►
From YouTube: Cloud Custodian Community Meeting 2021-08-17
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 on our github repo: https://github.com/cloud-custodian/community/discussions
To get an invite to the meeting join the google group and you'll receive one via email: https://groups.google.com/g/cloud-custodian
A
All
right
welcome
everybody.
It
is
august
17th
2021-
and
this
is
the
cloud
custodian
community
meeting
the
public
meeting
that
we
have
every
week
with
the
custodian
community
to
go
over
all
sorts
of
stuff.
I've
put
the
url
to
the
notes
in
chat
and,
if
you're
reading
this
on
the
internet,
we
make
sure
that
we
link
to
those
notes
either
in
the
youtube
description
and
you'll
end
up
on
our
github
page
with
all
the
notes
to
the
past.
So
we
have
an
agenda
for
today.
First
we're
going
to
do
introductions
and
announcements.
A
So
first
we
do
introductions
if
it's
your
first
meeting
or
you
haven't
been
attended
in
a
while,
and
you
want
to
say
hello,
now's,
your
chance
to
say
hello.
This
is
obviously
optional.
So
I'll
just
go.
First,
I'm
george
castro.
I
work
at
saclet,
I'm
a
community
manager
here
for
clock
custodian
and
I
help
organize
things
like
this
meeting
and
all
sorts
of
other
stuff.
Is
there
anybody
else
that
wants
to
say
hello.
A
B
All
right,
let's
have
eric,
go
first
hi,
I'm
eric
hansell.
I
work
at
metadata
solutions.
B
Obviously
I
know
aj
from
working
there
as
well,
and
I
used
to
be
on
a
different
team,
more
of
a
lease
engineering
role
and
then,
after
some
time
I
needed
a
career
change.
So
I
moved
into
the
cloud
team
and
they're
using
this
tool,
so
I'm
just
finding
ways
to
learn
it
and
aj
reached
out
and
invited
me
here.
So
thanks
again,
aj
for
everything
and
good
to
see,
you
again
awesome
welcome.
C
Yeah
welcome
eric
thanks
aj
for
getting
him
to
come
on
over
here.
My
name
is
liz
pronouns.
Are
she
her
and
I
am
a
developer
at
the
kit
at
stacklit
reporting
live
from
san
francisco.
D
A
Awesome
thanks
all
right.
If
you
want
to
follow
along
the
notes,
I'm
I
usually
share
those
and
as
we
have
the
meeting
I
just
type
them
as
as
we
go
so
first,
I
have
a
few
announcements.
I
want
to
get
some
events
out
of
the
way.
First,
so
first
of
all
cloud
custodian
day
is
coming.
A
This
is
the
day
before
kubecon
cloud
native
con
and
you
can
click
there.
Register
to
attend
attendance
is
free,
it's
both
physically
there
in
la
which
is
still
happening.
I
think
and
virtual
as
well,
so
we
have
a
plan
if,
if
kubecon
ends
up
going,
hybrid
or
virtual
will
be
doing
that,
and
we
also
have
a
call
for
papers
happening
that
ends
at
the
end
of
the
month.
So,
if
you're
using
cloud
custodian
in
any
kind
of
clever
way,
you
want
to
share
some
best
practices
with
the
community
and
things
like
that.
A
We
highly
encourage
everyone
to
submit
a
cfp
and
then
liz
zumer,
and
I
will
look
at
it
and
hopefully
put
together
a
schedule
of
awesome
content
for
everyone,
so
they
can
check
that
out.
Also
a
reminder,
since
we've
had
some
questions,
you
do
not
need
to
be
registered
to
kubecon
to
attend
this
event
at
all.
I
know
some
people
might
or
might
not
attend
kubecon
depending
on
the
cost.
So
this
event
just
happens
to
be
co-located
in
the
same
city
really
close
to
where
kubecon
and
cloudnativecon
are.
A
But
it
is
a
separate
process,
so
you
don't
have
to
pay
any
actual
space,
different
fees
or
anything
like
that
to
attend
this.
This
is
a
purely
standalone
event.
Does
anybody
have
questions
about
cloud
custodian
day?
That's
going
to
be,
let's
see
what
what's
the
dates
on
those
sorry.
