►
From YouTube: Carvel Community Meeting - March 31, 2022
Description
Carvel Community Meeting - March 31, 2022
We meet every Thursday at 10:30am PT. We'd love for you to join us live!
This week we had some great discussion covering ytt, kctrl, and a demo on what the team is working on with app commands. Check out full agenda here: https://hackmd.io/F7g3RT2hR3OcIh-Iznk2hw#March-31-2022-Agenda
A
Hi
everyone
welcome
to
this
week's
edition
of
the
carmel
community
meeting.
Today's
date
is
march
31st
2022..
If
you're
watching
this
recording
on
our
youtube
channel,
we
welcome
you
to
come.
Join
us
live
we
meet
every
thursday
at
10,
30
a.m.
Pacific
time
it's
just
an
opportunity
for
you
to
come
and
meet
other
members
of
the
community
meet
the
maintainers.
Listen
in
what
we
are
working
on.
A
You
can
ask
questions,
provide
feedback
if
you
have
anything
that
you
need
help
with
any
of
the
karmal
tools,
and
you
would
like
to
have
some
some
live
discussion
on
it.
It's
a
great
way
to
to
do
that.
Otherwise,
you
can
find
us
on
our
kubernetes
slack
workspace
channel
on
the
carville
channel
or
on
twitter
at
karbal
underscore
dev
at
any
point
that
you're
interacting
with
us
or
any
members
of
the
community.
A
A
A
B
Yeah
sure
nancy
yeah,
so
this
is
just
a
little
a
leg
up
here.
If
you've
been
digging
into
ytt
like
pretty
quickly,
you
find
that
you
want
to
use
overlays,
and
then
I
don't
know
when
I
first
did
that
I
was
pulling
my
hair
out.
So
I
was
hoping
to
help
folks
have
a
healthier
head
of
hair
by
showing
a
few
of
the
things
that,
like
I've,
learned
along
the
way
like
if
you
just
knew
that
that
can
really
help.
B
B
We
get
to
walk
someone
along
the
way
of
developing
their
like
kind
of
make-believe
deployment
and
what
they
would
use
overlays
for
and
what
that
would
look
like
and
how
to
debug
them,
how
to
how
to
read
them.
Stuff
like
that
so
yeah.
I
hope
it
helps
out
and
makes
that
part
of
the
journey
a
little
bit
easier
and
accelerates
your
learning
and
and
gets
you
back
to
the
point
where
you're
being
productive
with
the
tool.
That's
our
that's
our
goal.
C
B
A
D
A
Awesome
thanks
garrett,
so
for
for
content,
you
know
the
team
is
pumping
out
some
stuff
things
that
they're
passionate
about
things
that
they
want
to
share
regarding
carvel,
but
it's
not
just
limited
to
to
the
maintainers
here.
We
we
would
love
to
know
how
the
community
is
using
cargo
and
any
ways
that
they
want
to
share
their
use
cases.
A
So
you
know
how
are
you
using
cargo
with
any
other
tools
that
are
out
there
and
you
want
to
write
a
blog
post
just
share
a
demo,
anything
that
you're
comfortable
with.
We
have
this
signup
sheet
on
the
repo
that
you
can
add
your
name
to
and
the
topic
that
you're
going
to
be
covering
what
type
of
medium
you're
going
to
be
sharing
that
information
in
and
if
you
wish
to
have
your
your
twitter
handle
tagged
on
when
we
promote
it
through
twitter.
A
You
can
add
that
here
too,
if
there's
a
future
date
that
you
have
in
mind-
and
it's
not
listed
on
this
table-
feel
free
to
add
more
rows
and
add
that
date
in
that.
You
would
like
to
share
that
information
and
for
anyone
who
isn't
a
maintainer
and
is
going
to
be
participating
in
this,
we
would
love
to
send
you
a
t-shirt
or
any
other
type
of
karbal
swag
that
we
have
so
once
you
fill
it
out
I'll
reach
out
and
get
your
information
and
share
that.
Send
you
that
swag
as
a
thank
you.
