►
From YouTube: 20190619 scl cluster api
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
A
A
Nothing
perfect
is
the
enemy
of
the
good
and
from
what
I
see
from
this
proposal
is
it's
a
good
incremental
step?
It
doesn't
solve
everyone's
problems,
but
I
think
it's
it's
trending
in
the
right
direction
and
we
can
always
revisit
this
again.
My
intention,
or
my
thoughts,
at
least
from
leading
this
group,
will
be
that
we
will
iterate
as
many
times
as
needed
in
the
Alpha
cycle,
to
get
folks
into
a
state
that
they
feel
comfortable
with
before
we
go
to
beta.
A
That
seems
like
a
reasonable
approach
that
will
allow
us
to
have
a
long
tail
and
I
want
to
reiterate
that
it's
alpha
if
people
are
doing
production,
great
things
on
alpha
based
technologies,
I
highly
recommend
that
you
wrap
your
implementation
to
support,
churn
or
modifications
that
can
occur
so
again.
This
this
should
be
like
the
last
call.
I
think
time
out
by
next
week
seems
like
a
total
reasonable
time.
We
have
there's
been
three
weeks
time
frame
on
this
well,
plus
the
eons
of
time
we've
had
before
hands
I'm
looking
for
showstoppers
in
particular.
A
So
if
you
have
comments,
please
add
them.
I
also
wanted
to
give
a
brief
shout
out
to
constructive
feedback
and
criticism.
So
one
of
the
things
that
this
group
has
kind
of
languished
or
had
issues
with
is
people
have
lots
of
comments
and
which
is
good,
but
giving
constructive
feedback
is
even
better
right
so
and
recognizing
that,
even
though
you
might
feel
strongly
about
an
idea
that
it's
okay
to
implement
and
go
on
with
the
current
incantation
and
revisit
the
next
iteration,
so
I
want
to
give
a
shout
out
to
Doug
Yan
and
David.
A
B
We
kind
of
like
from
like
high
level
work
items
and
like
issues
and
I
try
to
capture
as
much
as
I
could
like
in
terms
of
like
individual
items.
We
can
break
this
down
even
further
like
one,
so
we
go
into
the
implementation
at
a
high
level
like
we
have
changes
at
the
API
types.
So
we
we
need
to
introduce
V,
1,
alpha,
2
and
then
work
to
kind
of
like
change
the
controller,
so
the
machine
controller,
the
machine
set
controller
and
machine
deployment.
B
B
A
A
C
A
So
there's
gonna
be
a
cost
associated
either
way
right.
One
cost
is
the
def
graph
and
the
the
other
cost
is
the
the
vendor
a
dosis.
But
if
it's
highly
contained
a
binary
artifact
that
is,
does
not
any
reverse
dependencies
then,
and
you
have
a
strict
clean
dependency
graph,
then
that's
even
better.
In
my
opinion,.
B
B
D
In
yes,
Pellegrino
a
slightly
different
topic,
I
don't
want
to
interrupt.
The
discussions
is
more
than
we
had
on
this
o.
If
you
have
a
different
topic,
feel
free
to
add
to
the
agenda
at
the
end.
Well,
it's
I
wanted
to
ask
about.
What's
the
other
providers
are
going
to
be
picking
up
these
new
types?
It's
dad
there's
a
rough
sketch
for
that
I.
D
B
That's
a
good
question
so,
like
it
all
kind
of
looking
depends
on
like
oh
we're.
The
first
iteration
of
the
Machine
controller
actually
gets
published,
because
that's
the
first
thing
that
we
can
actually
test
in
an
infrastructure
provider
I
would
assume
that,
like
some
of
these
pieces,
actually
can
probably
work
in
parallel,
but
until
we
actually
finalize
the
first
spec
or
draft
of
the
api's
for
me,
one
off
with
you
and
I
don't
think
we
can
start
working
on
the
infra
to
provide
yeah
Andy.
E
Thanks,
there's
going
to
be
a
period
of
time
where
master
is
broken,
most
likely
between
close
to
API,
proper
and
the
various
infrastructure
provider.
Given
that
you
have
to
modify
cluster
API
first
get
those
changes
merge
and
then
you
can
go
over
to
Kappa
and
the
others.
So
it's
going
to
be
a
process,
and
hopefully
everybody
can
be
patient
and
we'll
figure
it
out
as
we
go
on.
A
C
B
A
Any
other
questions
comments,
complaints
concerns
with
regards
to
the
breakdown
and
the
current
last
call
for
showstoppers
good
once
twice
three
times
all
right.
