►
From YouTube: 2022-07-14 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
All
right,
hello
and
welcome
to
the
by
weekly,
I
think
it
is
still
cross
playing
community
meeting.
It
is
july,
14th
2022,
my
name
is
dan
mangum
and
I
am
standing
in
for
jared
today.
I
have
not
been
around
in
these
meetings
for
a
while.
So
it's
great
to
be
back
and
see
a
lot
of
familiar
faces
around
here
and
I'm
going
to
go
through
the
agenda
pretty
much
as
usual,
but
feel
free
to
jump
in
if
any
patterns
have
changed.
A
Since
the
last
time
I
hosted
here,
but
before
we
get
started,
I
always
want
to
give
folks,
especially
new
folks,
an
opportunity
if
you
like
to
to
introduce
yourself
and
if
there's
anything
you
need
or
anything
you're
interested
in
or
kind
of
reasons.
For
for
why
you
showed
up
feel
free.
If
there's
no
one,
then
we
can
move
forward,
but
I'll
give
folks
a
bit
to
speak
up
here.
If
you
like.
B
Sure
so,
I'm
joining
for
the
first
time,
I'm
shibashi
I
go
by
shibu,
I'm
coming
from
adobe,
so
we
have
been
kind
of
looking
at
cross
plane
for
meeting
some
of
our
needs
doing
some
analysis
working
with
the
aws
team.
I
know
they
have
been
joining
this
forum
and
they've
been
keeping
us
updated,
but
I
think
we
have
some
interest
here.
So
we
want
to
kind
of
come
to
the
session
track,
how
the
how
the
progress
is
happening
and
maybe
hopefully
ask
some
questions,
take
some
help
and
contribute
as
we
move
forward.
C
Yeah
I
can
go
next,
I'm
an
escapers!
I
work
for
shopify.
I
was
looking
for
crossland
to
at
least
help
replace
terraform,
because
cosplay
and
its
declarative
format
is
quite
looks
quite
nice
in
our
opinion,
but
yeah
this
first
time
joining
the
meeting
for
me
as
well.
So
I'm
just
gonna,
listen
in
the
background.
A
little
bit
maybe
ask
some
questions
and
yeah.
D
C
D
A
Awesome
welcome
nema
and
chris
chris
always
has
some
good
stuff
up
his
sleeve.
So
I'm
sure
it's
gonna
be
an
awesome
demo
super
excited
as
well
all
righty.
I
also
dropped
the
link
to
the
doc.
A
You
should
have
commenter
privileges
with
that
link,
so
feel
free
to
add
anything
to
the
agenda
as
you
see
fit
and
we'll
make
sure
to
accept
that
change
and
cover
it,
as
we
go
through
also
make
sure
to
add
yourself
to
the
attendees
list
up
at
the
top
under
july
14th
as
well
with
that
I'm
gonna
go
ahead
and
start
sharing
my
screen
here
and
we
can
get
into
the
agenda
alrighty.
So
first
thing
we're
gonna
do
is
do
a
quick
milestone,
checkup
on
some
recent
releases.
A
A
Yesterday
there
was
a
cross-plane
runtime
release.
I
I
think
there
was
a
few
updates
here
there
they're
linked
here
in
the
release
notes.
I
know
there
was
also
a
change
to
some
structs
that
I
believe
came
from
this
pr
here
and
it
had
an
impact
on
the
the
cross
plane
released
yesterday.
A
I
don't
think
it
should
have
any
impact
on
downstream
providers.
I
don't
have
full
context
here,
but
I
can
grab
that
from
nick
later
on,
but
just
a
heads
up
that
I
think
there
was
a
decently
significant
change
here,
so
feel
free
to
take
a
look
at
that
and
reach
out.
If
you
have
any
questions
about
that,
I
believe
it
was
actually
strictly
additive,
but
it's
resulting
in
adding
some
things
that
core
crossplane
is
not
going
to
honor
so
anyway.
A
I
believe
that
should
all
be
good
for
folks
that
are
consuming
that
on
the
provider
side,
so
that
releases
out
so
feel
free
to
update
to
latest
there.
A
There
is
we're
a
little
late
on
the
cross
plane
release
for
v.
1.9
nick
is
running
this
one
and
I
believe
it's
pretty
close,
but
there
were
a
few
issues
open
that
he
was
just
tying
up,
but
you
can
follow
along
with
the
steps
required
to
release
here.
So
you
see
that
we
already
have
some
of
it
in
progress
and
I
believe
nick
was
targeting
getting
this
out
today.
So
we
should
see
this
get
closed
out
by
end
of
day
today.
A
I
believe
yep
just
a
call
out
here
to
the
community
calendar
as
well
for
folks
who
want
to
make
sure
to
have
things
on
their
calendar
when
events
are
happening
related
to
the
crosslink
community.
So
that
would
include
this
meeting
as
well
as
any
special
sig
meetings
that
are
going
on
around
things
like
composition
or
packages,
and
things
along
that
so
feel
free
to
take
a
look
there.
A
Next
we'll
go
ahead
and
move
into
milestone,
planning
and
I'm
gonna
jump
over
to
the
road
map
board,
and
I'm
not
certain
that
we're
completely
up
to
date
here,
but
we
can
take
a
peek
at
some
of
the
things
that
were
first
included
in
v,
1.9
and
and
walk
through
those
a
little
bit
feel
free
to
jump
in
if
I'm
covering
something
that
you
worked
on
and
want
to
give
some
more
context.
