►
From YouTube: Cloud Foundry for Kubernetes SIG [February 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).
B
Okay,
so
welcome
to
this
week's
cloud
foundry
on
kubernetes
special
interest
group
called
looking
at
the
poll
I
think
at
the
moment.
What's
leading
there
is
what
is
the
state
of
Cloud
Foundry
/
CF
for
kubernetes
you're,
a
questionmark
company
question
mark
I
meani
question
mark
CF
when
it
is
networking
question
mark
logging,
on
the
other
hand,
I
think
over
in
the
release
integration.
B
A
select
channel
Danielle
asked
the
question
around
what
I
think
is
kind
of
a
recent
addition
to
the
Cloud
Foundry
repository,
which
is
the
CF
on
on
kubernetes
project
and
I,
guess
the
the
positioning
or
like
the
headline
of
that
project
so
yeah.
Maybe
you
want
to
like
ask
your
one
or
two-sentence
question
around
that
project,
and
then
we
can
collectively
decide
if
that
is
maybe
also
something
that
we
should
talk
about
in
this
course.
Thanks.
C
B
D
Oh
thanks,
folks
and
I
suppose
I'm
coming
at
this
from
representative
use
of
customers
and
potential
customers
of
we.
If
somebody
is
going
to
adopt
cloud
foundry
on
communities
where
should
they
start
consuming
stuff
from
and
will
that
can?
Will
there
be
one
place
for
all
time
that
will
change
its
contents
and
change
its
shape
to
give
people
confidence
that
it's
a
decision
worth
making
and
I've
got
a
meeting,
hopefully
next
week
with
the
potential
customer
who's
asking
me
should
I
use
cloud,
foundry
or
kubernetes
I
want
to
say
to
them.
D
You
can
use
both
there's
a
thing.
That's
ready
start
using
this,
and
eventually
that
will
become
where
all
of
the
Clara
foundry
community
congregates
for
its
cloud
foundry
on
Cuba
Nettie's
efforts
without
having
say
to
them.
Are
you
gonna
pick
one
of
these
two
repos
and
this
one
isn't
ready
yet
that
one's
much
more
ready?
But
it's
not
gonna
be
the
final
thing,
and
then
you
have
to
switch
to
this
other
one
that
wasn't
one
or
two
sentences,
but
it's
the
best
I
can
do
what.
B
Great
great
summary,
I,
guess
and
I
think
like
it's
somehow
a
follow-up
of
last
times,
conversation
as
well
right
where
we
talked
about
like
qcf
and
the
activities
that
Zion
and
team
are
are
doing
and
how,
like
one
could
start
from
cube,
CF
and
then
probably
bring
in
more
or
about
kubernetes
native
components
into
that
and
I.
Think,
like
part
of
your
question,
Daniel
was
also
like
is,
is
that
discussion
and
and
the
outcome
of
that
discussion
still
well
it
or
like
that?
Does
that
project
change?
E
E
F
H
G
We
can
be
confident
that
they
will
continue
to
have
a
good
time,
even
as
other
things,
even
as
we
get
to
that
end.
Point
right,
I,
think
some
of
the
worry
validly
is
that
I
I
agree
which
I
don't
think
we
can
just
stop
and
keep
CF
and
just
evolve
to
simpler,
I
think
I.
Think
it's
really
hard
to
evolve
to
simple
I
think
we
actually
do
need
to
start
with
simple
and
get
there,
but
we
also
need
to
have
a
way
for
customers
who
start
oops
yes,
now
to
continue
to
have
a
good
time.
G
B
Think
I'm
going
to
make
it
even
more
difficult
for
everybody,
but
like
mid
long-term,
we
also
need
to
have
that
kind
of
plan
on
how
to
move
for
everybody.
That's
not
face
to
name
right,
but
right
then,
at
the
same
time
you
wouldn't
take
like
see
if
deployment
today
and
then
kind
of
start
to
I,
don't
know
change
it
over
time
to
kind
of
result
in
something
that's
minimally
and
kubernetes
native
I.
Guess.
I
It
sounds
like
the
undercurrent
across
these
separate
concerns
is
really
having
a
set
of
practices
and
tooling
for
migrating
between
environments,
because,
if
we're
not
going
to
do
in
place
migration
of
CFD
or
possibly
coops,
DF
or
CF
or
Kate's
or
what-have-you,
we
need
an
easy
way
for
customers
to
migrate
from
foundation.
You
know
a
to
foundation
B.
