►
From YouTube: 2021-02-15 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
has
started-
and
this
is
the
february
15th
2021
crossplane
community
meeting.
We
are
heading
into
the
1.1
release
in
a
couple
weeks
here.
I
think
we're
entering
a
feature
freeze
this
week,
we'll
talk
more
a
little
bit
more
about
the
timing,
but
so
yeah
we're
heading
towards
a
upcoming
release
soon.
A
So
let's
go
ahead
and
I
don't
know
if
the
project
board
view
is
going
to
be
super
useful
for
this
conversation
here,
but
maybe
we
can
just
do
some
overall
status
stuff
on
the
general
areas
down
here
like
we
normally
do.
I
think
that'll
probably
be
the
best
way
to
spend
our
time
here
sounds
good
all
right.
So
let's
go
ahead
and
get
us
started.
A
Dan
you've
been
focusing
on
a
lot
of
security
enhancements
and
such
fill
us
in
on
what
you've
been
up
to
and
what's
going
to
be,
shipping
in
1.1.
B
For
sure,
so
the
main
updates
to
security-
and
this
is
really-
I
guess-
adding
new
functionality
rather
than
you
know,
boosting
some
sort
of
security
aspect
of
our
code
base.
But
the
new
functionality
basically
allows
you
to
provide
credentials
to
providers
in
a
more
secure
manner
and
so
the
the
main
ways
we've
enabled
that
is
added
new
secret
sources
that
are
used
by
default.
So
any
provider
could
support
any
authentication
source
they'd
like
but
we'd
like
to
give
tooling
to
make
that
really
easy.
B
So,
by
default
now
all
consumers
of
crossplane
runtime
will
get
things
like
environment
variable
file
system
and
then
also
kubernetes
secrets,
which
have
been
supported
out
of
the
box.
B
So
we
now
have
a
guide
in
the
docs
and
we
did
a
tbs
episode
on
this
last
week
of
using
the
vault
sidecar
injector
to
put
credentials
into
the
file
system.
This
can
be
a
much
more
powerful
credential
supplying
method,
and
it
also
has
secrets
engines
for
the
different
cloud
providers.
If
that's
the
type
of
provider
you're
using,
we
don't
go
into
a
lot
of
depth
into
that
in
this
guide.
This
will
just
show
you
basically
how
to
store
things
in
vault
and
have
them
sync
to
the
file
system
of
your
provider.
B
So
really
it's
small
change
in
the
providers,
but
enables
a
lot
of
different
functionality
and
the
only
other
really
change
in
in
cross
plane
cross
plane,
for
that
was
that
we
now
support
setting
the
annotations
and
labels
on
the
pod
for
the
provider.
Previously
we
allowed
on
the
deployment
and
service
count,
but
the
controller
config
basically
has
a
spec.metadata
section
now
which
enables
things
like
sidecar
injectors.
So
that's
pretty
much
all
the
changes
but
they're
all
in
and
documented.
At
this
point.
A
Nice
with,
like
user
guides
and
stuff
too,
to
be
successful
in
trying
it
out
themselves,
nice
very
cool.
What
was
that
dan?
What
was
there
was
some
user
guide
that
I
was
asking
about
just
on
friday,
what
you
said
it
was
coming.
Oh,
it
was
kellen,
I
think,
for
the
bi-directional
patching
was
it.
B
Yep
and-
and
it
doesn't
look
like
kevin's
on
the
call
today,
but
I
know
kellen
mentioned
that
he
was
working
on
that
and
was
hoping
to
have
in
the
next
couple
days.
Maybe
we'll
talk
about
this
a
little
bit
with
the
release
schedule
part
down
at
the
bottom.
That's
the
kind
of
change
that
could
come
in
this
week
and
so
we'll
talk
about.
You
know
that
kind
of
thing,
and
then
there's
also.
I
want
to
call
it
one.
B
Other
thing
that
you
had
asked
for
jared
was
a
guide
around
kind
of
like
supporting
multi-tenant
scenarios,
which
some
other
folks
have
also
shown
interest
in
now.
So
I'm
working
on
getting
that
together
as
well
this
week.
A
Awesome
yeah.
That
was
definitely
a
guide
I
was
excited
about,
so
that
we
everyone
doesn't
have
to
come
to
dan
to
get
the
personal
one-on-one
support.
The
guide
will
be
really
really
useful.
I've
heard
a
couple
folks
asking
about
that.
It
seems
to
be
a
pretty
common
scenario
and
set
up
for
for
people
to
run
their
clusters.
A
Mulaf,
I'm
going
to
add
you
as
an
editor
right
now.
I
I
said
this
every
week:
I'm
just
going
to
do
it
right
now.
Let's
see
where
is
boop
editor
hey
feel
free
to
accept
your
changes.
Liberally
mafik!
You
have
been
trusted
all
right.
Let's
go
ahead
and
get
an
update
on
provider
acceleration
as
well
too.
I
know
there's
a
lot
more
coverage
across
providers
in
this
release.
