►
From YouTube: SIG Cluster Lifecycle 2020-08-11
A
A
Well,
okay,
so
I
I've
added
some
arguably
interesting
agenda
topics
for
today.
A
A
Okay,
so
the
first
one
is
registrations
for
kubecon
na2020
are
going
to
end
on
the
september
the
13th.
Apparently
I
was
a
bit
confused
whether
we
are
still
allowed
to
have
three
sub
projects
plus
the
sick
intro,
like
we
had
for
kubecon
eu
2020.
A
I
guess
we
should
just
ask
this
sentence
here
says
that
for
activities
such
as
super
projects,
we
can
have
sessions,
but
last
time
we
were
limited
to
three,
because
sequel's
life
cycle
has
many
super
projects.
A
So
I
guess
in
this
discussion
here,
can
we
figure
out
what
to
highlight
which
subproject
subproject
deep
dives
to
include
so
in
kubecon
eu
we
had
costly,
piano,
mini
cube
and
cube
adm.
B
Just
on
the
number
topic
that
does,
I
believe
the
the
sentence
is
correct.
I
think
that
there
was
a
debate,
for
I
guess
this
time,
like
europe
around
whether
it
was
going
to
be
this
policy
or
three,
and
my
understanding
was
that
the
this
is
the
correct
policy
or
this
is
the
desired
policy,
and
there
was
some
you
know,
grace
period
or
grandfathering
that
applied
to
europe.
So
this
is
not
the
first
time
we've
seen
this
policy
and
we
got
special
dispensation
more
like
a
one.
A
Oh,
so
we
will
not
be
able
to
submit
multiple,
deep
dives
this
time.
B
That
is
my
understanding.
We
can
always
ask,
particularly
as
it's
being
held
virtually
like,
I
believe
the
the
justification
previously
was
around
like
the
physical,
the
cost
of
a
physical
room.
I
don't
know
if
the
same
applies
here
so
I'm
reading
this
email
and
it
does
say
like
that
it
is
a
virtual,
a
virtual
coupon.
It's
confirmed
as
virtual,
so.
B
I
do
note
that
I'm
just
comparing
what
you
quoted
the
email
that
you
linked
does
have
this
section,
which
says
with
the
recent
change
of
the
event
going
virtual.
Our
goal
is
to
max
minimum
is
to
minimize
screen
fatigue
and
the
number
of
sessions
hold
it
hosted
virtually.
I
can't
read
today,
I'm
sorry
when
we
one
way
we
are
doing
this
is
to
allow
only
one
maintainer
track
session
per
cncf,
which
is
the
thing
you
quoted
and
then
they
added
an
or
there
is
a
note.
A
Okay,
I'm
going
to
set
an
email
to
ask,
I
guess
for
clarification.
In
the
meantime,
I
was
wondering
whether
we
can
pretty
much
try
to
prioritize
some
stations
and
if
we
have
three,
we
can
decide
like
a
pick.
The
top
three
today,
obviously
the
sig
into
as
well.
A
If
we
have
only
one
deep
dive
next
to
the
sick
intro,
we
can
pick
which
one
it's
going
to
be
hard.
But
you
know
tim
is
on
pto.
B
I
suggest
we
get
clarification
on
this
other
idea,
this
this
more
engagement,
if,
if
there
is
some
form
of
like
per
subproject
birds
of
a
feather
or
whatever,
it
would
be,
that's
organized
in
some
way,
it
might
be
that
we
use
our
maintainer
track
session
for
like
high
level
alignment
rather
than
focusing
a
rather
than
focusing
on
a
on
a
particular
sub
project.
A
A
Okay,
the
next
item
is
something
that
I
saw.
I
believe
somebody
internally
at
vmware
picked
me
about
it.
It's
the
adobe
that
is
currently
crosstalk
has
there's
a
pr
to
move
it
from
bash
to
golang
and
in
a
way
promote
it
to
the
cmd
folder.
