►
From YouTube: 2020-08-03 Crossplane Community Meeting
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
Okay,
the
recording
has
started-
and
this
is
the
august
3rd
2020
cross
playing
community
meeting.
We
will
start
with
our
milestone,
checkup,
so
0.13.
We
we
chose
to
skip
the
release
of
the
main
cross
plane,
repo,
at
least
for
the
mid
monthly.
You
know
release
for
july,
so
0.13
did
not
go
out
in
july.
There
wasn't
really
a
lot
of
any
code
changes
to
the
main
repo.
A
It
was
mostly
in
providers
and
documentation
and
other
stuff,
but
we
will
be
planning
on
doing
0.13,
for
I
guess
you
know
mid
month
released
in
august.
The
providers
did
go
out
last
week
or
maybe
a
week
actually
the
week
before
that.
But
since
the
last
community
meeting
the
we
did
a
whole
round
of
releases
for
all
the
providers
like
aws
and
alibaba
and
such
we
could
bring
up
the
board
here
for
where
we're
tracking
things
at
the
cross
plane
organization
level.
A
But
then
the
main
main
feature
areas
are
down
below
further
that
we
can
dive
into
more
details
on.
But
let
me
clear
some
things
off
my
screen
here:
they're
in
the
way
so
yeah
in
this
board
view
here,
dan
or
phil
or
anybody
are
there
any
issues
in
here
that
would
we
want
to
talk
about
or
bring
up
that
are
not
already
captured
in.
Like
the
you
know,
high
level
feature
status,
update,
section.
B
A
Is
that
like
does
that
fit
in?
Do
you
want
the
status
about
it
dan,
or
do
you
want
a
like
a
technical
discussion
of
it
at
the
in
the
technical?
You
know
section
of
the
media,
yeah
I'll,
probably
drop
it
in
technical
deals,
okay,
okay,
that
sounds
good
all
right,
so
then,
let's
move
ahead
then
to
the
status
for
some
of
the
major
features.
So,
as
we
mentioned
earlier,
nick
and
mobafiq
are
off
this
week.
C
Yeah
I
mean
so
casey's
putting
together
the
design
dock
for
that,
and
that
should
be
available
for
review
here
shortly
and
he's
also
working
on
kind
of
sketching
out
the
open
api
for
the
providers
and
we'll
be
probably
floating
that
in
the
provider
channel
in
cross
plane
slack
to
get
some
some
early
feedback
on
that.
So
those
are
the
two
things
that
are
in
flight
and
you
know
a
little
bit
slower
progress
with
summer.
Vacations
kind
of
you
know
intermixed,
but
yeah.
That's
still
going
on.
A
Cool
sounds
good.
I'm
looking
forward
to
seeing
that
document
all
right
and
then
cross-plain
agent
updates.
Is
anyone
aware
that
movafic
is
the
important
point
for
that
one?
But
anybody
else
can
answer.
B
Yeah,
I
can
answer
that.
There's
a
pr
open
that
folks
can
look
at
if
they
like
that,
basically
adds
the
initial
functionality
it's
doing
the
kind
of
basic
pulling
from
the
host
cluster,
and
it
also
has
the
controllers
for
both
running
it
in
the
local,
which
would
be
you
know
when
you're
in
the
same
cluster
as
well
as
remote.
If
folks
are
interested
in
that
they
can
take
a
pass
on
it.
I've
done
one
review
but
need
to
go
back
through
as
well,
but
it's
kind
of.
A
A
B
Yep,
so
there's
been
a
couple
different
kind
of
concurrent
efforts
around
this,
but
basically
saying
we're
still
working
through
it.
B
I
added
some
notes
around
kind
of
like
direction,
so
the
proposal
right
now
stands
with
package
being
a
single
type,
so
whether
it's
like
underlying
different
under
the
hood
you're,
still
using
the
same
crd,
there's
a
desire
from
a
user
experience
perspective
to
be
able
to
kind
of
distinguish
the
creation
of
these
two
different
types
of
packages,
one
being
a
provider
which
most
people
are
familiar
with
and
the
other
being
a
configuration
which
would
be
kind
of
composition,
types
that
you'd
use
to
bootstrap
a
platform
or
add
to
your
apis.
B
Basically,
that
you
exposed
and
there's
a
desire
to
kind
of
like
separate
those,
even
if
they
look
very
similar.
B
One
way
to
do
that
would
be
to
have
a
single
package
type
and
then
just
have
kind
of
like
cli
tooling
around
that
and
you
know,
get
a
package
and
you
could
say
get
configurations,
but
then
have
it
filter
packages
by
their
type
or
something
like
that.
