►
From YouTube: 2021-10-07 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
B
B
B
There
we
go
all
right,
recording
has
started,
and
this
is
the
october
7th
cross
plane
community
meeting.
So
we're
still
filling
out
the
agenda
a
little
bit
here
and
folks
are
more
than
welcome
to
add
things
that
they
want
to
talk
about.
So
we
can
keep
filling
out
the
agenda
as
we're
going
along.
A
B
Let's
get
dive
right
on
in
here
we
had,
I
had
1.4.1
as
a
release,
but
I
think
that
was
actually
from
the
last
community.
So
I
I
don't
think
there
has
been
a
recent
patch
release
for
for
cross
plane
has.
Has
there
been,
there
hasn't
been
one
route
for
that
too,
or
anything
else
has
there
moaf
or
dan
or
anybody.
B
It
again
has
there
never
been
any
like
recent
upstream
cross
plane
releases
since
the
last
community
meeting
right
window,
401.
A
C
B
B
You
know
a
code
and
feature
freeze
on
that,
so
we
can
keep
we'll
take
a
look
at
the
1.5
board
and
really
the
main
main
importance
here
is.
We
can
talk
about
some
of
the
obstacles
and
some
of
the
risks,
but
we
also
want
to
talk
about
hey.
Is
there
anything
that
folks
are
working
on
the
community
that
they
need
some
help
on
or
need
a
push
to
to
help
get
in
for
this
release?
That's
towards
the
end
of
the
month.
B
Oops
all
right
yeah,
so
any
you
know
we'll
talk
about.
Most
of
the
team's
focus
has
been
on
the
terror
jet
work,
obviously
so
that
which
is
a
huge
part
of
upcoming
releases
but
in
terms
of
general
upstream
cross
plane
the
things
coming
on
here.
Is
anybody
want
to
raise
a
flag
or
bring
up
any
any
risks
or
things
that
they
need
help
with
for
1.5.
B
All
right,
I
know,
there's
some
other
pr's
linked
later
on
in
the
agenda
that
we'll
get
to
also
so
we
can.
We
can
dive
into
those
there,
but
if
any,
anybody
from
the
community
that
wants
to
raise
flags
on
1.5
features
are
more
than
welcome
to
in
this
format
or
on
slack,
that's
always
always
available.
B
Then
yeah,
maybe
we
could
just
go
ahead
and
dive
into
some
of
the
updates
for
all
the
providers
and
terra
jets
and
surrounding
ecosystem
around
core
cross
plane
so
off.
If
you
want
to
get
us
started
here
with
updates
on
terra
jets,
dive
on
in.
C
Yeah
yeah,
so
we
have
now,
like
you,
know,
more
excellent
features
and
in
fact,
right
now,
once
you
configure
your
object,
the
target
resource
should
be
fully
exam
compliant.
So
one
of
the
things
was
late.
Initial
design,
initialization,
which
alpert
has
implemented.
That
is,
like
you,
know,
generic
and
using
reflection,
library,
we're
testing
a
few
like
you
know,
different
classes
of
resources
and
we're
fixing
some
exceptional
cases
there.
C
But,
like
you
know,
for
most
resources,
it
works
as
expected,
and
the
second
thing
is
cross
resource
referencing,
which
hassan
has
integrated
the
the
tooling
in
cross
plain
tools
into
telegen,
because
if
you
want
to
use
the
reference
resolver
generation,
you
also
need
to
add
the
reference
selector
first
to
the
schema
and
that's
what
what
we
do
intelligent.
C
So
once
you
say,
like
you
know,
hey,
this
is
the
field
that
will
reference
reference
that
type
we
first
add
the
field,
reference
and
selector
fields
and
then,
like
you,
know,
put
the
command
to
make
crossband
tools,
part
work
and
connection
details.
Secret
is
again
hassan's
update.
I
I
think
the
last
two
connection
is
secret
and
sensitive
input.
Hassan
can
give
an
update
about
those
two.
D
Yeah
sure
yeah,
basically
one
of
the
problems
that
we
have
identified
in
the
like
previous
technical
release
of
like
terajet
based
providers,
was
like.
We
were
keeping
tf
state
in
the
annotation
and
tf
state
could
include
sensitive
information
and
basically,
we
need
to
keep
this
sensitive
information
in
a
in
a
like
kubernetes
compatible
way
to
make
sure
that
they
land
on
secrets
etc.
D
And
you
know
like
crossplane,
already-
has
the
notion
of
connection
details
secrets,
so
it
it
is
the
number
one
thing
that
we
would
need
to
solve,
like
the
fields
that
marked
as
sensitive
should
go
into
connection
details
secret,
if
like
their
outputs,
but
sometimes
they
could
be.
The
inputs
like,
for
example,
for
a
database,
a
you
database,
password
could
be
master.
Password
could
be
a
an
input,
and
in
this
case
we
should
take
that
input
from
a
secret.
D
So,
basically,
in
a
single
pr,
we
kind
of
handled
all
different
cases,
including,
like
reading
connections,
see
secrets
from
tf
state
and
storing
them
or
purchasing
them
in
humanities,
secrets
and
generating
the
required
schema,
etc.
So
it
was
one
of
the.
I
think,
big
changes
that
I
I
spent.
My
most
most
of
time
till
the
last
community
meeting
yeah-
that's
that's
pretty
much
it
about
connection
details.