Governance
is
cloud
as
code
day
with
cloud
custodian,
not
cloud
custodian
day.
I
was.
A
If
you,
if
you
it
has
it,
has
all
that
stuff
and
it
has
a
little
schedule
here.
The
cfp
closes
at
the
end
of
the
month
and
then
we'll
announce
the
schedule
on
september
1st,
and
it
has
all
the
goodies
here
and
if
you
have
any
questions
you
can
ping
myself
liz
or
umir.
Anyone
have
questions
about
that
one
and
then,
of
course,
we
have
the
cloud
custodial
101
workshops
we're
doing
these
once
a
month.
A
You
can
come
here
and
register
for
those
as
well
and
if
you
click
through,
let
me
just
give
you
an
example.
I
wanted
to
show
everyone
the
drop
down
just
in
case,
because
I
know
it
can
be
confusing
sometimes,
but
you
just
follow
that
and
then
you
should
be
fine.
We
should
drop
down
on
this
one
here.
It
is
right.
The.
E
A
E
Not
yeah,
this
is
the
last
one
of
the
series,
so
we
need
to
start
a
new
series.
Oh
okay,
are
we
doing
that?
Do
you
have
a
updated
link?
Actually,
this
is
the
same
link,
but
we
don't
have
a
schedule
because
in
our.
A
You
all
right,
so
those
will
those
will
spin
back
up
in
a
minute.
So
let
me
just
update
the
notes
and
then
do
we
know
when
they're
gonna
resume
again.
E
A
E
No
that
I
guess
the
biggest
thing
is,
if
you
could,
you
know,
sign
up
for
cloud
custodian
day
join
us,
and
I
know
there
were
some
questions
around
that.
Do
you
have
to
be
part
of
kubecon
while
it's
co-located
at
kubecon,
the
entry
is
not
based
on
a
kubecon
entry
ticket
or
something
like
that,
and
we
would
love
to
hear
you
know,
talks
from
end
users.
So
if
you
have
some
please
submit
if
you're,
if
you
have
ideas,
you
want
to
bounce
off
I'll,
just
type
in
my
email.
E
A
Okay,
all
right
so
moving
on
rest
of
the
agenda,
regular
weekly
doc,
sprint
sessions.
This
is
where
on
fridays,
myself,
liz
and
carl.
Can
I
hang
out
on
chat
and
try
to
do
a
pull
request
on
fixing
little
things
in
the
docs?
We
try
to
do
that
once
a
week
to
kind
of
learn
how
the
process
works
and
things
like
that.
Unfortunately,
I'm
in
michigan
and
I
lost
power
last
friday,
so
I
was
not
able
to
attend
liz
and
carl
did
you
have
a
chance
to
meet
anything
for
the
group.
C
Sorry,
I
was
muted
yeah,
we
met
and
we
talked
about
how
I
believe
it
was,
and
I
have
notes-
and
I
I
see
that
you're
also
in
here
carl,
but
I
think
we
talked
a
lot
about
how
the
documentation
around
and
I
think
I.
F
Mode
and
regarding
cultural
mode
and
how
some
some
parts
of
the
documentation
are
spread
out,
we
got
repeated
information
spread
apart
regarding
lambda
boats.
F
A
C
Because
yeah,
because
I
believe
the
what
kicked
off
this
conversation
too,
was
the
one
about
adding.
I
don't
think
I
was
there
for
it,
but
adding
delay
right
and
that
that
was
one
of
the
things
that
you
can
include
as
part
of
that
event.
Mode.
A
Yeah,
that's
my
catch-up
is
actually
is
I'm
this
friday,
I'm
planning
on
adding
notes
for
the
delay,
while
you
all
structure
the
notes
and
then
we'll
we'll
go
from
there,
but
we'll
figure
that
out.
A
If
anybody
in
this
call
is
interested
in
joining
us
feel
free
to
ping
me-
and
I
will
send
you
all
the
information
that
you
need-
we're
basically
just
hanging
out
every
friday,
finding
a
place
of
the
docs
that
needs
help
and
they're
just
making
little
fixes
here
and
there,
and
then
we
plan
on
doing
that
every
single
week
to
just
help
the
project
get
better
over
time.
