►
From YouTube: 2020-04-14 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
Okay,
the
recording
has
started-
and
this
is
the
April
14th
2020
cross
playing
community
meeting-
will
start
first
as
usual,
with
milestone
checkup
we
just
at
the
end
of
last
week.
We
did
a
set
of
official
version
releases
or
minor
releases
for
all
the
packages
that
the
cross
plain
community
maintains,
so
the
providers
for
AWS,
GCP
asher
and
then
the
sample
stacks
for
them,
and
then
the
WordPress
app
were
all
updated
to
the
versions
shared
on
the
screen.
A
Phil
is
updating
the
road
map
with
a
given
work
in
progress,
pull
requests
right
below
thanks
for
opening
that
fill,
and
that
will
capture
some
of
the
features
that
we
are
planning
on,
including
in
the
0.10
milestone,
which
we'll
talk
in
more
details
about
as
we
go
through
the
rest
of
the
agenda
today.
Dan
did
you
have
any
comments
on
on
the
provider
packages
releases
from
late
last
week
or
on
any
of
the
things
in
0.10,
not.
B
Too
much
we'll
talk
about
the
kind
of
reason
for
releasing
those
later
when
we
talk
about
backup
and
restore
one
difference
this
time,
and
it's
usually
not
a
huge
burden.
Releasing
these
specific
packages
it's
more
when
we
do
the
full
release
of
crossplane
and
then
the
corresponding
packages,
and
that
sort
of
thing
that
it's
more
complicated.
But
one
thing
that
did
make
it
easier.
This
time
is
that
we
are
moving
to
using
doxiadis,
though
for
the
documentation
for
CR
DS,
which
will
automatically
pick
up
our
new
versions
and
provide
documentation
for
them.
B
Whenever
we
cut
releases,
so
we've
been
able
to
remove
some
of
the
basically
manual
updates
on
the
release
branches
for
providers.
That
was
just
changing
like
version
numbers
and
that
sort
of
thing
the
tool
is
still
in
pretty
alpha
stages.
I'd
say
it
does
run
on
crossplane,
which
is
a
cool
dog
food
scenario.
B
A
Know
there
will
be
a
fair
amount
of
feedback
that
comes
in
from
you
know.
Having
that
be
the
API
documentation,
you
know
main
line
hosting
scenario
where
people
when
they
want
to
learn
about
the
types
that's
exactly
where
they
go
to
that's
the
only
places
publish
to
this
is
that
big,
great
way
to
drive
feedback
into
that
experience
as
well
Phil
any
comments
on
the
your
roadmap
update,
PR
or
are
straight
forward.
C
Yep
got
it
and
if
you
just
scroll
down
to
0.9,
so
you'll
see
all
the
stuff
that
we
got
done
there
of
note,
the
no
sequel
database
claims
are
coming
in
so
I
think
Archer.
You
got
the
viewer
open
for
that.
So
thanks
for
adding
that
kubernetes
app
resources,
a
piece
of
objects
to
the
remote
cluster
and
then
basically
we
got
like
a
new.
You
know:
selinux
style.
You
know,
restrict
core
API
groups
for
her
cross
playing
stack
manager,
kind
of
allows
us
to
lock
things
down.
C
Unless
you
want
to
open
those
up
and
then
basically
the
ability
to
you
know
schedule
directly
to
kubernetes
target
in
the
same
namespace,
so
that
if
you
have
clusters
that
need
to
get
populated
in
there
for
for
scheduling
those
that
that's
now
supportive
and
as
part
of
the
terminology
shift,
we
basically
added
a
package
type
to
the
optimal.
So
we
have
like
provider
stacks
apps
and
add-ons.
A
B
C
The
official
package
types
they're
a
ton
of
work
on
kind
of
getting
the
the
template
stacks
hardened
up,
as
well
as
support
for
private
repos
and
robotic
on
credentials,
as
well
as
a
huge
doc
overhaul
that
damn
legs.
So
thank
you
for
that
and
then
with
0.10,
basically
continuing
to
percolate
down
the
enhanced
terminology
for
providers,
tax,
apps
and
add-ons,
we
have
a
v1
beta
one
quality,
conformance
doc.
That
was
in
kind
of
a
final
review
right
now.