Here,
that's
gonna,
be
pretty
exciting.
I
think
in
casey
is
not
on
the
call
izzy
so
muaf.
C
Yeah,
I
can
give
you
update
about
the
aws
stuff.
So
now
we
have
the
so
there
are
a
few
things
that
are,
that
needs
to
be
written
like
manually
like
late
initialize,
which
did
not
like
ack
didn't
have
support,
so
we
have
a
pr
now
that
has
a
support,
so
it's
generated
automatically
and
also
is
up
to
date,
almost
completed.
There
are
a
few
quicks
there
that
I'm
trying
to
figure
out
and
also
the
reference
field
generation
is
also
in
pr
stage.
C
So,
like
you
know,
with
all
these
getting
it,
I
think,
like
you
know,
we
will
have
less
and
less
custom
code
that
is
required
to
be
written
manually
and,
like
you
know,
right
now,
the
like,
after
these
steps,
there
is
this
external
name
annotation
that
we
might
be
able
to
somehow
let
the
developer
choose
a
field
for
that
property
to
be
used.
That
will
also
eliminate
a
lot
of
hooks
that
you
need
to.
C
You
know,
take
care
of
in
your
custom
code,
so
these
are
like
ack
side
and
we
have
new
managed
resource
api
secrets.
Manager
has
landed
thanks
to
benedict
it's
it's
pr's
merged
today
and
it's
not
ack
generator.
So
that's
what
I'm
trying
to
get,
but
I've
tried
to
figure
out
today,
like
you
know
how
to
convert
it
to
generated
one.
There
are
a
few
things
that
we
need
to
add
to
code
generator,
but
yeah
it'll
be
in
the
next
release.
A
Awesome
and
then
in
terms
of
casey's
work,
you
know
to
do
provider
acceleration
with
doing
some
code
generation
wrapping
around
terraform.
We
did
see
a
very
cool
demo
from
him
that
he
showed
us
on
friday
for
some
of
the
vsphere
vmware
vsphere
stuff
that
he
had
generated.
So
he
had
types
for
like
tags
and
categories
and
vms
and
all
sorts
of
cool
stuff
like
that,
so
that
he
was
able
to
show
some
controllers
running
and
reconciling
that
using
the
vstr,
visor
client
and
setup
that
he
was
using
there.
A
So
that
was
that
was
super
cool.
I
don't.
I
just
literally
know
about
that
demo.
I
don't
know
if
there's
more
details
about
the
work,
they're
filled
that
you
might
be
on
top
of.
D
Yeah,
so
you
know
we
kind
of
did
a
sync
up
with
the
lod
from
cdks
and
stuff
on
the
multi-language
support
and
kind
of
looking
at
two
approaches,
so
one
is
kind
of
like
a
code
first
approach.
You
know,
like
cube
builder
and
that
kind
of
a
thing
where
you
kind
of
create
your
structs
and
then
kind
of
go
from
there.
D
But
in
terms
of
the
amount
of
tooling,
that's
required
to
really
get
to
multi-language,
there's
just
kind
of
gaps
there
in
terms
of
providing
a
good
experience
around
that
so
kind
of
landed
on.
D
You
know
having
an
xrd,
yaml,
first
type
of
an
experience
and
then
being
able
to
generate
you
know
crds
from
that
and
then
import
those
into
you
know
your
your
packages
and
stuff
and
then
use
those
from
within
the
kind
of
running
you
know,
cd
gates,
you
know
composites
behind
the
the
kubernetes
api
line
there,
and
so
you
know
a
lot.
It
has
a
prototype
cdk
its
operator,
and
so
you
know
you
basically
need
to
accept.
D
You
know
the
input
you
know
xr
schema,
and
so
you
know
there
would
be
like
tooling.
That
would
generate
that
and
then
ideally
generate
a
scaffold
around
all
the
methods
that
you
would
need
to
implement.
So
then,
at
that
point
you
know
you
could
just
say
you
know,
here's
the
definition
of
my
xrd,
whether
that
be
done.
You
know
in
cdk
it's
like
we've
not
shown
you
know
with
the
crossplane
cdks
project,
which
is
in
crossplane
control
right
now
or
just
manually
authoring.
D
You
know
your
xrdml,
then
running
it
through
a
command
and
then
that
would
basically
generate
you
know,
and
you
know
an
operator
that
would
be
like
fill
in
the
blanks
kind
of
method,
experience
and
so
we'd
have
to
look
at
like
round-tripping,
support
and
stuff
like
that,
but
in
general
it
feels
like
that
could
work
fairly
well
for
whatever
target
language
you
want
and
yeah
it's
still
kind
of
an
idea
phase,
so
we're
kind
of
sitting
on
the
api
design
number
two.
D
A
lot
has
provided
feedback
on
that
on
that
pr
in
the
cross
playing
cdk,
it's
repo
and
yeah,
and
if
other
people
have
you
know,
questions
or
comments.