A
As
a
binary
artifact,
I
dropped
a
small
review
that
I
think
that
if
we
are
doing
that,
we
should
add
some
duplication,
notices
to
the
the
binary
artifact
itself
or
maybe
in
the
readme
as
well,
because
this
is
kind
of
conflicting
with
the
work
of
the
quest
around
us
project,
I
was
hoping
that
we
can
get
feedback
from
the
su
project
here
on
this
particular
pr.
B
They
should
know,
I
have
some
background
knowledge
I
didn't
realize
they
were
going
to
submit
it
to
open
source.
I
guess
it
makes
sense.
The
as
I
understand
add-on
manager
is
the
gce
unit
tests
and
I
believe,
gke,
and
I
think
this
is
motivated
by
gke
in
particular,
wanting
a
like
better
instrumented.
You
know
more
reliable
add-on
manager.
B
So
that's
why
I
wasn't
sure
what
it
was
going
to
go
upstream,
and
this
is
intended
very
much
as
a
minimal
minimal
rewrite.
I
I
mean
I
guess
I
mean
obviously
I'm
in
favor
like
given
the
choice
between
open
source
and
not
open
source
I'll
pick
open
source.
B
B
Regardless
it
feels
like
this
can
go
in
to
the
like,
as
a
sort
of
short-term
like
it's.
This
is
a
band-aid
on
some
of
the
problems
of
bash
is.
Is
why
they're
doing
this
so
like?
I,
I
don't
think
it's
a
problem
for
it
to
like.
I
don't
need
to.
I
wouldn't
say
we
shouldn't
improve
the
bash.
I
don't
manage
this
just
because
of
cluster
add-ons.
This
is
obviously
a
bigger
improvement.
B
I
will
certainly
have
a
look
at
this
and
see
whether
there
is
potential
for
reuse
but
like,
for
example,
I
wouldn't-
I
wouldn't
insist
on
them
going
into
a
different
sub
project,
because
the
usage
of
this
is,
as
far
as
I
know
today,
limited
to
well
an
open
source
to
gce
ede
tests,
and
only
you
know,
yeah
the
gce
kubernetes
tests
that
don't
use
cops
and
don't
use
cluster
api
and
don't
use
you
know
the
other
tool.
A
Yes,
I
mean-
I
agree
with
this
as
long
as
we
have
these
deprecation
notices
in
the
in
I
mean
if
you
agree
with
my
comments,
you
can
comment
on
this
part.
We
have.
We
have
to
have
deprecation
notices
inside
the
readme
in
the
like
some
somewhere
around
the
main
entry
point
of
the
binary,
the
suggestions
to
move
this
to
a
separate
subproject
conflates
with
the
questionnaires
project,
or
maybe
we
can
have
it
as
a
temporary
solution
to
move
it
there
again.
Everything
should
be
somehow
explicit.
A
The
intentions
should
be
explicit.
I
think
the
the
main
motivation
here
to
move
it
sorry
to
transition
this
to
to
golang
is
that
they
want
to
remove
the
bus
dependency
and
for
the
costa
radon's
manager
image.
They
want
to
move
this
image
to
be
destroyless.
B
C
Right,
I
just
want
to
say
it
wasn't
clear
whether
the
people
who
maintained
this
thing
had
offered
to
join
the
the
add-ons
group
sub-project,
which
I
think
would
be
a
positive
outcome
or
whether
the
suggestion
was
that
they
kind
of
throw
the
code
over
the
fence
and
the
people.
The
existing
people
in
the
add-on.
Subproject
look
after
it,
which
I
think
is
not
going
to
fly.
A
Yeah,
I
think
again,
the
main
problem
here
is
that
the
people
who
are
doing
this
change
do
not
know
of
the
existence
of
the
quest
around
the
super
project
or
sequence
life
cycle.
They
just
want
to
do
this,
refactor
that
benefits
tomorrow,
some
sort
of
a
security
aspect
of
the
kosterano's
image
itself.
Well,
the
legacy
cluster
manager,
adam
costrado
manager,
but
again
brian.
If
you
have
comments,
just
drop
them
on
the
pr.
C
A
B
A
I'm
kind
of
confused
why
there
is
a
meeting
of
cigarettes
on
friday.
Let
me
check
yeah.
