►
From YouTube: 2019-09-03 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
All
right,
the
recording
is
started,
and
this
is
the
September
3rd
2019
crossplane
community
meeting.
We
have
a
big
milestone.
It's
coming
up
very
soon,
so
we
will
dive
into
a
bit
of
a
review
and
status
update
on
the
0.3
milestone
and
to
see
where
we
are
with
that
and
their
status
and
risks
associated
with
that.
Let
me
bring
up
the
participants
list
here
to
see
who
all
is
on
the
call.
A
It
looks
like
we're
missing
a
couple
of
Engineers
that
have
items
assigned
to
them
in
0.3,
but
we
will
we'll
do
our
best
to
get
an
accurate
status.
So
I
took
a
pass
this
morning
through
the
0.3
milestone
and
tried
to
clean
up
some
of
the
issues
there
and
get
things
into
a
the
current
status
as
reflected
in
reality,
and
so
one
of
my
biggest
observations
from
going
through
the
0.3
milestone
this
morning
was
that.
A
There
are
a
lot
of
the
number
the
quantity
of
items,
issues
on
the
board
right
now
is
pretty
high
or
in
the
milestone,
but
there's
when
I
took
a
dive,
a
deeper
dive
into
them.
I
noticed
that
about
10
of
them
are
actually
epics
the
things
that
the
epic
level,
which
you
know,
have
associated
features
with
them,
but
don't
themselves
necessarily
have
work
associated
with
them.
So
these
epics
here
are,
you
know
they
have.
You
know
a
spot
in
the
milestone,
but
the
real
work
is
for
them
as
being
tracked
by
all
these
individual
issues.
A
That's
one
observation.
The
next
observation
I
had
is
that
the
bulk
of
the
engineering
work
here
is
around.
You
know:
resource
connectivity,
bottling
networking,
connectivity,
types,
reach
of
the
cloud
providers
and
being
able
to
you,
know,
provision
and
configure
those
networking
types
using
crossplane
in
order
to
create
a
securely
connected
example
of
application
talking
to
a
database,
6,
etc.
That
was
a
huge
focus
for
this
milestone,
and
so,
although
have
pool
requests
open
in
are
converging
towards
getting
ready
to
merge
semester
and
the
that
covers,
probably
amongst
those
three
poor
requests
across
the
three
cloud.
A
So
the
quantity
of
issues
on
the
board
really
concerns
me
this
morning,
but
these
hurt
in
terms
of
where
things
are
in
progress
and
how
we're
tracking
things
I
am
NOT
I
feel
fairly
fairly
confident
about
where
we
are
and
how
many
things
are
about
to
close
soon
I
mean
yeah.
We
can
always
have
the
discussion
about
work
in
progress,
limits,
etc,
but
you
know
the
we
can
do
that
as
a
retrospective.
This
may
be,
you
know,
after
the
milestone.
A
A
B
A
Man-
okay,
let's
I,
don't
see
Javad
on
the
call
to
give
us
an
update
on
AWS,
but
I
I
did
see.
The
pull
request
over
the
weekend
hear
that
Javad
has
all
of
the
days
like
seven
or
eight
different
types
to
represent
resource
connectivity
in
AWS,
because
AWS
doesn't
like
necessarily
to
do
things
very
in
a
simple
way,
but
Javad
has
controllers
and
types
for
all
of
those
now
and
I
believe
since
they're,
following
the
manage
reconciler
pattern
that
they
are,
you
know
fully
featured
as
well.
A
With
all
the
crud
operations
I
saw
the
so
Nick
provided
us
a
feedback
on
the
pull
request
and
shibata's.
It's
creating
it.
I
think
the
big
thing
that's
missing
here
for
AWS
is
as
test
still
so
we
may
need
to
do
some
some
shuffling
around
to
be
able
to
pick
up
the
tests.
The
tests
work
for
AWS,
because
there's
a
strong
possibility
that
Javad
is
going
to
be
out
the
rest
of
the
week
and
maybe
next
week,
someone
next
week
as
well
for
a
wedding
that
he's
not
only
involved
in
it's
his
wedding.