I
would
be
happy
to
engage
with
you
kind
of
either
in
github
there
or
on
crossplane
slack,
and
you
know
we'll
probably
have
a
little
bit
more
time
to
to
focus
on
this
after
one
one,
and
so
it's
still
kind
of
in
its
current
state,
with
cross-playing
cdks
for
the
1-1
release.
A
So
yeah
so
then
composition.
I
think
that
you
know
we
have
a
some
of
the
bigger
announcements.
Are
you
saying
something.
C
Yeah,
so
there
are
like
a
few
pr's
that
have
landed
like
kind
of
bigger
pr's
one.
Is
that,
like
a
lot
of
users
from
community
asked?
Is
that
connection
secrets
published
by
the
composite
now
would
be?
C
You
would
be
able
to
choose
a
field
from
composite
resource
to
be
included
and
with
the
bi-directional
patching
we
have,
for
example,
you
can
say
hey
like
take
the
ip
address
from
status
of
this
resource
and
patch
it
to
status
of
composed
resource,
and
then
in
the
connection
circuit,
you
can
use
that
ip
address
so
that
you
don't
you're
not
anymore.
Like
you
know,
tied
to
what
provider
managed
resource
implementation
gives
you
in
the
connection
circuit,
you
can
choose
any.
C
So
that's
kind
of
cool,
and
also
we
have
the
nick
updated
the
composition,
introduction
which
is
like
you
know.
You
know,
you
know
much
better
place
now
so
yeah
I
think
there
are.
There
were
a
few
like
you
know,
smaller
pr's
as
well,
and
some
bigger
ones
like
bi-directional
patching.
We
have
talked
about
it.
You
know
last
comment,
comment
meeting
so
like
overall,
we
have,
we
have
pretty
big
features
in
composition
for
1.1,
without
breaking
any
apis.
A
Yeah,
I'm
definitely
interested
for
any
other
potential,
like
user
guide
updates,
or
you
know
like
we
were
already
talking
about
that.
Kellen
is
hopefully
putting
together
for
the
bi-directional
patching.
So
I
think
that's
one
thing
that
you
know
I
haven't
always
kept
up
with
all
the
feature:
updates
or
additions
to
composition
as
a
whole,
and
so
as
a
composition
author,
you
know
having
it
very
easy
to
understand.
What
are
the
features
that
I
can
use
like?
A
A
What
can
I
do
with
composition
when
I'm
building
one
out,
you
know
because,
like
sometimes
I'm
reading,
like
a
long
guide
or
if
I'm
looking
like
a
large
example
and
like
there's
like
a
comment
of
explaining
like
a
type
of
transform
or
something
like
that,
I'm
like
okay,
cool,
like
I
now
know
that
I
can
do
that
because
I
scanned
you
know
the
entire
thing
and
got
into
the
depths
of
it.
But
maybe
some
like
summary
of
here's,
all
the
you
know,
capabilities
that
you
have
with
composition.
C
Yeah
yeah,
definitely,
I
think
we
will
we
can
have
like
you
know
more
examples,
maybe
like
simple
examples
to
show
us
some
of
the
new
features
that
would
also
help,
and
additionally,
we
have
nick
also
opened
a
pr
but
like
required
patch
policy.
C
We
have
that
we
had
this
problem
with
one
of
the
composition,
where
you
take
the
name
space
of
the
claim
that
the
user
crea
that
users
created
and,
for
example,
you
can
choose
a
provider
config
or
some
other
resource
with
that
name
or
like
using
the
naming
space
of
the
claim.
Basically,
but
this
didn't
reliably
work
because
patch
might
you
know
start
before
the
claim
referendum
space
is
populated,
so
this
will
enable
us,
like
you,
know,
hey
if
namespace
is
not
their
weight
before
rendering
the
resources.
A
Yeah
yeah,
definitely
it's
becoming
quite.
It
was
already
quite
useful
and
now
like
these
extra
additional
features
to
it
like
filling
out
and
fleshing
out
more
scenarios,
you
know
discoverability,
for
those
scenarios
is
definitely
important
as
well
too
cool
phil.
You
already
already
were
talking
about
multi-language
sport
right.
You
already
got
into
that.
A
Of
hopped
ahead
there,
okay,
I
got
it
yeah,
that's
what
I
was
like
I
was
listening
to
and
I
was
like
wait
hold
on:
where
are
we
in
the
dark?
Okay?
Let's
that's
that
on
there
and
web
hook.
Support
for
providers
is
that
is
that
you
yeah.
C
Yeah,
so
for
a
long
time
like
you
know
a
lot
of
users.
B
C
The
community
have
asked
for
web
books,
public
validation
and
stuff,
and
we
have
like
other
use
cases
such
as
immutable
fields
and
other,
like
you
know,
ux
improvements
and
we
finally
decided
that
we
want
to
you
know,
invest
in
that
area.
So
I
am,
I
will
be
working
on
a
one-pager
for
replica
support
this
week
and
hopefully,
like
you
know
in
I
think,
crossplay
1.2
we
will
have.