C
Yeah
thanks
assange
and
in
addition
to
that
we
used
to
like
it
is
somewhat
like
unrelated.
We
used
to
store
the
tf
state
in
annotation,
which
also
had,
like
you
know,
common
sensitive
fields,
so
the
connection
details
secret,
stuff
removed
that
need,
like
you
know,
to
have
that
information
in
the
ts
state,
but
we
also
removed
it
from
the
annotation
completely
so
right
now,
if
the
tf
state
file
is
gone,
we
are
reproducing
it
from
like
you're
using
spec
status
and
also
the
content
of
the
connection
detail
secret.
C
So
we're
not
relying
on
any.
You
know
secondary
source
of
truth,
like
tf
state
it
just
like
you
know,
other
providers,
the
custom
resource,
it's
a
resource
itself,
the
ammo
that
is
stored
in
each
cd
is
our
like.
You
know
one
and
only
source
of
truth.
C
In
addition
to
that,
we
have
added
a
we
have
implemented
the
memory
based
bookkeeping
for,
like
you
know,
old
operations,
and
before
that
we
used
to
have
like
you
know,
crosspin
locks
and
terraform
locks
as
well,
and
with
the
experience
that
we
got
during
testing
and
development,
we
decided
to
switch
to
a
memory
base
where
we
store
all
the
operation
metadata
in
a
map.
C
Essentially-
and
this
allowed
us,
like
you,
know,
to
be
more
flexible
with,
like
you
know,
debugging
testing
and
also
not
tied
to
the
the
not
not
type
plug
you
know,
file
locks
and
we
also
removed
the
terraform
locks
as
well.
So
right
now,
if
you
stop
and
start
the
terajet
provider,
it
will
just
go
on
with
this
operation
and
if
you
even
delete
the
workspace,
completely
it'll
still
work.
So
so
that
is
like
you
know.
C
That
was
like
you
know,
one
of
the
big
changes
that
I've
that
I've
worked
on
and
with
that
we
introduced
a
few
more
methods
on
the
workspace
client.
C
Specifically,
the
synchronous
calls
so
right
now
for
every
resource
you
can
choose
to
use
either
sync
or
async
mode
in
the
controller,
and
that
means
that,
for
example,
for
vpcs
we're
using
synchronous
mode
where
it
is
blocking
until
the
apply
operation
is
completed
and
then,
like
you
know,
takes
the
state
and
like
does
the
necessary
operations
for
resources
that,
like
you
know
whose
creation
takes
long
like
database
and
cluster
users,
like
developers
can
say,
like
you
know,
hey
use,
async
api
for
this,
and
we
recommend
it
like
network
resources
to
creation,
takes
like
no
more
than
a
minute,
and
in
that
case,
like
you
know,
we
we
issue,
like
you,
know,
async
calls
and
we
keep
track
of
those
those
operations
and,
like
you
know,
let
you
know
about
the
operation
in
the
conditions
of
the
like.
C
You
know
resource
but,
like
you
know,
details
and
errors
and
all
that
stuff.
So
I
think
these
are
like
you
know
the
main
things
that
we
have
worked
on.
So
I
think
the
best
news
for
this,
like
you
know
as
a
result
of
this
recent
sprint,
is
that
we
are
like
you
know,
xrm
compliant
completely,
even
like
you
know
better
than
in
some
like
it's
for
better
than
some
resources
native
providers.
C
So
we
are
going
to
really.
We
are
going
to
cut
it
for
release
in
the
next
coming
days.
Right
now
we
are
thinking
about,
like
you
know,
capping
the
number
of
resources
that
we
generate
to
50
in
azure
and
aws
because,
like
you
know,
because
of
the
bugs
and
like
you
know,
other
changes
that
we
make
the
custom
configuration
object.
Interface
could
change,
which
means
like
you
know.
C
We
have
to
go
back
and
update
the
custom
config
of
every
resource,
so
we
were
thinking
of
capping
it
at
50
and
but,
like
you
know,
I
believe
after
the
alpha
release,
like
you
know,
we're
going
to
have
a
meeting
and,
like
you
know,
discuss
better.
These
resources
can
actually
exist
in
native
provider
or
they
should
you
know,
stay
in
their
own
provider.
C
But
until
then,
like
you
know,
alpha
release
will
be
open
for
testing
by
the
community.
I
mean
even
today
you
can
test
it,
but,
like
you
know
after
this
will
be,
it
will
be
like
a
properly
tagged,
complete
release
that
you
can
use
so
yeah.
That's
that's.
I
think
all
I
got
for
telejet.
B
Awesome
yeah,
those
are
pretty
pretty
exciting,
updates
there
and
I'm
really
really
excited
for
you
know
the
balcony
citizen
for
people
to
be
able
to
get
their
hands
on
it
and
try
it
out.
So
that's
that's
really
really
cool
progress
there,
and
so
we
will
make
an
announcement
and
and
kind
of
you
know,
make
sure
that
people
on
slack
and
twitter
know
when
those
alpha
providers
are
ready
and
that
people
can
start
really
kind
of
taking
them
for
for
a
child
there
and
getting
all
the
feedback
too.
B
B
Cool
I
wanted
to
call
out
also
in
terms
of
recent
releases,
that
the
first
release
for
provider
git
lab
came
out
about
a
week
ago,
or
so
so
you
know
great
work
to
yon
and
the
rest
of
the
team
there
that
have
been
focusing
on
that,
and
so
it's
really
nice
to
to
have
this
as
I
like
this.
