►
From YouTube: 2023-02-23 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
recording
has
started,
and
this
is
the
February
23rd
2023
cross
plain
community
meeting.
I
will
drop
a
link
directly
in
the
chat
again
for
some
of
the
folks
that
are
just
arriving
to
the
agenda
document,
where
it
is
open
for
folks
to
add
suggestions
for
anything
that
you
all
would
like
to
talk
about.
A
So
we'll
start
from
top
to
bottom,
and
the
first
note
to
make
is
that
there
was
a
recent
cross-plane
runtime
release
for
v1.19.1
that
fixed
an
issue
with
external
secret
stores,
specifically
The
kubernetes
Secret
store
implementation,
which
is
the
default
around
ownership
of
connection
Secrets,
so
that
is
out
and
available
as
1.19.1
and
then
I
think.
A
Let's
roll
that
into
the
discussion
about
a
patch
release
for
1.11.1
of
core
cross
plain
as
well
so
Bob
brought
this
up
because
and
Bob
is
driving
a
fix
for
a
particular
issue
that
is
affecting
a
number
of
people
that
are
testing
the
new
Alpha
feature
for
composition
functions.
A
So
that
is
the
main
thing
that
I
think
that
we
would
want
to
be
releasing
a
patch
for
so
that
the
early
adopters
and
the
folks
that
are
starting
to
use
composition,
functions
and
getting
us
some
great
feedback
can
can
get
around
an
issue
that
is
affecting
it
has
fairly
broad
reach
that
I've
seen
a
lot
of
people
run
into
so
Bob.
Do
you
want
to
tell
us
about
that?
B
Yeah
I
mean
the
fix
is
basically
just
doing
what
Nick
had
suggested.
It's
a
one-liner
that
I
cannot
take
credit
for
and
I'm,
not
even
entirely
sure
that
I
understand
but
reading
through
it.
It
made
sense
to
me
to
the
best
of
my
abilities,
and
so
it's
basically
just
from
what
Nick
was
saying:
it's
basically
changing
the
the
way
the
patching
is
handled.
So
it's
not
it's
not
appending.
It's
completely,
replacing
the
the
resources.
B
Now
he
did
mention
something
about
it
would
prevent
I
think
he
said
it
would
prevent
attributes
from
being
deleted,
but
I'm
not
sure.
That's,
that's
a
huge
deal
in
comparison
with
with
the
benefits
of
the
change.
So
like
I
say
this
was
more
just
picking
up
what
Nick
had
suggested
as
a
fix
and
and
verifying
that
it
worked
for
the
scenarios
that
we
were
seeing
and
then
just
trying
to
get
it
pushed
out
there.
So.
A
And
Bob
I
will
I
will
say
that
doing
as
Nick
suggests
is
not
not
commonly
going
to
be
a
bad
Sebastian.
C
A
Enough,
okay,
it's
me
and
then
so
yeah
Nick
is
on
vacation
this
week.
He
is
on
some
PTO,
but
so
I
guess.
The
status
of
this,
though,
is
that
he
approved
the
pr
here
suggested
that
it's
worth
backboarding
as
well.
So
I
think
that
you
know
we've
got
Nick's
approval
on
the
pr.
You
know
it's
a
good
idea
to
go
and
backboard
it.
A
So
I
think
then
to
continue
getting
this
out
into
the
hands
of
folks
that
are
testing
composition,
functions,
which
we've
heard
a
lot
of
feedback,
and
it
seems
like
there
are
a
lot
of
folks
on
those
that
we
can
roll
go
ahead
and
roll
a
patch
out
for
this.
You
know
this
week
now
without
waiting
for
Nick
to
get
back.
I
think
that's
totally
reasonable.
A
So
then
let's
then
talk
about
what
else
might
be
included
in.
Oh,
it's
all
right.
So
you've
already
got
the
the
back
Port
label
here
so
that
when
we
merge
this
PR
that'll
already
automatically
open
a
backboard
PR
and
we'll
get
that
into
the
release
to
1.11
Branch.
We
can
roll
forward
with
that.
A
So,
let's
talk
about
then
other
things
that
might
be
included
in
a
back
a
patch
release.
1.11.
A
your
PR
was
the
only
one
that
was
open
still
and
these
other
ones
have
already
been
merged,
I
believe
into
1.11
right,
or
at
least
they
have
backport
PRS
open.
So.
B
Yeah
they've
been
merged
into
they've,
been
merged
into
main
or
Master,
but
not
into
the
into
the
branch.
Yet
and.
C
A
The
Bob
release
this
is
yeah
yeah,
so
the
run
the
release.
Bob.
No
seriously
do.
What
do
you
want
to
take
a
stab
at
running
a
patch.
A
We
could
partner
on
that
today.
If
you
have
time
for
that,
we
could
we
could
totally
partner.
You
know
it's
not
a
requirement
like
I
could
run
it.
If
you
know,
if
you
don't
have
time
for
that,
but
if
you
wanted
the
opportunity
to
run
it
since
it's
literally
your
face,
that's
yeah
that
seems
reasonable
to
me,
okay,
so
yeah.
Any
notes,
then,
on
these
these
fixes
too
Bob
that
you
want
to
share.