C
We
will
have
like
you
know,
web
hooks
for
especially
the
api
changes
that
might
be
breaking
so
that
we
were.
We
are
more
free
to
you,
know,
bump
the
versions
from
we
want
better
one
to
be
one
or
even
better,
too
or
so.
A
A
C
For
example,
like
you
know,
crd
scheme
allows
us
for
such
things
like,
for
example,
field
validation,
enums
and
all
this
stuff
with
cube
builder,
but
we're
not
able
to
implement
our
own
custom
validations,
for
example,
right
now
in
composition,
when
you
fill
out
the
base,
if
you
made
it
make
a
typo,
it
is
accepted
until,
like
you
know,
you
try
to
render
it
and
resource
cannot
be
created.
C
A
Oh
got
it
okay,
I
I
see,
I
see
you're
talking
about
now
there,
where
books
really
come
in
handy
and
I
was
like
validation
stuff
got
it
yeah
that
any
deck
and
that
definitely
helps
improve
the
authoring
experience
like
when
you're
building
your
your
competitions
and
then
having
less
to
debug
and-
and
you
know
like
navigate
through
the
system
when
something
doesn't
work.
That's
that's
awesome.
Nice.
B
Yeah,
the
as
part
of
I
was
on
duty
last
week
and
one
of
the
tasks
that
we've
had
kind
of
like
floating
around
for
a
while
and
then
now
that
on
duty
has
started,
we
bumped
it
into
that
queue
was
a
rate
limiting
and
back
off
for
controllers.
B
So,
right
now,
in
all
the
different
providers,
we
basically
have
like
a
short
and
long
wait,
and
if
we
have
an
error,
that's
not
due
to
talking
to
the
api
server,
then
we
basically
use
the
short
weight
and
otherwise
we
use
the
long
weight.
So
that's
kind
of
how
we
have
that.
You
know
every
minute,
reconciliation
when
things
are
healthy
or
we
have
like
every
30
seconds.
B
These
are
configurable,
but
those
are
the
defaults,
for
you
know
re-cueing
on
an
error,
like
you
know,
an
error
from
the
aws
api
or
something
like
that.
So
we'd
like
to
have
some
sort
of
exponential
back
off
there,
which
controller
runtime
actually
implements
a
default
controller
rate
limiter.
B
Well,
I
don't
know
I
didn't
write
the
code,
I
just
interpreted
it
so,
but
this
kind
of
describes
what's
going
on
there
and
one
of
the
main
issues
that
we've
had
with
this
is
we
hit
errors
like
pretty
frequently
in
our
different
reconcilers,
and
we
want
to
be
able
to
re-queue
faster
and
the
back
off
pretty
quickly
for
the
default
implementation
is
going
to
go
up
to
like
16
minutes.
That's
the
max
delay.
B
I
think
it's
you
know
1640
or
something
like
that,
which
is
not
what
we
want
for
the
type
of
errors
that
we
see
in
crossplane,
but
it
would
make
sense
right
if
you
continuously
like
can't
hit
the
api
server,
because
that's
something
we're
expected
to
do
every
time,
and
so
basically
the
default
doesn't
make
sense
for
us,
which
is
why
we've
used
these
hard-coded
but
overrideable
values,
and
so
the
idea
here
is.
We
want
to
make
this
configurable
for
consumers
and
also
use
some
sane
back
off,
for
you
know
the
defaults
for
providers.
B
So
basically,
what
this
is
proposing
is
here
in
the
each
controller
we
have
a
bucket
a
token
bucket
rate
limiter,
which
basically
says
you
have
a
max
burst
value
for
the
number
of
re-cues.
You
could
do
it
one
time
for
all
items
reconciled
by
a
controller.
So
if
you
have
10
000
rds
instances,
for
instance,
the
default
burst
here
is
100
for
the
token
bucket
rate
limiter,
which
means
that
controller
can
only
re-queue
100
times
before
it
start.
B
It's
going
to
start
backing
off
that
value
and
delaying
those
requeues
longer
and
then
for
each
item.
It
just
has
a
basic
exponential
back
off
and
that's
what
goes
up
to
the
16
minutes.
B
One
of
the
things
that
we've
talked
about,
which
is
in
the
issue
or
maybe
the
pr
that's
linked
in
the
agenda
today,
is
doing
instead
of
a
token
bucket
rate
limiter
for
the
each
controller,
doing
an
overall
provider
token
bucket
rate
limiter,
which
basically
means
that
we
can
make
guarantees
like
we
will
only
hit
the
aws
api
at
a
maximum
of
like
10
times
per
second
or
once
per
second
or
something
like
that
across
all
controllers.
B
And
so
we
could
pass
that
down
into
there.
So
you
can
kind
of
see
some
of
the
conversations
that's
happening
here.
This
pr
is
actually
pretty
small.
It's
basically
just
returning
errors
instead
of
returning
re-queue
afters,
but
we
want
to
also
have
a
default.
