybe
after
the
stock
or
whatever
it'll
grow
our
understanding.
But
given
what
we
know
now,
what
are
some
different
ways?
We
could
sort
of
architect
the
whole
systems,
and
this
gets
more
into
the
technical
bits
and
lesson
to
it.
Sort
of
the
these
are
facing
bits
of.
C
G
D
A
I
will,
while
reading
through
doc,
I,
actually
found
it
interesting
that
you
have
this
notion
of
like.
Could
we
have
something
like
a
manifest
C
or
D
in
there,
which
I
think
is
kind
of
clearly
user-facing
in,
like
the
terminology
that
the
Julie
said,
I
didn't
actually
think
about
those
like
when
I
thought
about
series
in
the
context
of
tree
I
need
more
thought
about
like
orgs
and
spaces,
and
probably
also
roots,
and
then
I.
Think
like
this,
the
special
one
which
is
like
the
app
there
I
left
a
comment
in
the
doc.
A
Basically
asking
I
can
imagine
like
how
an
app
would
get
like
certain
properties
and
associations
to
other
objects.
But
how
would
you
envision
like
the
app
it's
actually
being
transferred
for
to
to
where
they
actually
need
to
end
up
in
in
order
for
k2
to
do
its
job?
Would
that
be
like
a
large
finery
field
inside
the
C
or
D?
Or
is
it
a
reference
to
something
else,
or
how
would
you
envision
that.
C
Yeah
I,
don't
think,
we've
gotten
as
far
as
you
know,
scoping
out
exactly
what
an
app
sort
of
C
or
D
or
manifests
here
D
would
look
like,
although
I'd
like
us
to
start
getting
there
soon.
You
know
at
a
high
level,
we
don't
include
the
upload
when
it's
orchestrated
via
the
CLI.
It's
a
number
of
steps
in
the
CLI
manages
and
upload,
and
then
it
then
it
puts
some
app
resources
into
the
API.
Imagine
we
could
do
something
similar
with
Kate's,
where
we
could
have.
C
You
know
the
CFC
Li
manage
getting
the
bits
somewhere
available
and
then,
whenever
it
creates
an
app
C
or
D
it
would
it
would
create
it
with
a
pointer
to
it
wherever
those
application
bits
are
so
I
imagined
be
pretty
similar.
Talking
to
you
today,
I
think
the
discussion
of
user
facing
versus
infrastructure
is
is
really
interesting
and
you
know
I
think
one
thing.
We've
talked
about
a
little
bit
but
we'd
like
to
at
least
be
able
to
explore.
C
You
know
CF
proper,
that
we
bring
together
and
you
go
through
the
CC
api
and
get
them
get
the
full
experience.
So
that's
another
area
we've
been
sort
of
thinking
about
as
well
like.
How
would
we
break
these
down?
It
probably
had
some
complications
where
you
know
some
cds
might,
in
some
cases
be
a
direct
user
interface
but
in
other
cases,
they're
intended
to
be
sort
of
managed
by
a
higher
controller
tier
and
at
least
at
first.
So
that's
one
angle,
we've
at
least
been
sort
of
exploring
and
it
comes
up,
is
really
sorry.
F
So
we
have
a
kind
of
kubernetes
team
here
at
Sousa
and
one
of
the
architects
there
has
been
working
on
on
a
spec
for
organizations
and
spaces
based
on
the
Cloud,
Foundry
notions
of
organizations
and
spaces
and
obviously
they're
separate,
but
they
are
meant
to
do
exactly
what
orgs
and
spaces
do
for
Cloud
Foundry.
But
he
wrote
that
proposal
just
from
a
kubernetes
point
of
view,
so
they're
completely
reusable,
regardless
of
Cloud,
Foundry
and
I-
think
he's
on
vacation
now,
but
I
pinged
him
to
present
his
proposal
here
in
or
sake
at
some
point.
C
Yeah
that
would
be
super
awesome
to
hear
about
that
and
think
about
that.
One
thing:
the
team
looked
into
as
they're
exploring
you
know,
thinking
about
orgs
and
spaces
is
really
interesting
because
they're
they're,
yes,
a
core
component
of
Cloud
Foundry,
but
to
that
they're
they
are
this
integration
point,
and
so
we
found
that,
for
example,
some
of
our
services
teams
in
lieu
of
orgs
and
spaces
existing
are
creating
sort
of
shadow
series.
C
So,
like
this
concept
of
at
least
knowing
about
them,
is
pretty
interesting,
you
know,
and
then
you
extend
it
and
there's
this
amount
of
okay.
