►
From YouTube: 20200114 sig cluster lifecycle
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
Apparently,
there's
there's
some
constraints
that
exist
at
coop
cons
and
we
sing
coaster.
Lifecycle
has
had
a
number
of
separate
talks
that
they
traditionally
have
for
any
given
cycle
during
good
con,
but
Nancy
in
trying
to
prune
things
down
realize
that
we
shouldn't
have
been
doing
what
we've
been
doing,
even
though
they're
doing
it
for
years.
So
I've
been
talking
with
Dan
and
I
need
to
talk
with
the
CNC
at
Forks
again
to
figure
out.
A
What's
a
good
policy
that
we
should
try
to
adhere
to
so
far,
I
threw
out
the
idea
of
having
possibly
two
to
three
talks
per
person:
cluster
lifecycle
for
deep
dives,
having
one
intro
talk,
which
is
good
in
general,
because
it
highlights
all
of
the
sub
projects
across
the
lifecycle,
but
then
having
two
or
three
deep
types
and
we'll
have
to
do
a
rotational
policy
on
our
side.
So,
as
we
go
to
different
coupons,
we
talked
about
the
different
sub
projects,
as
you
rotate
them
around.
A
Now
that
that's
kind
of
a
little
bit
thorny,
because
that
doesn't
mean
we
always
will
get
a
talking
slot.
So
there
might
be
other
options
or
opportunities
for
us
to
Bob
sessions,
or
it's
some
other
shared
space
and
how
for
us
to
do
face-to-face
things.
I
want
to
pause
there
for
a
second
and
see
what
other
people's
thoughts
or
opinions
are.
This.
B
B
We
started
over
lunch
and
then
we
broke
off
into
a
table
and
I
think
we
really
had
some
very
like
good
discussions
and
we
all
got
to
know
each
other
a
little
bit
better
and
I
thought
that
was
really
Oh
in
some
ways
better
than
a
talk
in
terms
of
like
understanding
everyone's
coming
throughout
from
I,
don't
know
if
it
would
have
worked.
If
there
had
been
three
three
or
four
times
as
many
people
for
example,
so
it
at
that
size.
It
worked
pretty
well,
but
it
was.
B
The
actual
cook
time
it
was
done
during
coop
con,
but
it
was
like
a
Tuesday
lunchtime
if
I
recall
correctly,
so
it
was
a
random
day
for
lunch.
We
met
for
lunch
that
was
outside.
They
were
kind
of
copters
and
stuff
flying
overhead,
so
we
moved
inside
and
we
snuck
into
a
room
which,
as
this
meeting
is
not
being
recorded,
I
will
say
was
the
ambassador's
lounge
and
we
used
that.
A
Think
I
think
that's
actually
a
reasonable
of
timing.
I
actually
kind
of
like
the
logistics
around
that
of
having
like
it,
had
a
hawk
lunch
face-to-face
conversations,
because
that
lunch
period
is
fairly
large
and
it's
a
good
excuse
to
sit
down
and
chat
together.
I
think
that,
from
a
logistics
perspective,
the
idea
that
we
actually
had
like
a
room
that
had
things
like
you
know,
race,
boards,
or
something
where
people
can
sketch
ideas.
A
C
A
A
C
A
So,
like
the
the
main
I
think
for
all
the
talks
like
if
we
get
two
or
three
accepted
for
deep
dives,
that
we
should
try
to
highlight
the
the
other
things
that
are
ongoing.
As
you
know,
as
things
come
out,
usually
the
intro
talk
is
always
in
the
first
day.
So
the
intro
talk
is
the
big
awareness
for
people
who
might
not
know
or
might
want
to
be
like
new
to
the
space
and
want
to
contribute
I.
Think
that'd
be
a
great
time
for.
A
So
I
like
that
idea,
the
folks
have
other
thoughts.
Are
there
any
like
I
know
some
people
for
some
people,
the
getting
the
talk
on
the
maintainer
track
has
always
been
like
an
excuse
for
their
employer
to
send
duquan,
which
is
sometimes
I,
sometimes
try
to
rotate
it
around
just
so
that
it
makes
sure
that
people
get
a
chance
to
go
because
I
know
that
budgets
are
tight
and
stuff
like
that
and
I.
Think,
because,
if
you're
actively
contribute
to
this
project,
trying
to
get
people
to
go
to
gouken
is
a
good
thing.