That's
good
for
most
providers
to
use
so
anyway,
something
to
keep
in
mind
should
be
a
better
experience
with,
like
fixing
having
resources
reconcile
faster
when
they
run
into
issues
as
well
as
potentially
not.
B
You
know,
like
rate
limiting
the
azure
api,
which
I
think
we
saw
or
if
a
controller
is
behaving
poorly
or
something
like
that
guarding
against
those
cases.
So
hopefully
we
can
get
that
in.
This
is
kind
of
decoupled
from
the
cross
plane
release
right,
because
it's
in
cross
plane
run
time
and
only
relates
to
the
different
providers.
So
we'll
we'll
see
about
that,
but
hopefully
we'll
be
in
with
the
next
round
of
provider
releases.
A
Very
cool
dan,
it
seems
like
a
really
good
nice
write-up
here
as
well
too.
I
can
only
imagine
that
by
the
end
of
typing
this
up
here,
that
your
your
hands
were
warmed
up
again
after
you're
done.
A
What
do
you
use
for
for
your
blog
hosting
or
a
theme
or
styling
or
whatever.
B
I
use
hugo
with
github
pages
and
I
don't
know
what
theme
this
is.
I've
had
the
same
one
for
like
five
or
six
years
now.
Oh
interesting,
are
these.
B
Exactly
which
is
actually
very
annoying
if
it's
a
code
block
and
a
link,
you
can't
really
it
bolds
it,
but
you
can't
really
tell
the
difference
too
much
but
interesting.
A
Cool
yeah,
very
cool,
very
cool
right
up
here,
all
right
and
then
cluster
api
integration.
There
was
some
interesting
stuff,
including
the
live
stream
last
week,
and
some
updates.
B
Yeah,
so
this
wasn't
on
the
agenda.
I
guess
for
or
not
the
agenda
the
the
board
for
stuff
to
do
this
release
necessarily-
and
it's
not
necessarily
really
coupled
to
this
release
either.
But
it's
something
that's
just
kind
of
come
out
of
necessity
with
the
place
that
the
cluster
api
project
is
at
so
watching
the
live
stream
and
maybe
reading
some
of
those
issues
will
probably
give
you
more
context.
B
I'm
not
going
to
go
through
all
here,
but
essentially
there
are
specific,
concrete
types
in
cluster
api
for
each
provider
and
then
kind
of
abstract
types
at
the
core
cluster
api
level
and
those
you
know,
through
a
reconciliation
process
of
the
two
separate
controllers
kind
of
get
bound
together
to
result
in
provisioning,
a
kubernetes
cluster.
B
Those
different
providers
are
basically
kind
of
like
cross
plane
providers,
but
extremely
scoped
and
like
touch
a
lot
of
different
apis
in
the
same
reconcile
and
that
sort
of
thing
and
so
yeah.
So
it
can
be
like
kind
of
difficult
to
maintain
those,
and
you
know
they
really
don't
want
to
duplicate
the
work
that
we've
already
done.
Most
of
the
cluster
api
folks
are
just
interested
in
you
know,
being
able
to
build
a
a
kubernetes
cluster
right,
so
they
have
more
expert
expertise
on
the
control
plane
side
of
that.
B
So
what
we'd
like
to
do
is
switch
out
some
of
those
concrete
resources
with
crossplane
xrds
that
represent
kind
of
like
their
concrete
implementation.
So
they
have
like
a
gcp
cluster
type
that
provisions
like
a
network,
sub
network,
firewall
et
cetera,
and
you
can
kind
of
imagine
how
we
could
abstract
that
with
composition.
A
Awesome
dan,
very
cool
yeah-
I'm
really
excited
about
that
initiative
as
well
too.
It's
just
because
of
the
or
not
just
because,
but
one
of
the
reasons
is
because
of
the
upstream
and
impact
that
it
has
in,
like
the
awareness-
and
you
know,
adoption
and
such
of
crossplane
into
the
upstream
kubernetes
community,
so
making
headway
there
with
you
know
those
that
type
of
folk-
and
you
know
the
you
know,
contributors,
kubernetes,
etc
is
really
really
cool.
E
Yeah,
I
have
a
question
on
on
the
cluster
api,
so
I
know
that
crossplane
providers
already
allowed
to
provision
clusters
in
some
of
the
clouds
like
aks,
cks
and
and
in
gcp,
so
do
expect
also
some
convergence
at
the
level
of
the
actual
definition
of
the
cluster
resource
object
itself
or
right.
Now
is
more
about
this
lower
level
resources
in
the
infrastructure
that
you
expect
to
be
somehow
imported
in
the
cluster
api.
B
Yeah
so
I'd
say
generally
more
in
the
latter
of
those
two
camps,
but
primarily
right
now
we're
trying
to
just
replace
the
providers
with
the
crossplane
providers.
So
essentially
you
can
do
everything
that
you
do
with
cluster
api,
but
just
installing
cluster
api
and
just
installing
crossplane,
and
so
that's
the
goal
for
the
time
being.