But
there's
been
more
of
a
push
to
have
them
separate,
potentially
because
of
we
may
treat
them
in
some
different
ways
and
et
cetera.
B
B
I
added
some
open
questions
as
well
that
we're
still
kind
of
working
through
some
of
these
are
related
more
to
package
format,
so
that
would
be
like
opinionated
directory
structure
and
sub
modules
versus
dependencies
which
that
one's
probably
worth
talking
about
a
little
bit.
There's
you
know
different
tools,
especially
infrastructure
provisioning
tools.
Have
the
ability
to
either
depend
on
other
package
types,
so
you
can
think
of
a
configuration
depending
on
a
provider.
B
You
know
if
it
composes
aws
resources,
it's
probably
going
to
depend
on
provider
aws,
but
then
configurations,
depending
on
other
configurations,
could
look
more
like
I'm
having
a
sub
module.
So
think
of
that,
like
a
a
sub
chart
in
helm,
which
is
obviously
different
than
just
having
the
thing
installed
separately,
it
bundles
it
all
together,
we
don't
know
how
much
of
a
desire
there
is
for
folks
to
be
able
to
do
that.
But
we've
been
thinking
about
what
that
would
look
like
likely.
B
You
could
do
it
just
by
referencing
that
thing
and
then
have
it
kind
of
compiled,
since
all
of
these
are
resulting
in
oci
images
eventually
anyway,
but
that's
just
one
of
the
things
we're
working
through
and
then
the
last
thing
would
be
the
installing
arbitrary
yaml,
basically,
the
the
reason
why
we
would
allow
that
would
be
so
that
you
could
actually
create
crs
not
just
crds,
and
that
includes
infrastructure
definitions,
et
cetera.
So
right
now
the
proposal
stands
with
you're
able
to
create
crds
infrastructure
definitions,
infrastructure,
publications
and
compositions.
B
If
you
could
do
arbitrary
yaml,
obviously
you
could,
you
know,
create
a
definition
of
a
database
and
then
create
an
instance
of
it
as
well,
and
if
we
were
to
do
that,
we'd
probably
make
the
package
manager
configurable
so
that
when
you
actually
run
the
package
manager,
you
could
specify
these
are
the
types
I'll
allow
you
to
make.
So
if
you
don't
want
people
to
be
able
to,
you
know,
install
arbitrary
things
that
may
create
anything.
B
Then
you'd
want
to
restrict
that
when
you
run
the
package
manager,
so
anyway,
that's
kind
of
a
lot
of
stuff,
but
still
trying
to
work
through
this
model,
because
we
think
it's
pretty
critical
for
the
future
of
crossplane
and,
as
with
last
meeting,
there's
still
ongoing
development
work
happening
in
that
repo.
In
my
github
org
and
another
thing
that
we're
doing
there
is
working
on
some
developer
tooling.
B
So,
if
you
have
thoughts
about
you
know
things
you
like
to
see
so
specifically
around
like
linting
and
building
packages,
and
that
sort
of
thing
definitely
feel
free
to
reach
out
about
that.
A
Nice
standard
did
you
have
an
ata
or
some
clear
blockers
that
need
to
be
overcome
before
the
design
pr
can
be
resolved.
B
It's
not
really
like
getting
it
merged,
wouldn't
really
mean
anything
at
this
point.
So
even
if
we
kind
of
went
ahead
with
it,
I
don't
think
we're
clear
on
the
full
model
we
want
to
have
so
there
isn't
clear
blockers
besides
kind
of
some
of
those
open
questions
which
I
guess
are
not
clear.
B
A
Yeah
that
sounds
great
okay
and
then
you
know,
there's
definitely
work
ongoing
in
the
on
kubernetes
runtime
repo
as
well.
Does
anybody
on
the
call
want
to
catch
us
up
on
that
as
well
like
ryan
or
harry.
D
Or
I
can
I
can
give
some
updates.
We
have
a
lot
of
new
developments
in
the
past
two
weeks
on
the
home
repo,
so
we
have
at
least
like
seven,
eight
pretty
significant
prs
going
in
and
we
don't
have
any
any
major
development
pending
though
so
it's
it's
in
a
relatively
stable
shape.
Maybe
we
will
give
it
another
two
to
two
weeks
or
a
month,
then
we
will
try
to
release
0.1
version,
we're
now
at
0.0708,
something
like
that
yeah
by
that
time.
D
C
What
are
the
plans
for?
I
guess,
finalizing.
The
move
of
the
app
config
controller
out
of
core
and
into
the
home
run
time
is
that
it
sounds
like
that's
what
happened
in
0.13,
but
do
we
have
like
a
date
for
that.