Would
that
be
sufficient
to
help
sort
of
ease
these
concerns
between
the
different
projects
and
how
we
bridge
the
gap
between
them.
G
D
D
If
we
end
up
in
a
situation
where
somebody
picks
to
use
the
you
know,
say,
cube
C
F
now
and
then
it
is
presented
that
that
repo
ceases
to
be
and
you're
going
to
use
this
other
one
and
then
you're
gonna
get
some
professional
services.
People
in
to
do
the
migration
for
you
or
you're
gonna
use
this
tool
to
migrate
from
the
thing
you
were
using
to
the
brave
new
world.
Maybe
it's
you
know
the
repo
that
certifcate
becomes
I,
think
that
will
do
damage
to
people's
trust
in
the
brand.
D
For
you,
then,
there's
no
undermining
of
trust
and
faith
in
the
in
the
brand
and
no
kind
of
unpleasant
surprises
when
someone's
like
hey,
you
told
me:
I
should
be
using
this
thing
and
that
it
would
be
great
and
that
cloud
foundry
on
communities
was
ready
and
now
you're
telling
me
that
I
have
to
give
you
more
monies
handled
migration
for
me.
So
a
sense.
C
You
just
touched
are
more
about
like
a
distribution
of
Cloud
Foundry
I
mean
this
would
be
in
a
similar
question,
setting
like
a
the
ember
of
little
people
like
now,
postulating
that
everything
that
is
done
for
a
CEO
for
cage
needs
to
integrate
like
perfectly
in
ops
man,
because
this
is
the
way
currently
their
customers
update
and
install
the
platforms
right
like
to
me.
This
seems
like
a
distribution
concern,
I
mean
and
they'll.
Yes,
like
open
source.
G
C
G
E
An
interesting
thing:
we
had
a
CFC,
a
light
plug-in
at
some
point
that
uses
the
API
to
download
the
staged
artifact
as
well
as
source
code,
as
well
as
all
setting
for
all
apps
and
just
through
the
API.
You
could
migrate
things
just
throw
a
number
out
there,
like
80%
of
everything
you
had
on
boundary.
I
think
we
could
get
it
higher
services
would
always
always
be
a
problem,
but
yeah.
H
Yeah
I
mean
we've
been
thinking
about
those
kinds
of
aspects
of
the
problem
as
well
and
honestly,
the
the
apps
seem
like
the
easy
part
to
move
around,
and
it's
it's
more
of
where
that
Cloud
Foundry
environment
connects
to
the
outside
world
via
the
service
like
system
of
record
for
service
instance,
citizen
service
findings,
or
also
splitting
ingress
traffic
across
environment.
You
know
reasonable
way,
so
yeah
I
think
like
we're.
As
Adrian
mentioned.
H
You
know,
it
seems
like
this
could
be
seven
capabilities
and
rules
that
solves
a
bunch
of
different
problems
associated
to
CF,
environment
fragmentation
and
we're
we're
getting
interested
and
more
concretely
pursuing
some
of
those
approaches
to
address.
At
least
like
you
know,
the
eighty
or
ninety
percent
of
this
problem,
space.
J
J
But
just
to
know
that
that
our
distribution
will
follow
cube
CF,
like
our
goal,
is
to
make
cube
CF
pass
without
bumps
for
a
customer
to
follow
or
for
Cloud
Foundry
user
to
follow.
Now
that
doesn't
mean
there
can't
be
other
paths,
but
we
will
always
try
to
make
that
the
path
of
least
resistance
and
to
be
able
to
consume
any
upstream
cube,
idiomatic
components
that
are
created.
J
So
we
will
adapt
the
project
to
use
everything
that
is,
that
is
new
from
upstream
and
the
component
teams,
and
we
are
going
to
try
and
let
the
the
Boche
part
of
that
wither
away
as
soon
as
we
can
take
out
pieces
of
it.
So
that
is
our
intention
that
the
operator
Rolanda,
the
fact
that
cube
CF
relies
on
an
operator.
It
may
always
rely
on
an
operator,
but
it
would
at
some
point
rely
on
a
much
simpler
operator
that
is
just
in
in
charge
of
managing
the
sea,
far
lifecycle.
J
So
that's
our
engineering
goal
as
a
community.
Anyone
can
get
on
board
that
as
long
as
we
don't
have
a
situation
where
one
project,
one
integration
point
project
is
chosen
over
another.
So,
if
we're
still
in
that
state,
we
will
continue
to
forge
ahead
and
we'll
sort
of
try
and
get
mindshare
from
the
community
by
making
that
the
best
way
now
there's
various
opinions
about
on
whether
we'll
be
successful
in
that
because
of
the
design
and
stuff,
but
that's
from
a
product
side
on
souza
product
side.