I
must
be
misinterpreting
what
dim
said
yeah.
I
don't
have
it
on
my
calendar.
I
know
it's
some
normally
on
thursdays,
I
believe
yeah.
So
it's
thursday
on
the
13th
is
the
the
main
bi-weekly
civil
architecture
meeting.
B
That
seems
not
unreasonable,
but
it
should
be
made
explicit
and
we
should
probably
shouldn't
just
delete
the
bash
code.
We
should
probably
deprecated,
as
you
say
or,
if
is
cigar,
is
the
right
place
to
make
a
decision
that
we
can
remove
it
without
any
notice,
as
it
were.
B
A
B
Yeah,
I
mean,
I
think
they
would
have
to
if
effectively
they
would
have
to
maintain
both
images
for
the
deprecation
period.
B
The
bash
script
would
probably
be,
as
I
understand
the
policy
if
they
are
going
to
follow
the
letter
of
the
policy
and
then
the
bash
script
would
effectively
be
unused
untested,
except
in
maybe
in
older
versions
of
the
gce
unit
tests
and
then
after
the
application
period,
but
they
would
have
to
backport
security
fixes
continue
to
update
for
security
fixes
and
then,
when
the
when
the
deprecation
period
had
passed,
I
imagine
we
would
delete
the
delete
the
code.
A
Okay,
so
yesterday
I
joined
the
working
group
naming
meeting.
That
was
the
first
meeting
they
had.
It
is
about
removing
offensive
language
for
kubernetes
from
kubernetes.
A
If
you
click
this
link
here,
you
can
see
the
meeting
agenda.
What
was
discussed,
I
I
something
that
that
was
stated
is
that,
if,
if
somebody
has
questions
they,
they
should
just
use
the
mailing
list
for
this,
and
I
believe
we
as
a
group
have
some
questions
that
we
can
ask,
because
we
have
some
projects
that
use.
A
You
know
language
that
is
considered
offensive.
Yet
I
guess
we
should.
We
can
define
some
questions
and
then
I
can
send
the
email
to
the
working
group
naming
so.
The
first
question
I
had
is:
what
is
the
preferred
timeline
for
making
breaking
changes?
So
if
you
have
a
ga
feature
like
a
flag
that
is
uses
like
master
or
a
taint
that
is
used
master,
what
what
is
their
recommendation
in
terms
of
timeline?
A
Can
we
follow
the
kubernetes
application
policy,
or
can
we
do
whatever
whatever
we
want,
or
should
they
suggest
something
else
like
hey
in
a
month
from
now,
we
are
looking
for
something
that
we're
looking
for
a
mass
of
cleaning
of
language.
Something
like
that.
So
this
this
is
the
first
question.
I
think
we
should
ask.
A
Questions
justin.
B
Sorry,
I
felt
like
particularly
perfect
right
my
hand
at
this
point.
I
guess
I
I
agree.
That
is
the
probably
the
most
pressing
question
I
think.
Rather
than
asking
the
preferred
timeline,
the
preferred
timeline
is
it
was
yesterday
right
like
ideally
like
this
would
already
have
been
done
and
we
wouldn't
have
this
problem.
I
think
that
it
might
be
useful
to
anchor
it
in,
like
you
know,
we
believe
the
node
role
master,
to
be
an
example
of
problematic
language.
B
Are
you
happy
for
us
to
drive
the
technical
process
and
like
we're
gonna,
follow
the
deprecation
period,
and
are
you
okay
with
that
policy?
And
are
you
okay
with
us
coming
up
with
a
technical
suggestion?
B
Like
I
don't
wanna,
I
don't
wanna
tell
them
how
to
do
it,
but
if
they
aren't
telling
us
how
how
to
do
it
technically,
then
we
should
we
should
offer
to.
We
should
offer
a
technical
solution
right.
It
doesn't
sound
like
they
are
looking
at
the
agenda.
It
wasn't.
It
didn't
look
like
technical
discussions
of
how
to
go
about
renaming
flags
or
how
to
go
about
renaming
labels,
but
I
wasn't
there.
So
I
don't
know
if
that's
accurate.