Well,
if
you
are
doing
that
and
we
do
have
CFC
artis,
you
know
how
do
we
keep
these
things
in
sync
and
I
finally
see
early
MOT.
The
organ
space
CID
model
is
a
really
interesting
place
where
it
starts
showing
where
we
need
to
make
some
architectural
decisions.
C
You
can
imagine
today
really
quickly.
You
could
come
up
with
an
organ
space,
er
D
and
an
operator
that's
interacted
with
the
existing
CF
API
and
reconciled
those
two
things
together.
So
users
might
create
organ
space
CDs
and
then
we
can
make
those
true
in
various
foundations
and
then
that
will
arrange
sort
of
looks
like
C
or
D
CF.
C
But
now
we've
got
this
two
really
different
ways
of
interacting
with
C
or
DS,
and
so
try
and
understand
like
hey,
actually
can
we
and
we
have
a
little
more
consistency,
even
if,
in
the
long
run,
that
might
take
a
little
longer
to
get
those
things
together
and
I?
Think
that's
what
we're
trying
to
learn
a
fair
bit
about
in
some
of
these
explorations.
F
Looking
at
this
proposal,
because
it
wasn't
at
all
biased
by
cloth,
foundry
right
I
mean
the
concepts
were
come
foundry
based,
but
not
biased
about
how
cloud
foundry
works
at
all.
So
it's
interesting
that
you
could
set
up
organizations
and
spaces
in
kubernetes
and
have
them
be
very,
very
useful
without
cloud
foundry,
so
it's
pretty
I
love!
You
can
come
in
here.
Maybe
this.
F
E
B
B
C
Yeah
but
that's
a
great
topic
and
another
one
that
gets
really
interesting
to
dig
into.
Is
this
mapping
of
orgs
and
spaces
to
namespaces,
and
you
know:
there's
there's
these
figuring
out
how
to
do
it
even
with
all
they
have
now
and
then
there's
you
know,
I
I
think
we
have
hopes
and
ambitions
that
we
can
have
sort
of
a
multi
cluster
representation
of
apps
on
Cloud
Foundry
at
some
point
in
the
future.
And
how
does
that
play
in
and
is
that
we're
thinking
about
that?
D
Ireenie
is
spiking
what
it
would
look
like
if
you
read
me
if
the
interface
between
Kathy
and
arena
versus
CID
that
that's
really
a
kind
of
architectural
CID,
if
you,
if
you
like
in
those
two
things
just
it
kinda
like
the
route
CID,
it
would
be
a
lot
easier
for
the
integration.
If,
instead
of
the
convergence
loop
being
in
CC,
we
have
to
PR
any
have
to
kind
of
figure.
All
that
stuff
out
I
mean
there's
a
convergence
sequence.
Ece
Cuban
SES
is
really
good
at
convergence
loops,
it's
basically
a
big
converge,
asleep.
D
Let's
move
the
converging
sequence,
so
that's
more
of
a
kind
of
architectural
thing,
but
as
Zack
says,
if
we
had
a
CID
there,
then
maybe
that
actually
does
kind
of
give
the
opportunity
for
people
to
build
other
things
on
top
of
it.
Maybe
it
makes
it
easier
people
to
extend
the
system
without
PR
ain't
was
maybe
there
are
some
things
that
come
out
of
that
anyway.
Maybe
there's
a
user
facing
that
comes
out,
I
think,
but
they
yes,
yes,
but
I
certainly
wouldn't
see.
Our
thing
is
trying
to
expose
kind
of
user
facing
thing.
D
Oh,
yes,
we
have
a
story.
What
we're
we're
thinking
just
just
basically
and
we're
gonna
just
support
in
the
Iranians
face
you
telling
us
what
the
namespace
should
be.
So
we're
not
going
to
worry
about
what
the
right
mapping
is,
but
then
Cathy
will
be
able
to
set
that
mapping
correctly.
So
if
it
decides
there
should
be
a
different,
namespace
Olga.
Definitely
the
space
Cappy
could
then
decide
that,
but
we
have
the
arena
side-story
it's
about
four
or
five
down
by
cog.
I
think
this
is
pretty
small
story
for
us.
Yeah
is.
C
I
think
the
closest
we
have.
Let
me
share
that
there's
we
shared
something
out-
probably
I,
don't
know
six
to
nine
months
ago
on
the
mailing
list,
around
a
Multi
cluster
experience,
we'd
like
to
expose
to
applications,
and
so,
if
you've
had
a
chance
to
read
s,
claw
I'll
link
it
here
again
in
the
chat,
real
quick,
but
to
be
clear,
it
doesn't
get
into
things
like
architecture
and
things
like
like
the
how
it's
more
of
the
you
know.