That's
the
way
we're
going.
J
You
guys
can
all
see
that
it
is
a
it's
worth
saying
and
I
should
write
this
down
somewhere,
maybe
in
a
blog
so
that
it's
on
the
record.
We
are
putting
this
upstream
as
soon
as
we
hit
1.0,
which
is
very
soon
I.
Think
we
talked
about
through
the
week
of
February,
24th
I
think
we're
okay,
putting
that
directly
into
Cloud
Foundry
foundation,
I'll
talk
to
Chris
and
to
chip
about
that.
To
make
sure
that's!
Okay,
please
raise
objections.
J
If
that's
a
problem,
I
think
it's
only
fair,
given
that
cf4
Cates
went
directly
into
the
foundation
rather
than
the
incubator,
so
and
I
think
I
think
we
have
the
support
of
the
foundation
for
that,
because
we've
been
talking
about
it
for
a
long
time
and
it's
been
part
of
it's
been
associated
closely
with
the
quarks
Project.
So
that's
our
intention
just
so.
J
You
know
what
we're
doing
we
want
to
be
really
transparent
so
that
our
partners
know
what
we're
doing
so
that
our
customers
know
what
we're
doing
and
at
this
time,
when
it's
being
there's
perception
in
the
press
that
oh
cloud
foundry
doesn't
know
what
it's
doing,
we
need
to
attack
that
I
mean
you
might
have
seen
some
of
the
press
that
we've
seen
recently
about.
Oh,
what's
gonna
happen
now
that
pivotal
is
gone
to
VMware
or
maybe
nobody's
gonna.
Take
care
of
Cloud
Foundry
I
mean
we're
all
here,
we're
all
talking
about
that.
J
J
We've
acknowledged
the
objection
to
CF
Operator
being
really
more
of
a
Bosch
operator,
I'm
not
going
to
rename
it
right
now,
but
we
get
that
it
is
a
Bosch
operator
and
that
we
will
maybe
need
a
Bosch
operator,
but
it
should
be
separate
from
what
takes
care
of
keeps
yeah
so
sorry
to
drag
it
back
to
that.
I
just
wanted
everyone
to
be
clear
on
what
our
position
was
and
what
we
intend
to
do,
and
that's
yeah
just
to
provide
some
clarity.
There
I
don't.
G
G
So
you
can
install
that
and
upgrade
that
vsc,
IDs
and
and
one
should
live
on
top
of
the
other
there's
and
one
thing
I'm
a
little
concerned
about
is:
if
we
treat
them
as
two
separate
things,
they
can't
become
two
separate
things
and
because
actually
I
think
you
should
be
making
sure
that
the
direction
of
CF
for
cakes
goes
in,
doesn't
end
up
conflicting
with
that
and
again
I
worry
about
some
of
the
ytt
stuff.
Because
of
that
so.
J
Yeah
I
was
trying
to
say:
ytt
specifically,
would
slow
us
down
right,
but
it's
not
something
we
couldn't
cope
with.
We
we
talked
about
it.
I
think
last
time
certainly
helm
charts
we
can
consume,
we
should
be
able
to
consume
straight-up
kubernetes
yamo
other
templating
formats.
We
just
require
that
we
figure
out
how
to
import
it,
and
we
would
like
help
with
that,
if
that's
the
choice.
So
if
y
TT
is
a
choice
for
the
component
team,
we
need
somebody
to
come
and
chip
in
on
qcf.
J
C
G
G
The
one
remaining
kind
of
worry
I
have
about
about
the
kind
of
how
do
we
migrate
people,
starting
with
a
Bashaw
operates
and
ending
with
a
CF
operator,
is
if
the
thing
the
customer
starts
by
supplying
is
the
boss
manifest?
If
you
see
what
I
mean
and
if
we
have
to
keep
a
Bosch
manifest
working
over
the
time,
that's
quite
hard
to
migrate,
as
opposed
with
is
the
CID
with
some
top-level
fields
and
the
fact
that
that
becomes
a
manifest
is
whatever
that's.
What
happens
today?
G
C
H
Troy
I
want
to
say
thanks
for
the
update
about
qcf
and
your
plans.
I
was
wondering
I
mean
you
kind
of
seems,
like
you
have
this
kind
of
BIGBANG
introduction
for
cube,
CF
and
I.