B
Yeah,
so
I
think
that's
more
on
our
side
right.
They
already
have
the
application
configuration
controller
running
so
right
now
the
oem
kubernetes
runtime
can
run
as
is
right,
but
we
just
would
need
for
whatever
the
0.1
version,
for
instance,
to
run
it
alongside
crossplane
that
need
to
be
moved
out.
So
we
need
to
have
you
know
a
a
way
to
install
that
and
do
so.
B
So
I
think,
as
soon
as
we're
ready
to
have
that
home
chart
have
like
a
0.1
release
and
alpha
release,
then
we
can
go
ahead
and
remove
that
application,
configuration
controller
from
crossplane
and
then
bundle
the
helm
chart.
However,
we
see
fit
so
maybe
maybe
at
that
point
the
agent
or
something
like
that
will
be
available
to
have
kind
of
that
cross,
plane
app
helm
chart.
We've
talked
about.
D
C
Cool
and
then
I
guess,
we'll
be
looking
at
roughly
maybe
mid
month
for
the
0.13,
if
we're
going
on
the
monthly
is
that
would
that
work?
Well
for
you,
dan,
like
ahead
of
your
wedding
and
stuff.
B
Or
yeah,
I
I
think
we'll
need
to
see
how
much
how
many
changes
we
have
right
because
move
off
it
can
nicker
out
this
whole
week
and
they're
primarily
working
on
composition,
slash.
We
need
them
at
least
around
to
be
able
to
make
some
progress
there.
So,
assuming
that
we
can
get
some
stuff
in
I'd,
be
happy
to
do
that
either
right
before
or
right
after,
if
that
was
necessary,
that
would
still
that
would
be
towards
the
later
part
of
the
month.
But
you
know
not
not
super
late,
so
cool.
D
A
Yeah,
hopefully,
we
can
not
necessarily
like
single
single
resource
bottleneck
on
on
dan
and
and
his
wedding
for
making
those
changes
like
in
the
crossfit
main
crossband
repo
like
move
off
or
nick,
or
somebody
can
pick
those
up
if
the
time's
up
every
time's.
Well,
that
way,
definitely
yeah
cool
all
right.
Brian.
Thank
you
for
the
updates
as
well.
I
appreciate
it
you're.
D
A
Ready
of
course,
let's
move
ahead
into
the
community
topics
in
questions
section,
so
dan
do
you
want
to
do
you
have
anything
to
update
us
on
for
upcoming
tbs
episodes.
B
Sure
yeah,
so
we
didn't
have
one
last
week
just
because
we
had
quite
a
few
things
going
on
including
lots
of
recording
of
cubecon
talks
around
there,
but
we
have
some
cool
episodes
coming
up
so
as
as
an
early
preview,
I
don't
have
a
date
for
this
one,
but
the
cto
of
source
graph
is
interested
in
coming
on,
so
they
basically
recommend
for
production
deployments
deploying
source
graph
on
kubernetes
and
that
uses
I
think,
postgres
and
redis.
B
So
there's
actually
an
issue
open
on
the
source
graph
repo
right
now
that
I
opened
for
allowing
folks
to
be
able
to
use
crossplane
to
use
managed
versions
of
postgres
in
red
in
redis
right
now.
If
you
deploy
you,
can
you
know
like
a
lot
of
helm,
charts
and
that
sort
of
thing
you
can
supply
credentials
for
a
managed
service?
B
You
know
of
a
database
that
you'd
like
to
use
there,
or
you
can
just
run
them
in
cluster,
so
obviously
crossplan
would
be
nice
there,
because
you
can
create
those
resources
rather
than
having
to
create
them
out
of
band
and
then
bring
your
credentials
and
that
sort
of
thing
so
that
one
we
were
working
out
with
his
schedule
and
also
that
work
that
would
have
to
happen
for
that
to
go
into
effect
that
one
should
be
coming
up
pretty
soon,
probably
not
the
next
one,
though
next
one
we
are
probably
going
to
have
another
security
focused
one
sasha
from
upstream
kubernetes.
B
He
works
at
suse
or,
however,
you
say
that
I
I
never
know,
and
I've
been
working
with
him
on
the
last
release
cycle
for
adding
a
setcomp
operator
which
is
a
kubernetes
add-on.
Basically,
that
allows
you
to
put
set
comp
profiles
on
your
kubernetes
nodes
and
then
apply
them
to
pod
workloads,
and
so
we're
probably
going
to
look
at
how
to
use
that
alongside
crossplane.
