►
From YouTube: Cloud Custodian Community Meeting 2023-06-27
Description
Our community meeting is public and we encourage users and contributors of Cloud Custodian to attend! You can find the notes for this meeting in both GitHub Discussions and in HackMD:
- https://github.com/orgs/cloud-custodian/discussions
- https://hackmd.io/@c7n
Check out our Slack for more info! http://slack.cloudcustodian.io
A
All
right,
we
are
recording,
welcome
everyone
to
the
to
the
June
27th
Cloud
custodian
community.
Meeting
I
am
going
to
put
the
link
to
the
agenda
in
the
chat
and
then
get
a
screen
share
going
and
then
we
can
start
walking
through
looks
like
some
familiar
familiar
faces
here.
So
we,
the
the
introductions,
may
be
a
moot
point
today.
B
Hey
yeah,
so
I
have
a
note.
I
do
have
a
PR
that
I
was
hoping
to
review
just
pass
that
out.
A
Yeah,
actually,
we
may
be
able
to
jump
right
into
that
then,
because
the
I
mean
normal
normal
deal
with
intros.
We
still
have
the
note
about
moving
on
from
python
37
this
year,
I
talked
to
Kapil.
I
know
he
won't
be
able
to
make
today's
call
because
of
that
conference
the
Phenom
sacks,
but
he
did
mention
that
we're
looking
at
a
release
sometime
this
week
might
be
today
or
tomorrow,
but
sometime
this
week
we
should
get
the
next
release
out.
A
I
know:
we've
got
a
couple
PRS
in
anybody
have
questions
or
important
things
coming
up
before
we
start
diving
into
looking
at
specific
issues
or
PRS.
A
A
All
right
cool,
oh
hey,
Sonny,
so
you
jump
into
yeah
so
Ryan
that
was
you
mentioned.
You
had
one.
B
A
A
A
Gabriel
welcome
yeah,
thanks
for
thanks
for
looking
at
that
too
Ryan
just
to
confirm
what
was
going
wrong
with
it.
That's
good
everything
else
looks
cool
good.
Any
other
issues
or
PR's
folks
want
to
call
out.
C
Just
just
my
usual,
which
apparently
I
need
to
figure
out
how
to
test
some
utility
tidbits
to
get
the
code
coverage
up
to
get
it
to
pass,
pass
the
code
coverage
test
so
that
we
can
merge
it.
A
C
Got
it
you
got
it
exactly
where
it
needs
to
be
refactored
how
it
calls
labels
so
that
now,
instead
of
having
to
modify
core,
you
modify
the
individual
resource
with
the
with
you
just
add
a
function
to
the
individual
resource
that
to
refetch
so
that
all
it's
doing
now
is
if
it
gets
the
fingerprint
error.
It
checks
to
see
if
the
function
exists,
calls
The
Reef.
C
If
the
function
exists,
it
calls
the
function
to
refetch,
otherwise
it
fails
out,
and
so
that
actually
made
core
look
a
lot
better,
because
now
we
don't
have
funky
per
resource
logic
in
there
and
the
resource,
the
availability
of
or
the
ability
to
make
the
retry
is
determined
by
the
resource
and
not
buy
core
Now
by
The,
gke
Core,
so
yeah,
it's
it's
just
down
to
I,
guess
getting
writing
tests
for
things
that
apparently
aren't
covered
now.
A
Okay,
I
just
approved
another
run,
it
looked
like
it
was
stuck
waiting
for
something
was
that
just
did
you
have
test
updates
in
that
last
comment,
or
was
that
just
pulling
in.
D
C
A
All
right
take
a
look
and
see
I
know.
I
know
we
looked
at
some
of
the
stuff
before.
Let
me
look
and
see
if
we
can
get
anything.
D
C
A
Yeah
GP
bits,
but
but
for
the
testing,
maybe
we
can.
C
Get
I
mean
like
util's,
not
Pi,
it
is
a
one-line
change,
but
that
one
line
is
apparently
enough
to
throw
off
for
it
to
not
have
full
coverage
right
so
and
I
guess:
I
have
to
write
individual
tests
for
the
the
retry
function.