Remember
a
few
weeks
ago
you
saying
you
were
doing
some
cosmetic
cleanups
within
the
cube
CF
repo,
just
to
like
Souza's
specific
things,
the
reason
not
to
propose
it
just
for
incubation
now
I'm
assuming
to
go
into
the
a
prime-time
PMC.
Yes,.
J
So
the
the
thought
was
and
and
of
course,
that
spike
turns
out
hasn't
happened
yet
so,
where
that
work
still
has
to
be
done,
it'll
probably
be
down
to
you
Vlad
to
do
a
bunch
of
that,
and
also
we
want
the
docs
in
a
good
enough
shape
that
anyone
coming
to
it
for
the
first
time
can
write
from
the
upstream
project
use
Cloud
Foundry.
So
so
we
want
to
get
that
in
shape
as
well
we're
just
a
little
behind
on
that.
J
That's
all
and
we
could
move
it
in
a
sort
of
half-finished
state
into
the
incubator,
but
that's
just
N
and
then
two
weeks
later
move
it
to
two
Cloud
Foundry.
We
could,
but
then
it's
just
one
more
move
that
is
really
kind
of
pointless
it
is.
It
is
fully
baked
it's
working
right
now.
I
use
it.
That's
my
that's!
My
Cloud
Foundry,
it's
just
we
haven't
felt
because
the
lack
of
documentation
mostly
and
the
fact
there
might
be
some
Sousa
artifacts
still
in
there.
We
haven't
felt
comfortable
putting
it
upstream.
H
J
H
J
See
stories
yeah
so
and
then
that
raises
the
question:
does
this
does
keep
CF
fall
under
the
quarks
umbrella
because
then
it
would
inherit
that
same
privilege,
but
is
that
the
best
governance
model
for
it
I.
H
J
H
J
J
E
J
J
C
To
like,
if,
if
we
are
following
whatever
is
written
down
for
and
and
with
the
foundation,
that's
not
like
our
around
to
decide
this
stuff
right
yeah.
We
have
PMC
meetings
for
for
the
specific
things
so
like.
If
you
want
to
talk
about
this
I
guess,
that's
something
called
EQ
pants
I'm!
You
see,
okay,.
H
H
J
Do
re
to
hijack
the
agenda
again?
Did
we
do
you
want
to
quick
update
on
the
technical
project
manager
proposal
from
last
last
week?
This
was
UA
was
proposing.
First
of
all,
that
Souza
contributes
someone
to
the
release
integration
team,
which
I
don't
know.
If
we
have
that
person,
the
TPM
role
that
she
mentioned
we
may
have,
but
it
I
think
we
have
considered
putting
someone
half
time
a
half
time
resource.
If
that's
helpful,
we
still
got
to
talk
a
little
bit
internally
about
about
that.
We
have
a
person
in
mind
but
yeah.
We
can.
H
Mean
I
think
it
would
be
helpful
in
terms
of
coordinating
some
of
this
activity.
We've
also
been
talking
on
the
VMware
side
about
getting
some
more
staffing
in
those
kinds
of
rules.
Okay
and
I
think
that
they
they
may
have
some
time
to
dedicate
to
some
of
the
community
coordination
efforts
as
well.
Okay,
so
maybe
maybe
between
the
two
of
us,
we
could
get
a
full
technical
program
manager
coordinating
a
bunch
of
this
activity.
B
Okay,
so
that
may
be
a
good
point
in
time
to
kind
of,
as
Yun
pointed
out
that
he
like
was
referring
to
like
kubernetes
native
components,
with
his
number
four
question
around
CF
for
kubernetes
to
kind
of
try
and
see.
If,
if
we
might
have
folks
on
the
call
that
can
you
can
provide
an
update
on
where
we
stand
with
those
topics,
I
mean
it
might
even
start
with
kind
of
SCI
describing
what
is
in
that
repository
now.
That's.
F
Yeah
sure,
can
you
all
hear
me:
okay,
awesome,
so
I
want
to
I
just
prepared
this
simple
sort
of
visual,
and
then
we
can
can't
jump
into
the
repository
in
a
moment,
so
the
sort
of
recalling
a
pre-alpha.
But
what
do
you
see
right
now
in
the
repository
is
a
script
that
installs
basically
capi
irony
in
networking,
and
we
also
have
a
fork
versions
of
ua
that
we
are
using
to
make
CF
push
with
talk
or
app
work.
F
So
right
now
the
that's
one
of
the
capabilities.
We
have
it's,
not
a
full
CF
experience,
but
you
can
do
you
can
install
and
you
can
actually
push
a
docker
app
and
you
can
route
to
it.