B
Yeah,
so
this
one
was
an
issue
we
found
where
the
you
can
specify
the
poll
interval
as
a
parameter
to
the
Cross
plane
server
executable,
but
it
wasn't
actually
applying
that
whole
interval
to
the
claim
and
composite
reconcilers.
So
if
you
told
it,
you
wanted
a
poll
interval
of
five
minutes
or
ten
minutes
or
whatever
it
would
continue
to
do
a
one
minute.
Reconciliation
interval
on
the
on
all
your
claims
and
Composites,
and
so
this
is
just
taking
that
parameter
and
applying
it
to
the
claim
and
composite
reconcilers.
B
This
is
it's
an
out.
It's
a
it's!
An
artifact
of
some
work.
We've
been
on
a
drive
to
reduce
costs,
which
means
reducing
CPU
usage,
which
means
you
know
trying
to
draw
I
have
all
the
CPU
usage
we
can
out,
and
this
is
just
something
we
happen
to
find
when
we
were
doing
that.
Nice.
A
Okay,
yep
I.
Remember
that
one
and
then
this
oh,
this
is
the
output
columns.
Like
you
know,
control
gets
output,
sort
of
display
right,
yeah.
B
A
Okay,
great
so
then,
of
these
three
PRS,
this
one
has
not
been
yet
emerged
into
Master,
but
it's
been
approved.
We
can
merge
that
these
other
two
have
been
merged
to
master
but
then
have
outstanding
back
Port
PRS,
open,
which
then
we
can
improve
and
merge.
That's
no
problem,
so
these
three
then
look
like
totally
reasonable
to
get
into
the
1.11
branch
and
we
can
roll
out
a
patch
release
from
there.
A
Were
there
other
fixes
or
open
issues,
then
that,
were
you
know
near
either
already
ready
to
go
or
very
close
to
being
ready
to
go
that
could
potentially
catch
the
train
today,
I
don't
know
of
any,
but
if
anybody's
got
suggestions
we'll
take
them.
A
A
So
I
will
then
follow
up
with
Bob
to
see
you
know
what
makes
sense,
then,
in
terms
of
running
that
we'll
get
those
PR's
merged.
I
definitely
do
have
time
today,
Bob,
so
we
can.
We
can
do
that
together,
I'll
ping,
you
and
you
know,
do
feel
free,
of
course,
that
if
you
have
other
priorities-
and
you
know,
you've
got
other
things
going
on
with
your
day
job
today
that
you
know
this
is
not
something
you
have
to
take
on.
A
Right
I'm
still
on
morning
hours,
so
I'm
coming
up
behind
you
all
right,
so
that's
1.11.1
folks
can
reach
out
on
slack
after
this
meeting
as
well
for
any
other
things
that
might
want
to
come
up
for
subsequent
patch
releases.
A
Let's
move
on
then,
to
1.12
and
start
talking
about
that
release
the
we're
pretty
early
in
the
Milestone
still
within
the
first
third
of
it,
I
guess
we're
coming
up
on
the
for
closing
out
the
first
trimester
of
the
release.
So
we
do
have
a
number
of
things
identified
for
VDOT
1.12..
A
Let's
take
a
look
at
the
entire
board
instead
of
just
scoped
1.12
right
now-
and
things
are,
you
know
pretty
much
ordered
By
Priority
as
well
within
the
backlog
column.
So
there's
a
number
of
major
initiatives
that
we
have
talked
about.
As
both
you
know
within
you
know,
engineering
teams
at
upbound
and
within
the
community
here
in
this
meeting
as
well.
So
a
number
of
discussions
have
taken
place
about
the
priority
of
some
of
these
very
big
items.
A
So,
just
as
a
reminder,
you
know,
I,
don't
think
that
you
know
that
I'm
gonna
say
exactly
which
of
these
would
be
in
1.12
or
what
would
happen
yet.
But
these
are
you
know,
kind
of
top
priorities
and
biggest
things
for
the
project
that
we'll
be
getting
some
eyes
and
some
effort
in
the
in
the
near
future.
A
So,
first
of
all,
there's
a
number
of
things
that
are
currently
in
Flight,
so,
for
instance,
being
able
to
validate
compositions,
you
know:
do
some
upfront
shift
left
some
validation
of
your
compositions
by
a
web
hook
before
they
get
submitted
and
accepted,
and
then
Downstream
run
into
errors
that
are
difficult
to
diagnose,
that
is
in
progress
by
Philippe,
as
he
has
been
working
on
the
extensibility
story
for
secret
stores,
pluggable
secret
stores,
that's
making
a
lot
of
progress
and
is
is
I,
think
John
and,
as
you
can
give
a
little
bit
of
update
that
later
on,
observe
only
resources.
A
That's
I
think
that
does
that
design
been
ratified
and
completed
Hassan.
Can
you
remember
me
on
that.
D
Yeah,
basically,
at
the
last
discussion
we
create
and
aligned
with
the
final
shape
of
the
design.
So
it's
just
waiting
for
an
approval
and
in
the
meantime
we
have
started
implementation.
A
Oh
sweet
so
invitation's
already
getting
kicked
off
fantastic,
so
yeah.
That
is
another
one
of
the
like
top
priorities
for
the
the
project.