Is
it
because
it's
an
interesting
example
of
you
know
using
crossplane
to
manage
anything
with
an
api
right.
This
isn't
necessarily
something
you
would
think
of
for
a
traditional.
B
You
know
cloud
provider,
infrastructure
or
databases,
and
things
like
that,
but
you
know
you
can
integrate
in
in
a
full
picture,
full
perspective
of
software
delivery
to
manage
some
of
the
gitlab
entities
as
well.
So
it's
really
cool
to
see
this
first
release
out
there.
B
So
yeah
great
work
to
everybody
contributing
to
provider,
get
lab
awesome
all
right,
so
we'll
go
through
some
updates,
then
for
the
the
other
major
cloud
providers,
aws,
gcp
and
azure.
So
let's
see,
I
guess,
aaron
and
if
you
wanna
talk
about
aws
for
a
minute
here.
A
Yeah
yeah
I
I'll
go,
I
mean
I
can
go
first
because
I
spent
most
the
last
week
on
it.
So
we're
doing
big
pushes
to
get
a
lot
of
long-standing
older
prs
merged
in
there's
a
lot
of
resource
generation
that
happened
over
the
last
month,
or
so
that's
just
been
waiting
for
review.
A
C
Yeah
yeah
yeah,
so
this
week,
like
you
know
I
was
like
you
know,
I
I
focused
on
provider
aws
the
prs
and
but
like
mostly
I,
I
started
with
sdk
v2
update
because
it
has
been
a
long
running
issue
that
we're
still
using,
like
you
know,
zero
point
something
version
of
aws
sdk
and
that
held
us
back
with
implementing
specific
features
that
the
community
wanted
and
one
example
was,
like
you
know,
eks
capacity,
type
and
things
these
fields
are
just
not
like
available
in
the
old
sdk
version,
and
this
update
had,
like
you
know,
huge
changes
from
like
a
programmatic
interface
perspective
and,
like
you
know
some
community
users,
like,
I
believe
and
their
v.
C
I
don't
remember
mr
stein,
but
and
also
steve
camalo
like
you
know,
they
did
a
lot
of
work
and
I
merged
the
pr
and
I
can
open
fixes,
like
small
fixes
to
make
it
like
you
know,
compile
and
stuff.
So
that
is
done
and,
like
you
know,
right
now,
we're
using
the
latest
and
like
now
aws
sdk
v2
in
the
manually
written
resources
for
the
generated
ones.
C
We
rely
on
ack
and
that
uses,
like
you
know,
another
version,
an
addition
that
I
yeah
I
I
already
mentioned
but
like
eks
now,
has
the
missing
features
that
isn't
that
are
introduced
in
aws,
sdk,
v2
last
version
and
in
the
coming
days
I
have
some,
like
you
know
some
more
work
to
do
with
the
community.
Push
that
I
wasn't
able
to
like.
You
know,
focus
as
much
as
I
would
like
to
so.
C
A
Yeah
there
was
a
question
in
slack
providers
about
cutting
the
next
release.
Do
you
have
an
eta
on
that.
C
Yeah
I
haven't
had
the
chance
to
write
an
answer
but
yeah.
I
believe
we
are
yeah.
We
can.
We
can
make
release
any
day
and
I
feel,
like
you
know,
by
the
end
of
next
week,
we
can
do.
We
can
do
one
because,
like
you
know,
I
will
focus
more
on,
like
you
know,
getting
more
ps
burst,
or
I
don't
know
I
I'm
kind
of
like
an
on
defense,
because
we
could
also,
like
you
know,
make
a
release
today,
for
example
with
the
sdk
v2
to
de-risk
the
next
releases,
so
yeah.
E
Sooner
the
better
on
my
vote
because
they're
outside,
but
not
what
you
haven't
listed
here,
is
the
aurora
clusters
completely
broken
in
the
current
release.
A
E
E
No,
it's
like
you
will
yeah,
I
mean
the
workflow.
Is
you
create
a
cluster
and
you
associ
and
you
create
an
instance
and
those
are
two
are
associated
together?
Those
that
does
not
work
today
and
there's
a
new
kind
that
was
introduced,
which
is
db
instance
under
rds
that
solves
for
it
and
it
all
works
beautifully
on
the
master
branch,
but.
D
A
C
Yeah
yeah
I'll
just
I'll,
just
like
you
know,
go
over
and
see
like.
You
know
how
many
pr's
that
we
have,
like
you
know
very
close
to
merging
because
like
now,
if
you
release
today
like
now,
they
will
have
to
wait
for
another
release
so
I'll
I'll
look
over.
Like
you
know
the
prs
and,
like
you
know,
we
can
have
a
release
tomorrow
with
all
the
new
changes
that
we've
got,
because
I
believe,
like
now,
many
pr's
also
landed
thanks.
Aaron,
like
you
know
his
distribution
effort
from
the
community.
C
B
Yeah
thanks
for
for
bringing
that
up,
jillian,
that's
that's,
definitely
a
good
point
and
you
know
one.
Maybe
we
should
consider
as
well
like
with
functionality
that
is
broken.
You
know
we
can
think
a
little
bit
more
about.
You
know
patches,
for
that
of
you
know,
doing
patch
releases
and
getting
the
functionality
that
the
community
is
depending
on
or
you
know
has,
has
worked,
and
now
it's
regressed.