So
we'll
see
some
underlying
stuff,
like
some
of
the
sys,
calls
that
the
crossplane
controllers
make
and
that
sort
of
thing
and
applying
appropriate,
appropriate
profiles.
E
Nice
that
sounds
awesome
dan
with
the
sourcecraft
one.
You
know
that
seems
like
another
instance
of
a
distributed
application
that,
like
the
existing
helm
chart
or
whatever
deployment
of
it
into
a
kubernetes
cluster,
uses
in-cluster
services,
but
then
giving
them
the
option
to
use
managed
services
instead.
Similar
to
that,
I
o
for
the
red
hat
kris
have
been
working
on.
Are
we
seeing
any
patterns
start
to
emerge
there?
B
So
I've
have
noticed
that
it's
super
common
to
specifically
have
postgres
and
redis
as
the
managed
services
required
and
also
s3
buckets.
It
seems
to
be
extremely
common
actually,
but
you
know
in
general
those
are
things
that,
like
you
know,
they
kind
of
come
with
cross
plane,
there's
not
really
a
specific
packaging
up.
That
would
be
reusable
across
them
that
that
comes
immediately
to
mine.
The
one
thing
that
could
come
up
is,
you
know
with
composition.
B
Obviously
we
can
define
like
abstractions
across
them,
so
we
could
publish
kind
of
general
abstracts
and
we've
talked
about
this
before,
with
as
claims
get
phased
out,
publishing
with
cross
plane,
optional
kind
of
like
default
compositions
that
are
like
you're,
probably
going
to
want
to
use
these.
So
you
know
maybe
at
some
point
we
have
sort
of
like
a
postgres
redis
composition
that
we
publish
along
with
crossplane
or
we.
B
We
have
a
separate
repo
right
where
we
have
a
collection
of
these
that
are
really
easy
to
install
they're
kind
of
like
getting
started
compositions,
and
then
you
know,
folks,
like
source,
grapher
or
koi
could
could
easily
use
those
if
they
fit
their
use
case.
B
But,
along
with
that,
we
also
want
to
make
it
super
easy
for
them
to
be
able
to
write
customize
their
own
and
and
make
their
own
abstract
resources
that
are
then
satisfied
by
by
crossplane
primitives,
so
that
that's
something
that
kind
of
harkens
back
to
some
of
the
developer.
Tooling.
In
my
mind,
I
want
that
to
be
super
easy
for
folks,
like
sourcecraft,
to
be
able
to
add
that
functionality
to
their
helm,
chart
or
deployment
manifest
without
having
to
do
a
ton
of
work
or
have
a
lot
of
specific
donating
knowledge.
E
Yeah
that
makes
a
lot
of
sense,
dan,
it's
and
then
it's.
I
think
that
they're,
probably
right
that
there's
probably
less
patterns
of
reusability
on
the
consumption
side,
maybe
and
some
and
more
about
patterns
of
setting
up
the
api,
composition
and
stuff
like
that
for
the
provisioning
of
the
infrastructure
resources.
E
Continue
helping
the
community
and
ease
adoption
of
crossplane,
especially
as
they're
common
use,
cases
that
are
people
are
getting
real
use.
Out
of
that
sounds
really
cool
sweet
phil
you
added
composition
to
v1
beta1
back
in
the
milestone
section.
Did
you
have
an
update?
You
wanted
to
share
a
lot
with
that
or
just
wanted
to
get
into
the
doc.
C
B
C
In
addition
to
kind
of
the
animal
based
composition
that
we
have,
that
actually
would
be
able
to
execute.
You
know
you
know
javascript
or
whatever
languages
that
are
supported
there
and
effectively
expand
the
contents
of
a
composition
to
support
those
variable
resources,
but
that's
basically
been
open
discussion
and
then
also
the
the
home
provider
for
use
in
compositions
has
landed
the
design
for
that.
C
So
there
was
some
discussions
on
the
associated
issue
and
that's
now
been
approved
and
landed,
and
I
know
that
for
many
folks
who
wanted
to
you
know
have
like
a
cluster
config
for
like
eps
clusters
and
this
kind
of
a
thing
and
then
drop
some
post
provisioning
config
on
it
with
helm.
You
know
that
they
wanted
to
be
able
to
do
that
inside
of
a
composition
and
the
home
provider
is
moving
forward
to
support
that
so
kind
of
exciting
stuff.
There.
B
Yeah
one
thing
I
just
want
to
add
to
that
is
there's
been
by
hassan,
has
been
working
on
this
and
he's
actually
made
quite
a
bit
of
progress
in
the
crosslink
contrib
repo
and
I'm
going
to
grab
the
link
there,
because
I
don't
think
it's
in
there
where
that's
where
the
provider
helm
is
right
now.
