►
From YouTube: 2019-07-09 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
is
started,
and
this
is
the
July
9th
2019
crossplane
community
meeting,
so
we
will
go
ahead
and
first
start
off
with
some
of
the
recent
progress
on
the
0.3
milestone.
So
let's
take
a
look
at
that,
so
the
items
in
the
0.3
milestone,
we
had
done
an
effort
to
scope
out
some
of
the
biggest
quality
issues
related
to
experience
and
reliability,
stability
performance
things
like
that,
and
then
we
also
have
a
number
of
big
design
issues
that
we
want
to
tackle
as
well.
A
A
I
don't
think
Nick
was
in
the
participants
list,
okay,
so
yeah!
So
some
of
these
these
designs
here
they're
some
of
the
meteor
parts
of
you-
know
what
we
want
to
accomplish
with
the
cross
plane
with
Crossman
as
a
platform
over
the
next.
You
know
end
of
the
year
and
into
next
year,
over
the
next
couple
of
releases
and
they're
the
riskier
items.
A
So
we
want
to
front-load
those-
and
you
know,
flesh
out
some
of
the
details
around
those
stories
early
on
before
we
get
too
far
so
I
think
we've
talked
about
a
fair
amount
of
these
before
so
I.
Don't
think
we
need
to
go
into
detail
about
them
now,
but
the
point
is
that
we're
front
loading
these
riskier
designs
now
and
then
also
taking
a
focus
on
some
of
the
major
quality
and
experience
related
issues.
A
There
are
a
few
issues
that
do
not
yet
have
assignees
I'd,
say
more
than
more
than
half
of
the
issues
riveted
that
are
still
open
in
the
milestone.
Do
you
have
a
signees
with
them?
So
there
are
some
opportunities
to
take
up
some
of
these
unassigned
issues
as
we
try
to
complete
this
milestone
over
the
next
next
number
of
weeks.
A
I
believe
that
these
0.3
projects
board
is
open,
I'm
sorry
is
updated
and
there
are
a
couple
of
things
that
are
currently
in
review
that
we
will
talk
about
later
on
in
the
agenda,
and
we
have
a
lot
of
things
currently
in
progress
right
now.
I
think
one
of
the
items
that
when
I
was
reading
or
revisiting
the
items
that
we
currently
have
in
the
0.3
milestone
one
of
the
ones-
that's
you
know,
I
think
is,
does
have
some
risk
associated
with
it
and
we
may
not
have
clear
ownership
of
that.
A
A
There
are
a
few
pieces
of
it
missing
and
when
we
release
0.2,
we
had
to
do
some
of
those
steps
manually
to
run
the
the
build
packaging,
publishing
and
promotion
phases
from
a
developer
laptop.
So
we
would
like
to
you
know,
integrate
that
in
further
with
our
continuous
innovation
environment
so
that
we
have
the
official
build
server
running
the
official
release,
build
and
publishing
it
and
doing
all
those
work
having
a
fully
automated.
But
it's
I,
don't
think
anyone
has
started
work
on
that
right
now.
B
A
Actually
have
a
whole
lot
of
that
implemented
in
the
build
sub
module
that
you
know
we
have
incorporated
into
the
cross
plane
repo,
you
know
things
like
packaging
the
container
and
to
contain
an
image
and
pushing
it
to
the
image
repository
and
pushing
artifacts
to
github
and
all
that
sort
of
things
they're
actually
pretty
much
completely
automated
they're.
What's
missing
here.
Is
the
pipeline
integration
to
actually
call
those
those
those
make
file
entries
with
the
rights
you
know,
credentials,
etc
from
the
release
server.
So
most
of
the
actual
work
is
done
here.
A
B
A
A
A
All
right
so
I'm
gonna
move
to
the
next
agenda
item.
Then
so
we've
been
talking
about
a
concept
that
we're
currently
calling
stacks
for
a
little
bit
now
and
I
wanted
to
more
formally
introduce
this
in
the
community
meeting
here,
just
kind
of
as
what's
something
that
we're
exploring
and
trying
to
understand
better
to
offer
a
better
user
experience
around
the
cross
flane
platform,
but
there
will
be
much
more
worksheet
as
you
dive
into
it
and
to
flesh
it
out
and,
to
you
know,
present
more
high-level
designs
around
this
in
the
coming
weeks.
A
But
essentially
you
know
this
this
idea
of
what
we're
calling
the
stacks
for
cross
plane,
just
real,
quick
to
kind
of
summarize
it
for
everybody
is
that
you
know
we
it
takes
so
the
idea
of
what
we
call
what
we've
called
extensions
so
far.
It
takes
that
a
little
bit
further.
So
for
extensions.