Any
questions
on
this.
A
C
G
Think
this
has
come
up
before
a
couple
times
in
issues
just
the
whole
idea
of
fleshing
out
some
of
the
making
the
the
ability
to
have
more
dynamic
tags.
I
don't
know
if
anybody
on
the
call
has
any
particular
needs
for
this
or
or
opinions.
I
think
one
of
the
things
that
has
come
up
for
adding
what
we're
trying
to
interpret.
I
think
it
was
pulling
resource
details
into
the
tag
where,
if
we
have,
if
we
match
on
a
specific,
I
think
it
came
up
with
slow
logs.
G
If
I
recall
something
else
where
you
match
a
match:
a
specific
resource,
name
or
there's
something
some
piece
of
the
name
and
you
want
to.
You-
want
to
grab
a
piece
of
that
and
then
set
a
tag
based
on
it,
just
support
for
a
little
bit
more
dynamic
tagging,
and
we
don't
support
anything
like
that.
Now,
kapil
may
be
able
to
correct
me.
I
think
gcp
does
and
with
aws
we
don't,
but
this
could
be
a
big
change.
G
I
just
don't
know
I
don't
know
where
we
want
to
go
with
this
good
place
for
a
for
a
call.
H
So
azure
in
gcp
have
some
notion
of
interpolation,
where
you
can
basically
take
the
resource
value
and
use
it
as
part
of
the
tag
value
that
is
currently
shared
between
gcp
azure.
It
does
not
exist
for
aws
for
aws.
It
means
switching
out
from
bulk
tag.
Api
calls,
on
the
basis
of
some
that
this
type
of
customization
existing
and
devolving
back
to
individual
api
calls
per
resource,
which
can
be
a
significant
differential.
But
I
think
this
use
case
has
been
heard
enough
that
we
should
try
to
at
least
enable.
G
Yeah,
okay,
and
for
I
mean
for
this
specific
issue,
it
looks
like
he's
looking
at
at
using
now,
which
I
know
we
we
interpolate
in
some
actions
that
one
I
think
we
could
still
go
both
with
right.
Yeah.
H
G
Yeah
yeah,
so
I
guess
that's
so.
Would
we
try
to
then
bulk
tag
where
possible
and
then
support
the
optional
resource
attributes
and
and
switch
over
to
the
individual
tagging,
just
as
needed
there
to
preserve
api,
where
we
can.
H
Yeah
effectively,
we
need
a,
I
would
look
at
it.
I
think
the
perspective
that
would
keep
the
code
change
minimal
is
just
changing
the
back
size
on
to
one
effectively.
If
we're
doing,
research,
tags
or
sorry
doing
resource
attribute,
says,
tag,
values.
I
H
G
That's
true
that
would
still
end
up
working
through
universal
tag,
then
right,
because
it
would
just
it
would
just
be
universal
attack
with
more
buckets.
H
We're
talking
about
changing
universal
tag,
we're
just
talking
about
sourcing
tag
value
and
I
think
ty's
brings
up
a
toddler
tj.
Sorry,
I
forgot
what
what's
your
preference
todd
thanks
the
I
think
we
can
regroup
the
values
and
we
can
make
the
batch
size
it'll
be
an
additional
loop,
because,
right
now
we
do
uniform
batches
through
the
whole
set
of
resources.
In
those
cases
we
would
have
several
batches
and
have
to
match
the
batch
so
speak
and
in
the
worst
case
the
batches
are
size.
One.
G
H
You
mean,
like
the
tag
array
type
of
thing.
Sorry.
G
H
Far
as
I'm
thinking
I,
I
also
have
not
reviewed
that
part
code
recently,
but,
as
I
know,
the
main
difference
is
just
the
tag
may
have
some
additional
legacy
variants,
but
from
a
functional
perspective
they
should
be
roughly
equivalent.
H
So
individual
actions
do
like
set
yeah
set.
Logging
on
an
alb
has
a
bunch
of
additional
bags
of
variables
in
this
case
place
for
the
resource
tag
capability.
It
would
just
be
injecting
the
resource
tag
itself.