So
if
you
actually
go
and
take
a
look
at
that,
he's
got
instructions
for
developing
locally
and
doing
all
that
kind
of
thing.
B
A
Yeah
and
when
that's
available,
we
should
definitely
make
some
blog
post
noise
about
that,
and
I
mentioned
to
people
that
functionality
and,
like
the
use
cases
around
it,
to
get
it,
get
it
out
there
all
right
cool
thanks
for
that
edition,
filling
in
yeah.
So
then
a
quick
cncf
update,
so
the
onboarding
has
been
going
on
for
after
our
acceptance
into
the
sandbox
project.
So
I
think
we
just
have
a
couple
more
things.
A
The
the
we
started
signing
the
and
following
up
on
the
trade
work,
trademark,
transition,
paperwork
and
legal
stuff
with
legal
stuff.
That's
gonna
take
a
little
bit:
it's
not
a
quick.
You
know:
email,
email
agreements,
it's
a
little
bit
more
to
it
than
that,
and
then
the
domain
transfer
as
well,
and
then
the
google
analytics
as
well.
I
think
it's
something
that's
actually
not
on
this
list,
but
those
are,
I
think,
the
few
scraggler
stragglers
that
we're
trying
to
to
turn
to
wrap
up
but
other
than
that
yeah.
A
That's
all
I
had
in
my
list,
yeah
so
other
than
that
things
are
going
smoothly
once
everything's
done
there,
and
you
know
we
are
being
treated
as
a
real
cncf
project
with,
like
a
you,
know,
kubecon
maintainer
booth
and
sessions
for
the
north
america,
one
and
stuff
like
that.
So
all
is
well
there.
A
Which
brings
me
to
kubecon
eu
so
friday
was
the
the
due
date
for
recording
of
your
talks
and
submission
of
those.
Did
everyone
get
those
in
phil
you
and
you?
You
all
got
your
zen
dan.
You
got
yours
and
me
and
harry
got
ours
in
as
well.
A
C
Totally
so
super
excited
about
the
talks.
They
were
great,
it's
always
great
to
kind
of
collaborate.
You
know
with
everybody
in
the
community,
so
yeah.
I
think
we
actually
got
like
almost
the
very
last
recording
slot
too
so
you're
advised
to
get
in
early
on
the
schedule
and
get.
C
A
That's
what
that,
but
I
one
of
my
better
takes
for.
I
don't
think
it
was
the
cross
playing
in
an
om
talk,
but
by
rook
talk
one
of
the
better
takes.
I
did
like
the
cathedral
next
door
like
the
bells
and
chimes
went
off
like
it
was
like
boom
boom
boom.
Oh
right
in
the
middle
of
a
good,
take
yeah.
C
Yeah
we
asked
if
we
could
re-record
in
general
and
they're,
just
like
you
get
the
one
shot.
A
Right,
that's
that's
the
pleasure
of
doing
recording
on
your
own
because
you
can
do
it
as
many
times
as
you
want
so
nice,
okay,
cool,
yeah
and
harry.
Thank
you
for
the
work
in
getting
everything
uploaded
for
our
talk.
Man
appreciate
that
yeah
no
problem,
sweet.
Okay,
then!
So,
then
the
we
get
a
booth
at
the
for
being
a
cncf
project.
A
We
get
a
booth
and
it
is
intended
for
maintainers
to
staff
it,
and
so
I
had
started
a
conversation
on
the
crossplan
slack
with
all
of
the
10
or
so
maintainers.
We
have
they
so
quick
note
on
this.
Is
it's
not
super?
This
is
the
first
time
they've
done
a
virtual
session
right,
there's,
definitely
lots
of
loose
ends
and
things
that
weren't
fully
finalized
and
smooth.
A
So
the
timeline
for
this
is
actually
really
really
short,
so
we're
supposed
to
be
signed
up
and
everything
by
wednesday
for
people
that
want
to
you
know
be
at
the
virtual
booth
and
talk
to
people
coming
by
and
answer,
questions
and
stuff
like
that.
We're
supposed
to
be
signed
up
and
registered
by
wednesday,
but
actually
since
they're
having
technical
difficulties,
they
we
need
to
get
people's
names
and
email
addresses
in
by
end
of
day
today
to
so.
A
If
you
want,
if
you
have
interest
in
staffing
the
cross-play
and
cncf
booth
for
any
time
period
at
all,
it's
not
really
a
set
schedule
that,
like
we
need
to
fill
out
completely,
but
we
do
want
to
get
some
good
coverage
for
it.