You
know
to
get
that
out
sooner.
So
that's
a
good
reminder
and
thank
you
for
bringing
that
up
julian.
E
You
know
I
don't
know,
and
I'm
gonna
take
the
I
was
on
leave
when
it
all
got
released.
So
I
was
just
getting
my
hands
on
it.
To
get
an
example
impul
together
for
our
team
and
saw
that
it
got
fixed
in
maine
and
there
was
a
new
kind
introduced.
So
it
may
never
have
worked.
B
All
right
yep,
so
we
will,
with
the
big
push
yeah,
so
thank
you
aaron
and
muafik
for
getting
a
lot
of
new
features
merged
in
there
and
then
obviously
thanks
to
the
community
and
for
contributing
all
these
new
features,
but
that
push
there
like
yeah.
Let's,
let's
definitely
do
what
we
can
to
get
it
released
out
soon
and
get
all
this
functionality
out.
There.
B
All
right
hassan:
do
you
want
to
give
us
an
update
on
gcp.
D
Yeah
sure,
actually
we
don't
have
a
lot
of
updates
on
gcp
sites.
There
were
a
couple
of
pr's
that
got
merged
and
I
would
say
there
are
two
fixes.
Actually
yeah
there
are,
there
is
one
fix
and
two
two
new
features.
One
is
topic,
has
now
a
reference
to
to
kms
key
thanks
to
and
the
like,
auto
generation
of
of
reference
resolvers,
and
we
have
like.
D
We
started
supporting
gcp
cloud,
sql,
backup,
point
in
time
requiring
recovery
configuration
and
also
there
are
a
couple
of
minor
fixes
and
documentation,
updates,
etc.
B
Yeah
it
hasn't
signed.
Is
there
an
anticipated
release,
then
for
gcp
with
some
of
these,
this
new
functionality.
D
Not
really,
actually
we,
like
till
the
last
provider,
gcp
release
it
like
didn't,
take
too
much
time.
So
maybe
we
can
do
a
release
with
the
like
new
pro
new
crosstrain
release,
but
I
think
there
is
no
major
feature
missing
on
the
last
release.
I
would
say.
B
Awesome,
let's
see,
let's
keep
our
our
eyes
open,
then,
for
anyone
from
the
community,
that
is,
you
know,
waiting
for
his
functionality
or
taking
or
wants
to
take
a
dependency
on
it,
and
we
can
consider
doing
a
release
of
an
earlier
time
frame
if
there's
demand
from
the
community,
for
it.
B
Okay
and
alper,
you
you
are
here
as
well.
Do
you
want
to
give
us
an
update
on
azure.
D
F
F
You
know.
Maybe
something
worth
to
mention.
Is
we
have
elevated
the
effects
of
some
recent
cross-plane
runtime
changes?
This
was
brought
to
my
attention
by
moffat
and
you
know
there
are
certain
resources
on
the
asia
side.
You
know
that
we
create
with
an
asynchronous
api
like
the
postgresql
server,
and
you
know
during
the
creation
phase,
we
need
to
we
store
in
the
current
implementation.
F
The
feature
object.
It's
a
json
object
by
the
way
returned
by
the
azure
sdk
in
the
status
conditions
of
the
managed
resource
and
with
the
recent
changes
on
the
cross,
plane,
runtime
site,
we
will
no
longer
be
able
to.
You
know,
modify
the
state
of
the
managed
resource
during
the
creation
phase,
and
you
know
this
is
right.
Now:
a
blocker
for
provider,
asia.
F
Before
we
can
update
the
cross
plane,
runtime
version,
I
did
some
investigations,
debugging,
etc
to
fully
understand
the
situation
and
its
effects.
Another
maybe
interesting
thing
is,
you
know
for
teleget
based
providers.
F
As
we
talked
in
the
previous
community
meeting,
we
were
performing
some
scaling
experiments
and
you
know,
since
the
last
community
meeting
we
have
also
performed
some
scalable
experiments
for
comparison.
You
know
with
the
native
provider.
This
means
that
you
know
we
have
done
some
scaling
experiments
for
the
native
provider.
Also,
details
are
in
that
command
yeah.
So
this
is
my
update.
F
Yeah
I
mean
during
these
experiments
we
take
various
measurements,
you
know
collected
from
controller
runtime
and
also
you
know
from
our
providers
and
the
api
server.
The
kubernetes
api
server
itself.
F
Yeah,
so
basically
you
know
I
did
not
try
to
fully
utilize
by
the
way
these
experiments
are
performed
on.
You
know:
e2
standards
on
an
e2
standard
machine
in
a
gta
cluster,
so
the
idea
was
not
to
stress
test
provider
azure,
but
you
know
we
were
testing
telephone
based
providers
up
to
100
resources
etc,
and
you
know
we
were
observing
how
they
behave,
and
you
know
I
just
went
with
250
managed
resources
for
provider,
asia,
and
you
know
it
hardly
hits.
F
D
F
You
can
see
a
histogram
of
you
know.
We
have
defined
a
metric,
you
know
in
which,
when
you
create
the
manage
resource,
you
know
it
gets
this
creation
timestamp
and
then,
when
it
gets
to
the
radius
state,
it
gets
its
status
conditions
of
type
ready.