A
C
Absolutely
so
all
functionality
has
been
implemented
that
PR
is
still
in
draft,
because
I
have
to
manage
reconciler
tests
that
I
am
finishing
up
this
morning,
so
I
just
want
to
make
sure
that
wasn't
merged
before
those
were
completed.
But
it's
ready
for
review.
There's
a
couple
of
follow-up
things
that
I'll
probably
open
issues
on
that
can
be
addressed
in
the
future,
just
in
terms
of
how
we're
doing
some
of
the
defaulting
and
that
sort
of
thing
for
a
sure.
But
as
far
as
the
network
connectivity,
it's
all
been
tested
and
ready
to
go.
A
A
No,
it
does
not
look
like
he
is
so
I'll
give
an
update
for
the
Stax
CLI
and
his
wordpress,
the
wordpress
back.
So
daniel
has
implemented
a
command-line
interface
experience
or
viq
control,
plug-in
to
initialize,
build,
publish,
install
and
uninstall
a
a
stack
which
is
great
to
have
that
full
it
like
a
full
experience
there
to
you,
know,
get
it
off
the
ground
and
running
and
then
develop
it
and
publish
it
and
then
actually
install
it
into
a
cluster
as
well.
A
So
that
is
a
good
progress
to
feel
like
really
good
about
that
there
and
the
wordpress
stack
is
fully
implemented
and,
I
think
really
what's
left,
there
is
taking
that
wordpress
stack
and
you
know,
working
with
across
the
rest
of
the
cloud
provider
implementations
to
get
the
documentation
default
resource
classes.
You
know
everything
there
to
be
a
experience
all
around.
So
that's
kind
of
gonna
be
a
bit
of
a
shared
responsibility
to
make
sure
that
we
are
updating
our
documents
and
our
examples,
etc.
A
So
we
can
make
it
as
smooth
as
awesome
for
the
0.3
release
and
then
also
make
sure
we're
not
duplicating
work
across
those
as
well.
Since
you
know,
Susskind
wants
to
have
again
there
shouldn't
have
again
in
that,
as
well
being
the
owner
of
the
wordpress
stack.
So
it
seems
to
be
in
a
good
place.
Marcus
you
had
a
big
PR
merge
on
Friday.
Do
you
want
to
give
us
an
update
about
the
stack
manager
sure.
B
Like
a
resources
demo,
a
group
yanil
UI
schema
gamal
to
become
annotations
in
the
deployed
CRD
so
that
when
a
stack
has
been
installed,
the
the
CR
DS
that
are
managed
by
that
stack.
You
can
get
a
little
more
information
about
them
about
the
stack,
that's
that's
controlling
them,
and
so
that
leads
to
probably
the
next
item
in
here,
which
is
it's
a
comma
list.
I
think
on
your
bulletins
right,
yeah,
packaging
annotation
is
insecurity
which
leads
us
to
that.
A
B
A
A
Hopefully
they
meet
in
the
middle
right,
awesome,
okay,
so
then,
let's
so
that's
a
full
status
update
for
what
you
know.
Everyone
is
working
on
and
this
year
about
three
released
here
as
we're
trying
to
converge
and
and
finish
by
the
end
of
this
sprint.
The
I
found
identified
some
of
the
risks
here
for
the
milestone.
I
see
two
fills
on
the
call.
Now
we
double
our
fill
productivity.
That
would
be
awesome.
All
right,
I'm
gonna
assume
that's
the
plan
of
record
going
forward.
A
Two
fills
all
right,
so
the
the
first
risk
on
here
is
that
you
know
he
had
wanted
to
update
all
of
the
infrastructure,
a
cloud
provider
stacks
to
pour
to
their
controllers
to
using
the
managed
reconciler
pattern.
Some
of
those
have
been
updated,
but
that
is
a
lower
priority
than
the
resource
connectivity,
functionality
and
user
guides
experience,
etc.
The
porting
reporting
of
those
controllers
is
a
non
function.
Change.
It
does
not
have
user
impact,
so
it
is
a
lower
priority
compared
to
the
other
ones.