So
that's
where
we
are
today,
I
think
where
we
are
right
now
is
we're
working
with
the
teams.
So
these
are
the
six
teams
that
we
identified
for
to
integrate.
F
We
working
very
closely
with
cross
team
pairing
with
them
exclusively,
so
we
are
working
copy,
especially
we
are
hoping
to
get
K
pack
integration
so
that
we
can
support
the
foals.
The
real
CF
push
using
the
K
pack
and
Bill
packs
and
we're
also
working
on
figuring
out
the
install
experience.
How
do
we
package
default
databases
and
the
blobstore
as
we
we
do
with
CF
deployment?
What
are
the
you
know?
F
How
can
we
do
the
batteries
included,
installation
experience
and
then,
last
but
not
least,
in
this
we
are
also
working
on
building
pipelines
and
this
will
not
only
build
pipelines
for
us
as
an
integration
team,
but
also
for
the
teams
so
that
they
have
a.
We
have
a
full
feedback
cycle
between
us
and
them.
So
that's
our
goal
in
the
next
two
to
three
weeks
and
then
we're
gonna.
You
know
as
a
bit
as
we
release
the
Alpha,
which
we
expect
to
do
it
in
Feb
time
for
him
and
the
fab.
F
You
know
some
of
the
capabilities
you
will
be
able
to
do
is
you
know,
you'll,
see
a
push
and
then
teams
will
have
a
full
feedback
cycle
and
then
we're
gonna
chip
away
individually
with
thinking
about
cats.
Okay,
what
are
the
things
and
cats?
For
example
DEA?
There
are
still
tests
from
DEA
which
may
not
be
required,
for
we
may
have
to
be
remove
it.
F
We
are
probably
thinking
about
some
of
the
tests,
but
they
will
just
have
to
think
of
you
know,
just
look
at
them,
review
them
and
see
what
what
stays
and
what
goes
start
thinking
about.
They
won't
experience
and
then
sort
of
chip
away
capabilities
and
then
to
day
two
and
then
so
on
and
so
forth.
So
that's
a
very
high
level
thought
around
how
we're
progressing
from
where
we
are
today
to
where
we
want
to
go
it
towards
GA.
F
C
I
feel
like
like
somewhere
along
the
way
you
mentioned,
like
cutting
things
like
the
a
specific
test
on
cats
right
and,
if
you're
not
only
talking
about
test
Suites,
but
we
are
talking
about
real
functionality
at
some
point
in
time
the
decisions
might
be
necessary
to
really
cut
existing
scope
to
say
there
might
be
existing
use
cases
in
Cloud
Foundry.
We
will
not
support
or
not
support
at
first
in
the
coordinators
base
book.
C
F
I
think
I
expect
that
to
be
a
proposal,
I'll
go
out
to
the
community
in
terms
of
okay.
These
are
the
tests
that
we
identified.
The
use
cases
are
not
supported
at
all,
or
this
is
just
a
relic
that
we
should
be
planned
to
remove
in
in
the
Boche
wall,
but
you
just
never
got
around
it.
So
I'd
expect
that
to
go
out
as
a
as
a
proposal,
and
then
we
can
collab
on
that.
H
I
know
there
was
so
I
think
there
was
a
PR
that
the
networking
team
sent
in
to
cats
where
they
were
introducing
some
conditionality,
because
when
envoy
is
acting
as
the
HTTP
proxy,
instead
of
the
go
router,
it
has
some
slightly
different
behavior
in
terms
of
like
passing
through
or
modifying
some
of
the
query
parameters,
and
so
even
even
something
like
that
could
end
up
being
a
change
that
it
doesn't
break
every
app
on
the
platform.
But
it
breaks
one
app.
That's
expecting
that
consistency.
H
H
Cool
is
it
really
detest,
sir
Diego
tests
and
sorry
I
believe
there
is
d8s
and
also
Diego.
There
might
be
a
compatibility
test
for
apps
that
were
staged
on
the
DA's,
like
that,
the
droplets
can
still
run
on
Diego,
it's
possible
yeah,
that's
the
kind
of
thing
that
maybe
could
come
fall
off
at
this
point.
F
So
I'm
gonna
share
the
repository
so
just
to
give
a
glimpse
on
that.
So
we
updated
the
instructions
a
bit
more
helpful
for
folks
to
install.
But
as
I
mentioned
today
you
could
only
run
see
if
docker
based,
apps
and
I
think
the
big
next
big
milestone
for
us
is
capi
integrating
with
KPAC
and
bill
packs.