Sorry,
the
resource
itself
into
it,
with
some
ability
on
doing
that,
jamie's
path,
expression
to
extract
that
value
and
potentially
to
set
a
a
missing
value
on
it.
That
expression
turns
up
nothing.
H
I
I
think
I
was
just
thinking
that
there's,
depending
on
what
you're
doing,
there's
a
different
set
of
interpolation
values
that
you
can
use,
and
I
suspect
that
maybe
even
some
of
those
resources
that,
where
you're
not
doing
any
specific
interpolation,
it
doesn't
support
things
like
now
at
all,
because
we
don't
do
any
interpolation
right.
You
just
provide
a
value,
and
so
maybe
that's
where
some
of
the
same
consistency
is
coming
from
that
is
mentioned
here.
H
So
now
there's
a
couple
variables
get
done
at
a
policy
level
and
now
is
one
of
those,
so
it
should
work
university
across
the
whole
subgang
and
that
top
level
interpolation
is
actually
has
some
awareness
of
the
action
level
vocabularies
around
interpolation,
so
it
can
pass
those
through,
but
almost
all
the
top
level
variables,
like
all
the
top
level
variables
should
work
universally,
and
that
includes
now.
We
also
have
some
extended
date
formatting
around
now
that
allows
for
gay
arithmetic,
which
is
probably
not
documented,
but
I'll.
H
Leave
that
aside
the
but
yeah
now
should
work
universally.
It's
mostly
just
some
actions
have
extended
vocabularies,
which
that
top-level
interpolation
accounts
for
and
passes
through.
A
If
you
wanted
to
plop
something
there
all
right,
moving
on
to
the
next
topic,
moving
this
meeting
to
a
bi-weekly
cadence,
so
we
had
two
short
meetings
in
a
row
and
unless
anyone
thinks
otherwise,
I
feel
like
we
can
move
to
every
other
week
for
meetings.
So
you
know
we
kind
of
have
less
arrows
but
more
wood
behind
each
one.
I
guess
this
is
saying
anyone
have
opinions
on
this.
A
All
right
I'll
go
ahead
and
adjust
the
schedule
accordingly,
then,
on
the
calendar
and
just
as
an
fyi
the
way
the
calendar
works
for
the
meeting
is,
if
you
join
the
google
group,
four
cloud
custodian
you'll
automatically
get
the
invite
to
this
meeting
and
all
the
changes,
and
things
like
that.
So
I
apologize
ahead
of
time
for
the
adjustment
span
that
you're
about
to
get.
A
I
see
it,
I
will
snag
that
all
right
and
with
that
reminder
too
that
we
do
have
time
set
aside
at
this
meeting
for
demos,
and
things
like
that
and
aj
has
graciously
volunteered
to
do
our
first
demo.
So
how
you
looking
there
aj
yeah.
I
think
I
think
we're
doing.
Okay
I'll
try
to
try
to
share
here.
Okay,
let
me
stop
sharing
here
we
go
and
I'll
fix
the
notes.
While
you
do
your
business.
A
B
G
All
right,
let's
see
if
we
can
make
it
make
it
bigger.
Okay,
is
that
is
that
suitably
big?
Thank
you,
sir
okay
yep
cool.
So
this
the
whole
idea
for
this
demo
then
came
out
of
some
questions
that
we
got
in
gitter
about
when
people
are
working
on
policies,
how
they
can
see
where
what
a
given
the
value
filter
takes
a
james
path,
expression
and
just
for
for
reference.
G
G
These
big
type
james
path,
expressions
against
it
to
to
perform
pretty
deep
matches
and
when
people
are
getting
started
with
those
filters,
even
when
you're
not
getting
started
when
you're
just
working
with
a
pretty
gnarly
filter,
you
want
to
keep
working
on
that
and
keep
focusing
a
little
bit
at
a
time
against
resources
that
you
already
have.
G
So
I
just
wanted
to
go
through
a
couple,
quick
ways
that
you
can
do
that
and
I'm
also
curious
anybody
feel
free
to
jump
in
with
questions
or
with
your
own
ideas
on
the
fly
here
with
how,
when
you're
going
through
building
up
filters,
how
you
do
it
because
one
way,
let's
see
one
way
that
I
would
do
it
here-
is
I'll,
usually
start
with
something
like
I'll
just
run
a
policy
with
no
filters
at
all.