With
this
timestamp,
you
know
using
the
delta
there
we
can
see
how
long
it
took
for
the
manage
resource
to
become
ready.
F
After
you
know,
we
had
provisioned
it,
so
the
histogram
you
see
belongs
to
you
know
how
the
times,
how
the
you
know,
time
to
readiness
measurements
you
know
are
distributed
for
250
resources
and
by
the
way
these
are.
You
know,
reconciled
simultaneously.
B
If
all
the
dig
like
the
deep
dive,
you've
done
into
performance
is
really
really
useful
out
there
to
inform,
you
know
where
we
need
to
be
making
improvements
or
investments
to
you
know
to
increase
for
the
performance
or
the
you
know,
scale
at
which
it's
operating
and
then
all
this.
The
data
here
is,
is
really
really
useful
too.
F
Yeah
and
also
you
know,
mentioned
many
improvements
on
the
you
know
terajet
side,
as
you
all
heard,
so
we
are
also
planning
to
you,
know:
re-run,
similar
experiments
to
see
our
performance
gains,
and
you
know
our
bottlenecks
and
the
points.
The
new
points
for
which
you
know
we
can
plan
for
improvements.
B
All
right
so
lots
going
on
those
are
all
the
updates
about
where
the
the
community
and
team's
been
focusing.
So
I
think
I'll
open
the
floor,
real
quick
for
any
other
comments
on
releases.
You
know
pr
or
not
guys
we'll
talk
about
vrs
later,
but
in
the
upcoming
releases
and
milestones.
If
anybody
wants
to
make
a
comment
before
we
move
to
the
community
topics.
A
B
Cool-
let's,
let's
hop
on
in
to
the
community
topics,
then
so
you
know,
as
always,
the
first
section
is
just
kind
of
highlighting
some
recent
content
from
the
the
community
and
around
crossplane.
B
There
are
two
different
write-ups
recently
from
from
folks
in
the
greater
community
that
were
interesting,
mark
and
abhishek
have
both
written
cool
cool
blog
posts
around
some
of
their
experiences
with
crossfit
and
some
of
the
things
that
they
they
were
building
like,
for
instance,
this
one
that
mark
wrote
was
around
aks
clusters
in
museum
to
deploy
resources
to
it.
So
that
was
pretty
interesting
if
you
folks
want
to
read
up
on
that
one
and
then
also
the
gke
cluster.
B
One
that
I
wish
I
could
focus
on
was
pretty
interesting
in
addition
to
the
delicious
looking
popsicle
at
the
beginning
of
the
article
there
so
a
couple
of
write-ups
from
community
folks
to
share
there,
and
then
I
know
that
victor
had
done
some
collaborations
with
armada
and
straya
as
well
too.
So
those
are,
I
don't,
have
links
to
those
I'm
not
sure
if
they're
posted,
maybe
to
their
youtube
channels
of
those
of
those
companies.
B
But
those
were
definitely
interesting
content
as
well
that
maybe
we
can
get
links
to
them
added
to
the
attended
documents,
but
victor
has
been
working
on
those
and
delivering
some
interesting
stuff
there
and
then
so
kubecon
north
america
is
next
week.
So
a
lot
of
folks
will
be
involved
with
that.
B
I
think
a
lot
of
folks
are
going
to
be
doing
it
virtually,
but
there
will
be
a
bit
of
presence
dan
will
be
at
the
in
la
I
think,
at
the
conference
there
I
think,
grant
will
be
there
also.
So
some
folks
from
the
community
will
be
there
in
person.
We
have
a
crossband
office.
Hours
session
is
probably
one
of
the
bigger
events
for
us
there
at
cubecon.
B
So
on
wednesday,
at
1
30,
there
is
an
office
hour
session
where
you
know
the
set
of
maintainers
will
be
there,
we'll
be
presenting
on
cross
plane
and
answering
questions,
giving
some
demos
and
stuff
the
audience.
If
it's
anything
like
coupon
eu
was
where
we
had
like
200
people
show
up
there
to
participate,
it'll
be
a
wide
range
of
familiarity
and
experience
with
crossplane.
B
So
we
will
be
doing
some
introductory
content
for
the
folks
that
are
maybe
newer
to
crossplane,
but
we'll
be
able
to
deep
dive
into
a
lot
of
stuff
as
well
it'll,
be
largely
community
driven.
The
questions
that
come
in
so
folks
can
attend
and
participate
and
maybe
even
help
out
with
answering
questions
in
the
chat.
Also,
so
there's
definitely
some
opportunities
to
get
involved
there
and
really
excited
to
see.
B
You
know
meet
more
folks
from
the
community
and
be
able
to
collaborate
and
explore
it
together,
so
that
would
be
pretty
fun
so
that
that's
definitely
a
big
event
next
week,
I
guess
six
days
from
now
on
wednesday
at
1
30.
pacific.
That
is
where
the
conference
is
held
in
person
and
then
there's
a
couple
of
talks
that
are
focusing
on
crossplane
and
from
crossband
contributors
as
well.
So
I've
linked
all
of
them
here
in
the
agenda
document.
B
So
you
can
add
those
to
your
your
schedule
for
kubecon,
so
mauricio
is,
is
doing
a
talk.
That
is
pretty
interesting
around
some
of
the
tools
that
he
wished
had
been
available
when
he
was
building
a
sas
a
couple
years
ago.