So
we
can
do
that
you'll
see
of
push
experience
that
we
have
custom
to.
F
Anything
specifically
want
to
call
out
that
we
are
using
cap
as
a
way
to
install
cube
annuities,
so
cap
allows
us
to
manage
the
lifecycle
of
the
of
the
installation.
So
when
you
do
capital
eat,
it
deletes
all
of
the
Kuban
of
these
objects
that
we
have
installed
by
a
cap
similar
to
helm
and
then
we're
using
white
tea
tea
for
templating.
F
Yeah
anything
that
comes
to
your
mind,
you
want
me
to
go
deep.
Here's
the
install
script
that
does
most
of
the
installation
and
we're
gonna
continue.
You
know
a
trade
on
this
as
each
and
every
team
will
bring
in
you're
using
a
package
called
vendor.
So
what
this
does
is
very
similar
to
how
you
pull
in,
like
you
know
dependencies
so,
for
example,
here's
a
meta
controller,
here's
the
sea
of
kate's
and
working
copy
and
we're
gonna
continue
to
add
additional
components
like
logging
and
metrics,
which
is
missing.
F
One
thing
I
do
want
to
call
out
that
when
you
do
a
CF
push,
if
you
install
this
today,
if
we
do
see
if
perch
the
app
will
be
running,
but
you
will
get
an
error
back
because
we
are
missing
the
logging
and
metrics
component,
but
the
app
is
actually
routable.
So
just
keep
that
in
mind
when
you
take
this
for
a
ride.
C
F
It's
too
early:
we
still
don't
have
all
of
these
system
components
integrated.
Yet
so,
once
we
are
at
a
maybe
our
beta
release,
maybe
we
can
consider
that
once
we
have
a
you
know,
sort
of
a
cat's
running
and
once
we
have
more
stability
and
confidence,
I
think
that
would
be
a
good
place
to
start
thinking
about
how
we
want
to
distribute
the
components
and
the
workloads.
Okay,.
F
So
so
we're
thinking
about
alpha
release
and
a
Feb
March
early,
so
beta
release
is
probably
gonna
be
a
month
or
two
later,
but
as
we
chip
away
in
the
next
coming
weeks,
we
will
have
a
more
visibility
on
those
dates.
So
so
right
now,
it's
just
a
estimate,
but
as
we
get
closer
to
alpha,
we'll
have
more
concrete
dates
for
the
Vado.
J
A
F
So
we
are
looking
at
cab
for
variety
of
things.
It's
likely
experience.
It
provides
very
similar
to
Bosch
right
the
Bosch
when
you
deploy
and
upgrade
it.
It
allows
you
to
manage
the
entire
lifecycle.
So
at
this
stage
right
now
we
are.
We
believe
cap
is
a
right
tool
for
that,
but
we're
open
to
suggestions
if
there
are
any
other.
You
know
tooling
out
there
that
could
help
us
manage
that
life
cycle,
because
we
were
looking
at
as
a
you
know.
J
J
J
I
Yeah
I
think
it
would
be
great
to
explore
having
the
ability
to
have
sort
of,
as
we
briefly
mentioned
earlier
in
the
call
the
base
set
of
Cloud
Foundry.
You
know
deployment,
specs
and
other
resources
that
are
not
templatized
and
then
different
distributions
or
projects
could
then
add
templating
layers.
On
top
of
that,
so
you
know
config
Maps
secrets
overlay
settings
that
you
may
have
on
the
base
temp.
I
But
so,
if
we
eventually
get
to
that
stage,
then
I
think
that
may
well
your
fear
of
the
dependence
on
K
for
teens
tooling,
because
the
base
deployment
specs
would
not
actually
have
any
templating
at
all.
So
the
out-of-the-box
experience
may
not
be
quite
as
rich,
but
it
will
be
more
extensible
downstream
right.
G
They
say
this
is
why
I
worry
about
why
I
think
template
thing
in
the
base?
Is
it's
your
template
at
the
top?
We
can
do
all
sorts
of
things
at
the
top.
If
your
template
in
the
bottom,
everyone
has
to
go
through
the
white
c-team
stuff
and
it's
gonna
end
up
non-native
and
it's
gonna
be
a
pain
and
I'm,
not
sure
we're
at
the
point
where
we
need
it.
Iii
think
I.
G
Think
I
I
would
like
us
to
try
to
knock
our
heads
against
not
doing
any
templating
and
and
see
what
breaks
and
then
maybe
we
need
to
add
some
templates
me
higher
up.