Isn't
that
great,
because
I
think
that's
mostly
what
the
if
I,
if
I
look
at
it,
the
I
think
that's
most
of
the
changes
that
it's
flagging
on
in
gke?
C
C
I'll
have
to
figure
out
exactly
what
it's
what
it's
flagging
on,
but
it's
incredibly
easy
to
now
add
fingerprint
or
not
fingerprint
label
functionality
to
something
that
isn't
a
gke
cluster
or
just
a
standard,
compute
node
as
a
proof
of
concept,
I
added
it
for
images.
Awesome.
A
C
A
Okay,
cool,
so
that
seems
good
is
anyone
else
on
the
call,
particularly
familiar
with
with
the
Google
side,
with
the
gcp.
C
A
Well,
we'll
try
to.
It
may
be
one
of
those
things
where
it
just
picks
a
village
to
get
a
gcp
right
and
then
we're
all
learning
something.
So
thanks
again
for
sticking
with
it.
A
Cool
others,
anyone
else
I
know
I
saw
something
you
came
into.
Did
you
you
have
anything
or
sending
British
I
knew
there
were.
There
were
a
couple
things
coming
through
recently
on
I.
Think
we
one
or
more
of
us
got
pinged
on.
Oh,
the
the
mailer
stuff.
There
was
something.
D
Yeah
there
was
one
fear
that
I
was
trying
to
get
some
review
on
and
a
couple
is
out
for
Finn
up
so,
but
if
you
have
time,
if
you
are
able
to
look
at
the
SNS
notify
action
message
to
longer,
that
would
be
helpful.
A
A
D
And
when
you
have
like
a
lot
of
metadata
around
resources,
so
we
had
a
scenario
where
every
bucket,
in
an
account,
had
multiple
life
cycle
rules
which
caused
the
256
KB
threshold
to
which,
like
crossed
that
threshold
with
250.
So
we
like
lowered
it
down
to
100
on
our
end
and
then
it
didn't
go
through
either
and
I
think
this
was
something
that
couples
wanted
to
Implement
from
the
dynamic
that
sizing
point
of
view.
D
The
pr
is
still
the
tests
need
to
be
done,
but
I
think
potentially
the
logic.
If
that
makes
sense,
I
should
be
able
to
go
ahead
and
write.
Some
tests.
A
B
A
Sounds
right
and
certainly
take
a
look,
but
that
sounds
like
what
we
were
talking
about
doing
anyway,
and
no
so
you
I
I,
know
there's.
We
would
still
need
tests
for
this,
but
you've
you've
run
it.
No,
no
weird
surprises
just
in
and
what
you've.
D
D
Yep
not
that
I'm,
aware
of
but
yeah
I
would
love
to
have
some
of
you
guys
also
look
at
it.
D
A
Sure
we
cover
the
packed
stuff,
yeah,
so
cool
thanks
for
working
on
that
one
I
did
want
to
point
out
this
one
too
I
know
Kent
had
had
a
PRN.
This
looks
pretty
cool
I
think
we're
going
back
and
forth.
A
I
I
have
I
own
a
review
on
the
latest
round
of
changes,
but
just
in
case
anyone's
using
any
of
the
folks
who
use
mailer
just
curious
if
anyone
was
using
internal
mailer
methods
or
doing
any
kind
of
particularly
advanced
stuff
with
mailer
that
might
that
might
break
due
to
a
refactor.
This
looks
like
all
not
changing
anything
with
the
way
that
policy
interfaces
with
mailer,
but
some
of
the
under
the
cover
stuff,
which,
if
anyone's
looked
at
the
mailer
code,
it
can
be
a
little
bit
naughty
to
get
through.
A
Sometimes
this
is
trying
to
make
it
a
little
bit
clearer
seemed
like
mostly
good
changes
to
me,
but
anybody
anybody
work
with
Miller
code
much
or
use
these
internal
bits.
A
E
E
A
We
are
in
the
middle
of
doing
some
issue,
PR
review
if
anybody
else
has
specific
questions
or
issues
or
PR
as
to
to
call
out
for
the
group.
Now
is
the
time.
B
I
actually
do
have
one
question
on
that:
PR
that
we
had.