One
of
those
tools
is
crossplane,
so
he'll
be
focusing
on
that
a
little
bit.
Some
of
the
folks
from
google
like
megan
and
shabir.
There
will
be
showing
a
couple
of
different
ways
to
configure
and
you
know
manage
your
kubernetes
resources.
B
Crossplane
is
one
of
the
the
tools
that
they'll
be
diving
into
there,
so
that'll
be
some
interesting
content
also,
and
then
our
very
own,
stephen
and
dan
got
talks
accepted
at
kubecon.
B
Stevens
is
interesting
because
it's
gonna
be
focusing
on
code
generation
of
which
we're
doing
a
lot
of
in
our
in
the
crossband
project
and
the
work
that
everyone
was
talking
about
with
terajet
and
then
the
native
code
generation
with
the
cloud
providers
and
partnerships
there
there's
a
whole
bunch
of
interesting
stuff
about
code
generation
in
general,
how
it
works
in
in
the
ecosystem,
how
you
know
cube
builder
controller,
runtimes,
use
it
and
then
how
crossband
is
using
it.
B
So
that'll
be
a
good
talk
and
then
dan
is
collaborating
with
john
from
google.
That
is
one
of
the
contributors
or
maybe
founder,
I'm
not
sure
for
the
go
container
registry
project
and
they'll
be
talking
about
some
very
interesting
uses
and
usages
of
container
registries,
so
that'll
be
a
cool
talk
to
listen
to
also
so
add
those
to
your
schedule
and
lots
of
possibilities
for
those
next
week
at
cubecon.
B
A
B
A
I
think
we
will
be
actively
engaging
with
this
community
in
the
future
I
joined
today's
call
mostly
to
learn
more
about
this
community
meeting,
just
discovered
it
so.
B
Awesome
yeah,
it
was
super
nice
to
meet
you
and
definitely
we're
excited
to
hear
more
about
the
you
know
what
you're
building
and
compositions
there,
and
if
there's
any
that
stuff,
you
can
share,
there's
always
opportunities
to
to
like
either
share
it
on
slack
or
in
like
the
github
discussions
area.
There's
like
a
show-and-tell
thing
there.
So
it's
always
cool
to
see
what
what
folks
in
the
community
are.
Building
and
kind
of
sharing
that
knowledge
is
really
really
cool.
A
Yeah
I
mostly
cosplay
care
for
a
couple
of
sessions
for
bitcoin
and
service
meshcon,
so
I
will
be
on
tuesday
I
can
post
and
advertise
our
sessions.
B
Yeah
yeah,
if
you
yeah,
do
do
that.
Definitely
so
folks
can
learn
about
them
and
we
can
put
links
to
in
the
agenda
document
here
if
you,
if
you
want
to
throw
those
on
there.
That's
awesome.
Congratulations
on
this
talks.
B
Right
on
all
right,
then
yeah,
so
that's
everything
for
kubecon.
I
think
so.
There's
some
other
other
agenda
items
that
are
actively
being
added
right
now.
So
jillian,
do
you
want
to
bring
those
up
as
you're
as
you're.
E
Yeah
sure,
sorry,
I
was
just
making
more
context
to
that
to
that
last
one
versus
like
read
this
and
this
the
first
one
we're
looking
for
more
community
feedback
on
the
composition,
revision
management.
So
I
presented
a
problem,
a
new
who's,
also
from
my
organization
and
actually
my
boss.
You
know,
suggested
an
idea.
E
Nick
had
you
know,
commented
on
here
saying
that
we're
kind
of
looking
for
some
more
feedback
and
output
to
see
what
other
people's
opinions
are
or,
if
they're,
having
any
issues
or,
if
they've
even
reached
to
the
stage
of
implementation
for
us
as
we're,
you
know
actively
developing
and
we
manage
multiple
kubernetes
clusters
more
than
multiple.
E
Then
some
sort
of
solution
is
necessary
for
us
and
we're
just
looking
to
see
what
other
people
have
come
across,
and
you
know
so
that
we
can
kind
of
come
together
and
agree
on
what
this
should
look
like
before.
Any
solution
is
developed.
B
And
jillian
was
have,
you
may
have
already
done
this.
I
bet,
but
adding
a
link
to
this
call
for
feedback
to
slack.
So
folks
see
it.
There
would
be
good
if
you
haven't
already
done
it
or
if
you
have
another
reminder
there
is,
doesn't
hurt
at
all.
E
I
think
yeah
I'll
put
yeah
I'll
post
it
back
to
say,
hey,
come
take
a
look,
because
we
definitely
want
to
make
some
movement
on
that,
as
it
is
highly
important
for
us
as
we
try
to
get
gain
adoption
and
things
like
aurora
clusters
and
things
like
that,
where
you
do,
we
have
to
use
the
composition
revision
feature
in
order
to
successfully
manage
it.
Something
like
this
is
important
and
for
our
success
in
getting
crossplane
adopted.
E
Yeah,
the
second
one
is
regarding
resource
references
of
dependencies,
so
even
oath
had
sent
a
message
in
slack
and
then
I
think,
move
off.
You
referenced
us
to
an
issue
related
to
to
this
and
about
how
we,
you
know
patching
from
common
data
sources.
E
This
is
obviously
a
real
interest
and
I
know
it's
made
some
progress,
but
not
a
ton
and
just
kind
of
curious
on
how
we
can
help
move
things
like
that
forward,
especially
based
off
of
the
problem
statement.