But
if
you
start
with
it,
it's
really
hard
to
get
away
from
it
and
I.
Think
a
lot
of
people
who've
gone
in
with
him
over
a
period
of
time
have
realized
the
template
thing
close
and
it's
bad
enough
to
do
it.
I
One
of
the
challenges
that
we've
seen
with
some
of
these
components
as
they've
been
brought
unto
safer.
Kubernetes
is
many
components,
are
not
written
in
a
way
that
they
could
accept
configuration
via
config
maps
or
files
in
some
ways.
Right
or
you
know,
they're
sort
of
configured
they're
expecting
a
very
specific
way
of
being
configured,
so
it
made
it
much
easier.
I
You
add
those
into
the
pod
spec
definitions
and
things
like
that,
so
I
can
see
in
some
ways
the
way
to
see
that
the
templates
look
now
may
not
be
the
final
stage,
because
some
of
these
are
in
service
to
get
a
working
control
plane
by
some
definition
of
the
word
such
that
we
can
have
a
base
to
iterate
on
and
make
it
more.
You
know
duction
ready
and
sensible
all
that
sort
of
stuff.
G
A
D
The
main
distribution,
the
people,
the
kind
of
primary
members
of
the
foundation-
is
there
a
sense,
I'm
guessing-
maybe
probably
not
on
the
for
the
people
on
this
call,
but
the
tension
between
getting
stuff
done
quickly
and
getting
wider
spread.
Adoption
and
engagement
from
the
CN
CF
ecosystem
developers.
D
There's
a
tension
between
those
suit,
like
you
could
use
all
of
the
k14
stalling
now
neighbor,
all
you
selves
to
go
quickly
to
be
able
to
get
chef
Achatz
out
getting
in
the
hands.
Customers
that
you
already
have
that's
gonna,
as
people
have
mentioned,
alienate
folks
that
kind
of
come
into
the
ecosystem.
What's
all
this
weird
tooling
that
we've
never
seen
before
and
it
will
be
like
Bosch
all
over
again,
the
alternative?
Is
you
don't
use
any
of
that
tooling
that
helps
you
go
faster
and
then
there
are
commercial
implications
from
that.
D
Do
does
anyone
think
it
will
be
possible
to
kind
of
come
to
a
collective
understanding
of
where,
on
that
spectrum,
the
community
wishes
to
go.
For
example,
you
could
make
that
more
concrete
by
saying
when
something
is
GA,
we
will
have
not
included
all
of
these
tools
like
we
use
these
tools
to
get
going,
and
then
there
will
be
a
kind
of
deadline
by
which
we
decides
to
just
be
issuing
native,
vanilla
and
template
two
cubes
Hamill
is
that
sort
of
thinking
that
might
be
useful
and
might
be
achievable.
D
H
Know
I
think
we've
been
wrestling
with
that
issue
a
lot
around
like
do
we
adopt
tools
that
we
don't
like
or
that
have
serious
deficiencies,
but
they're
used
or
do
we
build
tools
that
work
and
work
to
our
standards,
but
then
go
through
the
loop
of
getting
adopt?
Some
of
that
we
ship-
and
this
is
like
this-
we
should
yeah.
H
A
H
And
as
we've
been
coming
together
with
the
M,
where
we
we
know,
there's
a
lot
of
other
opinions
and
firmly
here
about
that
and
we're
trying
to
chart
a
course
through
those
kinds
of
recommendations,
as
as
we
get
integrated
into
those
sets
of
concerns,
so
I
think
there's
been
some
Nvidia
you're,
making
a
broader
push
for
one
of
the
K
14
schools,
but
I
wouldn't
say
that's
hundred-percent
definite.
Yet
I
may
be
a
thing
that
we
actually
like.
We
look
for
feedback
on.
You
know
we
can't
make
it
now
and
expect
that
to
necessarily
pertain.
H
Even
in
six
months,
it
may
be
worth
putting
more
muscle
behind
the
cape
or
to
those
tools.
If
we
think
that
they're
the
right
approach
going
into
the
helm
is
garbage,
you
know,
if
you
want
to
do
templating.
This
is
a
better
way
to
do
it.
Q
cut
will
apply,
is
unreliable,
like
you
know,
there's
there's
as
as
declarative
as
we
want
kubernetes
to
be
they're
still
ordering
dependency.
Is
there
still
bootstrapping
that
needs
to
happen,
and
this
is
a
tool
that
is
trying
to
strike
a
better
balance
there.