A
So
I
think
that
you
know
there's
definitely
going
to
be
some
of
those
controllers
to
do
not
get
poured
it
in
a
0.3
milestone
the
CIC
D
pipeline
examples
a
fill
key.
Do
you
want
to
speak
to
that,
because
I
I
don't
have
a
strong
understanding
of
what
you
know?
What
would
be
necessary
to
be
able
to
do
those
sort
of
examples?
I
have
a
feeling.
A
A
A
Back
down
it
too,
you
I
felt
you
could
keep
working
on
your
audio
there
and
we'll
try
to
sync
up
with
you
as
soon
as
you
get
that
back
up
to
date,
I
have
a
feeling
that
you're
working
on
getting
the
right
microphone
connected
since
you
have
multiple
microphones
in
your
setup,
all
right.
So
the
move
to
the
next
item.
For
now
the
documentation
and
user
guide
updates.
Those
are
those
are
declared,
those
as
a
risk
as
well,
because
the
scope
of
it
is
fairly
broad
and
wide-reaching
we
want
to.
A
We
had
had
Dan
hashed
in
here
drove
a
an
ish
ative
or
last
couple
a
couple
weeks
ago
around
you
know
getting
discussion
going
for
what
would
be
the
entire
overhaul
of
our
documents.
Documentation
user
guides
would
be
they've
been
in
the
same
state
they've
been
in
since
the
initial
you
know,
taking
the
repo
public
back
in
2018,
zero
got
one
release
was
so
there's
a
whole
lot
of
improvements
that
we
identified
there.
That
we'd,
like
to
kind
of
you,
know,
incorporate
and
improve
overall
the
entire
documents
and
user
guide,
initial
experience,
etc.
A
So
the
scope
is
kind
of
broad,
but
we
do
have
the
the
right
attention
and
resources
on
that,
as
well
as
each
engineer
that
owns
one
of
the
cloud
provider
stacks
finishes
there
networking
connectivity
that
that
will
be
a
main
focus
for
them
as
B
as
updating
the
user
guides
and
making
sure
that
that
is
a
smooth
experience.
So
we
have
the
right
resources
on
it,
but
scoping
it
to
the
set
of
attainable
changes
that
make
some
measurable
improvements
and
progress
forward
guides.
A
E
A
E
I
mean
I
think
that
that
issue
goes
into
a
bit
of
detail
on
that.
But
the
basics
are,
you
know,
pipeline
examples
for
like
Jenkins
pipeline
examples
for
gitlab
standard
and
then
like
a
get
ops
example,
maybe
like
we've
works,
and
so
what
that
would
look
like
is
effectively
I'm
like
a
getting
started.
You
know,
doc
that
says:
hey
like
here's,
three
popular
forms
of
CI
CB-
and
you
know
this
is
basically
what
it
looks
like
to
use
cross
playing
in
each
of
those
environments.
A
E
I
think
it's
something
that
could
be
entirely
self-contained
cause
plane
repo,
and
it's
literally
just
you
know
like
here's,
you
know
what
your
pipeline
look
looks
like
you
know,
first
Andrew
pipeline
and
Jenkins,
and
then
this
is
you
know
the
the
one
you
know
we're
two
commands
that
you
drop
in
a
script
and
here's
what
the
script
supplier,
here's
the
individual
cross,
plane
commands
and
often
to
give
you
integration.
So
I
think
that
that
would
be
pretty
straight
ahead
to
do.
Okay,.
A
Then
that
sounds
good
yeah
definitely
is
I
would
identify
it
still
as
a
risk,
since
it
hasn't
been
addressed
yet,
and
you
know
the
up
thing
or
user
guy,
it's
a
documentation,
I'd
say
is
probably
a
higher
priority
than
that,
since
that's
it's
so
user
facing,
and
you
know
what
the
primary
experience
that
people
have
with
cross
plane.
So
it
said
yeah
I'd
say
it
definitely
declare
it
still
as
a
risk.
A
You
know
all
at
the
same
time
and
I
think
that
we're
going
to
want
to
do
a
retrospective
about.