A
But
I
know
we
got
the
one
pager
move
off
was
owning
for
plugable
web
hooks
merged
during
this
window,
and
so
it
looks
like
that
is
in,
and
I
believe
this
just
includes
just
the
design
dock.
No.
Actually
it
includes
as
well
integrating
into
the
package
manager.
So
this
might
actually
be
yeah
1.7,
so
this
one's
a
little
bit
old,
it
looks
like
a
number
of
these
are
a
little
bit
old.
A
Let's,
let's
move
over
to
some
of
the
in-progress
ones,
then
so
I
see
the
cross-plane
scalability
one.
This
has
been
top
of
mind
for
a
number
of
folks
here
and
this
one
specifically.
Let
me
go
ahead
and
pop
this
open,
see
if
there's
been
any
updates.
A
I
know
nick
has
also
been
driving
this
for
folks
that
aren't
familiar
with
the
effort.
It
is
related
to
how
the
api
server
performs
when
we
have
a
bunch
of
crds
installed.
A
So
there
is
a
number
of
things
that
just
went
in
recently,
so
let
me
find
the
pull
requests.
I
think
yesterday.
A
There
was
an
update
to
the
package
manager.
This
won't
be
in
v
1.9,
but
it
parallelizes
installing
crds
and
so
we'll
we'll
have
a
number
of
go
routines
going
there
and
adding
crds
to
the
cluster
during
package.
Install,
I
know
bulat,
who
is
a
this,
was
actually
his
first
contribution.
I
believe
so
pretty
awesome
contribution
for
our
first
one
here.
I
I
believe,
he's
doing
some
testing
to
reflect
some
of
the
the
improvements
here.
A
I
will
say
that
this
is
not
going
to
be
in
this
1.9
release
right,
so
this
was
merged
after
code
freeze.
So
keep
that
in
mind
when
taking
a
peek
at
any
of
this
and
I'll
hop
back
over
to
that
roadmap.
A
All
right,
I
see
these
two
ack
issues
in
progress
here
is
anyone
on
the
call
familiar
with
the
the
work
that's
going
on
there
if
so
you're
more
than
welcome
to
comment,
but
I
don't
have
context,
I
don't
think
move
off
is
in
this
week,
so
I'll
refer
to
anyone
else
who
wants
to
give
an
update
if
possible,.
A
Alrighty,
well,
we
will
leave
that
one
off
for
today
and
we'll
let
move
off
circle
back
around
and
provide
any
updates
and
slack
on
those
if
necessary.
A
All
right.
I
also
see
custom
compositions
here,
so
I
will
hop
back
over.
Take
a
look
at
that.
I
know
a
lot
of
folks
on
this
call
are
familiar
with
custom
composition
and
we
can
see
if
there's
any
updates.
A
This
is
one
of
our
most
desired
issues.
I
think
here
and
so
I
know,
there's
been
some
progress
along
this
since
nick
isn't
here,
I
don't
think
there's
much
to
update,
although
I
will
say,
there's
been
some
really
nice
kind
of
like
experience,
reports
or
things
that
folks
would
like
around
this
functionality.
A
I
know
this
is
top
of
mind
for
folks
it
did
not
land
in
1.9,
so
I
am
guessing
we'll
need
to
bump
that
over
to
1.10
here,
but
does
anyone
have
any
updates
specifically
on
this
or
any
thoughts
or
questions
about
the
custom
composition,
effort
right
now.
C
So
how
does
this
different
from
normal
compositions?
I
haven't
looked
at
the
issue
itself:
okay,.
A
Yeah,
so
I
can
give
a
very
quick
overview,
but
essentially
the
normal
compositions,
or
you
will
or
compositions
by
default,
have
a
pretty
constrained
language,
which
is
by
design
we're,
definitely
not
looking
at
introducing
like
control
flow
and
some
of
those
things
into
the
composition
vernacular.
A
A
What
custom
composition
allows
you
to
do
is
kind
of
bring
your
own
functionality
as
needed,
and
there
is
an
open
design
for
how
to
do
this.
I
believe
I
don't
know
if
this
is
closed
out
right
now,
but
you
can
take
a
look
look
at
it.
It
is
linked
from
the
pr
for
more
information.
A
Implement
a
mechanism
for
out
of
tree
external
secret
stores,
so
it
looks
like
we
have
a
linked
pull
request
on
this
one.
But
let's
go
ahead
and
pop
it
open
and
see
where
it
is
at.
A
So
external
secret
store
support
has
been
added.
It
is
an
alpha
feature
and
I
believe
it
was
included
in
1.8.
So
it's
been
out
for
some
time
I
might
have
been
in
1.7.
Even
what
this
is
looking
at
is
making
this
plugable
so
making
it
such
that
you
can
support
kind
of
arbitrary
external
secret
stores.
A
I
know
there
is
a
one
pager,
that's
still
open
here
and
it
looks
like
it
does
have
approval
from
nick
and
muvafik.
So
that's
two
cross-plane
maintainers,
so
I'm
guessing.
This
is
in
a
pretty
good
state
here.
But
if
folks
are
interested
in
this
feel
free
to
take
a
peek,
did
anyone
have
any
questions
or
comments
about
the
plugable
secret
store
model
here.
A
Last
one
here,
it
looks
like
support
patching
from
common
data
sources.
A
A
I
don't
think
we
have
any
updates
on
this
one
currently,
but
I
know
it's
top
of
mind.
I
don't
know
if
the
custom
composition
work
is
going
to
impact
this
at
all.
It
does
look
like
they're
yeah,
so
there's
a
little
bit
of
guidance
on
how
you
can
work
around
some
of
this.