H
A
E
Wanted
to
ask
if,
if
there
are
deficiencies
in
upstream
project,
why
not
try
and
fix
those
deficiencies
upstream?
Why
build
something
new
if
there's
a
hound
efficiency
that
you
found,
which,
by
the
way
I'm
still
skeptical
about,
because
if
you're
making
components,
cube
native
helm
should
be,
should
be
pretty
easy,
but
anyway
yeah?
If
there
are
deficiencies,
why
why
not
try
to
fix
them?
I.
H
Think
actually,
these
are
a
great
example
of
this,
where
we
realize
that
you
need
they
CR
D
to
be
created
in
the
API
before
you
can
use
it.
So
there's
a
like
fundamental
sequencing
issue
there
and
there's
an
issue
on
on
the
pel
repo
for
held
3
saying
like.
Can
you
solve
this
issue?
I
believe
the
current
state
of
ends
to
say
no
to
make
a
separate,
helm,
template
and
chart
for
your
CR
DS,
so
they're
pushing
that
ordering
concern
out
of
hell
into
the
structure
of
any
projects
they're
using
and.
J
G
G
The
thing
I
was
going
to
say
is:
there's
a
rule
in
science
fiction
books.
You
only
allowed
one
divergence
from
reality.
If
you
have
you
pick
one
thing:
that
is
your
thing
and
the
reader
will
let
you
get
over
there
in
this
little
bit.
If
you
have
five
and
then
the
reason
it
just
doesn't
make
sense,
there
are
no
rules
in
this
universe
and
it
doesn't
work
and
I.
Think
I.
Think
it's
okay
for
us
to
say
he
is
our
opinion
where
it's
difference
and
we
just
we
just
can't
have
everything
be
different.
G
They
were
already
asking
people
to
make
a
leap
into
some.
Some
areas
and
I
think
yeah.
You've
kind
of
got
a
budget.
You've
got
a
you've,
got
a
weirdness
budget
that
you're
asking
people
to
you
know
take
on
this
new
cognitive
load,
so
everything
has
to
kind
of
pay
its
cost
in
weirdness
and
I'm,
not
I'm,
not
convinced
that
ytt
does.
That
needs
to
begin
this.
That's.
G
D
G
G
G
I
A
question
with
that
too,
though,
which
is
the
overall
I'd,
say
experience
and
product
direction
of
Cloud
Foundry
in
general,
in
terms
of
kubernetes,
because
if
we
look
at
CF
as
it
is
now
is
definitely
an
appliance
that
is
installed
onto
a
crepinette
YZ
cluster.
It
does
not
have
you
know,
custom
resource
API,
it
does
not
write.
You
know
from
the
application
containers
you're
not
able
to
hit
the
kits
API
other
things
like
that.
I
So
in
terms
of
a
customer
who
wants
a
kubernetes
native
experience,
regardless
of
our
template,
incurring
distribution,
it's
not
going
to
be
kubernetes
native.
So
if
that
is
who
we
are
targeting
with
our
distribution
decisions,
we're
not
only
not
going
to
succeed
with
that
customer
segment,
but
we're
also
going
to
have
a
less
expressive
experience
for
our
own
needs
in
our
own
installation.
So
I
think
we
also
need
to
figure
out
who
are
we
actually
targeting
with
that?
If
we
are
going
to
take
it
to
that
sort
of
CN,
CF
portfolio
extreme,
so.
G
So
we
need
the
the
the
fact
that
it's
gonna
run
on
top
of
kubernetes,
in
my
view,
is
necessary
and
not
sufficient.
We
also
need
to
use
the
fact
that
we
we
no
longer
have
to
spend
effort,
writing
our
own
scheduler
right.
We
need
to
use
that
to
evolve
the
UX
as
well
to
be
more
native,
and
there
should
be
more
CR
D's
in
there.
If
you
want
them
right,
there
should
be
escape
actors.
If
you
want
them
and
we
need
to
start
one
of
the
reasons.
G
G
B
Hey
folks,
so
looking
at
the
time,
I
think
one
minute
past
scheduled
timing,
I
feel
like
we
then
kind
of
started
discussing
a
couple
of
items
that
kind
of
hopefully
make
it
them
to
the
proposal
list
for
for
next
time.
So
I'm
going
to
open
on
the
voting,
then
for
for
next
time,
a
couple
of
days
before
the
meeting
and
as
always,
tools
be
prepared
to
be
called
in
for
ireenie
1.0
demo,
because
that's
one
of
the
few
items
that
I
think
we
haven't
ticked
off
so
far.