C
What
is
the
experience
we'd
like
to
provide
so
from
that
angle,
we're
sort
of
thinking
about
it.
We
don't
talk
about
Isis
eggs,
specifically
we're
more
focused
on
just
the
high-level
concept
of
there
are
apps.
There
are
clusters
we'd
like
to
have
these
experiences
and
get
them
validated
with
the
community,
so
I'll
link
that
here
and
you
know
that
can
have
at
least
that
part
of
the
conversation,
but
as
far
as
the
implementation
part
of
that,
but
something
I'm
really
eager
to
get
to.
C
But
my
thinking
right
now
is
for
one
view
we
have
to
get
to
some
GA
of
scheduling
any
apps
on
a
cluster
before
he
can
do.
Multi
cluster
and
then
to
we're
really
trying
to
figure
out
how
intertwined
this
idea
of
CR
DS
and
our
ability
to
distribute
things
on
multiple
clusters
really
are.
It
might
be
a
lot
easier
to
do.
Multi
cluster
work.
If
we
have
some
base
layers
here,
these
that
represent
our
application,
it
might
not
so
so
I'm
really
eager
to
get
to
multi
clusters.
C
A
A
Speed
just
quote
unquote,
pure
kubernetes
based
workload,
so
that,
like
folks
that
are
using
the
CF
API
to
deploy
their
apps
could
still
interact
with
other
kubernetes
stuff
on
the
same
cluster
and
and
like
not
be
able
to
to
actually
interfere
with
Cloud
Foundry
control
claim
components
because
they
are
running
on
on
it's
completely
different
cluster
elsewhere.
C
Thanks
I
think
it's
definitely
a
use
case
that
that
I've
been
hearing
from
you
know
from
folks
in
that
dock
I
think
I've
seen
that
comment
and
then
I
think
even
you
know:
VMware
customers
also
that's
one
of
the
first
things,
I
hear
from
them,
and
so
I
meant
about
the
right,
I.
Think
in
my
head.
That's
somewhat
related
to
the
multi
cluster
work,
but
it's
nothing
we're
thinking
about
you
know,
maybe
maybe
there's
a
fast
path
to
some
of
the
different
use
cases.
I.
A
Think
I'm
bringing
it
up
here
because
one
thought
that
we
had
is
like
what
is
the
minimum
set
of
things
that
you
need
to
have
on
that
remote
cluster.
So
to
speak
like
apart
from
the
extra
lap
and
if,
let's
say
the
the
C
or
DS
need
to
live
on
that
very
same
cluster
than
you're,
probably
limiting
yourself
to
like
certain
deployment
scenarios,
whereas
if
the
OSI
or
DS
would
be
called
remote
enabled
or
could
kind
of
live
on
a
different
cluster,
then
I
think
you
would
have
less
less
restrictions
but
complicate
things
as
well.
A
F
I'm
not
sure,
if
you're
interested
in
cube
CF
with
Diego,
but
that's
exactly
the
use
case
of
that
spike.
That
I
was
telling
you
on
about
with
a
control
plane
living
in
one
cluster
and
then
cells
living
in
in
the
other
cluster
I
mean
if
you
could
take
a
look
like.
If
you
have
thoughts.
That
would
be
very
useful
because
it's
like
you're
you're
talking
about
the
exact
same.
A
And
then
probably
need
the
other
thing
that
that
I
found
very
interesting
reading
through
the
docks
was
a
discussion
around
like.
Where
does
the
persistency
actually
live?
Like
cloud
controller
database
versus
etcd
versus
sinking,
the
true
versus
like
one
one,
API
path,
kind
of
owning
an
object
and
the
other
one
just
allowing
to
do
to
read
stuff?
So
I
found
it
interesting
that
you
kind
of
came
up
with
those
various
different
like
layouts
and
persistency
options.
C
Yeah
at
that
point
and
I
think
you
know
it
helps
to
sort
of
anchor
those
or
ground
them
on
some
some
very
specific
use
cases.
You
know
I
think
at
least
what
stood
out
to
me
so
far
is
the
most
interesting
sort
of
decisions
we'll
have
to
make.
Are
you
know
we
really
talked
about
that
org
space
idea
and,
like
you
know,
maybe
there's
a
quick
path
if
we
layered
it
above,
but
then,
if
long-term,
we
want
it
sort
of
below,
you
know,
maybe
there's
a
way
to
migrate,
some
on
the
other.
C
We
think
that's
a
relation
case
and
something
that
we
have
to
like.
Try
and
figure
out
the
right
thing.