A
Looks
like
this
one
was
closed
here
so.
A
Let's
see
if
there's
any
yeah,
it
looks
like
it
was
just
closed
in
favor
of
the
aggregate
issue.
So
I
don't
see
any
movement
on
this
one
right
now,
but
I
don't
have
full
contact
so
once
again,
if
anyone
wants
to
weigh
in
on
that
feel
free.
A
So
let's
go
ahead
and
take
a
look
at
some
of
the
highlights
for
1.9
and
we
can
just
pop
these
open
and
quickly
look
through
them.
So
this
one
is,
I
believe,
referencing
the
paralyzing
install
effort
that
we
were
just
talking
about.
A
So
I
already
ran
through
that.
But
if
folks
want
to
take
a
look
at
how
that
was
implemented,
basically
we
just
split
off
a
few
go
routines
when
installing
and
do
it
in
parallel.
So
that's
been
updated,
shout
out
again
to
blue
lot
for
that
great
contribution.
A
All
right,
so
it
looks
like
this
one
was
fixed
if
you
have
circular
dependencies
within
a
given
resource
and
I
believe
that's
actually
what
the
the
pr
we
looked
at
earlier
was
covering
here.
A
So
I
don't
think
any
of
the
folks
who
worked
on
this
are
in
office
this
week,
because
most
of
them
are
on
holiday
in
turkey.
But
I
believe
this
does
give
you
some
policies
to
be
able
to
resolve
references
here.
There's
some
pretty
good
detail
in
the
pr
body.
But
if
anyone
here
wants
to
add
any
more
context,
this
one
feel
free
as
well.
E
Yeah
so,
for
example,
in
provider
aws,
it's
now
possible
to
reference
security
by
itself.
In
the
security
group,
for
example,
we
have
some
needs
from
the
guys
who
are
using
aws
glue.
For
example,
then
you
need
security
groups
for
self-referencing
the
same
security
groups,
and
I
think
we
also
include
this
crossplan
runtime
in
the
latest
provider
aws
release
as
well.
So
it's
possible
now,
since
let's
release
an
aws
provider.
A
Awesome
thanks
for
that
chris
and,
as
we
mentioned
at
the
beginning,
we
do
have
a
formal
release
of
this
out
now,
with
0.17
on
cross
plane.
Runtime
so
feel
free
to
pin
to
that
one.
If
folks
need
to.
A
We'll
see
what
we
have
here,
patch
a
field
into
multiple
array
items,
so
it
looks
like
this
one
was
completed
by
john
edge.
So
thank
you
to
john
for
some
great
work
here.
I
don't
have
a
ton
of
context
on
this
one,
but
it
looks
like
the
issue
has
a
pretty
good
description.
A
A
Yes,
I
saw
this
one
as
well
looks
like
there
is
now
support
for
regular
expressions
in
patching.
I
believe
it
was.
A
Yep,
so
you
can
use
regular
expressions
now
in
string
transformation
in
your
patches.
Let's
take
a
look
at
this
one.
It
looks
like
it's
merged
and
yeah
once
again,
pretty
self-explanatory.
Just
the
ability
to
use
catcher
capture
groups
in
your
your
string
transforms
here
pretty
useful
for
things.
It
looks
like
the
canonical
one
that's
being
used
across
a
couple
of
these
is
doing
it
with
orange
from
aws,
which
can
certainly
be
useful.
A
Alrighty
so
once
again
that
1.9
release
should
be
coming
out
pretty
soon
here.
Hopefully,
today
now
we're
going
to
jump
into
provider
releases
and
it
looks
like
we're
starting
off
with
provider
aws
christopher.
I
imagine
you
have
some
pretty
good
context
on
this
one.
Is
there
anything
you
wanted
to
cover
in
this
release?
A
E
Think
the
best
thing
was
what
we
introduced
was
from
manabu
he's
also
in
the
call
I
think
he
implemented
that
we
can
detect
and
partition
and
regions
when
you're
using
service
account
authentication
in
non-aws
partitions,
for
example,
it's
possible
now
with
the
provider
and
unblocks
a
few
folks.
E
This
was
very
nice
and
other
things,
I
guess,
was
a
lot
of
things
fixing
and
adopting
things
after
our
latest
sdk
updates
that,
for
example,
lsd,
cache
version
handling
is
not
possible,
and
so
on.
A
Yeah
awesome
yeah
those
sound
like
really
really
important,
updates
there,
especially
the
one
with
the
partition
in
region
and
irsa.
So
thank
you
all
for
the
great
contributions
there
looks
like
this
was
a
number
of
folks.
First
contributions,
so
shout
out
to
all
of
y'all.
Did
anyone
else
want
to
give
any
other
updates
on
this
provider?
Aws
release.
A
A
Anyone
involved
in
this
release,
who
wants
to
give
an
update,
looks
like
yuri.
You
might
have
had
a
contribution
here.
D
Yep,
so
it's
a
very
small
release.
It's
basically
driven
by
one
of
our
bound
customers-
and
I
stabilized
couple
of
resources
to
unblock
our
customer
and
one
of
the
nested
status
field
is,
is
a
change
that's
spanning
over
all
resources?
It's
basically
when
you
have
a
part
of
the
spec
like
frequently
measure
its
identity
and
it's
like
an
inline
set
of
fields.
We
had
an
issue
that
this
inline
set
of
fields
was
not
properly
properly
populated
in
status,
so
there
is
no
way
to
automatically
consume.
For
example,
identity
automatically
generated
principles.