Next
up,
PSA
for
final
grooming,
this
Friday
so
currently
on
the
call
we
have
I,
don't
know.
What's
the
total
come
here,
I
lost
track.
We
have.
B
A
So
some
people
did
get
the
updated
invite
for
the
code
walkthrough
and
just
for
another
PSA
is
the
code
walkthrough
is
next
week.
Tuesday.
Is
that
correct,
Vince,
yeah
and
noon
PST?
Okay
and
above
that,
above
the
notice?
The
general
notes
is
a
listing
of
sort
of
some
of
the
high-level
ideas
that
we
want
to
go
through
with
the
code
walkthrough.
If
you
have
recommendations
or
things
you
want
to
see
or
know
about,
please
add
that
to
that
list
and
we'll
try
to
get
that
addressed
as
well
as
part
of
the
code
block
there.
F
F
A
A
fair,
totally
fair
statement,
my
apologies,
the
problem
currently
is
trying
to
find
in
it
just
fine
trying
to
find
a
time
with
this
group
of
people
is
actually
been
very
challenging,
but
I
think
for
we've
been
recording
them
and
posting
them.
But
one
of
the
things
too,
as
well,
is
even
if
you
can't
attend
grooming.
We
do
mark
a
bunch
of
issues
as
have
wanted
and
there's
standard
operating
procedure,
which
is
actually
documented
for
sig
cluster
and
life
cycle,
which
says
just
because
a
person
has
been
default.
A
Lea
sign
something
doesn't
mean
that
they're,
the
only
person
that
can
work
on
it.
If
you
want
to
work
on
that,
just
say,
raise
your
hand
on
the
issue
or
if
it
affects
you
and
you're
very
interested
in
it,
and
you
can
do
this
in
a
synchronous
fashion.
So
I
will
put
a
link
to
the
protocol
that
we
typically
follow.
It
is
well
documented
and
it's
worked
across
the
sig
for
several
cycles,
and
that
way
you
can
asynchronously
chime
in
and
issues
that
you
want
to
work
on
and
coordinate
with
the
default.
A
Good
once
twice
three
times,
alright,
next
up
I
want
to
do
a
PSA
with
the
proposal
that
came
in
from
a
said,
leave
I
have
not
personally
had
enough
time
to
read
it,
so
it
is
a
pretty
well
crafted
proposal
from
first
blush,
so
I'm
I
think
the
ask
is
for
both
maintainer
is
as
well
as
you
know,
the
peanut
gallery,
which
is
a
very
large
peanut
gallery,
to
take
a
look
at
that
proposal
and
maybe
ace
if
you
are
around
next
week,
we
can
address
any
questions.
Comments,
complaints,
concerns
at
that
time.
Okay,.
G
A
A
My
intent
is
to
kind
of
rally
all
the
issues
and
backlog
and
come
up
with
a
sort
of
more
comprehensive,
cleaner
story
with
on
cluster
cuddle,
possibly
within
the
be
1
alpha
2
timeframe,
but
at
least
I
haven't
well-documented
with
what
we
want
to
do
or
to
have
a
roadmap
for
cluster
cuddle.
That
is
sane.
Currently
it's
been
it's
very
thorny
and
there's
a
lot
of
legacy
issues.
So
if
you
are
interested
in
this
work,
please
let
me
know-
or
please
chime
in
on
the
issues.
A
F
A
I'm
gonna
try
and
put
them
into
bins
or
buckets
and
then
try
to
create
one
umbrella
issue,
which
kind
of
outlines
all
the
bins
and
buckets
in
how
a
potential
approach
to
a
change
things.
Given
the
change,
that's
coming
and
B
what
alpha
so
I'll
try
to
create
an
umbrella
issue.
That
kind
of
summarizes
all
these
things
together,
but
it's
going
to
be
a
there's.
A
there's
gonna
be
a
huge
number
of
issues
associated
with
that
umbrella.
D
A
So
Chuck
had
requested
us
to
make
a
final
call
with
regards
to
a
question
we
had
with
in
the
Cappy
repo,
which
is
related
to
a
kind
deployer
as
a
default
test
employer
inside
of
the
Maine
happy
repo.
The
purpose,
in
my
mind,
would
be
to
turn
key
modify
in
any
API.
Modifications
would
have
a
turnkey
example
test
employer
that
would
have
to
also
be
modified
in
order
for
us
to
make
changes,
and
it
also
serve
as
a
reference
example,
in
my
opinion,
for
any
downstream
providers.