A
A
A
This
okay,
so
I
know
I
know
we
had
some
discussions
around
staffing,
which
is
process
related.
Do
we
we
want
to
ask
this
question
like?
Can
the
working
group
help
with
basically
finding
people
to
do
the
work,
or
can
we
tackle
the
work.
B
Ourselves,
I
I
think
it's
a
very
good
idea
to
ask
for
help,
because
otherwise
their
work
risks
becoming
sort
of
an
unfunded
mandate,
and
they
will
produce
a
long
list
of
things
and
be
frustrated
when
nothing
happens
and
like
they
need
to
have
some
notion
of
the
cost
of
these
things
so
that
they
can
prioritize
and
or
prioritize
or
rally
right.
Like
rally
the
troops,
the
person
said,
I
apologize
or
gather
enough
people
to
to
do
things.
A
D
Yeah
I
mean,
I
think
I
think
what
justin
is
saying
makes
entirely
too
much
sense.
You
know,
I
think
we
should
first
talk
to
them
to
figure
out
like
yeah.
D
What
how
do
they
feel
about
the
technical
changes
and
you
know,
are
they
confident
to
go,
make
them
themselves
or
you
know,
could
they
use
some
assistance
from
us,
or
you
know
just
what
I
guess
figuring
out
where
that
level
is
and
then
we
can
start
to
say
all
right
either
we
need
to
propose
some
people
to
help,
or
maybe
they
need
to
find
some
people
but
yeah.
I
agree
in
general
with
what
justin's
saying.
A
Yeah,
basically,
the
the
group
is
not
particularly
technical,
obviously
they're
going
to
get
some
help
from
technical
people,
but
it
it
kind
of
feels
to
me
that
somebody
is
going
to
come
to
a
repository,
create
a
ticket
and
leave
it
there
maybe
help
what
is
going
to
be
on
it.
Maybe
not,
but
it
feels
to
me
like
that
they
want
the
maintainers
to
do
the
work.
A
This
is
a
an
important
discussion
to
have
like
if
the
maintainers
don't
have
the
bandwidth
to
make
such
a
breaking
change
like
what
do
we
do?
Who
is
going
to
stuff
it?
That's
that's
the
problem
and
if
they,
if
they
cannot
stop
it,
then
it's
just
nobody's
going
to
change
it.
Maybe.
D
Right
I
mean
at
some
level,
regardless
of
how
technical
or
non-technical
they
are.
They
need
to
be
brought
in,
at
least
to
the
shallow
end
of
the
technical
pool,
because
you
know
it.
I
think
we
all
agree.
These
naming
changes
are
good
and
we're
all
kind
of
set
to
make
them,
but
it
can't
just
be
like
you
know
it
can't
just
be
thrown
over
the
wall,
which
is
kind
of
what
you're
talking
about
a
little
bit.
D
You
know
we
have
to
make
sure
they
understand
that
when
they
ask
for
a
naming
change,
some
of
them
are
going
to
be
easy,
some
of
they're
going
to
be
really
tough
and
and
they
need
to
be
willing
to
stick
through
the
process
until
the
end.
You
know,
even
if
they're
not
doing
the
technical
work,
but
just
to
help
guide,
you
know
to
guide
us
and
to
be
there.
You
know
to
understand,
what's
happening.
A
Yeah,
this
is
also
related
to
the
first
question
if
they
give
us
some
sort
of
a
a
deadline
like
hey,
this
has
to
be
done
next
month.
If
the
complexity
is
too
high
and
we
don't
have
staffing,
obviously
again,
we
cannot
complete
the
work,
even
if
we
are
the
maintainers
who
want
to
do
the
work
so
yeah
it's.
D
A
Yeah,
to
my
knowledge,
when
you
use
something
like
I'm
going
to
type
it
in
the
agenda
document
as
an
example,
when
you
do
something
like
that,
this
is
the
command
to
type
the
ticket.
As
work
group
working
group.
E
A
Somebody
is
going
to
periodically
go
over
the
existing
tickets
and
add
them
to
a
project
board
at
github.
So
that's
how
they
are
going
to
monitor
the
work.