D
A
Awesome
thanks
yuri
sounds
great
alrighty
and
then
it
looked
like
we
might
have
had
a
provider
jet
aws
release
as
well.
If
I
can
get
this
pulled
back
up.
E
Here,
oh,
I
think
we
had
no
release.
I
think
the
community
wanted
to
have
the
release,
but
at
the
moment
we
have
or
the
the
maintainers
have
no
time
to
cut
the
release.
I
think
the
last
release
is
from
march
and
we
had
a
lot
of
prs
merch
since
march,
and
also
a
few
open
things
to
get.
For
example,
securities
security
group
and
security
group
rules
running
with
the
provider,
and
we
wanted
to
have
the
prs
reviewed
merged
and
cut
the
new
release
in
the
provider.
E
A
Cool
awesome,
I
assume
that
some
of
those
folks
are
our
folks
who
are
out
this
week
in
turkey,
but
I'll
make
sure
to
surface
that
as
well.
Thanks
for
the
the
incremental
availability
there
chris,
that
sounds
super
useful,
at
least
so
folks
can
test
it
out.
At
the
very
least.
A
All
righty
so
with
the
community
topics
and
updates
here
it
looks
like
we
have
a
podcast
with
victor,
so
victor
making
the
rounds
as
usual.
This
looks
like
a
a
pretty
good,
deep
dive
here.
I
think
they're
talking
a
little
bit
about
both
cross
plane
and
it
looks
like
just
cross
plane
focus.
So
that
sounds
awesome
specifically
looking
at
get
ops,
which
I
know
is
a
important
use
case
for
a
lot
of
folks.
A
So
I
haven't
gotten
a
chance
to
listen
to
this
one,
but
I
usually
always
check
out
anything
that
victor
is
involved
in
and
it
looks
like
it's
a
pretty
lengthy
one
too.
So
I'm
sure
there's
a
lot
of
great
stuff
here
and
appreciate
the
transcription
as
well.
That's
awesome.
A
Let's
take
a
look
at
that
very
nice
styling
here,
so
it
looks
like
sasha
is
giving
an
overview
of
infrastructure
self-service.
So
it
looks
like
this
is
both
practical
and
philosophical
to
some
extent.
So
nice
awesome,
it
looks
like
goes
all
the
way
into
packaging,
and
things
like
that.
This
looks
super
useful,
nice.
A
Alrighty
have
one
with
gke
and
crossplane,
also
by
the
way.
If
anyone
who
wrote
any
of
these
posts
is
on
the
call,
please
feel
free
to
speak
up
and
I'll
definitely
defer
to
you
to
cover,
but
it
looks
like
managing
a
gke
cluster
from
crossplane
on
k3d,
so
kind
of
that
local
cluster
managing
the
cloud
cluster
there.
So
this
looks
also
quite
practical
as
well,
so
definitely
encourage
checking
that
one
out.
A
Yes,
I
I
did
read
this
one.
I
thought
this
was
a
great
one
talking
about
some
multi-cloud
scenarios
with
crossplane
plane,
and
so
it
talks
a
little
bit
about
having
abstractions
and
being
able
to
target
underlying
different
concepts
goes
through
talks
about
xrs
xrc's.
Pretty
extensive
post
here
would
definitely
recommend
for
folks
they
even
go
into
talking
about.
You
know
how
we're
using
oci
images
and
that
sort
of
thing
and
the
benefits
of
that
so
definitely
recommend
this
one.
A
Alrighty-
and
I
think
this
one
was
just
published
recently
from
autodesk
and
jesse-
wrote
this
one
up
and
there's
this
is
based
on
a
talk.
I
believe
that
jesse
and
nick
gave
a
while
back.
This
looks
like
it
might
be
actually
a
transcription
here,
but
there's
some
really
interesting
information
included.
A
I
haven't
gotten
a
chance
to
read
through
this
one,
but
I
am
familiar
with
the
talk
comment
that
was
at
the
aws
open
source
symposium,
which
sounded
like
had
some
really
neat
stuff.
A
All
right
looks
like
chris,
you
added
this
one.
Did
you
have
anything
to
add
on
this
one.
E
It
was
in
the
community
that
he
wrote
the
book
about
intent,
automatization
with
kubernetes
and
crossplane,
and
he's
looking
for
folks
who
want
to
read
the
book.
You
can
ping
him
in
the
community
and
you
get
the
book
for
free
and
you
can
add
your
thoughts
about
this
awesome.
A
All
righty
so
some
updates
on
lfx
mentorship
program.
So
I
am
more
familiar
with
the
second
than
the
first
and.
A
So
it
looks
like
this
is
just
linking
to
the
content.
I
don't
have
much
to
share
here,
but
I'm
sure
jared
and
move
off
could
give
an
update
as
well
as
their
mentee
as
well.
I
don't
see
any
of
them
on
the
call
here,
but
basically
the
idea
here
is
just
to
detect
whenever
we
have
a
breaking
change
in
a
crd
from
version
to
version
and
do
that
in
an
automated
fashion.
A
So
you
don't
have
to
you
know
parse
through
release,
notes
and
things
like
that,
and
then
the
one
that
I've
been
a
little
more
involved
in
here
is
verifying
that
our
pulling
from
private
registries
is
working
well,
so
I've
been
working
with
jared
in
parole
here
and
parole
has
done
some
awesome
work
getting
back
into
our
cross,
plane
testing,
so
I'll
go
ahead
and
open
this
up.