I
got
filters
all
commented
out.
Let's
say
I've
got.
G
I'm
filtered
to
two
roles
here,
so
I
have
a
couple
roles
and
I
might
look
at
look
at
roles
that
come
back
and
say.
I
think
the
question
we
got
on
getter
was
trying
to
look
for
specific
principles
under
the
assume
policy
document
here:
the
trust
policy
for
a
role.
So
you
can
see
that
that
information
is
multiple
levels
deep
here,
so
you
run
a
policy
with
with
no
filters
first,
and
you
can
see
the
kind
of
data
that
you're
looking
at
and
then
from
there.
G
G
Right
here,
we've
we've
got
the
all
the
data
here
you
can
just
try
to
add
filters
here
and
build
them
up
a
little
bit
at
a
time
and
see
what
kind
of
matches
you
get
back.
If
you
get
comfortable
enough
with
the
james
pass
syntax,
then
that's
probably
okay,
you
might
do
fine
with
that.
There's
a
games
path,
cli
that
you
can
run
against
your
output,
I'll
pop
that
link
in
the
chat
as
well
jamespatch
cli.
You
can
run
against
your
resources.json
and
kind
of
explore
some
of
those.
G
If
you
give
it
a
couple
brackets,
it
will
it'll
flatten
down
and
start
pulling
some
things
out
that
way,
and
then
you
can
keep
digging
a
little
bit
at
a
time
and
see
the
kind
of
stuff
that
you
get
back.
So
you
might
start
and
say.
I
want
to
look
at
the
assumer
policy
document
and
you'll
get
that
back
for
each
each
entry
in
your
file.
And
then
you
might
look
at
the
statement
and
each
each
time
you're
running
through
the
cli.
G
You
get
rid
of
this
little
first
part,
because
this
is
you're
running
against
the
whole
set
of
resources
and
whereas
in
custodian
you're
going
resource
by
resource,
so
you
can
do
the
rest
of
it
and
that'll
help.
You
build
up
a
filter
one,
but
I
find
that
that
can
be
the
jp
cli
can
be
pretty
useful,
but
it's
also
you
can
end
up
getting
into
a
bit
of
a
wacky
loop.
So
in
the
getter
channel
again
a
couple
probably
a
couple
years
ago,
kapil
mentioned
hey.
There's
this
tool
out.
G
There
called
busy
data
and
it's
pretty
neat
just
kind
of
an
off-handed
comment:
hey
it's
pretty
neat
for
looking
at
resources
from
your
custodian
policies,
so
data.org,
jp
or
jpterm.
I
don't
know
jb
term
say
we
got.
We've
got
new
tools
here,
but
so
this
this
may
all
be
moot
or
we
can
add
jp
term
to
it.
There
may
be
other
ways
to
do
it.
G
Jp
term
sounds
vaguely
familiar,
but
with
busy
data
I
can
go
pop
into
an
output
directory
and
pull
up
some
resources,
and
then
I
see
the
nested
data
with
one
resource
per
row
and
there's
no
way
to
obviously
get
the
to
run
james
path,
expressions
against
these,
but
you
can,
but
I
did
add
and
I'll
put
a
link
here.
I
thought
how
bad
would
it
be
to
write
a
plugin
to
do
this
and
it's
really
not
bad.
G
G
Look
at
what
what
you
get
out
of
that
and
each
time
you'll
see
the
the
expression
as
the
column
heading
and
you'll
see
the
results
in
there
in
here.
This
is
going
to
be
a
list,
so
you
say:
oh
there's
a
list.
There's
principles
in
here,
so
maybe
I'll
add
something
else
and
what
a
principal
and
each
time
it's
like.
Oh
well,
now,
there's
there's
aws
in
there,
so
you
can
keep
kind
of
iterating
on
that.
G
G
G
Does
this
oh
and
compel
just
like
jp
term,
that's
great
go!
Look
at
that,
but
the
whole
idea
here
is
between
the
jp,
cli
or
or
just
exploring
with
your
own
intuition
or
busy
data
plugins
or
jp
term.