That's
a
really
great
reference
for
what
it
takes
to
make
controllers,
be
one
beta
one
quality.
C
We
added
backup,
restore
support.
So
if
you're,
using
tools
like
bolero
that
you
have
a
clean
backup
restore
with
the
ability
to
adapt
existing
external
resources,
which
is
super
important,
we're
looking
to
add
basic
versioning
and
upgrade
support
for
both
cross
plane
and
the
stacks,
as
well
as
a
resource
composition.
And
then
we've
been
working
on
an
experimental
support
for
Omega
types
in
the
cross,
plane
core,
and
so
you
know
cross
claims.
You
know
a
fairly.
C
A
Cool
thanks
for
getting
this
updated
here,
Phil
we'll
take
a
look
at
that
PRN,
add
feedback.
The
link
to
the
roadmap
update
is
in
the
agenda
document
here,
so
feel
free
for
folks
to
weigh
in
and
and
provide
any
comments
or
feedback
all
right.
Let's,
let's
move
ahead
into
the
community
topics
section.
So
the
first
one
on
the
list
here
is
is
from
Nick
who
is
not
able
to
attend
this
timeslot
that
we
have
normally
held
this
community
meeting.
A
So
one
know
as
we
continue
to
grow
the
crossplane
community-
and
you
know
we
have
more
interaction
and
collaboration
with
some
other
communities
as
well,
especially
with
the
ohm
community.
Did
we
in
order
for
Nick
to
be
able
to
join
these
meetings?
Who
is
driving
a
lot
of
initiatives
and
the
crossplane
project
and
then
be
kind
of
aligned
a
little
bit
as
well
with
what
my
understanding
of
the
time
slot
or
time
area
that
works
well
for
folks
would
like
the
own
community
meetings.
The
Nick's
proposal
is
to
move
this.
A
This
cross-play
meeting
from
9:00
a.m.
Pacific,
to
which
normally
is
to
10:30
a.m.
Pacific,
to
push
it
back
an
hour
and
a
half
I,
don't
think
we've
had
a
lot
like
a
ton
of
traction.
We
can't
you
know,
go
back
on
at
all
for
this
current
time.
Slot
I,
you
know,
think
that
this
probably
will
work.
But
if
anybody
has
a
major,
you
know
a
major
issue
with
that
new
proposed
time
slots
to
please
either
speak
up
now
or
on
the
crossplane
PRN,
which
Nick
has
made
this
proposal
official
on
14
21.
A
D
D
C
A
B
So
last
week
we
had
open
policy
agent
on
tauron
who's,
one
of
the
maintainer
zand
cofounders
of
the
project.
There's
a
really
great
episode.
The
project
integrates
super
well,
I.
Think
one
of
the
things
that
we've
heard
people
kind
of
talking
about
when
using
crossplane
is
how
to
do
permissioning
and
how
to
allow
different
people
in
an
organization
access
to
provisioning,
different
resources
and
OPA
fits
and
perfectly
for
that,
and
it's
really
easy
to
get
up
and
running
alongside
cross
plains.
B
So
there
is
some
documentation
about
how
you
can
do
that
they
came
out
as
a
result
of
the
show.
Also,
a
few
folks
in
the
OPA
community
are
looking
at
building
some
tighter
integration,
so
there
should
be
some
more
news
coming
there
if
anyone
on
the
call
is
interested
we're
also
looking
at
developing
sort
of
like
a
a
cookbook
of
common
OPA
policies
that
you
may
want
to
use
when
deploying
crossplane
and
we'll
continue
to
kind
of
develop
that.
C
A
E
B
So
our
back
is,
and
we've
seen
actually
some
of
the
architecture
of
cross
plane
kind
of
designed
with
this
in
mind
is
based
on
at
the
object
level.
So,
like
that's
one
of
the
reasons
why
we
had
a
separation
of
like
kubernetes
targets
and
Cuban
eyes,
applications
and
that
sort
of
thing
to
be
able
to
establish
control
on
who
can
decide
what
clusters
can
be
provisioned
to
in
that
sort
of
thing.
Opa
is
a
lot
more
granular,
so
you
can
write
pretty
much
policies
to
do.
B
Provision
from
this
cluster
with
size
greater
than
twenty,
or
something
like
that.
So
you
can
also
do
comparison
to
objects
when
evaluating
policies.