The
most
highly
demanded
features
from
the
community
and
then
now
in
the
backlog,
where
there
hasn't
been
a
lot
of
work
started
on
them.
But
there
are
proposals
out.
There
is
discussion
on
them
and
I
would
like
to
see
progress
during
the
1.12
time
frame
is
for
so
this
issue
right
here
is
around.
A
A
So
that's
one
obvious
outcome
of
that,
but
another
outcome
of
that
is
that
it
would
then
enable
us
to
do
a
a
better
solution
for
across
resource
references
which
then
unblocks
this
proposal,
potentially,
which
is
not
ratified
yet,
but
a
proposal
for
being
able
to
break
apart
some
of
the
massive
providers
into
smaller,
more
manageable
chunks,
because
I
know,
there's
been
a
number
of
struggles
with
the
number
of
crds
in
the
cluster.
A
So
one
proposal
has
been
to
break
them
up
and
something
to
unblock
that
would
be
able
to
reference
resources
across
providers
in
that
way.
So
all
those
are
top
of
mind
and
top
of
priority
here
and
will
likely
get
some
some
sort
of
traction
in
1.12.
I'm,
not
sure
exactly
to
you
know,
can't
say
exactly
what
will
be
delivered
yet
during
there.
But
those
are
absolutely
things
that
are
of
importance,
there's
another
another
set
of
items
that
are
also
important.
A
That
will
likely
like
composition,
revisions
moving
those
to
B1
I
think,
that's
probably
something
that's
likely
for
review
1.12,
but
we
could
discuss
that
a
little
bit
further
and
also
promoting
environment
configs
to
to
V1
beta
one
that
it
could
be.
That
could
happen
in
in
1.12
as
well,
since
the
initial
feedback
that
I'm
seeing
so
far
is
that
it's
you
know
working
fairly
well
and
as
intended,
I
haven't
seen
major
issues
come
up
with
it.
I
haven't
seen
proposals
for
big,
read
rights
of
the
API
or
the
experience.
So
that's
something.
A
That's
definitely
possible
and
I'll
follow
up
with
Max
on
that
one
for
getting
that
in
1.12
as
well,
but
yeah.
Those
are
the
big
priorities,
a
lot
of
things
in
progress
right
now,
we'll
get
in
we'll
dive
into
a
few
more
details
on
some
of
these
issues
in
just
a
second
but
as
an
overall
v1.12
look
here,
it's
looking
to
be
another
major
Milestone
like
1.11,
was
where
there's
a
lot
of
engineering
work,
a
lot
of
progress,
a
lot
of
things
that
are
highly
demanded
and
a
lot
of
progress
happening.
A
Okay,
well
yeah.
So,
as
I
mentioned,
we
are
a
third
of
the
way
through
the
release.
Pretty
much,
we've
got
a
lot
of
traction
and
progress
on
these
things
that
are
in
these
columns
right
now
on
the
release
board,
that's
linked
from
the
agenda
document
and
we'll
keep
making
making
progress
on
that.
I
know
that
I.
A
All
pretty
aligned
on
on
priorities
and
demands
there.
So
there's
a
lot
of
work
to
do
and
we
want
to
get
these
things
right,
but
yep
we're
a
lot
of
exciting
things
going
on
in
1.12,
I'll,
say
all
right,
so
we
can
move
on
then
I
think
to
the
providers
section
there
is.
This
is
Christopher
on
today.
I,
don't
think
I
saw
Christopher,
no,
so
I
think.
Since
the
last
community
meeting
there
was
a
new
release
of
the
Community
provider.
A
Aws
0.37.0
was
there
a
patch
since
then
I
could?
Yes,
there
was
a
patch
as
well
37.1,
so
we'll
conclude
that
in
the
agenda
documents
also.
A
Yeah
with
a
fix
for
sorting
in
the
open
search
domain,
which
probably
was
causing
some
later
knit
errors
all
right
and
then
we'll
go
through
quickly.
So
another
set
of
of
official
providers
from
from
upbound
that
have
been
released.
A
I
think
maybe
one
of
the
biggest
ones
to
note
is
that
there
is
a
new
provider
since
last
community
meeting
as
well
Azure
does
they
have
a
split
between
like
the
main
set
of
resources
and
then
everything
related
to
active
directory.
So
there
is
a
brand
new
Azure
ad
official
provider
as
well,
so
definitely
check
that
out
and
then
from
the
community
there
is
a
GitHub
provider
that
is
now
available
and
published
to
the
marketplace
as
well.
A
I
think
that
it
was
the
first
release
that
made
it
to
made
it
public
that
way,
all
right.
So
now
we're
on
the
topic
of
providers.
I
did
specifically
want
to
call
out
a
lot
of
ongoing
buzz
and
discussion
around
performance
and
some
of
the
beneficial
providers
like
the
provider
AWS
of
from
outbound
is
there's
been
a
lot
of
discussion
on
that
there
about
the
some
performance
concerns
and
so
I.
A
First
of
all
wanted
to
thank
everybody
for
participating
in
contributing
to
you
know
the
investigations
here
providing
some
of
their
evidence
and
observation
Etc,
and
so
that
was,
you
know,
definitely
helpful
for
kind
of
you
know
getting
this
investigation
going
and
kind
of
trying
to
helping
find
some
of
the
what
we
believe
to
be
some.