You
know
that
is
the
mechanism
in
which
you
can
teach
the
control
plane
new
functionality.
A
That's
not,
you
know,
built
into
the
cross
playing
repository
or
the
core
across
plane
functionality,
but
right
now,
if
you
want
to
do
that,
if
you,
for
instance,
if
you
look
at
the
gate
lab
controller,
it's
it
requires
a
full
fledged
controller.
Runtime
based,
you
know,
custom
controller
for
kubernetes.
You
have
to
need
to
have
go
code
experience
and
it's
fairly.
A
The
effort
for
that
is
fairly
extensive
and
so
where
we
want
to
be
able
to
have
a
lower
barrier
to
entry
or
a
less
complicated
skill
set
required
to
be
able
to
extend
crossplane
and
incorporate
new
services,
complicated
applications,
etc.
And
so
there
are
a
number
of
different
possibilities
here,
we'd
like
to
explore
and
the
key
here
will
be
around
getting
the
user
experience,
correct
and
smooth
and
intuitive.
And
basically,
you
know
lowering
the
the
barrier
entry
here,
it's
really
what
this
is
all
about.
A
And
so,
instead
of
writing
a
full-fledged
controller
to
you
know
manage
to
interact
with
the
communities.
Api
you'd
be
able
to
use
a
language
of
your
choice
such
as
Python
or
JavaScript,
or
whatever
it
may
be,
to
manage
your
application
and
just
focus
on
the
logic
of
your
application.
Cares
about
not
all
the
plumbing
et
cetera,
that's
associated
with
the
controller
and
then
another
thing
to
potentially
explore
would
be
a
whole
domain-specific
language
around
this.
A
You
know
to
capture
the
common
set
of
tasks
or
functionality
or
operations,
and
it
takes
to
deploy
and
manage
you
know
multi
cloud
applications
and
workloads,
Portability,
etc,
and
give
a
nice
common
experience
around
that.
That
does
not
require
a
whole
lot
of
new
concepts
or
cognitive
load
to
be
able
to
understand
how
to
manage
your
application
in
these
environments.
So
people
are
starting
to
explore
this
effort,
and
you
know
we
want
to
be
able
to
provide
a
nice
experience
to
be
able
to
help
grow.
A
The
crossplane
community
and
the
ecosystem
here
and
to
you
know,
incorporate
new
services
new
applications
and
continue
to
grow.
This
whole
community
so
wanted
to
introduce
that
in
this
meeting
here
and
just
kind
of
get
that
idea
start
to
simmer
and
as
more
designs
and
more
effort
goes
on.
It
goes
into
this.
We
will
continue
to
refine
this
idea
and
make
progress
on
it
all
right
in
the
community
topics
here.
A
Cross-Claim
community
ilya
spoke
about
how
to
extend
the
kubernetes
scheduler
for
multi
cluster
and
multi-cloud
workloads
in
any
agenda
doc.
We
have
direct
links
to
Ilyas
slide
deck,
as
well
as
the
YouTube
recording
and
then
Steve
talked
about
composability
for
cloud
native
applications
and
how
to
play
well
with
others
and
those
that
slide
deck
in
video
recording
are
also
here
in
the
agenda
documents
for
folks
that
are
on
the
call
that
were
in
attendance
at
Shanghai.
Were
there
any.
A
A
Phil,
maybe
you
maybe
you
had
some
observations.
C
C
And
men
wanting
to
you,
know
partner
within
us,
and
so
you
know
explored
what
that
looked
like
and
they
were
super
excited,
and
so
you
know
kind
of
as
we
make
progress
on
this
front,
you
know
being
able
to
you
know
kind
of
work
with
them
and
enable
them
to
join
the
community
would
be
really
fantastic.
So
you
know
one
of
the
things
we're
going
to
want
to
consider
is
how
to
kind
of
outreach.
As
far
as
you
know,
the
community
goes
kind
of
into
into
agent
stuff.
C
You
know.
Basically,
a
lot
of
folks
were
saying
as
soon
as
you
know.
We
deliver
kind
of
the
key
capabilities.
You
know
with
the
crossplane
services,
you
know
and
kind
of
make
that
available,
and
you
know
kind
of
like,
like
a
beta
type
release
that
it
would,
you
know,
be
super
issued
and
kind
of
picking
that
up.
Then
we
also
had
a
handful
of
you
know.
Folks
from
you
know
we're
both
like
you
know
just
traditional,
you
know
kind
of
user
contributors
that
have
contributed
to
kubernetes
and
Prometheus
and
other
open
source.
A
Cool
thanks
for
sharing
that
Phil
one
thing
I
was
kind
of
curious.