Here
you
know
things
like
the
vpc
and
security
groups.
These
are
things
that
we
would
love
to
be
able
to
patch,
because
they're
known
and
created
at
the
cluster
level,
but
are
definitely
referenced
by
resources
like
rds.
C
Yeah
yeah,
so
we
originally
like,
if
you
see
my
command
and
an
actual
open
pr,
the
design
dog
we
were
thinking
about
like
tackling
this
in
the
context
of
generic
references
and
one
of
the
reasons
that
we
were
planning
to
do,
the
generic
appearances
was
like
you
know
the
cases
where
we
wouldn't
have
references
in
build
time,
but
let
users
do
that,
but
the
main
problem
with
this
design
and
the
original
design
here,
like
you
know,
having
a
patch
for
a
common
data
sources,
is
the
security
implications
of
it.
C
For
example,
if
we
allow
the
bank
resource
to
have
to
be
able
to
fetch
data
from
like
you
know,
for
example,
secrets,
then
it
means,
like
you
know,
whoever
has
access
to
that
that
managed
resource
our
back
has
access
to
those
cps
like
you
know
that
provider
can
access
and
for
cross-plane.
C
That
means,
like
you
know,
if
composition
can
access
that,
like
you
know
secret,
then
it
means,
like
you
know,
anyone
who
is
able
to
like
you
know
create
a
claim
like
you
know,
they
would
be
able
to
essentially
access
to
the
secret
to
the
arbitrary
second,
which
is
because,
like
you
know,
crossbane
has
access
to
all
of
those.
C
So
these
are
like
you
know
the
main
security
concerns
and
the
the
big
initial
motivation
for
the
generic
references
was
actually
related
to
terajet
because,
like
you
know
before,
like
do
before
starting
the
telegraph,
we
were
thinking
that,
like
you
know,
cross
resource
references
would
be
really
hard
to
do
interest
in
a
genetic
way.
So
we
would
need
that,
but
it
turned
out
that,
like
you
know,
we
didn't
have
to
have
genetic
references
for
that
to
work
and
we
were
able
to,
like
you
know,
implement
strong
types
usual
cross
resource
reference
portal.
C
So
right
now
that
genetic
resource
references
pr
is
put
on
hold.
So
I
believe,
like
you,
know,
the
way
we
attack
this
problem
is,
like
you
know,
back
to
the
initial
idea
where
we
add
a
new
patch
type
and
we
need
to,
like
you
know,
think
through
about,
like
you
know,
with
security
implications
because,
like
you
know,
right
now,
all
patches
work
with
the
like
you
know
existing
composite
or,
like
you
know,
other
managed
resources
that
are
composed
in
the
same
composition.
C
C
You
know
the
the
big
concern
there,
but
I'll
I
I
can
like
you
know
I
can
leave
a
comment
about
like
you
know,
just
like
summarizing
this,
and
maybe,
like
you
know,
brainstorm
about
like
you
know
how
we
can
design
the
api
so
that
we
don't
like
you
know,
cause
like
indirect
access
to,
like
you
know,
stuff
that
users
do
not
normally
have
access.
E
Right
and
and
and
that-
and
that
makes
sense
right,
especially
since
you
know
from
a
cluster
perspective-
we
know
people
put
in
security
controls
so
to
prevent
things
from
leaking.
But
we
obviously
want
to
make
sure
that
we
improve
the
application
developer
experience
and
not
having
them
to
when
you're
composing
infrastructure
not
having
to
provide
things
that
are
known
to
the
cluster
and
are
expected
of
the
application
developer
to
use
when
they're
provisioning
a
resource
right-
and
you
know
again
calling
out
like
the
vpc
ide
and
the
security
group.
E
C
B
Awesome
all
right,
yeah,
great
great
feedback,
jillian
and
pushing
forward
some
of
those
features
there.
That's
really
really
helpful.
B
Okay,
next
thing
on
the
agenda
that
I
see
is
from
max
with
aws
pr
for
822.
G
Yes,
exactly
that's
from
us,
so
we
have
the
situation
that
aws
does
not
automatically
tag
the
subnets
anymore.
Starting
from
I
don't
know,
eks
version
1.2
or
something
so
we
have
to
tag
them
manually
and
because
we
already
have
a
subnet
resource.
And
if
we
spawn
a
number
of
communities
clusters
in
the
same
subnet,
then
we
cannot
edit
the
the
same
subnet
resource,
and
so
we
created
this
tag,
resource
and
yeah.
G
Just
to
so,
we
can
tag
some
resources
inside
the
the
like
some
ec2
resources,
but
this
turned
out
to
be
quite
a
big
pr
and
there
was
a
big
discussion
going
on
and
on
how
to
how
to
implement
this,
and
because
nothing
happened.
For
a
few
weeks,
I
I
was
wondering
maybe.
G
I
don't
know
if
we
can
discuss
what's
how
to
move
forward
on
this
yeah.
A
G
This
was
this
would
help
us
a
lot.
C
G
C
I
see
some
comments
from
nick
and
carl
regarding
that,
so
usually
I
mean,
I
think,
nick
also
wrote
about
that.