A
A
So
I
think
we'll
have
to
do
a
rotation
with
that
said,
what
I
wanted
to
do
down
below
is
like
we
will
have.
We
will
definitely
have
a
deep
dive.
I
do
think
that
SCL
chairs
should
be
the
ones
doing
the
deep
dives,
because
they
have
the
most
to
mean
experience,
knowledge,
etc,
and,
what's
going
on
as
well
as
like
they're
kind
of
the
ambassador
right,
the
chairs
should
be
the
ambassador.
A
The
the
public
figure
heads
that
people
go
to
there's
questions
comments
want
to
know
more
about
SEO
I'm,
not
going
to
keep
conning
you
so
I've
thrown
gloomier
and
Justin
under
the
bus
for
doing
the
trip
talk
this
year,
but
so
I
think
that
one's
filled
out
I've
already
submitted
that
honestly.
So
if
Justin
and
we're
didn't
get
the
email
that
one
is
coming.
A
B
A
Well,
it's
gonna
be
rough
cuz
like
some
projects.
We
do
want
to
highlight,
because
they're
so
transformative
right,
but
I
do
think
that
we
should
rotate
so
like
if
we
do
a
cuvette
DM
Talk,
for
example,
comedians
been
around
a
long
time
that
we
should
rotate
that
project
around
with
something
else
for
North
America.
A
B
A
So
I
really
do
want
to
cluster
API
is
front
and
center
for
a
lot
of
the
things
for
doing
nowadays.
I
do
want
to
highlight
that,
because
a
lot
of
the
sub
projects
below
below
it
could
potentially
leverage
it
right.
I
do
want
to
get
a
mix
of
cross
employers
working
on
those
deep
dives
I'd
like
to
have
to
or
possible.
So
it's
not
just
one
figurehead
company
trying
to
promote
their
agenda.
A
I
wanted
to
have
like
multiple
participants,
so
I
talked
with
I
talked
with
Raphael
and
Ross
about
Canadian,
deep
dive,
I
think
if
folks
they're
they're,
both
your
European
zone,
folks
I,
think
it's
good
for
the
you
folks
to
have
representation
there
and
being
gauged
so
I.
Think
that's
a
good
one.
Cluster
API
I,
don't
know
who
you
put
your
name
down
is
but
I
think
from
the
dear
said
he
wanted
to
earlier.
So
I
don't
know
who's.
There
can
only
be
one
be
more
employee.
Then
we
have
to
get
somebody
else.
I'm.
E
D
A
We
we
could
do
that.
We
haven't
set
up
by
a
criteria
for
how
we
want
to
evaluate
it.
I
think
we
need
to
figure
out
a
policy
for
sure,
so
we
so
we're
fair,
so
I
think
an
accident
for
me
is
to
start
to
like
as
well
as
Justin
Amir
is
to
start
to
craft.
What
is
the
rotational
policy
that
that
is
fair
for
the
group.
A
A
F
F
A
B
I
think
like
there
is
interest
in
like
hosting
the
face
to
face
around
some
of
the
topics
that
are
particularly
relevant,
but
yeah
III,
don't
want
to
take
like
also
with
a
deep
dive.
I
also
feel
bad
about
like.
If
someone
else
really
wants
a
slot,
I
don't
want
to
take
someone's
slot,
but
yeah
I
I
see
the
desire
to
be
like
an
ambassadors,
so
that
makes
sense.
B
A
Right,
yeah,
so
technically
the
way
the
maintainer
trap
works.
Is
you
get
an
intro
and
some
some
number
of
deep
dives
based
upon
what
sig
you
live
in
and
for
us
we
have
to
pick
between
two
to
three
sub
projects
to
do
a
deep
dive.
So
in
this
case,
if
we
did
kuba
DM
cluster
API
and
mini
cube,
that
would
preclude
us
from
you.
You'd
only
get
one
deep
dive
talk
we
could
not
submit
to
so.
Unless
you
get
your
talk
accepted
for
the
main
track,
there
would
be
no
other
deep
dive
session.
A
C
A
Say
submit
independently
because
there
was
like
a
weird
loophole,
but
that
I
was
looking
for
exit
strategies.
There
was
a
weird
loophole
that
existed
inside
of
Nancy's
email,
where
working
groups
are
technically
not
housed
by
saying
some
projects
are
and
working
groups
fans
things
as
such
a
working
group
could
get
its
own
talk,
independent
and.
C
A
A
Yeah
everyone's
going
to
switch
now
to
become
a
gallon
working
group
right.