Yes,
if
you
still
got
it
open,
so
when
I
looked
in,
it
doesn't
seem
like
this
has
an
Arn
type,
the
receipt
rule
set.
E
B
A
A
B
A
Yeah,
some
don't
so
I
might
have
to
look
for
that
and
make
sure,
because
this
this
gets
pulled.
This
gets
generated
from
the
docs.
So
if
we
don't
see
it
in
the
docs
when
we're
here,
when
we
regenerate
this
file,
we
won't
see
it.
So
we
either
need
to.
A
We
might
have
to
just
flag
that
as
a
as
a
skipped
Resource
as
a
type
without
an
arm
and
then
just
say,
I
think
it's
has
Arn
is
false,
so
that
it
won't
try
to
generate
one.
The
only
thing
is
I,
don't
know
if
we
need
one
usually
so
wait.
What
is
this
thing?
Resource
type
and
delete,
sometimes
where
we
end
up
needing
an
RN,
is
if
we're
using
the
bulk
tagging
API,
because
that
needs
one.
So
we
have
to
either
pull
one
from
wherever
it's
available
or
build
one
on
the
Fly.
A
But
if
this,
if
there's
no
reference
to
an
aren't
anywhere-
and
we
don't
need
it
and
we're
not
using
let's
see,
do
we
have
yeah,
we
don't
have
any
and
we're
not
pulling
any
kind
of
tags
or
anything,
and
we
might
have
to
just
tell
it
that
it
doesn't
have
an
iron
I
can
verify
that
we
can
try.
Let's
see,
can
you.
A
B
Yeah,
that's
what
my
comment
was
about
because
that'll
get
it
to
pass
the
tests,
but
there
is
still
that
Arn
type
on
the
resource,
I'm,
not
sure
about.
A
A
A
A
A
So
we're
back
to
the
PRS
and
issues
if
there's
anything,
let's
see
I
know
so
Robert
from
on
the
Oracle
side.
I
know
that
the
initial
support
for
the
the
Oracle
provider
landed.
Is
there
anything
from
your
side
going
on
with
that
as
follow-ups
that
you've
come
across
or
or
no.
E
I,
don't
think,
there's
any
follow-ups
on
that
initial
PR.
At
the
moment
we
are
looking
into
other
other
changes.
You
know,
based
on
some
of
the
initial
feedback,
that
to
follow
up
on.
E
A
Yeah
I
think
we
would
see
See
the
updates
when
they
come
out
and
it's
like
we
used
to
do
them.
We
would
do
releases,
live
sometimes
and
record
them,
but
I
think
what
the
last
changes
I
know
is
a
sunny
and
Kapil
to
both
work
on
some
of
the
stuff
recently
to
to
automate
away
a
lot
of
the
manual
work
that
goes
into
it.
So
it's
not
so
much
recording
now
I
mean
we
can
post
and
certainly
link
to
where
to
all
the
steps
I
mean.
A
E
A
E
A
Animated
stuff,
eventually
it
will
be
just
and
actually
for
this
release,
I'm
glad
you
asked
that,
because
for
this
release
we
may
be
able
to
just
follow
it
in
follow
the
action,
progress
and
GitHub
actions.
A
Sonny
may
be
able
to
keep
me
honest
on
that.
One
I
may
be
lying,
so
I'll
check
to
make
sure,
but
that
would
be,
and
that
would
be
the
idea
going
forward.
Is
you
would
just
see
a
log
like
you're
talking
about.
A
Okay
yeah,
so
this
one's
talking
about
going
through
and
building
everything
and
then
doing
the
release,
so
we
should
I,
don't
think
this.
This
flow
I
don't
think
we
use
this
flow
for
the
last
release,
but
this
may
be
the
first
one
where
we
do
so.
In
that
case,
we
would
watch
and
go
look
for
the
actions.
Yeah.
No,
no
workflow
runs
here,
but
this
would
be
where
to
find
it.
Once
we
start
using
this
action,
we
would
see
the
release
the
release
flow
here.
E
A
All
right:
well,
thanks,
I'll
get
the
notes
and
the
recording
published
and
try
live
streaming
again.
Another
time
thanks,
everybody
have
a
good
rest
of
the
week.