C
Like
you
know,
historically,
we've
been,
like
you
know,
adding
tags
fields
to
the
resources
themselves
to
have
a
consistent
experience
because,
like
you
know
the
essentially
the
reason
ec2
resources
doesn't
doesn't
have
those
tags
is
that
the
api
is
old,
so
they
had
to
add
a
new
api
call
in
order
to
enable,
like
you
know,
tagging,
for
example,
when
you
look
at
eks
cluster
apis,
they
already
have
the
tags,
so
that
was
like
a
deliberate
decision
that
we
made
and
terraform.
C
Does
it
as
well,
like
you
know,
including
the
tags,
map
or
array,
whatever
the
api
you
know
wants
there
as
a
as
I
feel
top
level
field
in
the
parameters,
so
that
users
can
change
them,
but
obviously,
like
you
know
that
that
brings
some.
Like
you
know,
complexity
in
terms
of
code,
like
you
know,
you
need
to
calculate
the
div
and
add
or
remove
things,
and
I
see
that,
like
you
know
from
your
comment
here,
that
subnet
resource
controller
doesn't
have
the
implementation
to
be
like.
C
Yeah,
yes,
I
think
one
I
think
the
the
option
here
could
be
like
you
know
actually,
like
you
know,
edit,
that
logic
to
be
able
to,
like
you
know,
get
a
div
and
and
add
and
remove
stuff,
like
you
know,
essentially
the
logic
that
you
implemented
here
could
be
moved
to
there
to
be.
As
part
of
that
subnet
controller.
C
I
think,
for
example,
vpc
should
have
like
the
full
logic
as
well
like
there
are
some
resources
and,
like
you
know,
some
functions
that
they
use,
that
you
can
like
reuse,
because
it's
all
in
the
ec2
package,
so
I
would
suggest,
like
you
know,
I
would
suggest
going
to
like
you,
know,
subnet
controller
and
see
whether
you
can
like
you
know,
enhance
the
logic
there
to
be
able
to
remove
as
well.
G
Yeah,
but
the
thing
is
that
we
we
cannot
edit
the
the
subnet
resource
after
so
so
the
subnet
resource
is
kind
of
static
and
it
was
you
know
if
we,
when
we
set
up
our
infrastructure
in
the
first
place,
and
so
it's
it's
kind
of
read,
only
more
or
less
and
then,
if
we
let's
say
later
on,
we
we
spawn
spawn
some
eks
clusters.
G
Then
we
just
have
to
tag
the
the
subnet
separately,
and
so
we
cannot
have
kind
of
because
it's
done
via
composition
and
I
think
we
we
cannot
just
override
or
add
a
attack
entry
to
the
to
the
subnet.
And
that's
that's:
that's
the
problem,
we're
having
gotcha
so.
G
Yeah
exactly
so,
our
idea
was
that
we
have
a
composition
with-
I
don't
know:
security
group
and
eks
cluster
and
set
up
all
the
the
things
that
are
necessary
and
then
also
have
a
tag
resource
that
just
adds
a
tag
to
the
subnet
that's
referenced
or
to
the
subnets,
because
we
have
multiple
for
multiple
availability
zones
and
just
add
this
one
tag
to
the
subnet
and
all
this
via
one
composition,
gotcha.
C
Yeah,
it's
it's.
It's
an
interesting
use
case,
yeah
yeah!
Let
me
let
me
think
through
and
like
you
know,
write
a
comment
there
and
I
also
need
to,
like
you
know,
read
the
rest
of
the
comments,
yeah
yeah
sure,
because
because
in
the
so
in
there
is
we
have
this
api
patterns
dock
that
explains,
like
you,
know
how
we
decide
to
make
a
resource
a
like.
You
know
a
resource
on
its
own
or
at
a
as
part
of
like
the
parameters
of
another
resource.
C
C
So
I
would
like
you
know.
Let
me.
A
C
A
C
This
thought
the
convention
stock.
It
is
like
you
know,
in
addition
to
the
krm,
like
kubernetes
api
combination
stock,
and
that's,
like
you
know
what
we
usually
follow
when
we
make
like
you
know,
api
decisions
like
that,
like
you
know
whether
it
should
be
a
single
resource
or
it's
parameters
of
another
one.
G
B
All
right,
yeah,
thanks
for
that
max
and
you're,
bringing
up
that
that
discussion
there
and
thanks
moffitt
for
for
weighing
in
on
with
your
perspectives
as
well,
too.
B
Okay
cool.
So
that's
everything
that
was
on
the
agenda
here.
A
couple
minutes
left
before
top
of
the
hour.
Did
anyone
have
other
items
they
wanted
to
bring
up?
I.
A
Saw
sergey
added.
B
Talks
for
coupon
and
related
related
zero
day
events
and
such
there.
So
thanks
for
doing
that,
sergey.
A
B
Other
things
to
bring
up
in
this,
what
we
are
all
together.
B
Cool
then
we
can
adjourn
and
we
have
some
follow-ups
for
some
of
the
topics
that
were
brought
up
today
and
we'll
keep
talking
on
slack
and
github
and
hopefully
we'll
see
some
folks
next
week
with
kubecon.
B
Yeah
yeah
some
folks
will
be
in
person
like
dan
dan
will
be
their
person
and
great
we'll
be
there
in
person.
I
might
I
live
in
san
diego,
so
I
might
pop
up
for
one
of
the
days
also,
so
we
can
yeah
coordinate
that
on
slack,
probably
and
then
at
the
office
hours
is
another
good
place
to
to
get
folks
together
at
the
office
hour
session
linked
here.