Of
course,
somebody
has
to
first
walk
the
tickets
and
it's
not
clear
if
the
maintainers
are
going
to
do
it
or
if
the
working
group
is
going
to
stuff
again
people
to
go
over
all
the
repositories
find
the
offensive
terminology
by
gripping.
A
A
I
guess
what
we
need
before,
of
course,
before
executing
any
change
is
that
we
need
the
replacement
warding
like
if
we
are
replacing
bus.
Obviously
we
can
use
control,
plane,
whitelist
denialist,
but
we
don't
we
don't
have.
We
don't
have
the
full
list
of
terminology
that
we
want
to
replace.
So
I
guess
one
question
that
I
had
is:
should
we
wait
to
the
list
for
the
list
to
be
finalized
before
we
proceed
with
the
changes,
or
can
we
just
go
and
change
master
all
over
the
place.
B
I
think
that's
a
good
question,
I
don't
know,
that's
a
question
for
them.
I
mean
I
feel
like
we
can
guess
that
master
is
going
to
be
on
the
list.
I
I
think
we
can
verify
that
with
them,
but
I
I
don't
see
it.
I
mean
it's
one
of
the
classic
examples,
so
we
could
start
figuring
out
what
we're
going
to
do
in
terms
of
like
if
we
were
to
follow
the
the
the
deprecation
process
and
do
all
of
that.
B
How
would
we
actually
do
this
and
because
it's
it's
both
in
labels
and
in
taints,
and
we
need
to
sort
of
figure
out
both
and
then
we
can.
We
can
put
that
in
a
dock
and
I
don't
even
know
if
there's
a
first
step.
We
do.
Maybe
we
start
tolerating
different
additional
taints
and
that's
something
we
could
do
in
the
short
term.
But
if,
in
our
long
term
schedule
this
might
be
a
you
know
like
a
nine
month
process,
and
I
don't
see
the
harm
in
starting
that
process.
D
Sorry
sorry,
jess
yeah,
I
mean
I
kind
of
like
the
idea
of
using
master
as
like
a
dry
run
example
and
just
produce
the
artifacts
that
would
come
out
of
this,
because
then
we
can
go
and
say
all
right.
We
tried
to
do
one
change.
We
tried
to
change
this
one
thing
and
here's
what
it
looks
like
here's
everything
that
came
out
of
it,
because
that
to
me
that
becomes
a
really
compelling
discussion
point
for
the
probably
for
the
working
group
at
that
point,
because
now
they
know
so
yeah.
A
But
personally
I
have
not
seen
a
working
group
and
kubernetes
be
disbanded
right
after
it
was
created.
This
seems
like
an
effort
that
is
going
to
happen,
but
I
personally
I
would
like
to
wait
for
some
process
and
warding
and
some
sort,
some
solid,
some
solid
definitions
to
be
present
before
we
execute
on
the
tasks,
because
I
I
don't
like
uncertainty,
I
just
want.
A
D
I
mean,
I
think
in
in
terms
of
doing
the
real
work.
I
I
totally.
I
agree
with
that
lumiere.
We
should
wait
for
them
to
you,
know,
kind
of
make
sure
everything's
written
in
stone
before
we
go
on,
but
I
think
I
think
us
providing
some
materials
about
what
it
looks
like
to
do.
This
might
help
them
to
make
some
of
those
decisions.
You
know
just
if
they
have
a
taste
of
what
it's
going
to
feel.
A
Because
I'm
sure
they
are
not
familiar
with
this,
and
this
is
going
to
be
a
technical
example.
A
Yeah,
I
just
I'm
going
to
try
to
emphasize
some
timelines.
Okay,
we
need,
like
I
don't
know
three
months
for
sorry
three
releases
for
the
application.
In
the
meantime,
we
have
to
introduce
a
new
parallel
label
things.
I
could
give
this
as
an
example,
so
others,
because
there
are
obviously
technical
people
joining
the
meeting.
Others
can
also
the
more
technical
people
can
also
comment,
maybe
provide
feedback
for
us.
