►
From YouTube: SIG Cluster Lifecycle - Cluster Addons 20210202
A
Right
welcome
everyone
to
the
cluster
add-ons
meeting
today
is
tuesday,
the
2nd
of
february.
Do
we
have
anyone
on
the
call
who's
entirely
new
wants
to
introduce
themselves.
A
B
B
A
B
B
Quite
a
few
folks
who
could
help-
I
don't
know,
maybe
if
the
bundle
stack
stuff
could
fall
into
that
category.
I
know
that
nick
was
working
on
it,
but
summer
of
code
is
a
long
ways
away
from
now.
We
kind
of
ran
into
this
a
little
bit
from
last
year,
where
we,
you
know,
thought
okay
well,
some
of
these
project,
maintainers
they
want
to
ship
operators.
B
Now
we
don't
have
the
tooling,
you
know
necessary
for
them
to
take
the
approach
that
we're
suggesting
and
summer
code
is
kind
of
a
long
ways
out,
but
with
the
amount
of
activity
that
we've
had
over
the
past
few
months,
we're
in
a
little
bit
of
a
trough
right
now.
So
maybe
it
would
be
good
to
be
thinking
about
what
summer
code
could
look
like
for
pushing
cluster
islands
a
more
uniform
kind
of
status
quo.
B
There
are
a
bunch
of
other
interesting
work
items
like
the
applications
api
that
we
talked
about
two
coupons
three
coupons
ago
and
yeah
I
mean
in
general.
We
have
things
that
we
could
help.
Somebody
do,
and
I
do
think
that
they
could
shape
up
to
be
the
sort
of
epic
kind
of
style
of
work.
That
tells
a
story
that
you
know
they
look
for
in
a
gsoc
proposal,
not
exactly
sure
what
the
timeliness
of
this
is,
but
you
could
start
looking
into
it.
A
Yeah,
I
just
had
a
brief
look
and
it
seems,
like
organizations
still
have
time
until
like
two
and
a
half
weeks
or
something
to
apply.
So
I
guess
that's
in
the
cncfs
realm
net
for
now,
so
we
actually
do
have
some
time
but
yeah.
A
I
think
it
would
be
excellent
if
we
found
like
a
really
nice
project,
or
maybe
even
two,
because
I
feel
we
have
like
many
very
experienced
people
and
within
cluster
add-ons
and
if
we've
identified
like
one
or
two
projects
of
things
that
are
still
missing
until
people
can
start
using
it
or
we
could
start
talking
about
what
we've
been
doing
more,
I'm
just
not
quite
sure
what
those
projects
might
look
like.
B
Yeah-
and
perhaps
this
could
be-
you
know,
one
of
those
projects
might
be
some
joint
effort
with
the
core
dns
team.
You
know
that
they
were
keen
on
working
with
us
previously
and
maybe
that
that
work
has
just
become
prioritized
or
something
so
if
we
kind
of
chase
them
for
what
they're
looking
for,
but
then
that
could
help
us
reboot.
Some
of
this
work.
B
2020
was
a
definitely
interesting
year,
for
I
think
the
kubernetes
project
as
a
whole
yeah,
so
just
trying
to
help
things
not
be
dropped
if
their
priorities
cool
well,
there's,
definitely
some
things
to
to
chase
down
there.
B
A
But
I
think
if
we
have
a
head
start
and
like
the
the
two
thing
or
like
for
two
of
these
things,
it
it's
actually
quite
clear
so
for
the
manifest
bundle
cap
we
know
who
to
talk
to
and
the
accordions
team.
What's
his
name
again
rajan,
I
think
there's
a
there's
a
couple
of
people.
We
can
already
ping
and
say
like
this
is
like
an
obvious
fit.
What
do
you
think.
B
Yeah,
perhaps
I'll
also
ping
in
that
slack
thread
to
see
if
nick
or
rajan
want
to
comment
on
either
of
those
items
potentially
being
a
good
path
for
their
to
stock,
then
yeah,
I
I
would
be
surprised
if
the
cncf
wasn't
applying
as
a
whole
as
they've
done
in
the
previous
years.