A
C
Yep
so
my
biggest
concern
is,
and
and
if
we
do
this
post
as
part
of
the
v1
alpha
to
work,
I,
think
some
of
my
concerns
go
out
the
window.
My
biggest
concern
around
doing
it
with
the
current
state
of
things
is
that
the
bootstrap
provisioning
is
tied
very
heavily
with
the
infrastructure
provisioning
today,
and
that
would
cause
us
potential
concerns
with
which
versions
of
the
cube
ATM
types
were
venturing
versus
the
controller
runtime
dependencies
that
we're
pulling
in
I.
Think.
A
Because
it's
localized
to
a
kind
deployer
and
you
eat
the
vendor,
unless
somebody
else
picks
up
that
particular
provider
would
be
a
non-issue
because
it'd
be
localized
to
that
specific
deployer.
That's
a
better
provider!
I'm,
sorry
I'm,
I'm
vernacular
is
getting
lost
here.
So
then,
during
from
the
main
Cappy
repo
who
wouldn't
affect
anyone
unless
they've
entered
in
in
a
crazypants
fashion,
the
provider
specific
stuff,
that's
underneath
there,
no.
C
A
A
C
A
E
We
may
want
to
consider
that
we
stop
been
during
cube,
ATM
types
or
other
API
types
from
a
given
repo
and
instead
think
about
copy
and
paste
and
maintaining
it
ourselves
so
that
we
can
have
multiple
versions:
v,
1
beta
1,
V,
1,
beta
2,
and
so
on.
Without
worrying
about
having
to
stick
to
what
controller
runtime
forces
us
to
import
just
as
a
possibility.
E
B
This
you
had
your
hand
raised.
Yeah
first
I
wanted
to
echo
that,
in
the
chat
like
expressed
interest
of
like
having
external
external
as
well
and
I,
say
like
I'm,
also
leaning
towards
external
and
giving
it
like
if
we're
trying
to
copy
the
access
well
like-
and
this
is
kind
of
like
also
perspective
like.
When
do
we
put
the
cubed
ian
bootstrapper?
B
If
we
put
it
in
the
Reaper,
then
the
type
my
also
be
another
repo
that
makes
sense,
and
they
will
like
free
Q's
with
that
story,
and
given
that
we
had
like
a
lot
of
issues
with
that.
A
Yeah
I
think
my
biggest
concern
and
which
has
been
a
consistent
concern
with
this
project,
which
is
why
I
kind
of
want
an
entry
is
for
speed
the
tricky
time
for
being
able
to
detect
modifications
to
the
API
and
detecting
their
breakage
across
a
given
provider.
So
that
way,
when
providers
have
a
guarantee
or
single
eventually
from
the
main
kepi
repo
that
they
have
a
reference
example,
that
is,
is
always
consistent
or
functional,
but
I'm,
not
the
one
who's
gonna
be
doing
most
of
this
work,
so
I
think
it's
fair.
B
Too
so
I
agree
that
the
nice
to
have
but
I
think
that,
like
it
will
be
a
short-term
gain
instead
of
a
long-term
one,
given
their
like
fixing
like
a
dependency
issues,
has
been
like
an
ongoing
thing,
especially
in
the
providers
right
now,
especially
made
obvious
provider.
When
you
have
a
controller
run
time,
and
then
we
have
to
override
the
cognize
dependencies
yeah.
A
I
think
the
the
death
graft
problem
is
for
real
yep.
So
it
sounds
to
me
I'm
being
outvoted,
which
is
totally
fair.
I
haven't
heard
a
one
besides
me
say,
plus
one
so
I
think
you
have
your
answer.
Chuck.
C
Yes,
I've
got
approval
to
go
ahead
and
do
another
batch
of
contributor
t-shirts
for
cluster
API.
If
you
haven't
already
seen
them
I
linked
into
the
cluster,
a
peon
logos
that
are
in
tree,
the
design
is
basically
going
to
be
on
based
on
those
and
I
want
to
limit
anybody.
Signing
up
to
people
that
haven't
already
received
a
cluster
API,
t-shirt
and
I
will
be
d
dripping
the
list
against
the
previous
list
as
well.
C
A
A
Great
once
twice
three
times:
okay
I
think
we
can
give
people
30
minutes
of
their
time
back,
which
is
great.
One
of
the
things
I
wanted
to
remind
folks
of
is
after
the
backlog
is
groomed
and
the
issues
are
in
place
as
well
as
the
as
well.
As
the
proposal
is
merged,
we
will
go
into
gogo
mode,
so
we'll
be
doing
continuous
screaming
as
part
of
this
call.