So
there's
a
lot
of
flexibility
there
and
that's
one
of
the
reasons.
I
think
we
want
to
develop
sort
of
a
cookbook,
because
not
all
that
is
apparent
right
off
the
bat.
So
if
we
had
some
good
examples
to
show
folks,
then
they
probably
provide
some
inspiration
for
writing
their
own
policies.
B
A
A
Gonna
see
and
I'll
tune
in
for
sure
cool
and
then
another
update
to
share
is
we
had
some
good
work
from
for
Marcus
added
crossplane
to
operator
hub.
So
now,
when
you
go
to
operator
hub-
and
you
know
you
look
in
different
topics
like
networking,
storage
and
things
like
that,
crossplane
is
now
available
there
as
one
of
the
choices
to
to
get
installed
in
you
know
open
ships,
clusters
or
in
other
environments
it
we're
using
operator
hub
to
kind
of
have
bringing
a
nice
ability
to
provision
and
manage.
E
A
There's
kind
of
two
different
things
here,
like
so
appropriate
hub
dot.
Io
is
kind
of
more
the
upstream
public,
accessible
way
that
you
can
use
operator
hub
from
any
kubernetes
cluster
cuz
like
when
you
click
install.
It
gives
you,
you
know
a
little
copy
thing
and
you
can
run
that
with
just
any
queue:
control
prompt.
It
doesn't
matter
what
the
underlying
cluster
is
and
then
this
packaging
format,
the
OLM
operator,
lifecycle
manager,
stuff.
When
you
do
the
packaging
work
to
show
up
on
operator
hub,
then
that
also
helps
you
natively
integrate
into
the
I.
A
Don't
know
what
I
can
remember
the
name
of
it,
but
there's
a
UI
and
a
control.
You
know
management
experience
that
is
on
open
chefs
clusters
specifically,
and
it
uses
the
same
thing
between
operator
hub
and
an
open
ship.
So
you
know
once
you
integrate
without
better
hub,
then
you
get
to
be
able
to
basically
get
included
in
the
native
openshift.
Ui
is
well
with
almost
no
work
extra
work,
I,
don't
think
so.
E
C
A
B
For
sure,
so
we
have
multiple
own
folks
on
the
call
today.
So
y'all
can
all
feel
welcome
to
weigh
in
as
well
of
course,
but
we
have
been
working
on
in
parallel
sort
of
add-ons
for
both
local
and
remote
scenarios
for
ohm,
so
the
remote
cases
where
you're
using
cross
plane
to
basically
provision
resources
into
a
remote
kubernetes
cluster.
It
was
either
provisioned
by
cross
plane
or
has
been
configured
to
be
accessible
for
cross
plane
and
the
local
scenario
is
where
you're
creating
things
in
the
cluster.
B
So
right
now,
it's
mostly
just
containerized
workloads
which
could
translate
to
deployment
so
that
either
gets
created
as
a
deployment
in
the
local
cluster
if
you've
enabled
the
local
add-on
or
it
gets
wrapped
into
kubernetes
application
and
deployed
to
the
remote
cluster
if
you're
using
the
remote
add-on.
We
also
recently
and
I,
probably
should
have
oh,
it
looks
like
it's
already
here:
we've
updated
the
service
tracker
app
to
be
compatible
with
the
new
spec,
and
this
is
actually
runnable
on
cross
plane.
B
A
Awesome
Dix
tin
for
that
updated,
I,
take
the
the
progress
in
both
you
know
the
own
community
and
in
general
and
then
across
plane,
implementation
of
them
spec
over
the
past
month
or
so
has
been
really
impressive.
So
I'm
really
happy
to
see
all
that
come
together.
It's
very
excited
to
see
those
demos
as
well
right.
D
D
This
is
first
one
second,
one
in
that
I'm
thinking
that
everyone
may
learn
to
add
some
more
demos
or
samples
for
the
own
support
part
like
a
wordpress
application,
and
maybe
we
can
think
about
a
role
to
trade
which
is
based
on
flagger,
or
we
can
even
do
the
detox
demo
based
on
that,
so
so
special
about
I
think
we
can
start
to
work
on
the
more
demos
and
documentation
for
the
support
on
call
spam
and,
of
course,
in
quotes
kind
of
repo.