Just
getting
that
cycle
of
I
have
the
data
I
need.
I
have
it
locally.
Let
me
explore
some
expressions
to
find
the
values
that
I
actually
care
about.
I
One
of
the
resources
I
want
to
narrow
in
on,
and
I
just
cut
and
paste
the
whole
thing
into
gmes
staff.org
in
their
ui
and
do
there
because
it's
real
time
updating
as
you
edit,
your
expression,
it'll
change
stuff.
But
it's
not
necessarily
if
you
give
it
an
invalid
expression,
nothing
changes
in
your
output,
so
it
can
be
confusing
on
whether
or
not
you
have
a
valid
expression
or
not,
but
that
that
tends
to
be
where
I
start
so.
Similar
idea
right.
H
Yeah,
not
everyone's
comfortable
pacing.
Their
data
on
the
web,
but
jamie's
path
terminal
is
effectively
that
same
experience
locally.
It's
especially
that
interactive
style
live.
It's
like
a
three-pin
view
where
left-hand
source
right
hand
is
like
results
and
top
is
basically
your
expression,
as
you
type
the
expression
it
auto
does
stuff,
and
it
has
the
same
issue
as
far
as
not
doing
anything
that
the
expression
itself
doesn't
know.
G
Yeah,
that's
good.
This
is
exactly
what
I
wanted
to
talk
about
on
a
meeting,
because
I
figured
other
people
that
have
different
ideas
on
it.
One
other
thing
that
I
wanted
to
point
out
here
is
that
the
value
filter
does
some
nifty
stuff
that
I
think
a
lot
of
times.
People
are
asking
about
because
they're
not
really
sure
what's
going
on
under
the
hood
and
something
like
when
we
do
a
tag
called
a
name.
G
That's
really
a
simplified
thing
that
from
the
james
path
side,
these
two
filters
are
kind
of
equivalent.
We're
saying
like
let
custodian
do
this
transformation.
This
look
up
under
the
hood
to
look
for
a
bucket
with
a
specific
name
tag,
or
you
can
just
go
right
into
that
tags.
Object
yourself!
That's
just
a
it's
just
a
pretty
gnarly
expression
to
get
there,
and
usually
you
wouldn't
want
to
do
this,
but
we
had
a
question
through
either
github
issues
or
getter
a
while
back.
G
How
can
I
tell
if
my
bucket's
name
tag
matches
the
actual
name
of
the
bucket,
and
that
was
one
of
those
cases
where
we
could
do
it
with
a
filter
like
this?
It's
just
pretty
awkward
and
a
lot
of
times
when
there's
a
when
you're
writing
a
pretty
awkward
filter,
there's
a
there's
something
other
than
the
value
filter
that
can
get
you
there.
G
So
that's
and
also
some
of
the
other
filters
add
their
own
annotations
to
a
resource
which
you
can
then
use
the
value
filter
to
look
up.
So
there's
a
lot
of
little
things
that
we
can
go
into
more
on
other
demos
or
other
questions.
H
F
H
Notion
of
an
annotation
where
you
say
cross
account
role,
and
it
adds
an
annotation
so
that
typically
either
for
reporting
purposes
or
for
actions
to
do
remediation.
They
can
take
some
action,
but
what
those
keys
are
for
purposes
of
users
using
also
using
them
in
their
policies,
is
particularly
clear
because
they're
not.
I
I'll
also
note
that
I
think
I
think
it
was
a
security
group
filter
or
something
like
that
that
I
helped
figure
out
how
to
solve
for
somebody
who
asked
about
it
or
where
are
you
trying
to
figure
out?
I
H
Yeah,
that's
a
great
point
like
jamie's
path
you
can
has
functions,
has
a
lot
of
pretty
interesting
capabilities.
It's
actually
built
into
both
the
aws
cli.
So
if
you
do
query
on
the
wwc
a
lot
you'll
have
jamiespa
data
for
azure,
although
I
think
they
have,
it
might
be.
I
think
it's
the
same
field,
but
you
can
do
functions
like
counts.
You
can
do
maps
projections,
lots
of
different
things.