You
know
root
causes
that
are
contributing
to
this.
A
So
thanks
for
everybody
for
doing
that,
and
then
I
wanted
to
call
out
specifically
too,
that
we've
kind
of
started
a
little
special
interest
group
around
this
issue.
Specifically
so
John.
Do
you
want
to
give
us
a
quick
update
since
you've
been
kind
of
driving
some
of
this,
and
you
know
what
we
expected
see
in
this
in
this
channel
coming
up.
E
Yeah,
so
what
we?
What
we
want
to
to
give
folks
is
two
different
tested
images.
So
I'll
give
a
quick
tldr
of
of
what's
going
on
here.
Essentially,
there's
there's
a
bunch
of
processes
being
forbed
when
we
make
calls
to
tear
from
CLI,
and
it
makes
calls
to
the
to
the
underlying
terraform
providers
and
under
certain
circumstances,
especially
during
their
Innovation
initial
resource
creation.
E
You
know:
there's
there's
a
higher
frequency
of
of
calls
happening
now,
there's
a
whole
bunch
of
rate
limits
and
stuff
in
place
that
but
there's
also
burstable
kind
of
like
scope
for
for
these
rate
limits.
E
So
it's
quite
complex
in
terms
of
just
making
sure
you
get
these
right
now
we're
going
to
release
two
two
images
for
for
folks
to
test
the
first
one
actually
uses
a
puts,
the
the
terraform
provider
into
kind
of
like
a
shared
mode,
where,
instead
of
forking
processes
for
every
call
to
the
provided
there
from
CLI,
can
just
make
a
one
call
to
it.
E
We
used
to
run
the
providers
with
that
enabled,
but
we
ran
into
some
some
what
we
believe
were
race
conditions
that
eventually
led
to
some
duplicate
resources,
which
is
why
we
we
moved
away
from
that
approach,
but
we
want
to
kind
of
like
see
if
that
approach
yields
sufficient
kind
of
like
gains
from
a
from
a
resource
efficiency.
E
You
know
what
we
expect
to
see
is
a
lower
resource
utilization,
but
likely
slightly
so
a
slower
resource
provisioning.
So
there's
a
bit
of
a
trade-off
between
the
two
that
we
want
to
make
and
a
couple
of
things
that
will
come
down
the
line
is
you
know
we
want
to
put
a
sizing
guide
out
for
folks
to
be
able
to
say,
like
hey,
if
I
plan
to
provision
100
resources,
you
know
here's
the
very
is
configuration
options
you
have
available
to
to
tweak
and
here's
some
recommended
kind
of
like
resources.
E
You
might
want
to
provision
your
node
on
and
so
on,
so
that
that's
kind
of
like
what
we,
what
we're
doing
on
there
in
terms
of
kind
of
like
what
is
the
the
proper
solve
you
know
like
we
don't
yet
have
a
final
solution
on
that
or
proposal.
We're
exploring
some
options,
we're
looking
into
various
things,
and
hopefully
you
know
in
the
in
the
next
few
days,
we'll
we'll
put
some
stuff
forward
and
get
feedback
from
the
community
and
and
ask
for
help
for
testing
and
so
forth.
A
C
A
I
believe
a
while
ago,
or
last
or
late
last
year,
in
up
Jets,
yes,
and
so
the
question
is
around,
it
seems
to
be
another
facet
of
of
this
particular
issue.
A
Let
me
see
yeah
yeah,
so
as
this
issue
is
described
Roberto
and
for
my
recollection,
I
think,
yeah
I
think
you
did
a
lot
of
good
investigation
yeah
into
into
a
lot
of
the
metrics
in
in,
and
you
know,
Deep
dive
into
into
the
performance
characteristics
there
and
yeah
I
think
that
I
could
say
that
there
is
some
some
relation
here
right
like
that
I
think,
there's
kind
of
a
general
sense
here
of
you
know,
there's
there's
a
number
of
resources
to
reconcile.
A
They
are
going
to
be
queued
in
order
to
reconcile
them.
Some
of
them
can
be
done
in
parallel.
You
know
in
in
the
past,
we've
done
more
parallelization
than
than
we
had
now
and
so
I
think
that's
you
know
like
this
is
this
is
one
facet
of
it
right
of
like
having
a
work
queue
that
builds
up
when
you're
not
doing
things
in
parallel
fast
enough,
and
then
you
do
more
things
you
do
to
parallel.
A
Like
the
faster
you
get
through
there,
the
more
times
you
forth
the
process
and
don't
reuse
the
existing.
You
know
resources,
the
more
your
consumption
will
explain
so
yeah.
This
is
all
somewhere
related
into
a
fairly
complex
performance.
Like
you
know,
perspective
overall
and
I
think
that
that
is
a
probably
a
good
reminder
here
that
you
know
I
know
the
the
way
that
Albert
is
diving
into
this
to
to
approach
this
I
think
is
going
to
be
fairly
holistic
as
well.
A
I've,
never
seen
our
prayer
approach,
things
in
a
in
a
myopic
perspective,
so
I
think
he's
taking
into
accounts.