A
So
we've
had
some
kind
of
like
periodic
tests
that
have
run
against
crossplane
and
various
providers
for
a
period
of
time,
but
they
got
stale
over
a
while
and
parole
has
actually
gone
in
and
started
to
update
these,
and
one
of
the
things
we
want
to
do
is
get
back
to
verifying
upgrade
paths
from
different
provider
skus.
A
And
so,
if
we
look
at
something
like
this
conformance,
it's
basically
checking
from
some
version
to
another
version
and
this
test
framework
we'll
be
able
to
run
just
those
upgrade
scenarios
for
each
of
these,
and
so
parole
is
getting
these
kind
of
back
into
good
health
here
and
also
adding
support
for
testing
against
a
private
registry.
So
that's
like
if
you're
using
ecr
or
google
artifact
registry
or
docker
hover,
what
have
you?
A
We
have
different
machinery
that
kind
of
honors
the
different
authentication
mechanisms
for
those
various
registries,
and
we
want
to
make
sure
that
that
doesn't
break
over
time.
We
rely
a
lot
on
go
container
registry,
the
underlying
library-
and
it
has
had
some
changes
over
time
that
has
broken
folks.
Also,
just
auth
in
general
is
not
standardized
as
part
of
the
oci
distribution
spec,
so
there
are
some
some
intricacies
to
interact
with
all
of
them.
A
Basically,
what
we
want
to
do
is
make
sure
that
we're
not
breaking
any
any
components
when
we
are
updating
that
library
or
other
parts
of
the
package
manager,
so
parole's
been
doing
an
awesome
job
leading
that
effort.
A
All
right
so
I've
seen
some
stuff
coming
by
here
with
the
fuzz
testing,
but
I
am
not
super
in
tune
with,
what's
up
to
date,
once
again,
nick
jarrod
and
move
off,
I
believe,
have
the
most
context
here,
even
though
I
am
tagged
and
I've
been
not
paying
too
much
attention,
but
it
looks
like
this
has
been
merged
at
this
point,
and
so
it's
basically
just
someone
working
on
adding
fuzz
testing
to
crossplane,
which
is
obviously
just
going
to
increase
some
some
quality
here.
A
A
All
right,
so
I
think
the
first
thing
we
have
up
here
from
a
demo
perspective
is
christopher,
so
I'll
go
ahead
and
stop
sharing
and
let
you
take
it
away.
E
E
Oh
okay,
okay:
this
is
my
first
code.
Okay,
I
had
a
call
or
talk
with
manabu
in
the
community
about
how
we
can
we
generate
in
our
company
our
composition,
for
our
attendance
on
the
clusters
and
for
a
few
resources.
So,
like
ec2
instances,
efs
user
kms,
rds
s3
buckets
whatever
we
generate
our
compositions
out
of
this
drds
from
the
provider
perspective.
E
So
that
means
I
can
show
it
afterwards,
more
detailed.
So
in
general
we
say:
okay,
we
want
to
generate
in
our
group
s3,
for
example,
aws
dkb
cloud.
We
want
to
have
the
name
bucket
and
we
want
to
use
from
the
provider
perspective
for
one
beta
one
resource,
and
we
have
here
in
our
workspace
the
provider
in
dot
work
directory,
and
here
is
the
path
to
the
crd
from
the
provider,
and
then
we
want
to
create
a
composition
named
bucket.
E
E
So
it
took
a
few
seconds
and
if
you
see
now
the
composition
and
the
definition
is
completely
auto
generated
out
of
the
crd
from
the
provider.
So
you
can
see
here
now
the
api
for
screen
for
three
schema
completely
auto
generated
out
of
the
crd.
So,
like
everything
and
also
we
add
a
few
things
afterwards.
So
let
me
go
to
the
end
of
the
file.
So,
for
example,
we
also
adding
objects
for
observing
things
and
that
we
can
get
the
whole
status
from
manage
resource
to
the
claims.
E
E
We
normally
have
on
the
claims
and
yeah
patching
all
of
our
managed
resources
with
this
and
then
the
interesting
part.
Now
is
the
parameter
things
because
we
generate
also
the
patch
set
for
all
of
the
available
fields
from
a
resource
that
we
have
everything
available
in
claims
and
can
patch
every
field
through
the
managed
resources
yeah.
E
So
you
can
see,
then
the
resources
here
all
the
patches
are
applied
here,
yeah,
the
other
thing
we
can
do
here,
for
example,
if
you
check
or
if
someone
changes
today,
a
composition
thing
we
can
also
have
here.
A
second
composition
call
it.
I
don't
know
bucket
for
two.
If
we
have
changes
under
the
hood
and
then
let
me
check
if
we
delete
the
things
here
and
make
generate
again.
E
Then
you
can
see
that
we
also
generating
know
the
new
composition
with
the
new
version
and
then
it's
available
for
tenants
to
test
our
new
composition.
We
have
because,
with
the
versioning
at
the
moment,
there's
a
bit
it's
a
little
bit
difficult
to
test.
Otherwise,
so
we
generate,
then
everything
on
the
new
composition
name,
what
we're
doing
for
this?
At
the
end,
we
have
a
small
goal,
tooling
available
to
fetch
the
things
here.
We
we
have
also
the
options
to
override
fields,
skip
fields
for,
for
example,
for
rds
instances.
E
If
we
find
a
field
called
kms
key,
we
we
also
generate
directly
also
the
k
s
key
in
the
compositions
and
making
the
reference
selector
together
automatically.
Without
that
tenants
needs
to
know
this,
and
then
let
me
check,
we
have
a
few
json
that
things
available,
so
we
have
a
few
functions.