A
C
So
powerful
is
some
good
content,
and
I
just
realized
that
I
was
totally
missing
an
opportunity
to
like
bring
it
up
here
like
that,
that
blog
post,
that
john,
that
you
posted,
like
that's
some
content,
that
I
should
probably
like
point
people
to
and
is
there
anything
else
that
I
should
point
people
at
some
particular
examples.
I
know
there's
the
kubernetes
ytt
library
that
could
be
a
good
resource
for
people
as
well,
but
just
wanted
to
get
some
feedback.
A
E
A
F
I
know
that
image
pack
is
going
to
be
released
next
week.
There's
a
problem
with
the
last
two
builds
that
we
we
made
a
change
and
now,
when
you're,
trying
to
mount
a
blog
into
two
registry
into
two
repositories
in
the
same
registry,
it
is
failing.
So
we
know
that
that's
an
issue
there's
an
open
issue.
There
now
we're
working
on
solving
it.
H
C
Sorry,
the
this
pull
request,
399
the
draft
on
modularity
in
ytt
when
to
use
starlark
modules,
data
values
or
a
private
library
is
really
interesting
to
me,
and
I
was
talking
with
rupa
and
dimitri
yesterday
about
some
of
the
other
types
of
guide
content
that
we're
working
on.
I
think
carrie,
you
and
john
are
kind
of
like
focused
on
this
guides
kind
of
story.
Right
now,
right.
H
C
E
So
we
have
been
hammering
the
features
you've
been
talking
about
about
for
a
while
into
shape
like
the
ones
which
sort
of
help
users
observe
apps
better.
So
that's
what
that's
about
so
I
was
actually
hoping
to
sort
of
give
a
sneak
preview
of
such
into
what
we've
been
working
on
so
yeah.
If
there
are
no
other
updates,
maybe
you
can
do
that.
E
Their
first
instinct
is
not
to
sort
of
look
at
the
app
crs
that
are
created
by
package,
installs
or
sort
of
get
the
package
installs
entire
resource
to
look
at
the
useful
error
message
and
we
sort
of
wanted
to
give
the
user
better
observability
into
what's
actually
going
on
when
an
app
or
a
package
installation
is
reconciling
so
yeah,
that's
what
this
is
about.
So
if
I
quickly
share
the
app
will
be
done
using
the
demo,
it's
it's
a
pretty
straightforward
app.
E
B
E
Yeah,
that's
the
app
that
I
was
talking
about.
It's
pretty
straightforward.
We
pull
a
bundle.
We
template
it
with
cable.
So
that's
to
has
the
image
references
and
deployed
using
cap
and
the
underlying
bundle
looks
something
like
this:
a
job
which
sleeps
for
15
seconds
and
and
then
it
sort
of
finishes
successfully.
E
E
So,
as
you
can
see,
we
show
that
we're
fetching
from
this
location
that
we
templated
successfully
and
then
as
and
when
cap
control
sort
of
updates
the
deploy
status.
We
also
surface
that
information
to
the
user,
so
this
sort
of,
instead
of
seeing
that
the
app
is
reconciling
the
user,
actually
sees
what
the
app
is
doing.
While
reconciling,
if
I
were
to
quickly
show
what
it
looks
like
if
it
breaks,
I'm
going
to
go
ahead
and
break
the
app.
E
Oh
actually
before
you
do
that
so
now
that
the
app
is
reconsidered,
the
user
was
to
run
the
status
command
again.
They
would
see
something
like
this,
where
we
say
that
these
stages
were
done
these
many
seconds
ago
like
I,
so
we
don't
see
anything
here
right
now
because
it
reconciled
again
and
there
were
no
changes
so
yeah,
that's
how
it
looks
like
once
the
app
has
reconciled.
E
What
exactly
is
going
on
like
why
exactly
the
app
has
failed
and
yeah?
So
actually
that's
about
it
like
it
would
be
great
to
get
some
early
feedback
on
this
we're
still
playing
around
with
the
information
we
want
to
surface
and
sort
of
tweaking
the
output
a
little
bit,
but
the
final
product
is
going
to
be
pretty
similar
to
this
like
yeah.
E
C
About
it,
this
is
really
nice.
I
love
the
use
of
coloring
in
the
output,
which
makes
it
some
and
it's
so
much
more
concise
than
like
using
cuddle
described,
especially
with
some
of
the
ergonomic
issues
that
are
there
with
crds
right
now
is
all
of
the
information
just
basically
a
pretty
print
of
the
things
that
are
available
in
the
app
object
like
in
the
status
field.