A
I
know
jordan
was
present
on
the
meeting
last
time,
so
they
have
to
be
more
familiar
with
some
of
these
complications
that
we
would
face.
A
So
any
more
questions.
A
All
rights
by
by
the
way
the
meeting
is
monthly,
I
believe
so.
I
think
it's
the
second
week
of
each
month
on
monday,
if
you
go
to
the
page,
that
is,
I
think,
if
you,
google
working
group
naming
for
kubernetes
you're,
going
to
find
all
the
links-
and
you
know
when
the
meeting
is
going
to
happen,
agenda
documents
and
so
on.
So
if
you
have
interest
you
can
join
the
meeting
and
discuss
concerns
during
the
actual
meeting.
A
Anything
all
right
moving
to
through
project
readouts,
I
wanted
to
add
a
quick
one
for
cuba.
Again,
I
don't
have
any
updates
for
kubernetes,
I'm
relatively
quiet.
We
are
still
in
code
freeze,
which
was
extended
due
to
issues
around
destabilization.
A
A
B
Yes,
I
see
something
she
is
here,
so
I
don't
know
if
sometimes
you
want
to
give
an
update,
but
I
sometimes
she's
producing
who
is
our
google
summer
of
code
participant
is
producing
some
great
prs,
there's
a
little
bit
less
than
a
month
left
or
less
than
a
month
left
and
our
goal
for
that
final
period
is
to
get
all
the
work,
merged
and
coherent
and
working
towards
a
blog
post
and
ideally
more
adoption
in
cops
or
android,
coupe
adm
and
some
ede
tests.
E
Yes,
yes,
it
is
yes
good
day,
everyone,
yes
to
that
great.
B
A
Okay,
cops.
B
Yes,
we
did
the
cops
118
release
and
so
it's
fairly
quiet
on
the
releases
front.
Now
we've
sort
of
caught
up.
I
guess
we
caught
up
at
this
point
and
there
are
some
bigger
features
landing
again,
but
they
are
targeting
cops
119,
which
is,
of
course,
you
know
at
least
two
or
three
months
out.
At
this
point
we
tend
to
release
afterwards
so
yeah
there
are
some
bigger
features
landing
in
particular
around
node
authorization
and
some
of
the
bootstrapping
of
of
that.
A
Did
you
manage
to
push
the
coordinates
operator
pr.
A
I
I
remember
you,
you
saw
some
problems
around
our
back.
You
wanted
the
airbag
to
be
simplified
or
something
like
that.
B
Yes
and
samsoshi
has
actually
done
that
and
spread
it
out
very
nicely,
and
it's
it's
led
to
a
wonderful,
like
thread
that
sort
of
that
we've
pulled
on
in
terms
of
you
know
like
this
notion
of
a
generic
operator
and
all
these
things,
and
really
I
just
need
to
rework
the
pr
to
the
cops
pr2
to
use
the
split
rbac.
B
But
essentially
the
observation
is
we
can
pre-create
the
r
back
as
part
of
the
installation,
the
other
that
the
add-on
will
need
as
part
of
the
installation
of
the
operator,
and
it
doesn't
cost
us
anything
because
the
operator
needed
to
have
those
permissions
anyway.
A
A
Basically,
they
come
with
a
wall
of
our
back,
and
you
know
if
you,
if
you
open
the
folder
of
an
operator,
use
you're
going
to
start
wondering
what
is
this
all
this
hardback?
Why
is
it
needed?
A
I
was
looking
for
alternatives
to
operators
and
which
is
you
know,
it's
just
an
exercise
at
this
point
I
was
checking
something
out
and
I
saw
that
people
I
saw
a
particular
blog
post,
but
I
don't
have
the
link
for
it
where
somebody
is
suggesting
that
they're
not
suggesting
that
operators
are
heavy
they're,
just
suggesting
how
a
simple
controller
can
be
written
with
clan
gold,
a
very
simple
configmap
as
a
target
of
modification
instead
of
using
crds
and
also
very
simple
airbag.
A
To
give
this
controller,
the
privilege
is
to
modify
certain
objects
in
the
cluster,
a
very
scoped
controller,
which
is
ultimately
like
an
operator,
but
the
operator
sdk
produces
something
that
which
is
much
more
complex.