About
too,
is
that
you
know
the
the
Asian
market
has
a
number
of
cloud
providers
as
well.
You
know,
like
the
the
big
three
North
American
cloud
providers
that
do
have
global
reach,
but
they're
largely
American
companies,
Google
and
Amazon
in
Microsoft
Azure.
You
know,
that's
one
thing
and
Crossman
has
support
for
that.
A
But
were
you
able
to
get
any
sense
of
some
of
the
bigger
Asian
cloud
providers
like
Alibaba
is
one
of
the
first
ones
I
can
think
of,
and
you
know
what
sort
services
they
have
or
if
there
would
be
much
value-
and
you
know
potentially
thinking
about
integrations
with
some
of
those
cloud
providers.
Yeah.
C
Is
Huawei
and
so
there's
Huawei
cloud
there?
Definitely
you
know
like
the
AWS
equivalent
in
China
in
terms
of
kind
of
scope,
but
all
you
Baba
cloud
I
would
say
is
probably
a
close.
Second,
you
know
kind
of
maybe
you
like
on
the
azure.
You
know
if
you
had
to
compare
the
two
in
terms
of
you,
know:
kind
of
maturity
and
stuff
overall
and
kind
of
maybe
adoption
in
terms
of
customer
base
and
then
tense
ad
seems
to
be
a
little
bit
smaller,
but
I'm.
Also
one
of
the
big
three
so
yeah.
C
They
came
by
the
booth
actually
all
multiple
times,
so
that
was
actually
really
great.
They
brought
their
co-workers
and
their
bosses
back
to
have
conversations
about.
You
know
like
I'm,
driving
gauge
mint,
so
yeah.
So
all
of
that
was
was
super.
Super
awesome
and
you
know
we
got
to
go.
I
know
them
a
little
bit
more
on
a
personal.
A
C
A
Cool
I
I,
don't
think
I
knew
that
Huawei
had
their
own.
You
know
public
cloud
offering
this
world.
It
was
that
large,
my
impression
had
been
that
Alibaba
was
the
the
biggest
cloud
offering
over
there,
but
I
hadn't
actually
looked
at
any
market
numbers
or
anything
to
back
that
up.
It
was
mostly
just
based
on
my
own
personal.
You
know
how
many
times
have
I
seen
the
word
mentioned.
Yeah
I
was
probably
very
unscientific
approach
to
judging
market
sure
yeah.
C
C
So
I
believe
so
yeah
I
mean
the
question
is
to
the
level
of
sophistication
and
integration
so
we're.
You
know
they
started
running
into
the
problems
as
we're
crossplane,
you
know,
starts
picking
up
in
terms
of
you
know,
integrating
minute
services
being
able
to
bundle
up.
You
know
stacks
in
a
meaningful
way.
C
In
the
cloud
that
I'm
providing-
and
so
it
just
provides
a
more
frictionless
experience
for
them
and
so
yeah,
they
definitely
like
saw
the
value
there
and
you
know
we're
interested
in
helping
kind
of
ad
providers
for
their
stuff.
So
as
soon
as
we
have
I
think
a
more
you
know
easily
accessible
and
totally
communicated
like
here's,
how
you
bribe
here,
here's
how
you
build
your
own
provider
like
if
we
added
like
a
blog
post
on
that
or
something
like
that.
C
I
think
the
fact
that
actually
like
we
could
use
the
reference,
especially
with
the
language
barrier
or
whatnot.
You
know
it
could
be
something
that
would
help
make
it
easier
for
them
to
kind
of
understand
what
they
would
need
to
do
and
how
to
contribute.
So
I
think
we
have
a
little
bit
of
work
to
do
there,
but
yeah
in
general,
really
great
great
feedback
and
a
lot
of
interest
very.
A
Cool
felt
very
cool
yeah.
That's
one
thing
that
you
will
being
on
the
floor
of
the
expo
halls
that
a
cube
con.
You
quickly
realized
how
many
kubernetes
distributions
and
kubernetes
kubernetes
offerings
there
are
there's
a
lot
of
them.
Ok.
Thank
you
very
much
for
your
observations
and
your
insight
from
Shanghai
appreciate
that
Phil,
the
next
cube
con
is
already
getting
ramped
up,
and
the
call
for
proposals
is
the
deadline.
Is
this
Friday?
A
Alright,
so
there's
a
couple
PRS
that
I
wanted
to
mention
here:
Daniel,
oh
we've,
since
we
have
two
Daniels
now
we're
we're
gonna
have
to
start
uniquely
identifying
each
so
hash.
Dan
has
a
a
pull
request
here.
That
is
the
beginning
of
the
implementation
for
default
resource
classes,
which
is
a
great
way
to
improve
the
user
experience
for
crossplane.