So
there's
nothing
like
tailing
logs
from
the
controller
here
or
stuff
right.
E
Yeah,
so
so,
right
now,
if
the
user
were
to
install
a
package,
what
they
see
is
that
the
package
is
getting
reconciled
till
the
package
reconciles
right.
So
we
are
hoping
to
sort
of
plug
this
mechanism
into
most
of
the
commands
we
have
where
we're
waiting
for
something
to
happen.
So
if,
when
a
package
is
getting
installed,
we
would
probably
show
what
the
underlying
app
cr
is
doing.
Instead
of
just
saying
that
the
package
is
reconciling
if
a
patch
package
is
getting
deleted.
E
Oh
I'm
so
sorry
package
install
is
getting
deleted.
We
will
tail
the
deploy,
logs
and
sort
of
show
the
deletion
process.
So
if
the
delete
fails
for
the
resource
right,
the
user
can
see
then
and
there
which
resource
did
not
get
deleted
or
couldn't
get
deleted
tree.
They
don't
have
to
figure
out
what
they
have
to
do
to
debug
their
app
crs
or
package
installs.
C
Has
there
been
any
taco
of
also
making
this
possible
to
just
exact
as
a
coupe
cuddle
plugin?
I
don't
know
if
that
would
improve
the
usability
at
all,
but
maybe
having
this
available
in
the
crew
package
manager
might
make
it
more
discoverable
for
people,
even
if
it
means
I
have
to
type
another.
Like
word.
C
J
Yeah,
it's
a
the
crew
package
manager
is
great
for
doing
things
like
this,
and
I
think
that
if
it
goes
into
crew
one
other
command,
that
could
be
very
helpful.
That
a
lot
of
the
things
like
the
rabbit,
mq
crew,
plug-in
or
the
cert
manager
plug-in
do
is
there's
actually
an
install
sub
command.
That
would
install
cap
controller.
J
So
someone
could
do
a
cube
ctl
case
ctrl
or
whatever
it
is
install,
and
that
would
install
cap
controller
similar
to
how
it's
done
with
like
rabbit
mq,
which
installs
the
rabbit
mq
operator
it
just
makes
discoverability
and
getting
started.
A
really
smooth
experience
and
crew
provides
the
baseline
github
actions
to
just
update
crew
index
with
any
new
release.
That
makes
it
really
easy.
They
have
a
template
that
you
can
just
utilize
for
yourself
and
it's
really
easy
to
add
there.
C
J
C
J
C
J
C
It
also
just
doesn't
work
sometimes,
because
not
all
ytt
overlays
are
valid
yaml
files.
So
then,
like
it
breaks
the
yaml
parser
like.
If
you
have
values
that
are
comments,
then
you
end
up
with
invalid
yaml,
because,
like
you're
interpolating
data
values
and
stuff,
so
the
yaml
parser
breaks
and
then
that
the
autocompletion
breaks
it'd
be
nice.
If
there
was
if
we
could
use
the
same
methods
here
to
autocomplete
the
keys
since
the
ytt
parser
knows
where
those
things
are.
B
Yeah
we've
wanted
to
have
an
lsp
implementation
for
like
a
couple
years
now,
and
it's
been
yeah,
it's
it's
a
little
bit
of
a
lift
to
get
started
with
that,
but
it
would
be
amazing
to
be
able
to
fold
in
both
the
type
information
that
we
do
have
available.
So,
if
ykt,
if
ytt
could
digest
it
can
accept
open
api
v3
schema.
B
It
already
has
a
way
of
representing
types
internally
and
understands.
What's
going
on
in
that
whole
file
that
yeah,
we
could
absolutely
provide
auto
completion,
not
just
for
the
structure
but
like
also
for
globals
that
are
available.
Other
functions
that
are
in
the
library
like
it'd,
be
amazing,
yeah,
but
yeah.
We
just
we
just
haven't,
had
it
public
compared
to
the
other
things
that
were
where
we've
addressed
so
far,
but
we've
been
absolutely
open
and
would
eagerly
support
that
kind
of
effort.