You
know
all
these
related
attributes
and
facets
of
it,
and-
and
this
is
a
good
reminder
of
that.
A
Sweet
so
yeah,
so
I
think
that's
as
as
I'm
just
mentioning
there.
Then.
You
know
we'll
be
following
up
soon,
then,
with
a
couple
different
images
to
for
folks
to
continue
testing
and
to
get
some
more
observations
on.
So
we
can
kind
of
make
sure
that
we
are
approaching
this
in
a
complete
way.
That
isn't
you
know,
moving
the
problem
somewhere
else
right
Jesse,
you
got
your
hand
raised.
F
Hey,
hey
guys,
sorry
I
came
to
my
video
on
him
on
the
phone.
You
know,
I
think
it
just
it
temporarily.
It's
very
interesting
that
I
was
asking
some
very
similar
questions
in
slack
about
the
community
provider
and
I.
Don't
think
any
provider
is
really
going
to
be
immune
to
you
know
some
of
the
bottlenecks
that
happen
due
to
the
the
resources
needing
to
be
reconciled
by
single
pods,
and
that's
just
the
nature
of
kubernetes
right
so
I'm
wondering
is
there
anywhere?
F
Does
anyone
on
this
call
and
I
can
I
can
actually
put
it
in
the
community
chat
as
well
that
that's
where
techniques
for
horizontal
scaling
have
been
explored?
You
know
folks,
who
are
familiar
with
what
Autodesk
is
doing
have
seen.
You
know
the
ideas
that
we've
thrown
around
about
you
know
multi-cluster
Solutions
and
distributing
resources
between
clusters
and
and
I.
Just
wonder
if
there's
anything
that's
been
at
all
formalized
in
that
fashion,
because
I
would
love
to
start
to
explore
those
options
as
well.
F
A
Yeah,
that's
a
really
good
point,
Jesse
and
you
know
I'd
like
a
kind
of
reminder
as
well
that
no
matter
no
matter
what
you
know
depth
of
performance
improvements
that
get
made.
You
know
with
the
investigation
we're
currently
doing
and
I
do
have
you
know
fairly
fairly
High
optimism
for
it
like
there
will
still
be
finite
limits
in
terms
of
like
how
many
resources
can
it
still
handle
right.
So
in
terms
of
being
able
to
horizontally,
like
you
know,
Shard
or
scale
in
some
way
to
distribute
the
load?
A
Amongst
you
know,
multiple
entities
that
are
capable
of
processing
them
I'm
definitely
curious
to
hear
if
there
is
some
experience
or
some
suggestions
on
that
as
well.
I
personally,
don't
have
have
something
to
point
to
right
away.
F
A
Yeah
yeah
and
so
yeah
I
wonder
if
yeah,
that
might
be
something
good
to
pose
to
like
the
general
Community
as
well
and
slack
and
see
if
there's
this
resources
that
come
out
of
that.
A
Would
expect
to
see,
as
we
keep
moving
forward
here,
to
see
to
see
progress
on
those
type
of
you
know,
scale
and-
and
you
know,
load
concerns
as
well
as
we
continue
maturing
and
making
more
progress
too.
D
D
That
was
mainly
for
authentication
results
if
I'm
not
wrong,
but
there
was
it
like
I
think
that.
A
Yeah
I
do
recall
that
issue
as
well.
Hassan
I,
don't
remember
any
recent
traction
on
it
or
you
know
recent
discussion
about
pursuing
a
path
on
that
one,
but
that
yeah
I
do
remember
that
the
existing
set
issue
as
well
feel
free
to
drop
that
in
the
chat
Hassan
as
we
as
we
as
we
keep
moving
forward
at
the
agenda
here
all
right.
So,
let's
move
on
then
to
vendor
updates,
so
genre
lets
you
take
away
updates
for.
E
Yeah
I
think
most
of
them
have
been
covered
already
in
the
initial
discussion.
E
I
will
maybe
just
mention
you
know
we
we're
lucky
to
have
a
new
team
member
have
joined
us
recently
and
he's
actually
going
to
as
his
first
issue
going
to
look
into
one
issue,
that's
been
uploaded
quite
a
bit
around
the
ability
to
ignore
changes
similar
to
terraform's
life
cycle,
if
no
changes,
and
so
that's
something
that
we
will
hopefully
make
some
additional
progress
on
soon
other
than
that
you
know
with
like
I
said:
we've
talked
mostly
about
things:
observingly
resources,
external,
seek
results
and
composition.
A
Yeah
I
think
that's
yeah.
These
issues
here,
observably
resources,
external
secret
stores,
competition
validation,
Web
book
yep.
We
did
talk
about
those
earlier
and
that's
great
they're
still
making
progress
and-
and
the
latest
update
you
gave
there.
John
is
actually
really
quite
interesting
about
being
able
to
ignore
changes.
I
think
that's
something
that
you
know.
We
had
heard
requested
a
lot
recently,
but
you
know
I
wasn't
personally
planning
or
knowing
about
you
know
some
some
upcoming,
like
traction
that
was
going
to
be
made
on
that
one.
A
So
I
think
that
will
be
pleasantly
surprising
for
a
lot
of
folks
and
that's
that's
definitely
a
great
issue
not.