So
it
will
make
better
for
users
to
understand.
D
Okay,
how'd
that
work
in
quotes
kind,
and
another
thing
is
that
we
may
want
to
add
a
documentation
about
the
data
model
of
a
course
can
I
found
these
things.
Actually,
today,
missing
from
the
coast
can
recall:
example
you
Petrik
many
project.
It
will
give
you
the
diagram
or
a
few
explanations
about
okay.
What
is
the
data
model
designing
this
project
I
feel
that
we
can
also
add
in
these
time,
where
I
mean
cost
that
bill
to
make
people
understand
better
about
okay,
how
close
can
use
its
own?
D
C
D
You
can
be
working
with
that
another
point
about
the
next
release.
I
want
to
add,
which
is
not
relevant
to
own.
That
I
still
feel
very
interesting
is
a
single
container
course
can.
I
I
just
think
are
very
used
for
user
cases.
For
example,
you
can
actually
deploy
deploy
content
at
a
single
container
again,
for
example,
Microsoft
SDI,
and
is
the
uncross
been
condemned
for
reading
Microsoft
cloud
services
as
low
as
a
ACI
based
application,
there's
even
no
kubernetes.
D
E
D
Ci
is
actually
a
service
container
instance
from
the
Microsoft
and
actually
every
cloud
have
these
kind
of
stories
like
far
gates,
like
Laura
and
for
Alibaba
card.
We
also
had
ECI
so
all
of
these
services.
They
are
not
they're,
not
kubernetes
best,
so
the
user
interface
is
a
container.
So
in
that
case,
if
you
can
deploy
cost
and
add
the
single
container
there,
you
can
have
very
cool
demo
to
see
okay.
Well,
you
can
use
any
style
API
on
the
platform
which
maybe
even
not
predicts
best
yeah.
D
A
Harriet
I,
like
that
idea
too,
like
we
had
had
an
issue,
we
opened
an
issue
for
that.
A
similar
type
of
idea,
where
you
know
the
packaging
of
CrossFit
could
be
such
that
it
carries
it
along
its
own.
You
know
kubernetes
api
server
entity
whatever
it
is
to
be
a
minimal
jelly
and
so
but
then
so
we
there's
an
issue
of
open
song.
A
That's
somewhere
I
may
be
to
the
500s
we
can
search
for
it,
but
but
I
like
the
application
of
that
idea,
that
you're
expressing
for
you
know
being
able
to
run
in
container
based
environments
that,
like
that,
the
cloud
fighters
support
like
far
etc.
That's
an
interesting
application
to
that,
because
you
know,
of
course
we
could
bring
along
everything.
It
needs
to
be
a
control
plane,
America
that
it
could
be
useful
and
you
know,
have
valuable.
A
D
A
E
So,
basically,
today
we
have
like
predefined
claims
like
MySQL
instance.
Kubernetes
cluster
claims,
with
the
composition
work
with
you.
What
we
enabled
users
is
that
they
will
create
a
like
the
infrastructure
definition
that
they
will,
like
you
know,
list
the
fields
basically
like
define
the
claim
their
own
claim
and
then
like
with
the
resources
class
like
compositions,
that
will
list
it
like
these
sources
that
they
want
to
bind
to
that
one
claim.
But
you
know
in
the
resource,
references
and
all
like
no
connection
secret
and
all
that
stuff.
E
So
basically
user
will
be
able
to
create
their
own
claims
if
you
IMO
instead
of
writing
Goku
and
it
will
be
able
to
bind
like
bind
to
more
than
one
resources
as
opposed
to
today.
So
there's
the
the
main
scenarios
that
we
want
to
enable.
But
that
is
that,
for
example,
when
you
have
a
DPC,
like
you
know,
BPC
setup
with
minimal
stacks
or
like
other
stuff
in
as
you
like,
when
you
have
to
create
a
V
net
role,
which
has
to
refer
to
the
MySQL
server.
E
But
you
don't
you,
don't
know
the
name
of
it
unless,
like
you
provision
it
beforehand,
so
in
that
case
you
could
actually
put
them
under
like
same
composition,
referring
to
each
other
and
then
one
column
will
create
both
of
them
like
referencing,
each
other.