B
We
also
did
a
demo
back
at
the
crossplank
community
day
with
the
equinix
metal
folks,
where
we
kind
of
shoehorned
cluster
api
into
a
cross-plane
provider
as
well,
and
so
you
could
imagine
that
the
xrds
that
we're
defining
to
replace
the
cluster
api
providers
could
go
along
with
a
cluster
api
crossplane
provider
and
they
could
be
connected
that
way
into
a
higher
level.
Composition
with
you
know,
cross
resource
references
and
that
sort
of
thing,
but
that's
really
far
down
the
line.
B
One
thing
I
will
say
is
you
know:
the
kubernetes
community
is
very
large
and
there's
a
lot
of
different
folks
and
even
within
just
cluster
api,
there's
a
lot
of
different
companies
and
organizations
that
have
different
things,
they'd
like
to
do
with
the
project,
so
the
you
know
by
nature
of
that.
The
cycles
for
getting
things
implemented
or
changed
or
updated
are
pretty
long,
so
we
kind
of
are
going
with
the
approach
of
bite
off
a
little
bit.
D
And
looking
at
the
like,
the
kappa
code
and
stuff
as
well,
just
like
all
the
little
you
know
bits
and
bobs
that
they're,
adding
in
having
like
a
cross
playing
composition
with
the
resources
that
we
have
just
seems
like
it's
such
a
perfect
fit
for
the
extensibility
story
there.
So
yeah
super
excited
about
this
is
awesome.
B
Yeah
yeah
vince
from
over
at
cluster
api
who's.
One
of
the
maintainers
was
mentioning.
You
know
that
folks,
basically
when
they,
when
they
are
doing
all
of
that
in
a
single
controller,
there's
no
way
to
like
customize
different
values.
So
they'll
like
randomly
have
to
add,
like
oh,
like
all
right,
we
need
to
add
a
number
of
subnets
field
to
the
aws.
You
know
cluster
type
or
whatever,
whereas
with
an
xrd,
someone
can
just
fork
that
and
literally
add
a
new
field
that
maps
to
a
new
subnet.
B
You
know-
and
I
was
kind
of
explaining
that
and
that
you
don't
you
not
only-
could
do
everything
the
cluster
api
providers
do,
but
you
could
also
make
it
much
easier
for
people
to
essentially
have
their
own
cluster
api
providers.
Just
you
know
forks
of
those
and
distribute
distribution
with
packages
and
everything.
But
you
know
we
don't
want
to
give
them
too
much
at
once.
B
C
C
B
Yeah,
I
thought
about
adding
a
section
here:
maybe
I'll
drop,
something
in
slack.
There's
a
crossplane
demo
that
we're
giving
on
wednesday
for
working
group
kate's
infra,
which
is
the
team
that
maintains
all
of
the
kubernetes
test
infrastructure.
So
like
the
gke
clusters
and
that
sort
of
thing,
if
you've
ever
interacted
with
prowl.
That's
where
that
runs
et
cetera,
there's
they
use
terraform
and
bash
to
kind
of
manage
that
right
now.
B
Yeah,
so
I
mean
like
the
right
they're
they're,
calling
the
gcp
api
right
from
their
their
local
machine,
so
they
execute
bash
scripts
etc
from
their
local.
So
that's
not
not
a
way
to
do
anything
right,
and
so
you
know,
with
crossplane
being
in
the
cncf
and
especially
with
it.
Moving
towards
incubation,
there's
a
lot
of
interest
in
integrating
there
and
so
we'll
be
giving
a
demo
of
what
that
could
look
like.
So
there's
the
idea
of
like
managing
the
test
infra
and
then
with
cluster
api.
B
A
Very,
very
cool
cool
all
right,
so
all
those
awesome
features
and
initiatives
that
are
in
the
1.1
release.
The
release
itself
is
coming
up
in
two
weeks
from
now,
so
we
have
so
today
we
will
cut
the
1.1
release
branch
and
then,
for
this
week,
is
feature
freeze
where
we're
not
expecting
any
more
features
to
be
landing
in
in
coming
in
we're
gonna
be
doing
some
bake
time,
settling
things
down
a
bit
and
so
no
more
features
this
week,
because
it's
feature
freeze
and
then
wait.
B
A
Cool
okay,
so
yeah,
so
no
more
features
with
lots
of
complexity
or
changes
coming
in
now.
But
you
know:
docs
changes,
user
guides,
changes,
bug,
fixes,
etc,
are
totally
on
the
table
and
we
can
start
stabilizing
and
you
know
polishing
any
rough
edges
to
get
us
to
the
finish
line.
And
then
next
week
is
code.
Freeze
where
you
know
we
wouldn't
want
to
be
taking
any
more
code,
changes
at
all
and
it's
like,
but
user
guides
and
docs
and
stuff
like
that.
We'll
be
testing
we'll
be
know.
A
Updating
the
guides
and
all
that
sort
of
stuff
is
the
plan
for
next
week,
but
yeah,
but
the
1.1
release
branch
happens
today.