B
A
Media
issue
for
for
one
of
our
newest
members
to
jump
into
so
very
exciting
about
that
one:
okay!
Thanks
for
that
update
on
what
I
found
Engineers
are
working
on
there
Jean.
So,
let's
move
into
the
community
topics
section
as
well.
It
looks
like
I
can
completely
drop
the
ball
in
the
content.
Section
normally
I
have
this
populated
out
with
blog
posts
and
in
YouTube
videos
and
stuff
that
has
been
floating
around
the
community
recently.
It
looks
like
I
completely
missed
doing
that
this
time.
A
So,
if
folks
do
have,
you
know
any
sort
of
content
related
to
cross-plane
and
the
community
and
ecosystem
and
such
feel
free
to
add
the
suggestions
to
this
I
will
come
back
after
this
meeting
in
retroactively.
Add
links
in
here,
but
I.
Don't
have
them
right
now,
but
there's
always
buzz
going
on
and
definitely
interesting
stuff
that
people
are
writing
and
printing
and
stuff,
so
lots
of
stuff
I'm
sure
we'll
get
in
there.
A
I
wanted
to
make
a
note
here
is
Philippe
on
the
call.
No
he's
not
I
wanted
to
make
a
note
here
that
our
security
disclosure
policy
has
been
ratified
and
merged.
In
so
we
have
rolled
that
out,
and
we
are
now
following
this
policy
for
any
security
security
dis
issues
that
need
to
be
disclosed
so
feel
free
to
take
a
take.
A
look
through
this.
A
The
the
overall
summary
of
it
is
that
we
have
a
couple
different
general
means
of
reporting
security
issues
in
a
responsible
way
to
the
Crossway
project.
One
is
that
we
have
a
mailing
list
set
up,
that
the
cncf
is
hosting
for
us
that
if
you
find
a
security
issue,
you
can
report
it
there.
A
It
is
not
a
public
list,
so
you
know
it
is
a
responsible
way
to
disclose,
and
also
something
I
think
is
really
really
neat
and
is
rather
new
is
that
we
are
using
the
security
advisory
functionality
that
GitHub
now
has
available.
So
if
you
go
to
the
security
tab
in
the
cross-plane
repository,
you
can
responsibly
and
privately
disclose
a
vulnerability
there
and
then
act
as
a
collaborator
and
contributor
to
the
fix
of
that
issue.
A
While
it's
being
worked
on,
it's
a
really
neat
feature
that
allows
you
know
private
collaboration
and
working
on
fixes.
You
know
outside
of
the
public
eye
before
they
have
been
publicly
disclosed
and
then
once
the
fix
is
available
and
rolled
out,
then
could
be
publicly
disclosed.
So
I,
really
like
that's
functionality
there.
Thank
you
for
Philippe
for
driving
this
policy,
getting
it
defined
and
working
with
us
to
to
roll
everything
out
there.
A
So
that
is
part
of
our
ongoing
effort
to
you
know
mature
the
project
get
ready
for
graduation,
all
that
sort
of
stuff,
and
now
we
have
like
a
fully
fleshed
out
security
policy
that
is
defined
and
being
adhered
to
right
now,
so
I
wanted
to
call
that
out
for
folks
and
feel
free
to
reach
out
with
any
questions,
if
you
have
any
on
that,
just
a
quick
kubecon
update
for
now.
A
So
you
know
we're
talking
before
we
got
started
here
about
who
will
be
in
Amsterdam,
and
some
of
our
North
America
friends
will
looks
like
we'll
not
be
able
to
make
that
trip,
but
we've
got
lots
of
friends
in
Europe
as
well,
which
is
great,
so
I
followed
up
with
the
cncf
to
flesh
out
the
content
that
we'll
be
focusing
on
at
the
contrib
Fest.
I
am
personally
super
excited
about
this,
one
that
you
know
we'll
be
hosting
a
90-minute
session,
which
will
essentially
have
two
different
tracks.
A
One
of
them
will
be
like
a
general
contributor
enablement
session.
Where
we'll
be
going
through.
You
know
setting
up
your
development
environments,
how
to
you
know,
build
and
test
code
for
cross-plane
and
kind
of
bootstrapping
or
fast
tracking,
of
the
journey
for
a
couple,
people
that
want
to
start
writing
code
for
crossblain.
So
we'll
spend
some
time
on
that
and
then
we'll
also
be
spending
time
on
a
couple
of
our
newest
features
that
came
out
in
the
1.11
release
and
by
the
time.
Actually
we
get
to
Amsterdam.
A
Maybe
some
features
for
1.12
too,
but
spending
time
on
walk
through
those
with
our
you
know:
live
environments
and
live
clusters
rolling
those
out
so
I'm,
thinking
of
like
environment,
configs,
I'm
thinking
of
composition,
functions
potentially
observe
only
resources.
If
we
have
something
you
know
by
Amsterdam
to
play
with
there,
but
basically
kind
of
fast
tracking
people's
adoption
and
consumption
of
those
resources,
kind
of
like
Hands-On
tutorials
going
through
and
using
those
new
features
together.
So
that'll
serve
a
couple
different
purposes.