C
I
see
yeah,
I
imagine
that
it's
like
one
of
those
niche
things
where,
if
you
don't
know
about
grammars
and
about
the
particular
like
interfaces
to
implement,
then
it
would
probably
take
like
months
just
to
get
started
yup.
That
is,
it
is
that,
okay,
who
wrote
the
ytt
highlighter
that
works
in
vs
code
right
now,
that
like
has
some
basic
functionality.
That
thing
is
super
useful
because
it
just
allows
the
stuff
to
work.
B
C
I
just
I
imagine
that
if
we
had
this
level
of
integration
with
a
an
active
kubernetes
cluster,
then
you
might
get
some
real
usability
games
over
similar
tooling,
like
jasonette
tonka
starlark
q,
like
that
that
feature
inside
of
the
red
hat
yaml
like
autocomplete,
plugin
that
bridges
the
open
api
spec
from
the
kubernetes
plugin.
It's
just
really
so
much
nicer
than
a
lot
of
stuff.
B
Yeah,
that's
a
good
pointer
and
it's
it.
It
speaks
to
like
if
you
take
a
structural
perspective
on
managing
these
phosphate.
If
you
like,
instead
trying
to
getting
away
from
yaml
like
embrace
the
fact
that
you're
in
yaml,
then
these
kind
of
things
are
more
possible.
Yeah
should
talk
more,
maybe
there's
a
way
that
we
can
find
a
way
of
carving
out
some
of
this,
or
maybe
I
know
that
those
plugins
can
work
in
layers.
C
Maybe
a
way
to
do
that
because,
like
if
the
grammar
that's
inside
of
the
existing
ytt
plugin
already
can
identify
a
key
then
you're
already
like
half
the
way
there.
Then
it's
just
finding
the
proper
call
from
the
red
hat
plugin
or
copying
some
code
from
the
red
hat
plug-in
and
then
sourcing.
The
open
api
definition
from
the
kubernetes
plugin.
C
It
seems
to
like
be
pretty
reliable,
but
as
soon
as
you're
like
executing
into
wsl
or
over
an
ssh
connection
or
using
vs
code
from
the
browser
for
some
reason,
whatever
loop
they're
running
to
grab
the
open
api
spec
from
the
kubernetes
cluster
that
you're
attached
to
seems
to
like
only
work.
Nebulously
like
after
you've
been
inside
the
project
for
a
couple
hours.
C
If
you
do
something
weird,
then
maybe
it
goes
and
pulls
a
spec
when
it
would
really
really
nice
if
you
could
just
like
control
shift
p
and
then
like
run
a
command
that
goes
and
pulls
that
thing
or
even
if
you
knew
where
the
file
location
was
to
like
drop
it
manually
yourself.
So
it's
like
really
a
bad
unreliable,
leaky
abstraction
for
such
a
powerful,
an
essential
feature.
So
there's
this
sort
of
an
other
story
there
where
it's
like.
C
I
would
really
like
for
that
to
be
fixed
and
I
want
to
know
what
is
necessary
to
go
and
fix
it,
because
it's
probably
a
trivial
thing.
If
you
have
the
context,
so
maybe
if
we
just
copied
that
code
and
made
it
better
inside
of
the
ytt
plugin,
then
you
wouldn't
depend
at
all
on
the
kubernetes
or
red
hat
yaml
ones.
And
then
you
could
just
own
your
destiny.
C
A
A
And
that's
that's
why
I
love
your
feedback,
because
you
know
you
get
to
play
around
with
it
and,
and
you
have
all
these
other
tools
that
you
work
with
alongside
of
it,
so
you
have
a
whole
other
way
of
viewing
it
and
it's
super
helpful
yeah.
Anybody
else
have
any
comments
to
what
lisa
or
anything
else.
I
want
to
share.
A
Okay,
well
thanks
everyone
for
attending
today's
edition
of
the
carville
community
meeting.
If
you
are
watching
this
from
home
again
like
please,
come
and
join
us
live
we'd
love
to
meet
you,
and
here
are
your
thoughts
on
things.
If
you
have
any
feedback,
you
wish
to
provide
yours
for
any
of
the
stuff
that
you
saw
here
today.