So
anything
that
goes
in
you
know
should
follow
the
backport
cherry
pick
process
that
we
normally
follow
to
get
things
into
the
1.1
branch
and
yep
and
then
march
2nd.
We
plan
to
do
the
release.
So
that's
that's.
Awesome
really
excited
to
get
the
1.1
out.
First,
first
release
after
the
major
1.0
before
the
holidays,
okay
and
so
dan.
A
You
did
two
live
streams
yesterday
that
we
all
got
to
witness.
Oh
sorry,
last
week,
two
live
streams
last
week
that
we
all
got
to
witness
in
experience.
You
also
did
one
with
vault.
B
Yeah,
I
think
I
mentioned
this
a
little
bit.
We
basically
just
went
through
that
guide
earlier
very
nice
of
nick
jackson
from
hashcorp
to
come
by
and
help
us
out
in
the
chat
and
stuff.
While
I
made
some
errors
and
configuration
and
that
sort
of
thing,
so
that's
a
fun
one
to
rewatch
and
nick
also
mentioned
that
he'd
like
to
come
on
for
a
future
show
so
so.
A
That'd
be
cool
nice
yeah.
That
was
a
fun
one
too,
with
the
live
debugging
and
the
the
chat
was
was
active
as
well
too.
That
was
super
fun.
A
A
All
right
all
right,
so
yes,
the
next
thing,
csf
incubation
proposal,
so
I've
been
putting
that
together
over
the
last
week
plus
or
so.
The
the
big
thing
that
we're
putting
together
is
getting
adopters
of
crossplane
together.
So
one
of
the
biggest
like
important
things
there
is
that
the
to
be
included
in
the
public
proposal
you
we
need
to
have
consent
from
from
folks
that
are
using
cross
planes
so
to
be
able
to
use
your
username.
A
Name,
logo,
etc.
We
have
to
have
consent
from
them,
so
the
I
put
together
a
survey
and
sent
that
out
to
folks
that
I
know
are
using
crossplane
and
getting
some
responses
back
on
that
one
we
tweeted
about
it,
put
it
on
slack
as
well
too,
but
we
should
send
that
out
again
or
just
send
like
get
this
to
people
that
we
know
are
using
across
plane
or
try
to
get.
A
As
many
of
these
you
know,
use
cases
and
adopters
documented
as
we
possibly
can,
because
that's
the
biggest
hurdle
for
incubation
to
be
acceptance
incubation
by
far
so
we'll
need
to
keep
following
up
on
that
and
getting
as
many
people
to
respond
as
we
can.
I
may
be
reaching
out
to
some
of
you
all.
A
If
I
know
you
have
a
good
relationship
with
somebody
that
I
haven't
gotten
a
response
from
see
if
we
can
kind
of
team
up
and
and
get
the
get
connected
with
them
and
get
some
get
some
of
the
data,
but
feel
free
to
share
this
link
here
liberally,
though,
with
anyone
or
any
channels,
you
want
to
it's
we're
taking
it's,
not
just
people.
We
know
we're
using
crossplane
anybody.
A
Who's
who's
you
know
wants
to
participate
in.
That
survey
is
more
than
welcome
to
all
right
and
so
there's
an
announcement
last
week
as
well
too,
that
the
next
cross-plain
community
today
will
be
hosted
by
the
cncf
itself
as
part
of
kubecon.
Eu
it'll
be
a
zero
day
event
there,
which
would
be
really
cool
to
be
kind
of
part
of
the
kubecon
atmosphere,
etc.
A
This
is
hopefully
the
last
kubecon
that
will
be
virtual,
and
you
know
the
one
after
that,
which
I
believe
is
scheduled
for
los
angeles,
just
up
the
road
for
me
here.
So
not
super
exciting
for
me,
but
maybe
exciting,
for
you
guys
but
yeah.
They,
hopefully
we'll
get
to
be
in
person
for
that
one.
But
this
one
in
may
4th
we'll
have
our
next
crossblank
community
day
there.
So
we
will
have
a
call
for
papers
call
for
proposals
out
phil.
A
Do
you
know
when
that's
when
that's
or
cfp
for
crossplaying
community
is
going
to
be
put
out
or
what
the
timeline
for
that
is.
D
Yeah,
I
gotta
sync
with
terry
on
that,
but
I
might
imagine
it'd
be
you
know
shortly
here.
So,
okay.
A
Cool
yeah
we'll
make
some
noise
about
that
too,
and
get
some
get
some
of
the
some
speakers.
Just
like
we
had
an
amazing
speaker
line
up
at
our
the
community
day
just
for
the
holidays,
so
that
was
really
really
awesome.
Hopefully,
to
get
some
of
those
speakers
back
and
get
some
get
some
first-time
speakers
to
an
opportunity
as
well
too,
but
yeah
so
that'll,
be
part
of
kubecon
on
may
4th
and
look
out
for
the
call
for
proposals
call
for
papers.