A
One
will
be
that
you
know,
will
help
get
people
using
them
and-
and
you
know,
being
successful
with
them
and
getting
some
of
the
the
issues
out
of
the
way,
but
also
we
will
get
feedback
directly
from
people
as
well.
So
I
think
this
will
be
a
pretty
high
bandwidth
high
value
session
together.
Where
we'll
be,
you
know
enabling
new
contributors
we'll
be,
you
know
using
new
features
together,
but
I
definitely
will
need
help
with
this.
So
I'm
running
this
session,
but
I
will
be
able
to
do
all
of
that
myself.
A
So
I'll
probably
be
looking
to
lean
on
the
maintainers
here
that
will
be
in
Amsterdam,
especially
some
of
these
feature
authors.
You
know
that,
like
wrote,
that
wrote
these
big
features
to
help
kind
of
make
sure
that
people
are
getting
unblocked
and
you
know
we're
getting
getting
obstacles
out
of
the
way
so
I'll
be
reaching
out
to
folks,
but
if
people
want
to
volunteer
to
contribute
and
help
as
part
of
this
session,
please
do
let
me
know
and
we'll
get
you
involved
so
that
you
can
be
there
in
present.
A
All
right
we've
talked
about
this
before,
but
yeah.
The
adopters
list
continues
to
be
updated
over
time,
thanks
Jesse
for
making
a
great
contribution
to
it
this
week
as
well.
But
if
you
are
using
cross
plain,
feel
free
to
make
an
edition
of
your.
You
know
your
name,
your
company
to
that
list
as
well.
A
Okay,
continuing
to
move
along
I
got
some
reports
today
of
some
unsolicited
spam
in
the
cross-plane
slack
from
from
an
individual
there
and
like
I,
wanted
to
have
a
reminder
here
that
in
general
you
know
spam
in
inappropriate
conduct
is,
is
not
part
of
what
our
what
our
community
guidelines
essentially
right.
A
We
do
adhere
to
the
cncf's
code
of
conduct
and
which
is
somewhat
broad,
but
you
know,
has
a
lot
of
of
room
for
people
being
comfortable,
and
you
know
feeling
good
as
being
part
of
the
community
and
having
positive
interactions
right
so
I've
reached
out
to
that
individual
to
to
stop
that
behavior.
A
But
I
wanted
to
give
a
reminder
to
folks
are
always
welcome,
please
to
reach
out
to
the
steering
community
and
we
have
a
steering
Community
Channel
there,
or
always
just
reach
out
to
me
directly
I'm,
always
happy
to
accept
any
sort
of
messages
coming
in
when
there
are
concerns
about
things
going
on
in
the
community,
so
yeah.
So
please
adhere
to
the
community
guidelines,
treat
each
other
with
respect
please
and
reach
out
with
any
concerns
directly
and
immediately
and
we'll
take
care
of
it.
A
I
think
Pete
is
not
actually
able
to
make
it
today,
so
we
won't
be
able
to
talk
about
this
one,
but
it
looks
like
there
is
an
issue
to
discuss
removing
having
a
master
docs
version.
So
I
don't
know
if
there
is
more
discussion
on
this.
That
folks
wanted
to
bring
up
now,
there's
absolutely
an
opportunity
to
asynchronously
discuss
this
with
Pete
on
the
issue
here,
but
at
least
did
want
to
bring
this
up
for
some
attention.
So
folks
could
go
to
that
issue
and
attribute
there
too.
A
All
right
and
then
I
think
maybe
the
final
agenda
item
here
that's
published
at
least
is
that
Bob
and
Hassan
I
think
you
two
will
be
the
most
useful
here.
So
this
is
an
issue
that
has
been
present
for
a
little
bit
and
I.
Think
there's
been
some
good
Research
into
it,
I'm,
not
sure
what
the
best
one
to
focus
on,
though,
is
but
yeah.
A
Essentially
sometimes
this
will
cause
extra
reconciles
when
there's
an
error
or
some
certain
conditions
happen
and
we'll
get
into
a
a
Reconciliation
Loop
that
is
occurring
more
frequently.
They
want,
which
would
be
using
more
resources
and-
and
you
know
maybe
eight
more
API
calls
the
hit
quotas
as
well.
So
this
is
definitely
an
issue
that
Hassan
has
been
focusing
on.
I
think
has
some
recent
updates.
A
D
Outlined
in
this
issue,
I
have
conducted
some
experiments
and,
like
I,
found
out
that,
like
there
are
some
certain
conditions,
just
like
you
mentioned
that
triggers
a
you
know:
reconciliation
loop.
Basically,
the
two
conditions
were
one
is
when
the
status
is
updated
with
a
fluctuating
conditions
which
triggers
like
when
we
update
status,
then
the
resources
killed
again
and
without
backup
or
or
yeah,
basically
without
exponential
backup
fee,
get
another
reconcile
and
get
into
that
Loop.
D
Another
reason
that
I
found
out
was
the
external
creates
pending
annotation
was
being
triggered
with
the
timestamp
changing.
So
basically
there
are,
you
know
two
conditions
that
I
found
out
that
could
trigger
this.