A
So
this
would
be
a
huge
contribution
really
excited
about
this.
There
was
you
know:
Nick
gave
this
a
very
thorough
review
yesterday,
but
there
was
one
part
that
I
wanted
to
talk
about
a
little
further
still
where,
let's
see,
if
I
can
find
it
real
quick
here
we
go
so
in
this
current
design.
We
have
a
default
reconciler
per
concrete
class,
and
but
it
seems
like
we
only
really
would
need
one
at
the
resource
claim
level
at
the
general
abstraction
like
Postgres
or
you
know,
object,
storage,
buckets!
So
in
your
explorations
here,
Daniel
did
you.
D
At
all,
I'm,
the
initial
implementation
was
basically
just
because
I
was
following
the
kind
of
shared
reconciler
pattern
that
Nick
had
established
with
the
different
concrete
types.
So
that's
kind
of
how
I
went
about
implementing
it,
but
I
totally
agree
that
you
know
there's
nothing
unique
about
the
individual
claim
type
or
they
sorry,
the
individual,
concrete
types,
and
rather
just
having
you
know,
for
instance
like
with
buckets,
which
is
the
one
I
should
implementation
for.
D
You
really
only
need
one
controller
for
all
buckets,
whether
it's
on
at
Azure,
AWS
or
GCP,
there's
not
really
anything
different
about
it,
because
we're
just
looking
for
a
a
resource
class
that
defines
that
kind
that
claim
type
as
default.
So
we
really
only
need
one
for
you,
as
you
suggested
there
and,
and
so
I've
actually
gone
ahead
and
started
some
implementation
on
that
and
then
I
tagged
Nick
there
just
cause.
He
had
had
such
a
thorough
review
that
I
wanted
to
see
if
he
saw
any
drawbacks,
because
I
didn't
initially
got.
A
It
got
it
yeah,
thanks
for
looking
into
that.
Yet
when
I
was
thinking
about
it
more
last
night,
the
I
was
I
was
arriving
kind
of
bit
a
thought
that
you
know
is
say
we
have.
You
know
if
this
is
for
s3,
if
you
know
a
sure,
storage
and
Google
storage,
if
they
all
also
made
one
of
these
reconcile
errs
with
each
one
of
the
reconcile,
there's
be
fighting
over,
you
know.
Basically
whoever
gets
to
it.
A
B
I
just
want
to
add
in
the
same
topic,
you
probably
should
consider
for
resource
classes
which
are
out
of
tree,
because
in
fact
you
leave
somebody
else.
Key
provider
or
digital
ocean
or
I,
don't
know
make
up
Oracle
cloud.
How
would
that
coincide
with
this
specific
single
controller?
Just
just
a
thought,
you
don't
need
to
talk
about
images,
yeah.
D
A
Yeah,
when
I
was
thinking
about
I,
think
that
you
know
a
new
claim
would
need
like
an
assay.
An
a
claim
type
was
added.
There
was
an
extension,
then
that
would
need
some
sort
of
default
controller
as
well,
but
you
know,
and
just
a
new
concrete
type
from
another
provider
I
think
would
still
be
covered
by
you
know
the
injury
claim
type
default.
Reconciliation
I've
learned.
D
D
A
Yes,
awesome
the
uptake,
so
that's
a
great
effort,
dan
or
hash
Dan
I,
think
that
that's
gonna
be
a
you
know:
enormous
improvement
to
user
experience.
You
know
around
like
discoverability
of
resource
classes
and
not
having
to
have
the
end
user
worry
about
that.
Just
give
me
a
my
sequel,
please,
that's
experience
are
enabling
here
and
I
really
like
that
a
lot.
This
is
great.
A
Don't
think
he's
on
the
call
here
today,
but
he
would
love
some
feedback
and
some
you
know
some
discussion
on
this
potentially
having
an
offline
meeting
to
discuss
more
about
this
pull
request
and
make
some
progress
with
it,
because
I
believe
this
will
service
some
of
the
foundation
for
some
other
designs.
So
anyone
who
is
interested
in
discussing
this
can
weigh
in
and
let's
get
something
scheduled
with
Nick,
to
push
this
forward
and
get
some
closure
on
it.
A
A
Okay,
then
we
will
go
ahead
and
wrap
it
up
for
the
week
thin
and
a
reminder
again
about
the
cube
con
San
Diego
call
for
proposals.
That's
Friday
so
definitely
get
to
that
before
end
of
day
on
Friday.
If
you
want
to
speak
at
here
in
San
Diego,
alright,
everybody
thank
you
for
joining
and
we
will
see
you
all
in
two
weeks.
Awesome
thanks,
I,
bye,
everybody
all
right.