The
other
one
is
this
idea
that
you
know
III,
think
our
assumption
is
that
we
see.
Are
these
a
lot
of
the
benefits
of
having
that
that
cait's
based
representation
is
that
we
can
integrate
with
other
caves
based
ecosystem
tooling
a
lot
more
easily,
because
that
that's
the
API
level
that
they
expect
to
interact
with
right
and
so
a
real
simple
use
case
might
be
auto-scaling.
C
And
so
now,
we've
got
this
this
this
out
of
sync
situation,
where
what
we
actually
want
is
what
the
autoscaler
told
us
right.
We
want
that
change
that
no
user
told
us
that,
through
the
front
door-
and
so
now
we
have
this
source
of
truth
issue
of
is
what's
in
the
case,
the
source
of
truth
or
is
what
we
had
in
CCTV
source
of
truth,
and
how
do
we
reconcile
those?
C
We
want
people
to
really
be
able
to
extend
and
take
advantage
of
all
that
and
have
that
work
with
their
their
Cloud
Foundry
based
things,
and
so,
but
it's
really
good
said
a
small
use
case
to
see
like
why
that
that
that
source
of
truth
location
becomes
really
important
to
get
right
and
really
important
to
have
the
right
sort
of
patterns
in
place.
So
that
you
know.
Maybe
we
want
that.
True
for
our
absolutism
want
that
true
for
our
outsi
or
needs
as
well.
B
C
D
D
E
Been
we've
been
looking
to
drive
out
as,
like
you
know
what
I
think,
there's
broad
consensus
that
we
want
to
drive
this
domain
forward.
But
what
things
are
you
know
going
to
be
maybe
uncontroversial
and
obviously
the
right
decision
to
proceed
with
and
what
things
are
looking
riskier,
either
from
a
technical
perspective
or
from
an
interface
perspective
that
we
could
make
some
more
reversible
decisions
or
experiments
around
as
we're
trying
these
things
out
so
I
think
yeah.
C
You're
right
I
mean
I
think
as
far
as
we're
going
is
forward
we're
starting
to
explore
right
now
right.
We
have
we've
done
as
far
as
let's,
let's
get
a
doc
and
share
with
the
community
around
use
cases
we're
aware
of,
and
then
you
know,
hopefully
the
you
know,
vendors
and
different
folks,
cancer,
chime
in
with
with
where
that's
this
gonna
go,
where
the
integrators
can
chime
in
and
so
we're
at
the
stage
of
trying
to
gather
that
feedback
is
essential.
A
I
mean
probably
related
to
that
I
think
one
one
thought
that
we
had
early
on
in,
like
the
whole
conversations,
was
like
how
to
actually
get
more
traction
from
the
kubernetes
side
of
the
house
store,
but
like
I
was
wondering
if
some
of
them
could
some
of
that
be
validated
by
folks.
That
probably
don't
have
to
see
if
background
I
don't
have
a
good
suggestion.
A
E
Think
another
attack
on
that
B
might
be
like,
with
a
at
least
in
some
form
of
utility
within
the
kubernetes
community.
More
broadly
I
think
we
we
don't
have
anyone
from
networking
on,
but
that
I
think
is
something
that
I
recall
is
coming
up
for
discussion
among
the
route
C
or
D
is
like.
Could
this
be
something
that
would
be
maybe
a
more
convenient
interface
even
over
things
like
ingress
or
risk
you
virtual
services
for
or
maybe
a
better
multi-tenant
interface
for
routing
to
a
variety
of
workloads
on
the
Cates
cluster?
E
Not
just
CF
apps,
it's
not
having
it
tightly
coupled
to
that.
So
that
might
be
a
future
topic
of
discussion
that
we
could
solicit
as
a
deeper
dive
into
the
routing
series
and
how
they
might
serve
not
just
as
a
representation
of
that
CF
resource.
But
if
there
are
lots
around
there,
greater
extensibility,
yeah.
G
I
know,
like
speaking
with
folks
who've
been
working
on
the
route.
C
idea.
Definitely
I
think
one
of
the
benefits
of
having
it
is
idea
that
you
can
essentially
swap
out
the
backend,
and
so
it
becomes
like
a
more
generic
interface
and
then
right
now
you
about
see
Rd.
The
like
controller
actuate
in
Ghana
is
creating
sto
resources,
but
like
in
the
future,
you
could
like
consider.
G
D
Actually,
talent
down
that
line,
I
still
have
at
the
back
of
my
head.
That
I'd
be
pretty
cool
at
some
point.