How
can
we
improve
that
process
where
we
have
blog
posts?
You
know
throughout
the
milestone
on
a
regular,
recurring
cadence.
Instead
of
you
know
all
coming
in
at
the
end,
so
I
think
that
you
know
it'd
be
good
to
you
know,
keep
the
excitement,
keep
the
momentum.
E
Totally-
and
we
were
discussing
that
a
little
bit
last
week
towards
the
end
of
the
week
and
the
thinking
was-
is
we
could
do
a
blog
post
that
kind
of
announced?
You
know
0.3
and
you
know
kind
of
like
covered
all
that
all
the
aspects
like
not
in
any
great
depth.
A
lot
of
that
will
be
in
the
updated
kind
of
vision,
doc
forever.
B
E
So
it'll
just
kind
of
call
out
the
highlights
and
maybe
kind
of
announce
that
it's
available
and
then
here's
where
you
can
go
to
learn
more,
but
then
still
wanting
to
do
those
more
in-depth
blog
posts
and
potentially
having
those
go
out.
You
know
it
was
like
a
follow-on
to
the
announcement
that
kind
of
dives
into
a
little
bit
more
detail.
Oh
yeah.
A
Yeah
that
makes
sense
cool
okay,
so
that
was
everything
that
I
had
on
the
milestone
itself.
Just
recap:
here:
there
are
a
lot
of
items
that
are
still
in
the
milestone.
A
lot
of
them
are
epics.
A
lot
of
them
are
in
review
right
now,
as
well
for
the
cloud
provider
stacks
and
their
resource
connectivity.
A
Such
there
are
some
items
of
risk
that
we've
called
out
here,
but
in
general
we
are
converging
towards
being
able
to
release
I'm,
hoping
that
with
a
big
push
here
on
getting
the
resource
connectivity
into
master,
you
know
updating
the
stacks
manager
with
the
final,
the
security
model,
etc,
and
then
a
hue
pass
and
push
on
the
user
guides
and
documentation
and
experience
for
crossplane
that
will
have
a
solid
0.3
to
put
out.
Hopefully
you
know
on
the
on
the
timeframe
that
we
were
hoping
for
is
well
at
the
end
of
next
week.
A
Alright,
so
I
wanted
to
give
a
quick
update
on
the
stacks,
the
clutter
stacks
themselves.
So
all
three
proud
writer
stacks
exists.
Now
they
all
have
their
separate
repos.
Every
all
functionality
has
been
split
out
into
those
individual
repositories
and
I
have
links
here
for
little
install
guides.
If
you
want
to
take
those
for
a
test,
run
right
now,
there's
a
note
here
that,
because
the
main
cross,
flane
repo,
has
not
been
updated
yet
to
remove
the
cloud
riders
from
that
repo.
A
If
you
wanted
to
test
out
these
stacks
for
a
particular
cloud
fighter,
then
you'd
need
to
install
crossplane
from
source
locally
from
a
fork
that
has
the
cloud
providers
removed.
That
will
obviously
not
be
a
requirement
at
the
end
of
the
milestone.
But
if
you
want
early
testing
and
to
take
these
cloud
Rider
stacks
for
a
little
a
little
test
bin,
then
yeah,
you
will
need
to
build,
install
across
plane
from
source
right
now.
The
where
the
next
step
is
the
official
builds
build
pipelines
via
Jenkins
etc.
A
E
Totally
so
at
a
super
high
level,
you
know
we
basically,
you
know
had
behavior
before
that
allowed
you
to
have
like
portable
claims
and
then
do
class
of
service
selection
using
kind
of
the
name
of
the
generic
claim,
who's
your
classes
fund,
and
so
that
was
really
great.
E
But
one
of
the
side
effects
of
strongly-typed
is
that
it
made
the
the
class
ref
in
a
claim
bound
to
a
specific
provider
like
GCP
or
AWS.
And
so,
if
you
omit
that
that
class
ref,
then
you
know
you
can
rely
on
the
single
default
class
of
service
that
you
know
like
the
administrators
created
for
you
in
your
namespace,
and
so
what
this
proposal
does
is
says.