E
All
of
our
tagging
things
we
want
to
have,
and
then,
at
the
end,
it's
yeah,
it's
a
jsonet
thing
where
we
get
all
things
together
and
gen
and
render
all
the
things
you
see
in
this
make
generate
calls
yeah
and
afterwards
we
have
also
tooling
available
that
we
directly
test
everything
what
we
generated
here
so
out
of
the
providers
we
using
the
examples
and
generate
then
the
examples
here
directly
to
use
it
in
our
cuddle
testing
suite,
and
we
can
confirm
that
our
generated
composition
is
really
yeah
usable
and
after
that
the
tenants
can
use
it
in
our
clusters.
D
Can
you
sorry
go
ahead?
I'm
sorry.
Can
you
customize
these
labels
or
these
labels
hardcoded
this
one
yeah,
this
yeah.
E
I
think
we
have
a
few
things
available
for
this.
We
need
to
find
it.
No,
I
think
we
have
to
generate
functions
available
at
the
end.
We
have
here
all
of
our
controlling
labels
and
generic
labels,
and
the
good
thing,
for
example,
is
if
we
introduce
a
new
one,
so
I
don't
know
crossplane
community
yeah.
We
add
this
here
and
then
we
very
simple
make
generate
again
and
then
you
can
see
now
that
all
of
our
composition
are
updated
directly
and
we
can
push
it
through
the
repositories
and
also
this
update.
D
Pretty
cool
this
is
really
interesting.
I
have
a
few
questions.
One
is
a
more
high
level
question
and
feel
free.
If
you
don't
want
to
answer
it,
but
I'm
just
curious
what
was
the
motivation
behind
building
this
tool.
E
So
in
general,
the
motivation
was
that
we
have
at
the
end
five
or
six
people
in
the
infrastructure
team,
but
400
developers
requesting
resources
from
us,
and
I
think,
with
the
first
resources
we
starting
generating
or
copy
pasting,
our
own
compositions
together,
and
then
we
found
out
that
I
don't
know
with
the
second
composition.
We
have
the
same
patch
sets
like
okay.
These
are
our
basic
labels,
our
annotations
whatever,
and
that's
why
we
started
this
and.
E
In
general,
you
have,
I
don't
know
more
than
400
namespaces
available
on
each
cluster
and
I
don't
know
they
can
pick
whatever
they
want
in
from
there
from
the
available
compositions.
Okay,.
D
Yeah,
it's
it's
interesting
because
you
know
we
are
working
with
also
with
folks
who
are
interested
in
single
cluster
multi-tenant,
setups
and
and
this
idea
of
wrapping
composite
wrapping,
managed
services
in
compositions
came
up
in
conversations
before,
and
it's
good
to
know
that
you're,
not
the
only
ones
actually
looking
at
you
know,
apartment
potential
solutions.
So
a
very
cool
other
thing
that
I
wanted
to
ask
you:
what
are
the
configurability
options
that
you
have
available
like?
D
E
Yeah,
what
we
can
do
today
is,
let
me
check
again
we
can
skip
fields
directly
in
the
generate
yummy.
So
if
you
know
that
the
name
of
the
fields,
then
there
is
a
then
you
can
skip
it,
then
it's
not
configurable
from
the
for
the
customers,
so
normally
we're
doing
the
things
with
rds
databases
or
with
resources
where
we
need
to
talk
or
thought
about
security
or
governance
related
things,
so
chems
key
encryption
whatever,
then
we
hide
those
fields
for
our
tenants
and
we
auto
generate
the
things
then
together.
E
So
we
found
out
that
in
provider
aws
in
the
early
phase,
it
turns
out
that
there
are
fields
available,
but
the
fields
are
not
in
the
crds,
so
we
decided
only
to
use
the
crds
because
at
the
end
this
are
everything
I
can
configure
in
the
resource.
So
why
not
using
vcrds?
Because
I
have
not
more
available.
D
E
Yeah,
the
good
thing
is
in
the
make
files.
We
have
the
versions
of
the
providers
configurable
and
then,
if
we
update
the
provider
we
see
directly
after
the
generation
what
happened
with
all
of
our
compositions
and
then
we
can
decide
if
they
are
breaking
changes
for
our
customers.
And
then
we
can
have
a
look.
What
we
want
to
do
or
if
it's
safe
to
go
forward.
D
Okay,
very
cool,
so
my
last
question
and
sorry,
I'm
just
taking
too
long
with
like
bombarding
him
with
questions,
but
this
is
super
cool
and
we
are
working
on
something
similar,
but
any.
What
is
the
prospect
of
open
sourcing
it?
Potentially?
I
believe
the
community
would
be
interested
right.
Obviously
you
are
you
have
users,
we
have
users,
there
are
other
folks,
probably
out
there
any
chance
that
it
would
go
into
like
a
cross-plain
community,
or
do
we
have
such
an
art
at
all?
I
think
we
do.
E
Right
we,
we
are
open
to
open
source
this.
I
we
had
a
few
talks
today
in
the
company
about
this,
it's
enough,
so
why
not?
At
the
end,
we
had
not
need
at
the
moment
to
open
source
it,
because
we
don't
thought
about
that.
Other
folks
wanted
to
use
this
approach,
but
no
we
had
to
talk
about
this,
so
I
think
we
need
to
do
a
few
assignments
in
the
company
that
we
can
do
it.
It
takes,
I
think,
one
or
two
weeks,
and
then
we
can
open
source
it
in.