But
if
you
have
any
interest,
please
let
me
know
so.
I
can
get
your
email
and
name
in
by
end
of
day
today
to
keep
going
with
that
process
there.
D
A
Yeah
sure
ryan,
yeah
and
you
know
we've.
This
is
the
first
virtual
one
they've
done
so
we're
kind
of
making
some
some
assumptions
here
a
little
bit
on
my
part.
But
basically
you
know
they.
They
give
us
a
a.
A
You
know,
virtual
booth,
that
we
can
hang
out
in
for
the
entire
day
that
the
you
know
the
four
days
of
the
conference
monday
to
thursday,
and
we
can
hang
out
there
for
the
day
and
I
I
don't
know
how
they're
bringing
traffic
into
the
booth,
but
they
basically
have
you
know
all
attendees.
A
The
conference
can
come
through
and
you
know
start
start
conversations
and
you
know
ask
about
what
crossplane
is
and-
and
you
know
we
can
talk
about
everything
the
project
does,
and
you
know
what
we
do
with
ohm
and
all
that
sort
of
stuff,
and
so
it's
based
like,
oh
from
my
understanding
of
the
the
responsibilities
of
it,
it's
basically
just
being
in
conversations
with
people
that
are
curious
about
cross
plane
and
representing
crossbane
being
an
ambassador
of
the
project
to
the
community.
D
Okay,
actually
thanks,
I
think
in
that
sense
I
will
be
interested.
A
Yep,
that
sounds
perfect
right.
Thank
you.
Yes,
it's
normally
a
super
rewarding
experience
like
the
the
most.
By
far
the
most
I
learn
about
what
what
we
should
actually
do
with
crossplane,
which
direction
we
should
go
is
like
the
massive
number
of
conversations
you
can
have
with
users
in
the
community
and
the
ecosystem
as
a
whole
at
cube
cons,
it's
really
really
rewarding.
A
Well
and
phil,
I'm
hoping
you're
going
to
want
to
be
at
the
booth
as
well
too.
Absolutely
yeah,
nice,
nice,
the
jordan
there's
confusion
about
if
you're
already
registered
as
a
speaker,
which
a
lot
of
us
are
because
we
we
got.
We
did
a
great
job
and
got
a
lot
of
talks
this
time
that
it's
not
clear
if
you
need
to
also
register
as
a
booth
person
too,
but
so
just
if
anyone
who
does
have
interests,
though
just
give
me
your
name
and
email
address
and
I'll
send
it
back
to
the
support
case.
A
I
have
with
them
to
try
to
figure
it
all
out.
So
yeah
I'll
include
you
to
phil
cool
thanks
all
right,
awesome!
Okay,
so
I
think
that
was
the
end
of
that
section.
Did
anybody
have
other
agenda
topics
before
we
move
on
to
like
prs
and
other
technical
discussion.
A
Topics
all
right,
then
we
can
go
ahead
and
move
on,
oh
and
also
andy.
Thank
you
very
much
for
the
support
you
gave
to
get
our
demo
working
for
our
for
me
and
harry's
talk.
That
was
super
useful.
I
appreciate
your
your
timely
help
with
that.
A
Awesome,
okay,
so
yeah!
So
then,
so
we
don't
have
any
specific
prs
listed
in
the
regular
pr
section,
but
we
do
have
one
that
dan
added
to
the
deeper
technical
discussions,
and
so
this
is
a
completely
optional.
You
know
in
like
bonus
section
of
the
meeting
that
maybe
dan
there
are
some
specific
people
you
want
there,
but
otherwise
this
is,
you,
know,
kind
of
off
the
rails.
Technical
conversation
that
you
know
folks
are
not
required
to
to
be
at.
B
Sure
so
I
just
wanted
to
bring
this
up,
because
it
was
an
interesting
use
case
that
chris
here
is
working
on,
adding
the
ability
to
do
s3
bucket
policies,
and
this
would
be
a
new
alpha
resource.
B
Basically,
you
know
with
any
of
the
the
policy
type
resources
on
aws
that
you
pass
in
basically
a
raw
string
that
just
gets
passed
on
to
the
aws
sdk
and
then
to
the
api
to
create
the
policy.
You
know
it's
usually
json
or
yaml,
or
something
like
that.
B
What
krish
was
wanting
to
do
here
was
have
the
actual
policy
itself
be
structured
data,
so
you
know,
obviously
the
the
json
or
yaml
has
its
own
kind
of
structure
that
aws
requires
for
the
policy.
The
advantage
of
having
this
structured
in
the
the
resource
would
be
that
you
could
then
have
references
in
your
policy
itself.