You
know
if
you
were
to
add,
or
probably
even
replace,
class
ruff.
Now
that
were
strongly
typed
with
a
policy
ruff.
E
Then
we
get
the
level
of
indirection
to
the
claim
kind
without
being
bound
to
a
specific
provider
by
the
class
refs.
You
know
API
version
and
and
kind,
and
so
it
basically,
you
know
kind
of
keeps
that
that
capability
to
both
have
portability
for
the
app
owner,
as
well
as
allowing
them
to
select
the
class
of
service
that
they
want
from
the
available
classes
of
sir
administrators
created
in
an
environment
right.
A
E
So
this
this
used
to
be
handled
all
in
kind
of
the
generic
resource
class,
because
there
was
a
single
one,
and
so
since
it
was
generic,
it
served
both
as
you
know,
kind
of
like
a
generic
way
of
referring
to
a
class
of
service.
And
then,
when
we
introduced
strongly-typed,
it
basically
pushed
out
the
strong
typing.
All
the
way
to
the
claim
and.
A
E
A
C
So
we
would
just
update
that
with
a
like
a
core
v1
object,
reference
kubernetes
object
reference,
and
that
would
be
the
provider
ref,
just
like
there's
a
class
ref
and
it
would
be
optional
and
then
it
would
also
involve
updating
the
default
class
wreckin
sylars
to
instead
of
looking
for
a
single,
a
single
policy
in
a
namespace,
and
if
there's
more
than
one
saying
you
know,
I
can't
reconcile
this.
There
can
only
be
one
policy
in
this
namespace.
It
would
look
for
a
a
set
well
if
it
gave
a
provider
reference
on
the
claim.
C
It
would
look
for
that
provider
reference
or
that
excuse
me
policy
reference
in
that
namespace
and
if
it
emitted
a
policy
reference,
then
we
would
have
this
kind
of
label
on
one
of
the
policies.
That
would
say
this
is
the
default
if
no
policy
is
specified
and
that
would
allow
you
to
still
get
that
omission
of
a
class
ref
and
a
policy
ref
and
just
fall
back
on
whatever
the
administrator
has
defined
got.
A
C
Exactly
and-
and
it
actually
turns
out
really
really
well
and
Phil-
had
a
great
proposal
there
of
doing
this
in
a
simple
way,
because
essentially,
what
policies
do
are
they're
taking
that
kind
of
like
class
of
service
and
just
creating
it
as
a
middle
layer
as
opposed
to
having
it
integrated
into
the
resource
class?
So
we
get
all
the
benefits
of
the
strongly-typed
resource
classes,
while
also
being
able
to
have
these
classes
of
service,
and
so
you
can
define
with
a
policy.
C
You
know
this
is
what
my
post
grows
large
instances
and
you
can
change
that
around
without
deleting
all
of
the
resource
classes
you
have
so,
let's
say
you
have
a
cloud
sequel
class
and
a
cloud
sequel,
Postgres
a
RDS
post
grows
and
you
can
basically
decide
which
one
of
those
is
your
large
without
deleting
and
recreating
a
new
one
that
has
the
name
postgrads
large
or
something
like
that.
So
I
think
it's
a
it's
a
much
better
kind
of
like
stronger
reference
pattern
there.
So
yeah
I
think
I
think
this
is
a
really
good
proposal.
A
Cool,
that's
that
yeah
I
think
that
it's
something
that
may
be
beneficial
here
is
that
it
seems
to
be
somewhat
scoped
in
the
implementation.
Then
it's
not
a
massive
change
to
be
able
to
incorporate
this,
and
it
seems
like
it
would
improve
the
user.
Experience
is
well.
A
We
were
kind
of
relying
on
defaulting
to
kind
of
take
care
of
a
lot
of
the
experience
for
us
in
three
so
sounds
like
we
should
still
consider
taking
in
40.3,
but
if
it
yeah,
we
can
continue
that
conversation
ASIC
and
the
issue
here
and
and
if
you
had
any
more
comments
that
you
had
on
slack
there
that
captions
up
here
in
the
issue.