We
could
check
with
nikita
or
whoever's.
B
Be
our
or
representative,
probably
yeah,
we're
right
on
that:
okay,
cool
yeah!
If
you
just
ask
what
we
should
expect
sounds
like
I
mean
from
what
I
remember
last
year,
I
think
talking
about
this
right
now
is
like
definitely
the
way
I
have
been
involved.
So
we
definitely
have
plenty
of
time.
A
Cool
and
in
terms
of
integrate,
integrating
with
cops
and
in
other
bits
or
the
add-ons
installer,
do
you
feel
like
there's
anything
there?
That
would
be
worth
looking
at.
C
I
will
say
specifically
for
cops,
I
think
we're
just
sort
of
trying
to
get
everyone
aligned
and
I
don't
think,
that's
a
great
match
for
a
g-suck
project.
If
we
can
get
that
alignment,
then
there
could
certainly
be
like
a
bunch
of
of
these
sort
of
adoption
and
those
sorts
of
things.
I
think
that's
a
great
idea,
I
suspect,
but
I
don't
know
that
it's
the
same
story
with
kubereum
or
cluster
api
right
like
if
we
can
lay
the
groundwork,
then
it
could
be
a
to
sort
of
clear
the
way.
C
I
think
it
could
be
a
good
project
for
someone,
but
if,
if
for
someone,
a
sort
of
gsoc
type
thing,
if
that
g7
person
has
to
spend
their
time
basically
persuading
people
that
this
is
the
way
that's
going
to
be
a
a
tricky,
it's
not
going
to
be
a
great
gsog
project
like
because
there
might
be
nothing
to
show
no
nothing
concrete
to
show
for
it.
At
the
end
of
that
project,.
B
The
another
avenue
here
is,
we
could
also
host
some
of
the
cross-cutting
concerns
out
of
the
flux
project,
flux,
being
a
valid
cncf
governed
thing.
B
The
manifest
bundle
stuff
would
easily
be
consumable
by
source
controller.
So
that's
something
that
a
gsoc
person
could
work
on
either
as
part
of
one
big
proposal
or
as
a
separate
one.
B
Yeah,
I
guess,
if
so
say
we
get
the
manifest
bundle
spec,
you
know
really
kind
of
in
a
good
place.
B
Now
we
have
something
where
we
can
implement
a
client
or
tools
like
we
know
how
the
docker
images
or
the
oci
images
should
behave,
how
they
should
be
labeled,
where
the
metadata
comes
from,
how
to
sort
them
inside
of
the
image
registry.
That
kind
of
thing
even
like
verification
of
signatures,
so
we
can
implement
those
features
to
consume
those
bundle,
images
inside
of
flux's
source
controller,
so
you
could
have
the
get
ops
flow.
You
know
where
git
isn't
the
source,
and
this
could
be.
B
B
Are
you
following
me,
daniel,
so
like
the
package
for
the
add-on,
you
know
could
be
bundled
in
the
image
and
then
similar
to
how
flux
source
controller,
supports
git
repositories
and
elm
repositories
and
s3
buckets
object,
stores
and
stuff.
We
would
support
installing
add-ons
and
other
kubernetes
resources
from
bundles
which
are
stored
inside
of
the
image
registry.
B
That's
something
that
a
gsoc
mentee
could
work
towards,
and
we
could
organize
that
in
ways
out
of
the
cluster
add-ons
project
or
out
of
the
flux
project.
B
That
you
know
would
certainly
be
a
good
extension.
It
obviously
there's
or
shouldn't
say.
Obviously
there
are
some
things
that
like
need
to
happen
before
you
know.
B
That
would
be
something
you
could
work
on
right,
because
nick
was
kind
of
owning
the
spec
of
the
like
the
first
real
stab
at
like
a
specification
for
bundles,
that's
implementable,
but
the
that
you
know
shouldn't
take
half
a
year
to
get
done
and
whatever
work
you
know
we
were
already
planning
beyond
consuming
bundles
inside
of
declarative
operators,
and
that
kind
of
thing
we
could
just
as
well
be
doing
the
flux
work
at
the
same
time
and
that
would
get
the
ecosystem
closer
to
having
a
way
to
share
manifest.