B
I
I'd
love
to
see
the
blog
post
and
I
think
yeah
sometime
sometimes
I
think
I
hope
you're
taking
notes,
because
this
is
great
a
great
topic
for
that
blog
post.
That
you're
writing,
which
is
like,
I
think
I
think
satoshi
has
has
managed
to
address
the
biggest
concerns
here.
B
The
the
need
for
the
operator
to
have
the
full
set
of
permissions
by
splitting
it
out
in
this
way,
and
it's
a
neat,
it's
a
neat
trick
and
I
think
we
could
do
more
along
these
lines,
I.e
putting
less
functionality
in
the
operator
in
terms
of
the
ability
to
repair
and
it
could
be
a
this-
could
be
a
dial
where
you
know
you
can
give
the
operator
more
permissions
and
it
can
repair
more
things
or
you
can
give
it
fewer
permissions,
and
it
can
only
monitor,
for
example,
that
is
probably
more
work,
but
I
feel
like
that,
reading
the
blog
posts
and
seeing
where
this
person
reading
bloggers
do
you
mentioned
lupemir
and
seeing
where
that
person
did
the
drew
the
line
or
would
decide
it
was
a
sweet
spot.
A
I'm
going
to
add
it
here
under
cops,
I
guess
yeah,
you
know
trying
to
scope
down
the
operators
to
to
be
to
not
gain
full
privilege
of
the
cluster
to
be
able
to
only
have
specific
privileges,
and
also
to
I
guess
maybe
this
is
a.
This
is
a
pretty
topic
like
our
operators,
heavy
by
design
or
like,
but
yeah,
I'm
going
to
like
this,
this
blog
post
again,
the
outer
of
the
box
post,
is
really
not
stating
that
this
is
a
replacement
for
operators.
B
A
Yeah
and
potentially
certainly
like
that
could
be,
you
know,
a
lightweight
version
of
an
operator
like
a
micro
operator
product
where
that
produces
these
simple
controls.
B
If
clango
is
that
heavy,
we
should
also
look
at
that
right,
like
glencoe
shouldn't,
be
that
heavy.
A
Yeah
all
right
any
other
comments
or
questions
for
cops.
F
Yeah,
it's
been
a
pretty
quiet
a
couple
of
weeks
we
released
a
point
release
last
week
for
various
sort
of
upgrade
related
bugs,
and
I
snuck
in
a
a
new
add-on
where
you
can
automatically
authenticate
all
your
pods
with
any
google
cloud
credentials.
You
want
we're
going
to
release
113
at
the
end
of
the
month
to
coincide
with
kubernetes
119
and
we'll
have
we'll
have.
We
should
have
another
point
released
soon
with
that
supports
the
release
candidate.
F
The
other
thing
is
for
our
office
hours,
we've
been
having
a
whole
saga
with
the
office
hours
zooming
link,
and
so
we
do
have
a
new
link,
but
I
have
no
way
of
being
able,
or
none
of
none
of
the
maintainers
have
any
way
of
taking
post
for
that
meeting.
So
we
still
can't
use
it.
F
A
A
So
I
think
I
got
the
days
right
so
it
was
so
the
third
of
august.
You
had
a
meeting,
but
you
cannot
host
the
meeting.
F
A
A
A
A
It's
we
don't
have
a
solution
for
that.
It's
really
a
country
backs
problem.
I
guess,
if
you
encounter
this
again,
what
you
can
do
is
just
drop
me,
a
message
in
the
sequester
lifecycle
channel
and
I
can
try
to
log
in-
and
you
know,
start
you
the
meeting
at
that
point.
You
should
be
able
to
use
the
host
key
that
I
gave
you
to
to
host
the
meeting.
F
E
A
Message:
okay,
I
had
a
really
cute
question
because
I
noticed
that
you,
you
guys,
have
a
survey
associated
with
each
release
like
a
mini
survey
yeah.
So
I
was
wondering
what
is
what
is
the
level
of
feedback
that
you're
getting
from
the
survey?