B
So,
as
most
folks
probably
know
with
any
field,
you
can
have
a
additional
reference
and
selector
field
to
reference
other
kubernetes
or
other
cross
plane
types
that
then
you
know
a
field
gets
pulled
off
off
of
them
to
populate
the
value
on
the
object,
doing
the
referencing.
B
So
if
you
had
a
policy
that
you
wanted
to
apply
to
an
s3
bucket,
for
instance-
or
maybe
in
this
case
since
s3
bucket,
is
part
of
the
actual
spec
of
the
object,
it'd
be
more
appropriate
to
talk
about
the
iam
role
you
want
to
give
permissions
to.
You
can
create
an
iam
role
as
an
object
with
crossplane,
but
the
policy
itself.
If
it's
a
raw
string,
you
know
you're
not
obviously
not
able
to
reference
that.
I
am
policy
in
there.
B
So
there's
a
couple
different
ways
you
could
handle
that
you
could
have
some
sort
of
templating,
so
we
talked
about.
Well,
maybe
you
have
like
a
you
know:
it's
basically
like
a
go
template
and
then
you
reference.
I
am
role
in
the
body
and
then
we
automatically
put
that
in
there
for
you
or
you
could
actually
define
the
schema
of
the
policy
itself
and
since
you
have
a
schema
for
that,
you
could
add
reference
and
selector
fields
and
then,
basically
you
know
marshall,
that
into
the
the
result
that
you're
sending
off
to
aws.
B
So
there
is
a
few
kind
of
like
complications
with
this
in
the
way
that
aws
structures
these
policies,
specifically,
you
can
add
in
conditions
that
have
kind
of
like
custom
functions.
Basically,
the
aws
policies
are
their
own
templating
language
that
then
they
do
some
interpolation
in
to
make
things
happen.
B
But
anyway
we
just
went
kind
of
back
and
forth
on
this
because
we
didn't
want
to
remove
the
ability
to
literally
just
copy
paste
over
your
policy
as
well,
because
we
imagine
some
folks,
you
know
would
like
to
be
able
to
do
that.
And
if
we
have,
you
know
a
bunch
of
like
custom
fields
in
there,
then
that's
not
going
to
work
so
we're
still
kind
of
working
through
it.
B
I
I
think,
maybe
for
the
alpha
version,
we
may
just
say
just
pass
in
the
string
like
we've
done
in
other
places,
and
it's
not
going
to
be
as
configurable,
because
you
can't
use
references,
but
it's
going
to
be
more
what
a
user
would
expect.
You
know
if
they're
using
terraform
or
the
aws
api
directly,
so
yeah
that's
kind
of
the
idea
here.
The
other
option,
just
if
folks,
are
kind
of
like
evaluating
what
they'd
want
here,
is
to
have
basically
like
a
raw
policy
field
and
then
have
the
structured
schema
as
well.
B
So
you
could
kind
of
pick
and
choose.
That's
not
really
like
high
fidelity
with
the
aws
api.
So
you
know
there's
there's
trade-offs
to
all
of
these,
but
since
this
is
an
alpha
resource
and
chris
is
going
to
be
using
it,
it's
kind
of
a
nice
place
to
experiment
with
this,
and
because
we
have
a
really
flexible,
referencing
kind
of
strategy
within
crossplane.
We're
able
to
do
things
like
this.
B
So
while
there
is
kind
of
like
a
a
paradox
of
choice
here,
it
is
nice
to
see
that
we
can
do
things
like
this
so
folks,
especially
folks
that
use
crossplane
or
on
this
call
or
see
this
later.
E
Yeah
thanks
for
explaining
that
dan
and
I
I
would
definitely
expect
that
as
folks
start
to
use
it,
that
the
real
use
cases
that
they're
bringing
are
going
to
strongly
inform.
You
know
how
this
should
be
shaped.
I
wouldn't
be
surprised
if,
like
the
initial
implementation,
I
think
which
is
pretty
much
what
you
were
suggesting.
It
would
be
just
you
know,
taking
the
the
raw
policy
and
just
starting,
you
know,
being
able
to
set
that
as
a
field
on
the
crd
and
then
pass
that
directly
along
to
amazon.
E
So
there's
kind
of
a
principle
of
the
surprise
thing
going
on
there
and
people
and,
like
I,
like
the
potentially
the
easy
on-ramp
here
for
people
to
use
this
functionality
of
just
taking
copying
their
existing
policy
and
throwing
it
into
the
crd
and
having
it
just
work
without
any
sort
of
manipulation
or
new
concepts.