H
It's
definitely
a
cool
language
and
definitely
worth
a
look
if
and
a
key
part
of
something
that
kassadin
exposes
as
its
workhorse
for
for
users
on
policies.
B
G
Good
call
well
thanks
I'll
hand
it
back
to
george,
then
we
can
continue
rolling
all
right,
any
any
other
questions
for
aj.
A
All
right
moving
on
to
some
pull
requests.
There
are
two
that
I
have
on
the
agenda
today.
Aj
these
are
yours.
I
believe
6837
make
normalized
tag
universal.
G
G
Yeah
that
one
I
I
mean
could
be
worth
discussing
here
just
because
it
is
changing
the
way
that
that
that
action
is
going
to
make
it
more.
I
mean
to
be
universal,
and
so
it
applies
for
more
resources.
I
think
it's
it.
We
had
a,
and
this
is
where
I'm
curious
about
the
the
feedback
from
folks.
We
had
a
normalized
tag,
action
it
had
a
the
schema
said
that
it
supported
a
replace
sub-action,
but
but
it
didn't
it
didn't
seem
to
do
anything.
G
I
guess
because
it
was
people
would
just
use
rename
tag
instead.
So
I'm
kind
of
stealing
that
replace
effect
here
and
trying
to
have
it
do
a
regex
replace.
I
don't
know
if
this
is
the
right
way
to
do
it.
If
somebody
has
better
ideas,
different
ideas
or
how
this
feeds
into
some
of
the
other
tagging
work.
That's
going
on.
H
My
only
thought
is
like
we,
like
normalized
tag,
has
say
a
couple
different
quotations
and
right
now
this
is
fairly
focused
on
value
and,
I
think,
there's
been
use
cases
around
key,
but
that
wouldn't
necessarily
be
in
this
action
but
yeah.
H
I
think
it's
maybe
worthwhile
to
take
a
step
up
like
it
doesn't
have
to
be
part
of
this
vr,
like
I'm.
Just
I'm
just
thinking
like
on
a
project
basis
like
maybe
it's
worthwhile
for
us
to
take
a
step
back,
take
a
step
up
and
be
like
what
are
all
the
different
use
cases.
People
have
around
dealing
with
tags
and
making
sure
we
actually
cover
off
of
them
rather
than
trying.
H
What
we
have
been
doing
is
effectively
piecemeal,
targeted
stuff
based
on
an
individual
use
case,
but
I
feel
like
there's
potentially
some
stuff
like
like
tag
key
normalization
and
other
things.
That
would
be
worthwhile
if
we
went
up
a
level
looked
at
all
the
different
use
cases
and
just
try
to
identify
what
our
headless
and
that
bar.
H
Yeah,
it
is,
but
it
is,
but
I
think
it's
a
useful
exercise
and
also
is
worth
while
to
make
sure
that
these
things
are
lined
up
to
where
they,
where
we
want
to
go.
Directionally.
H
But
I
don't
want
to
tie
them
as
a
strict
dependency,
but
I
still
think
it's
definitely
a
useful
thing
because
it
may
inform
what
we
merged.
So
I
wouldn't
want
to
like
get
it:
how
to
release
this
stuff
out
per
se.
If
we
potentially
realize
that
we
just
did
this
and
this,
then
we
don't
have
to
have
a
concern
for
users
as
far
as
policy
compatibility.
A
H
Priority
list
yeah
yeah
yeah.
I
totally
know
what
that
means
now.
Okay,
now,
that's
not
that
one,
I
think,
is
pretty
straightforward,
although
again
like,
if
we're
actually
looking
at
it
from
that
holistic
perspective,
you
know
I
mean,
is
really
what
we're
saying:
they're
just
tag:
normalized
yeah.
G
A
All
right,
so
that
seems
to
cover
that
anybody
else
got
any
pr's
or
burning
issues.
H
A
All
right
that
pretty
much
finishes
it
up
thanks
everyone
for
coming
we'll
be
publishing
the
notes
and
everything
onto
the
list,
as
well
as
the
youtube
video,
and
with
that
we'll
give
everyone
15
minutes
back,
see
you
in
september.
Yeah,
don't
forget
to
register
for
cloud
is
governance
day
cloud
code
is
governance
day
by
cloud.