Is
it
like
consistent
in
terms
of
how
many
people
respond?
Are
you
happy
with
the
number
of
responses
the
you
know
the
questions
like?
What
is
your
experience
with
that.
F
Yeah,
so
that's
one
of
the
things
we
actually
we
look
at
at
our
at
the
office
hours.
Basically
after
every
release,
we
basically
look
at
the
the
new
survey
responses
since
the
last
release
and
and
we
basically
it
spits
into
like
a
a
sheet
like
and
for
the
most
part,
it's
been
pretty
good,
the
big
things.
So
we
have
one
question
which
is:
that's
been
the
most
valuable
question.
That's
basically
guided
our
our
road
map,
which
is,
if
you
could
change
one
thing
about
mini
cube.
What
would
it
be?
F
And
that's
where
a
lot
of
the
pain
points
become
very
obvious
to
us,
and
so
that's
that's
where
trends
among
our
users
show
up.
We
get.
We
get
a
good
number
of
responses
because
we
do
link
it
kind
of
everywhere
and-
and
so
it's
it's
good
for
for
good,
like
if
we've
improved
something
then
it'll
show
up
in
the
feedback
and
if
there's
like
a
big
problem,
it'll
also
show
up
in
feedback.
F
I
would
say
in
terms
of
like
which,
which
users
like
it's
it's
split,
pretty
evenly
right
now
in
terms
of
like
what
platforms
are
used
just
because
americans
use
mac,
os
non-americans
use
windows
for
the
most
part
and
then
there's
always
the
power,
the
linux
power
users.
So
in
terms
of
that,
it's
split
pretty
evenly.
A
A
Yeah
yeah,
so
how
many?
How
many
like
service
submissions
you
get
per
release
on
average.
F
That's
a
good
question.
I
think
it's
in
the
order.
It's
like
it's
like
double
digits
like
30
or
so
like
it's
a
good
number
that
that
sort
of
adds
up
over
time
I
haven't
looked
at
the.
I
haven't,
looked
at
the
the
spreadsheet
in
a
while,
but
I
could
probably
find
it
somewhere.
A
I
see
I
mean
that
that's
that's
fine.
You
know
which
you
should
release.
You
can
get
good
information
about
the
users
you
know.
Even
if
there
are
30
people,
it's
still
a
good
signal.
Yeah.
We
we
created
the
sequence
lifecycle
survey
for
2020,
but
we
didn't
include
any
mini,
cube
questions
and
I
then
really
realized
that
you
guys
have
a
survey
which
is
pretty
much
every
release.
So
you
don't
need
the
main
sig
survey
in
a
way
but
yeah.
A
Maybe
some
of
the
questions
are
still
going
to.
Some
of
the
answers
are
still
going
to
interest
you
with
respect
to
multi-load
support,
and
you
know
stuff
like.
E
A
Yeah,
by
the
way,
I
I
pretty
much
took
something
like
this
question
and
I
added
it
to
costa
rica
iqbay
dm,
and
I
think
this
is
a
very
general
but
important
questions,
because
you
leave
the
users
the
freedom
to
write
whatever
they
want,
and
it's
quite
good.
A
Yeah,
so
thank
you
for
for
this
survey
that
gave
me
some
ideas.
I
guess.
A
Yeah
all
right
any
questions
for
bdcub.
G
It
could
be
yes,
it
was.
I
just
put
some
notes
there.
Quite
a
few
people
away
from
class.
They
don't
forget,
so
I
just
filled
that
in
I've
not
been
working
on
core,
so
I
know
we
and
I've
been
away
myself,
so
I
know
we
released
038
recently
and
we
are
working
zero.
Three,
nine.
I
can't
tell
you
much
more
about
that
at
the
moment.
A
Yeah
we
seem
to
have
a
lot
of
people
be
pitching
not
only
viable,
like
our
people
in
tech
in
general,
all
right,
I'm
going
to
show
you
no
questions
for
this
still
do
we
have
any.
A
A
A
What
thank
you,
everybody
for
joining,
debating
and
see
you
again
in
a
couple
of
weeks,
bye.