They
have
to
learn.
So
that
may
be
like
a
easy
acceptable.
E
You
know.
First,
implementation
of
an
alpha
resource
sort
of
thing
to
get
krish's
first
pr
in
and
have
useful
functionality
did
did
his.
So
that
seems
reasonable
to
me.
I
I
assume
that
longer
range
that
you
know
being
able
to
do
more
complicated
things,
and
you
know
referencing
fields
from
other
objects
and
stuff
like
that,
will
be
pretty
useful,
but
probably
not
in
the
first
implementation
did
chris
chris
have
from
his
particular
use
case
with
quay.
E
Did
he
have
like
a
strong
opinion
about
like
how
people
would
be
using
it
in
that
use
case?.
B
Yeah,
that
was
the
main
motivation
really
because
right
now,
the
s3
bucket
is
a
very
old
resource.
B
It's
probably
the
oldest
one
that
we
have
in
cross
plane
right
now
that
hasn't
been
refreshed
yet
so
it
needs
update
and
right
now
it
behind
the
scenes,
creates
basically
an
iam
user
and
then
propagates
the
connection
information
to
be
able
to
assume
that
to
be
able
to
use
the
bucket
back
from
like
when
we
were
doing
git
lab
demos
and
that
sort
of
thing,
and
so,
if
you're
using
that
resource,
you
need
to
basically
be
able
to,
or
even
if
you're,
not
if
you
want,
I
am
role
to
be
created
to
be
able
to
talk
to
this
bucket.
B
You
need
to
be
able
to
reference
that
right
if
you
want
to
do
this
through
composition.
So
if
you
want
to
be
able
to
create
like
a
a
koi
infrastructure
instance
that
creates
you
know,
postgres
redis,
s3
bucket
and
like
this
necessary
policy
and
that
sort
of
thing,
then
you
need
to
be
able
to
configure
this
this
policy
right
based
on
the
iam
role,
that's
created.
That
being
said,
you
can
get
around
this
once
once
the
s3
bucket
is
updated.
B
One
way
you
can
get
around
this
is
either
having
you
know,
the
name
of
the
iam
role
or
the
and
or
the
policy
surfaced
on
like
the
top
level
object,
so
you
can
actually
configure
that
and
it
flows
down
right
to
the
manage
resources
or,
if
you're,
just
taking
in
a
raw
string.
We
have
some
transform
functions
that
you
can
use
in
these
patches,
so
you
could
just
pass
in
just
like
you
know
the
I
am
role
name
and
then
that
could
pass
through
and
create
the
policy
with.
B
You
know
basically
like
a
formatting
function
in
there
to
replace
the
string
with
the
I
am
roll
name,
so
there's
a
couple
different
ways
to
get
around
it,
but
yeah.
This
was
definitely
motivated
by
the
work
he's
doing
with
the
koi
operator.
D
Helpful,
oh,
do
you
think
this
type
of
policy
document
could
also
then
be
used
in
the
im
roles,
and
I
am
policy
because
I
know
we
have
some
issue
right
now
around
trying
to
use
like
the
eks
clusters,
like
oidc
provider
for
service
accounts
and
propagating
that
info
from
the
cluster
seri
or
the
cluster
definition
to
the
role.
D
So
it
seems
like
this.
This
would
work
for
that
too.
B
So
are
you
talking
about
if
you
provision
an
eks
cluster
using
crossplane
and
then
and
then
propagating
that
information.
D
Yeah
to
a
I
am
role
across
playing.
I
am
role
because
right
now
you
need
that
oiac
provider
information
to
use
that
and
in
the
json
string.
That's
it's
pretty
hard
to
do
that,
but
it
seems
like
this.
This
type
of
definition
would
allow
that,
since
it's
more
concrete.
B
B
Perfect
awesome
yeah
I'll
circle
back
around
on
this
and
and
drop
some
comments
about
how
the
the
work
on
the
s3
bucket
may
be
related
there
as
well.
E
Awesome
thanks
for
for
weighing
in
on
that
kellen.
E
All
right,
yes,
so
then
I
think
that
was
the
last
item
here.
Did
anybody
have
anything
else
they
wanted
to
bring
up
before
before
we
adjourn
for
the
week.
E
Cool
so
yeah
thanks
everybody
for
joining
in
this
week,
and
we
will
find
you
on
slack
and
again
once
again
anybody
who's
interested
in
staffing,
the
booth
or
you
know
talking
to
users
and
such
at
the
for
kubecon
eu
virtual.
Please
let
me
know
and
I'll
be
happy
to
include
you.
Alrighty
thanks.
Everybody.