A
B
And
that
would
definitely
be,
I
think,
there's
plenty
of
structure
in
place
where
somebody
who
a
qualified,
gsoc
mentee
would
be
able
to
implement
that
kind
of
thing,
with
guidance
over
two
and
a
half
months.
B
Yeah,
maybe
float
float
the
idea
to
the
maintainers
team.
B
Obviously,
if,
if
we're
doing
something,
you
know
in
this
group
that
needs
some
kind
of
the
community
adoption
that
would
be
a
good
place
to
do.
We
might
also
want
to
let
the
argo
folks
know.
B
So
then
they
could
consider
you
know
supporting
that
kind
of
source,
and
at
that
point
you
know,
you've
got
a
really
a
very
meaningful
addition
to
the
ecosystem.
You
know
people
would
probably
start
adopting
that
at
least
for
internal
use,
I'm
not
sure
about
getting
bundles
into
artifact
hub
or
something
but
there's
so
many
ways.
You
know
that
you
can
go
here
so
and
all
of
it
is
like
something
that
you
could
just
start
working
towards.
B
B
And
if
we,
if
we
adopted
in
flux,
then
I
imagine,
argo
and
other
projects
might
be
quite
interested
in
doing
so.
Maybe
there's
some
avenue
for
techton
as
well
to
be
publishing
these
things
so.
B
And
the
more
of
these
kind
of
manifest
bundles
that
we
have
actually
packaged
as
images.
I
think
it
really
creates
a
justification
for
building
operators
that
reuse
those
packages.
Prakash.
Are
you
racing
raising
your
hand,
you
can
speak
up.
D
Sorry
I
was
not
able
to
unmute
yeah,
so
I
just
wanted
to
check
on.
Are
we
talking
flux,
v2
or
v1,
and
any.
B
B
There's
no
there's
no
additions
to
v1
anymore
inside
of
flux,
really,
okay,
yeah!
All
additions
would
be
on
the
new
version
of
the
controllers
and
apis.
Okay,.
B
Oh
well,
we
definitely
covered
quite
a
few
avenues.
Here.
Definitely
talk
about
it
more
since
we're
very
ahead
of
schedule,
considering
it's
a
summer
program.
So
for
sure,
thanks
thanks
for
the
brainstorming.
D
Just
to
close,
actually
just
to
close
on
that
you
mean
that
for
the
licensing
purpose
we
need
to
use
for
osa
the
flux
we
want,
or
is
it
to
come
for
the
g-shock
I
mean.
B
No
there's
no
there's
no
criteria
that,
like
forces
or
the
cluster
add-ons,
the
sub-project
that
we
have
to
work
directly
with
the
flux
project.
They
both
like
our
sub-project,
is
organized
under
kubernetes,
but
we
consider
the
whole
ecosystem
of
things
and
projects
like
flux
and
argo
are
also
cncf
projects.
B
So
the
cncf
really
does
like
to
see
interoperability
between
the
different
components
of
the
cloud
native
ecosystem,
so
making
kubernetes
work
well
with
you
know,
other
tools
that
are
interested
in
artifacts,
whether
those
artifacts
be
in
git
or
a
helm,
repository
or
an
image
registry.
B
B
B
There's
no,
like
real
licensing
concerns
I
mean,
since
all
of
them
are
cncf
projects
like
we
all
use
the
apl
to
license,
but
no
one's
like
forcing.
You
know
this
kind
of
interaction
between
the
projects,
and
I
did
mention
techton,
which
I
think
is
a
cdf
project,
I'm
not
mistaken.
B
Cool
well,
I
think
that
was
our
only
agenda
item.
We
don't
have
anybody
here
who
is
asking
for
updates
or
help
on
any
of
the
things
that
we're
being
worked
on,
so
I
might
just
have
a
short
meeting
today.
It's
great
to
see
all
of
you.