If
it
reenie
was
creating
Kane
ativ
resources,
rather
than
deployments
like
it
seems
like
such
a
such
an
obvious,
potentially
possible
thing
to
do
at
some
point:
I,
don't
think
it's
quite
there.
Yet
there's
I
think
I.
Think
at
the
moment
the
kind
of
K
native
cold
start
is
a
bit
prohibitive,
but
it's
certainly
something
you'd
want
to
be
able
to
do
at
some
point.
I
think.
B
D
Think
well
about
the
sin
issue
right
now.
I
think
I
think
we
plan
to
deprecated
instance
index
anyway.
We
yeah,
like
I
think
I
think,
to
really
take
k.
Nature
surfing
right
now,
if
you
put
K
major
at
the
bottom,
wouldn't
really
buy
you
that
much.
Are
you
a
little
bit
because
it
means
that
the
CRTs
we're
creating
were
like
addressable,
so
it
integrate
with
the
venting
nice
of
it
may
be
a
cut,
but
it
wouldn't
buy
you
that
much
what
it
would
buy.
You
potentially
is
service
as
in.
D
If
we
we
could
then
expose
the
user
facing
feature
of
not
requiring
you
to
set
instance
counts,
and
they
said
if
you
could
have
sort
of
like
an
auto
scale
flag
or
something
on
the
CF
push,
and
you
get
that
almost
for
free.
If
you
could
swap
that
back-end,
if
you
see
whatever
you
know,
there's
maybe
some
stuff
down
that
line
where
I
don't
think.
D
I'd
want
to
put
cane
a
screen
just
book
K
anything
because
actually
like
pretty
damn
complicated
and
you
pay
you
don't
it's
not
free
in
terms
of
the
kind
of
auto
scaling
you
there's
some
significant
cold
start
time
and
some
of
that
stuff.
But
if
you're
exposing
some
features
from
it,
it
could
be
pretty
cool.
C
Yeah
I
think
that's
definitely
something
you
know.
We've
chatted
about
is
this.
You
know
what
is
the
right
layer
of
abstraction
right
and,
and
we
have
this
nice
in
the
CF
Road.
What
we
think
is
nice
in
terms
of
an
app
and
a
process
layer
and
under
the
hood
you
can.
You
can
and
auto-scaling
to
those
things
and
but
the
other.
C
Unless
that
happens
then,
and
they
maybe
they
have
that
chance
to
sort
of
be
able
to
understand
users
workloads
in
a
way
that
gives
them
one
definition
and
we
can
generate
either
kid
native
I
see
air
to
use
to
get
serverless
or
you
can
create
stateful
sets
to
get
things
that
need
volume
mounts
you
know.
Maybe
we
have
the
chance
to
be
that
layer
that
can
help
decide
the
right
underlying
thing
to
for
the
user's
workload.
D
Yes,
potentially
I
mean
I,
think
I
think
you
can
do
volume
answer
kit
native,
but
not
you
can
put
a
volume
out,
but
you
won't
do
the
schedule.
What
the
won't
a
bit
but
yeah
I
mean
I.
Think
general
point:
this
is
like
pretty
fair
there's
there.
We
have
this
nice
high
level,
CRD,
I,
think
or
nice
high
level
content
of
an
app
which
is
much
more
of
a
high
level
user
facing
don't
mess
with
all
the
Amaral
stuff.
D
Yes,
this
is
that,
what's
the
next
question
not
really
taking
much
advantage
yet
because
that's
from
user
point
of
view,
you
can
get
a
single
new
feature.
Five
moves
to
kubernetes
other
than
the
kind
of
operator
experience
thing.
The
question
is
at
what
point
we
can
actually
start
exposing
some
of
these
features
based
on
this
stuff.
That's
why
it
gets
kind
of
fun.
D
B
B
A
C
C
There
will
be
some
more
communication
coming
out
soon
around
deprecation
plans,
and
then
that
happens,
but
just
want
to
get
it
in
people's
sort
of
top
of
mind
a
little
bit
that
will
have
a
new
major
version
of
the
CLI
coming
out
soon,
and
you
know
if
that
raises
any
concerns
or
questions
like
to
start
dressing.
Those
sooner
than
later,
cuz
we'll
have
this
new
thing
we'll
be
supporting
it.
A
Okay,
then
I
think
we
indeed
can
close
it
a
little
bit
earlier
thanks,
everybody
I'll
send
around
the
fall
for
next
time,
as
I
said
again
a
bit
earlier
than
this
time
and
try
to
kind
of
also
close,
the
poll
in
internet
for
people
to
actually
be
notified
that
input
you
have
from
joining.
This
call
with
that
thanks.
Everybody
talk
to
you
next
time.
Take
care.