D
A
D
A
B
A
God
I'll
come
up
all
right.
I
just
wondering
you're
cloning
down
the
repo
right
to
source
the
crds
from
there
yeah
right
yeah.
I
was
the
one
another
option
right
that
I
think
would
be
an
interesting
thing
to
explore,
because
I
mean
it's
somewhat
similar,
but
you
it
might
be
lighter
weight.
A
Is
you
know
in
the
like
integration
tests
for
all
the
providers
we
started
using
that
xp
extract
command,
which
you
can
point
at
any
package
and
that'll
just
like
spit
out
the
full,
a
single
file,
in
that
case
that's
gzipped
and
you
could
source
them
that
way
if
you
wanted
to
which
would
allow
you
to
go
straight
from
the
registry.
I
love
that
you're
using
the
crd
manifest
directly
instead
of
the
ghost
trucks,
though,
because
I
think
one
like
that's
the
ultimate
form
right.
A
So
it's
it's
good
to
target
that
and
two,
if
you
don't
have
access
to
the
the
source
right,
then
it's
nice
that
you
can
do
something
like
pull
the
image
and
yank
those
out
of
there.
So
love
what
you're
doing
there.
B
For
many
contexts,
many
contexts
and
specifically
when
you're
talking
about
multi-tenancy,
I
just
wanted
to
kind
of
ask
a
few
more
questions
around
the
context
you
have.
You
said
you
have
a
few
compositions
that
you
define,
which
is
used
by
multiple
teams.
Do
you
have
a
good
idea
of
how
many
compositions
you
are
dealing
with
right
now
on
your
clusters?.
D
E
We
have
so,
I
think,
yeah,
I
I
shared
it
a
little
bit
so
one
two
three
four:
five:
six
and
10
11
12
12
round
about
12
compositions
at
the
moment
available
for
our
customers
on
the
on
the
clusters
and
okay,
I
would
say
at
the
moment
we
handling
roundabout
per
cluster
5000
to
10
000,
manage
resources
from
this
com;
drive
competition.
E
For
so
in
general,
in
our
company
we
have
as
the
main
thing
provider
aws
running,
but
we
have
also
a
few
custom
providers
running
like
providers
ted
ba
for
set
scaler
vpn
stuff,
and
I
think
the
other
one
is
pagerduty,
but
yeah
we're
working
on
this
also
to
bring
the
compositions
available
for
our
customers.
D
B
Okay:
okay,
fair
enough!
So
do
you
apply
them
dynamically
to
the
cluster?
How
does
that
process?
Look
like
one
of
the
things
that
was
a
good
idea
is
kind
of
pulling
it
from
the
zip
files
can
automate
the
whole
end
to
end.
Do
you
apply
it
means
from
with
some
cadence
or
you
have
automated
end
to
end?
How
do
you
kind
of
run
this
in
a
continuous
way.
E
Okay,
I
will
shortly
answer
this.
I
think
in
general
we
are
packaging-
all
of
our
tenant
composition
together
to
configuration
packages
bringing
the
packages
to
ecr,
so
the
oci
things,
and
then
our
approach
on
the
clusters
is,
we
have
running
flux
and
flux
will
pick
up
the
new
version
and
roll
out
anything
and
the
rest
is
handled
by
crossblank
core
to
bring
out
all
the
things.
B
E
A
Awesome
right
on
chris,
that
was
awesome
as
a
reminder
for
folks.
These
videos
go
up
on
youtube
afterwards,
so
as
soon
as
they're
available
we'll
get
it
up
there
on
the
crossplane
account.
It
sounds
like
a
lot
of
folks
are
interested
in
this,
as
you
should
be
so
that'll
be
a
way
to
share
it
out
if
you're
interested-
and
I
wanted
to
one-
I
saw
jesse
jumped
in
here-
jesse
just
wanted
to
give
you
a
shout
out.
A
We
we
talked
about
your
post
there
a
little
bit
earlier.
Oh.
A
Yeah,
that's
in
the
the
agenda
there
and
then
also
craig
is
on
the
list
here
and
craig
opened
up
a
discussion
on
cross
plane
around
success
stories.
So
craig,
do
you
want
to
comment
a
little
bit
on
that
and
I'll
pull
it
up
here
for
folks
to
be
able
to
see.
C
Yeah
thanks
dan
and
hey
everyone.
It's
my
first
time
joining
a
crossblank
community
call
I'm
a
product
manager
at
upbound,
but
basically
I
just
wanted
to
promote
visibility
for
a
proposal
that
I'm
trying
to
spearhead,
where
I
think
the
community
would
benefit
if
we
had.
You
know,
broadcasted
out
some
success
stories
and
basically
case
studies
of
people
who
are
successfully
using
crossplane
in
production
kind
of
like
the
the
autodesk
item
that
was
on
the
agenda.
A
couple
bullet
points
up.
C
We
would
have
more
of
those
and
they'd
come
out
under
the
cross,
plane,
blog
and
maybe
link
back
to
the
main
crossplane
website,
so
that
folks,
who
are
you
know,
investigating
cosplaying,
can
see
these
use
cases
and
that
sort
of
stuff.
So
if
you
know
someone
who
fits
this,
this
description
or
you
yourself
qualify
for
such
a
case
study.
You
know
reach
out
to
me
and
and
let's
talk
and
see
about
getting
case
studies
up.
A
Awesome
thanks
craig.
This
looks
great
also,
I
love.
I
love
the
I
feel
like
the
discussions
on
github
are
like
the
most
underused
feature.
These
are
super
nice,
so
there's
also
a
number
of
like
really
good
questions.