So
it's
all
in
the
same
context,
that
would
be
really
useful.
Absolutely.
C
And
one
other
thing
I
want
to
mention
is
this:
is
one
of
the
big
benefits
of
breaking
on
to
cross
training
run
time
and
that
these
are
embedded
fields
that
are,
you
know,
embedded
in
all
the
different
claims
are
all
the
different
policy
types
and
just
making
that
update
once
in
crosswind
one
time
and
bumping
the
dependency
version,
then
we
get
all
of
that
built
in
and
you
can
imagine
how
that
could
look
across
specific.
You
know
managed
resource
classes
as
well
across
stacks,
so
that's
pretty
cool
functionality.
We
get
from
that
bull.
D
So
here's
a
brief
one
I
just
wanted
to
raise
the
kind
of
issue
that
the
way
a
stack
is
represented
on
the
registry
affects
the
CL
eyes.
Ability
to
do
image,
relocation
and
I
think
wanted
people
to
be
looking
at
this,
but
just
wants
to
kind
of
float
it
in
the
call,
just
you
know,
bring
up
the
whole
topic,
because
at
the
moment
the
stack
is
a
kind
of
stack
and
registry
is
a
kind
of
blob
which
isn't
really
easy
to
introspect
in
the
CLI.
D
D
B
Added
some
comments
and
slack,
there
should
probably
move
to
github
here,
but
the
link
that
he
provided
that
glim
provided
there,
the
scene
Aptos
guys
somewhere
and
there
they
reference
the
OCI
specs
about
annotations
and
that
does
look
very
suitable
for
what
FP
ml
is
doing.
The
the
comments
that
I
added
in
slack
were
essentially
that
the
well
the
a
PMO
file
may
be
useful
for
a
developer
to
be
able
to
to
annotate
their
stack.
B
D
B
It
could
push
the
image
and
then
use
commands
that
the
equivalent
dr
command
is
like
docker
manifest
or
no
docker
was
a
experimental
docker,
manifest
annotate,
and
then
you
provide
the
annotation
there
and
in
doctor
says
it's
an
experimental
feature,
but
it's
part
of
the
OCI
spec
and
there's
bunch
of
annotations
that
are
part
of
what
is
it
open
container
dot
work,
a
bunch
of
like
pre-registered
annotations,
that
kind
of
fit
almost
field
for
field,
what
we've
gotten
at
the
ml?
So
it's
it's
certainly
interesting
and
certainly
something
to
to
dig
into.
A
Yeah,
I
think
it's
very
useful
to
have
that
you
know
indexing
or
annotations
metadata
a
bit
availability
in
the
you
know,
distribution
there
or
the
on
the
OCI
registries
there.
So
that's
definitely
really
useful
and
that's
thank
you
for
sharing
that
link
and
information.
Glenn,
that's
I
think
that's
going
to
be
definitely
coming.
Andy
yeah.
D
A
E
A
Awesome
so
then,
in
the
PR
section
we
got
a
status
update
for
all
the
cockfighter
resource,
connectivities
back,
etc.
Prs
dan
did
you
have
more
that
you
wanted
to
make
a
comment
on
for
this
as
your
PR,
or
do
we
already
cover
it?
No
we're
too
covered
it,
excellent,
okay,
all
right.
So
that
was
the
end
of
the
agenda
here.
Were
there
any
other
items
that
folks
wanted
to
bring
up
right
now.
A
A
So
by
the
time
we
have
our
next
community
meeting
in
two
weeks,
we
are
hoping
to
get
our
0-3
release
out
into
to
the
hands
of
the
community
and
user
base,
and
we
are
converging
on
all
the
resource
connectivity
improvements
for
each
of
the
individual
cloud
providers
and
hoping
to
get
those
merge
to
master
very
shortly
the
we
have
individual
stacks
for
each
cloud
provider,
so
you
can
install
only
a
single
cloud
provider,
that's
all
you're
interested
in
and
the
other
functionality
like
stacks
tli
and
stack
manager.
Improvements
are
converging
this
well.