So,
basically
like
it's,
it's
pretty
similar
like
the
idea
is
pretty
similar
to
type
of
modules,
but
in
cross
plane,
and
also
like
a
few
more.
You
know
actually
like
a
few
significant
things
that
will
make
applications
easy
easier.
E
You
know
make
it
easier
for
applications
to
bundle
the
infrastructure
for
themselves,
for
example
like
in
one
like
yeah
like
today.
If
what
application
can
still
like
include
the
claim
llamo
in
the
app
bundle
or
a
cat
an
chart
but
corresponding
stuff
is
like
fully
configurable
like
you
want,
you
can
find
it
too
masculine
server,
plus
Weiner
throat
or
you
can
just
bind
it
to
cloud
SQL
and
the
VPC
and
the
service
account
like
GCP
service
account,
all
three
of
them
and
all
that
stuff.
E
I
mean
I,
think
Nick
will
polish
it
and
like.
There
are
a
few
more
features
that
he
will
add
that
we
talked
like
any
gradient
will
need
to
add.
But
I
don't
know
the
timeline
for
him
so
but
like
we
have
started
implementation
so
like
we
know
what
it's
going
to
look
like
and
yeah
and
in
this
PR
the
definition
and
controllers
right
now,
it's
running
like
we
are
able
to
generate
so
your
actual.
There
are
like
examples,
Yama's
that
you
can
check
out
and
try
so
like
the
generation
part.
E
A
A
So
this
helps
a
lot
with
disaster
recovery
scenarios
so
be
able
to
take.
You
know
regularly
recurring
snapshots
and
backups
using
something
like
a
full
arrow
and
then,
if
the
control
playing
or
your
committee's
instance
or
whatever
it
may
be
dies,
then
you'd
be
able
to
restore
that
backup
and
get
back
up
and
running
and
have
cross
plains
still
appropriately,
managing
everything
so
for
a
reliability
perspective
and
operational
perspective.
That
was
a
huge
win,
I
think
for
kind
of
maturing.
E
So
there's
actually
so
the
implementation
changes
to
be
made.
They
also
enable
the
case
where
you
have
the
github
repo
as
usual,
but
like
since
you
don't
you
know,
you
don't
store
the
status.
Even
the
itself
is
gone,
you
could
you
would
be
able
to
like
still
apply
from
that
like
it's.
It
was
possible,
like
this
was
possible
before
too,
but,
like
you
can
now
do
this.
E
Like
you
know,
without
losing
or
like
you
know,
with
actual
manage
resources
there
with
no
status
or
like
you
know,
basically
some
assurance
about
like
the
IDS
and
all
that
stuff.
So
it's
it's
not
it's.
It's
more,
like
you
know,
adopting
Aventis
combinations
better
to
enable,
like
you
know,
other
other
scenarios
as
well.
A
Yep
good
point
and
then
and
then
there
is
also
a
one
outstanding
issue
for
in
terms
of
the
stack
manager
and
the
packages
that
it
can
install
like
provider,
AWS
or
just
be
a
sure
and
different
stacks
and
such
there
the
bookkeeping
of
those
installations
is,
it
was
functional
when
our
story
is
complete
from
a
backup
and
all
the
managed
services
and
external
entities
etc
will
be
functional
and
working.
But
there
is
some
of
these
instances
get
orphaned,
so
garbage
collection
becomes
a
bit
more
of
a
manual
process.
A
Instead
of
you
know,
uninstalling
a
provider
and
all
the
resources
to
getting
automatically
cleaned
up.
There
be
more
manual
cleanup
after
that.
Some
of
the
references
get
getting
orphaned
during
the
restore
process.
So
that's
still
an
outstanding
issue,
1389
that
is
being
tracked
and
will
hopefully
be
providing
a
fix
for
that
soon
as
well,
but
that's
as
far
as
I
know
the
only
outstanding
issue
on
the
backup
and
restore
or
scenario
being
fully
implemented
and
reliable
all
right.
A
All
right,
if
there
are
not
any
more
topics,
then
then
we
can
go
ahead
and
adjourn
for
the
week
here
and
we'll
see
everyone
online
and
and
then
at
the
other
community
meetings
like
own
community
meetings
as
well,
so
we'll
continue
collaborating
sweep
up.
Thank
you.
Thanks
everybody
for
joining
good,
to
see
everybody
yeah.