The
cfp
coming
up
soon.
A
We
had
talked
about
the
community
on
duty
rotation
a
little
bit
because
we
had
instituted
that
I
think
everybody's
gotten
a
chance
to
be
on
duty
by
now.
Right
is
there
like?
We
don't
have
to
talk
about
it
for
too
much,
because
we
kind
of
talked
about
it
last
time,
but
are
there
any
high-level
lessons
or
experiences
that
are
worth
sharing
from
some
of
the
people
that
have
been
on
duty
so
far.
B
I
feel
like
that
it's
gone
fairly.
Well,
I
think
that
in
general,
we've
already
been
pretty
responsive
in
in
slack
and
that's
been
a
positive
of
our
community
and
but
I
think,
we've
done
a
better
job
of
like
distributing
who's
responsing
or
respond
responding
to
responding
to
different
folks
who
come
along.
So
I
think
that's
been
a
big
positive
for
sure.
B
One
of
the
things
I
definitely
want
to
reiterate
is
there's
a
lot
of
folks
in
slack
who
are
answering
questions
and
really
steering
people
the
right
direction,
who
are
not
specific
maintainers
of
cross
plane,
and
this
role
is
totally
open
to
everyone.
B
So,
if
you're
interested
in
doing
that,
you
don't
have
to
sign
up
to
do
a
whole
week
all
day,
which
is
kind
of
like
the
current
schedule,
but
if
you're
just
interested
in
and
kind
of
having
a
little
more
formal
of
a
role,
lots
of
you
are
already
doing
that.
So
I
would
love
to
have
more
folks
join
the
program.
A
Hold
on
sounds
great,
okay,
so
then
the
last
item
I
had
here
is
on
docker
hub
and
their
rate
limiting
throttling
stuff
and
how
we
had
applied
as
an
open
source
project,
which
we
are
I
so
I
didn't.
A
They
took
a
long
time
to
get
back
about
our
application
and
they
said
that
they
would
be
reviewing
it
a
couple
weeks
ago
and
they
hadn't,
so
I
pinged
them
just
this
morning.
No
at
the
end
of
last
week
was
it
for
for
an
update
on
it,
and
so
just
this
morning
I
did
get
a
response
saying
that,
because
the
project
is
backed
by
commercial
entities,
that
we
are
not
eligible
to
be
an
open
source
project
on
docker
hub,
that
seems
like
a
fairly
narrow
viewpoint,
they're
taking
and
I'm
gonna.
A
I
don't
know
if
they
have
an
appeal
process
or
you
know,
but
I'm
going
to
push
back
on
it
a
bit
here.
But
if,
if
that's
their
stance,
I.
B
A
To
an
alternative
registry
source
makes
a
lot
of
sense
to
me.
The
rook
project
has
started
seeing
rate
limiting
where
people
cannot
always
download
or
pull
an
image
and
that's
unacceptable.
So
I
I
think
we
should
be
making
moves
to
or
making
plans
to
move
to
an
alternative
registry
source
soon.
A
Didn't
say
who,
but
I
would
imagine
upbound
is
probably
where
they
were
where
they
were
looking
at
if
it
was
for.
C
A
Projects
in
general,
that
would
because
that
is
a
nonprofit
right
or
isn't
it
right?
So
yes,
so
I
I
don't,
I
don't
know
exactly
I
did
they
didn't
it
was.
It
was
a
form
response
or
an
automated
response.
It
was
not
specific
with
details
about
us
at
all,
so
I
will.
A
I
will
push
back
and
try
to
get
more
details
or
see
what's
going
on
there,
but
I
don't,
I
don't
think
they're
the
direction
that
they
seem
to
be
heading
with
this
policy
in
the
first
place
and
then
you
know
open
source
projects
like
they
belong
to
a
foundation
like
all
this
stuff
and
then
like
not
letting
them
like
this.
This
is
clearly
going
in
in
a
poor
direction,
so.
A
Foundation
owns
the
ip
owns
the
trademark.
All
of
that
so
interesting,
yeah
dockrub
is
not
on
a
upper
trajectory.
I
don't
think,
let's
just
say
that:
okay
cool!
Oh,
so
that's
I
think,
that's
everything
we
had
on
the
just
add
this
in
the
notes.
Were
there
any
other
agenda
items
if
folks
wanted
to
bring
up
prs
or
anything
like
that.
A
D
A
D
I
wanted
to
mention
it
wants
to
check
out
our
talk
at
the
utah
kubernetes
meetup,
that's
going
to
be
tomorrow,
so
you
can
google
for
that,
and
that's
all
virtual,
so
yeah
definitely
check
it
out.
A
Throw
that
in
the
old
agenda
there
buddy,
we
should
tweet
about
it
too,
that
that
sounds.
Are
you
give
me
the
talk,
phil.
D
So
ryan
was
doing
it.
Let
me
see
here's
the
link,
let's
drop,
that
in
the
dock.
Here
I
guess
community
topics
I'll
just
put
it
after,
though,.