Folks
ask
in
here.
So,
if
you're
ever
interested
take
a
peek
all
righty,
so
we
do
have
about
five
minutes
here
and
we
have
a
few
prs
to
discuss
or
that
need
attention.
So
I'll
open
up
this
one
here,
it
looks
like
a
number
of
other
folks
have
as
well
and.
E
I
can
I
can
make
it
very
short.
I
bumped
in
the
provider
jet
aws
the
provider
from
terraform
for
aws,
it
turns
out.
I
have
running
it
locally,
so
with
an
existing
resources
from
the
old
versions.
E
So
I
employed
a
few
things
and
then
you
see
that
for
every
resource
you
get
an
error
message
that
the
following
required
provider
is
not
installed,
and
I
have
definitely
no
idea
at
the
moment
what
I
need
to
do
if
I
simply
make
run
the
updated
provider
locally
on
my
machine
so
because
normally
I
not
install
the
providers
or
something
like
this.
I
have
no
ideas
and
I
need
a
little
bit
guidance
from
the
guys
who
are
familiar
with
the
jet
providers.
E
A
Gotcha,
I
am
not
the
the
person
to
ask
on
this
necessarily,
but
I
will
say
I
have
messed
with
the
jet
providers
a
little
bit
and
what
I've
done
in
the
past
to
kind
of
like
figure
out
what's
going
on
when
I'm
trying
to
run
locally,
is
just
jump
into
the
controller
or
the
image
building
for
the
controller
and
look
at
what
they're
doing
in
here
and
kind
of
try
to
like
map
over
to
that.
A
This
might
be
useful
to
like
examine
the
paths
here,
but
that's
kind
of
the
best
I
have
I
don't
know
yuri.
Would
you
have
any
insight
here
or
you're
the
closest
person
I
could
think
of
to
this.
D
Yeah,
it
looks
like
so
the
situation
is
the
following,
so
you
chris
already
have
kind
of
checked
out
an
isolated,
terraform
workspace.
We,
which
reference
is
the
old
version
of
underlander
from
provider.
After
upgrade,
you
already
have
a
machinery
which
points
to
a
new
version,
but
there
is
no
automated
way
to
actually
transfer
the
workspaces
underlying.
D
It
looks
like
at
least
in
this
case.
It
didn't
work.
I
I
didn't
see
exactly
this.
The
same
situation
before,
but
well.
The
safest
way
is
like
they
let
orphan
and
recreate,
but
it's
obviously
ugly
right
as
in
in
any
kind
of
real
real
environment.
But
if
you
need
a
fast
workaround
yeah,
they
let
policy
or
fun
is
the
way
it
should
recreate
the
stuff,
but
definitely
something
to
to
debug
ping
me
afterwards.
So
I'm
interested
to
look
into
this
one.
It's
it's
something
crucial
for
potential
further
upgrades
yeah.
Thank
you.
E
Be
cool
to
catch
up
together
to
to
share
a
little
bit.
The
knowledge
of
this
would
be
great.
A
Cool
sounds
good
awesome.
Thank
you
thanks.
Folks,
thanks
for
bringing
that
up,
all
right
looks
like
we
have
one
more
here
and
just
in
time.
Let's
take
a
peek
at
that.
I
can.
A
E
We
we
learned
that
the
hard
way
the
last
week
with
the
latest
versions
of
crossplane-
it's
not
possible.
If
you
have
two
composite
field
paths
in
a
patch
set
that
this
is
working,
so
you
can't
do
two
composite
field
paths
in
a
patch
set
and
the
problem
is
then,
if
you
patch
the
things
to
the
resources,
you
investing
a
lot
of
time
and
you
see
nothing,
no
debug
message
like
nothing
and
the
only
thing
is
to
make
the
two
composite
field
pass
patches
without
a
patch
set.
Then
everything
is
working.
A
Gotcha
gotcha
is
this:
is
this
a
regression
a
confirmed
regression
or
is
it?
Was
it
not
working
beforehand?
I.
E
Think
one
one
guy
added
in
this
issue
list
that
is
working
within
1.4
version
of
crossplane,
but
not
the
newest
one,
so
yeah
we
also
coming
from
one
four
and
updated.
I
think
the
last
few
weeks
to
one
five
one:
six
in
production
and
then
we
figured
out
that
all
of
our
two
composite
keyboards
not
working
anymore
a
few
years
it
yeah,
so
it
would
be
cool
if
someone
has
an
idea.
What
this
is
to
take
care
of
this.
A
Yeah
yeah,
absolutely
no.
This
looks
like
a
critical
issue
here.
It's
all
surfaces,
especially
with
our
new
release
going
out
today.
Hopefully
we
might
need
to
do
a
quick
round
of
patches
unless,
if
we
can
immediately
identify
this,
this
is
super
helpful
though
I
don't
know
if
callum
is
on
the
call,
but
thank
you
for
narrowing
it
down.
A
That
obviously
makes
it
much
easier
to
find
out
what
went
wrong
there
so
appreciate
that,
just
in
time
all
righty
yeah,
perfect
timing
appreciate
everyone
showing
up
today,
especially
folks
who
are
newer.
Thank
you
again,
chris
for
the
awesome
demo.
I
love
when
we
can
show
stuff
off,
and
the
community
meetings
also
appreciate
y'all
letting
me
run
this
meeting
again
definitely
have
missed
being
in
these,
so
it
was
great
to
see
all
y'all
and
thank
you
for
all
the
impact
you
have
in
the
crossland
community.