This
Behavior
so
for
solution
like
in
in
the
original
provider
AWS
issue
a
Community
member
proposed
using
the
the
predicates
that
is
existing
in
the
controller
runtime,
like
the
the
three
ones
that
is
listed
here,
which
is
Generation,
changed
annotation
changed
and
they
will
change
basically.
D
Change
changed
is
like
triggered
only
when
the
resource
spec
is
updated,
but
since
we
want
the
resource
to
be
reconciled
and
annotations
or
labels
changed
the
combined
these
three
with
an
orc
condition.
So
this
is
currently
waiting
is
a
pull
request
in
Cross,
plane,
runtime,
but
Nick
correctly
mentioned
some.
Some
concerns
around
that.
Basically,
there
are
some
conditions
where
we
update
the
status
of
the
resource
and
assume
that
that
update
will
trigger
a
new
recovery,
reconciliation
and
actually
I
was
while
I
was
working
on
the
observable
implementation
in
object.
D
I
found
I
I
wanted
to
test
this
with
object,
generated
providers
and
actually
you'll
realize
that
there
is
such
an
assumption
and
the
resources
never
getting
ready.
Since
you
know
we
didn't,
we
no
longer
reconcile
with
the
status
update.
So
overall,
the
the
proposed
solution,
as
is,
is
not
feasible,
so
we
should
either
change
those
assumptions
and
change
the
existing
implementations.
D
So,
instead
of
expecting
a
status
update
to
trigger
a
new
reconciliation,
we
we
would
need
to
make
it
explicit
by
changing
something
else.
Or
you
know,
we
would
need
to
tweak
the
the
predicate
here
in
a
way
that
we
explicitly
rule
out
the
cases
that
causes
this
issue
So.
Currently,
the
the
status
is,
it
requires
further
investigation
and
maybe
some
more
discussion
but
I
think
we
are
on
the.
A
Awesome
is
that,
thanks
for
that
detailed
update
there
and
then
Bob
did
you
want
to
add
some
commentary
as
well
about
your
observations
and
then
in
your
better
terraform
and
such
as
well.
B
Yeah
I
mean
first
big
thanks
to
Hassan
for
all
the
work
that
he
did
and
all
the
documentation
of
all
the
work.
It
was
really
easy
to
follow.
You
know
and
understand
what
the
problem
was.
B
I
I
basically
came
upon
the
same
problem
in
provider
terraform.
You
know
in
my
hunt
for
CPU
Cycles
and
the
terraform
provided
the
terraform
AWS
provider
is
nice
enough
to
change
the
session
ID
in
the
error
message.
Every
time
it
fails,
and
so
you
get
a
new
status
update
or
sorry.
B
You
get
a
new
status
condition
update
on
every
observed
failure
because
of
that
one
string
that
is
different
and
so
I
leverage
the
work
that
sun
did
and
took
those
predicate
changes
and
just
applied
them
to
provider
terraform
since
I
know
they
that
provider
terraform
doesn't
have
the
same
up
jet
synchronization
issues.
B
You
know
the
assumptions
and
that
effectively
resolved
that
problem.
I
mean
it.
Obviously
it
still
fails
once
every
minute,
but
that
is
much
better
than
failing.
You
know
dozens
and
dozens
of
times
a
minute,
so
you
know
for
the
short
term
or
for
maybe
for
long
term
I.
Don't
know
you
know,
provider
terraform
is
is
quote,
unquote,
resolved,
I,
guess.
The
only
other
issue
that
I
was
concerned
about
was
the
fact
that
we
are
updating
or
or
cross-plane
runtime
is
updating
the
failed.
B
The
crate
failed
time
stamp
on
every
reconciliation
which
will
trigger
the
same
thing
because
of
The
annotation
update.
So
the
on
the
creation
side,
if
the
create
were
to
fail,
The
annotation,
update
timestamp
changes
and
that
triggers
another
reconciliation
which
triggers
another
failure
and
that
will
go
on
forever
as
well.
So
I
know
there
was
some
talk
about,
maybe
not
updating
the
timestamp
on
the
create
failure.
B
A
But
yep
that
sounds
that
sounds
good,
so
yeah
it's!
It's
definitely
an
issue
that
I'm
glad
we're
thinking
about
fairly
thoroughly,
because
you
know
when,
when
we
do
something
that
affects
core
reconciliation,
you
know
decisions
and
Behavior
like
within
the
managed
record
seller,
that's
saying
in
control
cross
plane,
runtime
that
absolutely
has
consequences
throughout
the
entire
ecosystem
of
of
cross
planner
providers.
So
it's.
A
Thoroughly
and
thanks
to
both
of
you
for
continued
effort
on
this
and
thinking
through
the
complexity
and
trying
to
get
a
great
solution
here,.
A
All
right,
so
that's
everything
that
was
on
the
official
agenda,
so
I'll
do
a
last
call
here
for
any
other
topics
that
folks
wanted
to
bring
up
before
we
adjourn.
A
All
right!
Well,
if
that's
everything
for
this
session
together,
then
let
me
thank
everybody
for
your
attendance
and
for
your
contributions
to
the
project
here,
so
we'll
keep
making
progress
on
1.12
and
on
all
the
issues
we
were
talking
about
today
and
yeah
great
to
see
everybody
today,
foreign.