The
whole
purpose
of
having
SIG's
own
some
projects
was
to
like
collapse.
The
org
chart
a
kubernetes,
pessimistic,
sprawl
right
and
now
now
because
they're
doing
this,
it
forces
other
things
to
happen.
People
are
gonna
be
by
the
system.
Eileen.
What's
your
email
address,
see.
F
A
A
A
A
A
B
Actually
agree
with
that
I
think
it
makes
a
ton
of
sense.
I
feel
like
it's,
probably
the
one
where
people
are
like
there
is
that
there's
the
most
change
I'd
say
or
the
most
like
user
facing
change,
maybe
so
at
least
at
least
for
in
North
America
right
based
on
current
progress,
like
that's,
not
gonna,
be
the
case
in
10
years.
Fingers
crossed
but
agree.
A
Agree,
I
think
once
once
things
in
matured,
the
GA
for
for
even
for
clustered
guy
I,
do
want
to
highlight
things
that
you
know
aren't
getting
enough
limelight
over
time,
because
I
think
that
people
gravitate
towards
the
shiny
right
because
it's
it
may
solve
a
set
of
problems
but
I
think
there's
there's
a
set
of
things
that
we
need
to
solve
as
a
community
that
are
always
glorious.
That
I
think
are
super
important.
A
All
right,
this
was
surprisingly
easier
than
I
thought.
It
would
be
so
we'll
highlight
this,
and
one
of
the
things
I
will
I
will
talk
with
Nancy
and
Dan
about
is
I
will
try
to
figure
out
logistics
about
face
to
face
sessions,
to
see
if
we
can
get
some
meeting
rooms
that
are
during
like
the
lunch
period
times
and
see.
If
we
can
try
to
get
you
know,
some
logistics
nailed
down
there
and
see
if
it'll
work
out
for
folks
I.
A
A
A
A
B
The
I
would
encourage
others
that
are
considering
this.
They
don't
yet
have
a
deprecation
policy
to
adopt
this,
to
think
about
a
deprecation
policy
for
old
code,
and
it
is
certainly
going
to
clean
up
the
cops
codebase.
A
lot
no
really
cares
about.
Kubernetes
is
less
than
or
equal
to.
1/6
at
this
time,
particularly
as
it
doesn't
get
security
updates-
and
we
know
that,
like
people
shouldn't
be
running
the
old
versions
and
yes,
we've
talked
to
them
and
they
know
this,
but
we
have
to
sort
of
deal
with
reality.
So
that
was
an
interesting.
A
So
just
so
we're
trying
to
have
some
conversation
piece
here
like
kuben
am
officially
follows
the
period
these
maintenance
cycle
for
lack
of
a
better
term
where
we
do
n
minus
three
right,
so
you
do
117
116
115
with
that
said,
like
we
should
probably
codify
this
into
a
document.
That
is
explicit
because
almost
invariably,
as
you
mentioned,
there
are
folks
in
the
wild
who
are
like.
My
starts
expired
like
we
did
this
on
purpose,
because
we
knew
you
were
going
to
try
to
run
this
forever.
B
B
Also
clarified
cops
has
like
a
soft
deprecation,
so
we
say,
like
you're,
running
a
version
which
you
really
don't
recommend,
but
if
you
pass
an
environment
variable,
you
can
continue
to
do
so.
Like
I
acknowledge
this.
It's
just
inconvenient
right
now.
It's
like
Friday
right
I
want
to
update
right
now
that
sort
of
thing-
and
this
is
a
hard
deprecation
where
we
actually
remove
the
code,
and
there
is
no
way
around
it
other
than
to
run
that
older
version
of
cops
I
think.
A
What
I
will
do
is
I'm
gonna
log,
an
issue
in
the
community
repo
as
an
action
and
for
us
to
start
to
formalize,
as
we
have
at
the
top
of
the
stack
their
fruit
topics
that
we
should
start
to
formalize.
Some
of
this
for
as
an
official
policy
of
the
state
just
to
prevent
us
from
getting
into
our
own
mired
and
hid,
support
hell
forever.
C
We
had
somebody
join
us
last
meeting
who
built
open
source
solution
for
packaging,
air-gapped
installation
of
kubernetes,
really
simple,
UX,
it's
built
on
top
of
kuba
diem
and
they
had
interest
in
making
sure
that
the
add-ons
work
that
we've
been
putting
together,
it
would
be
integratable
yeah
and
they
were
super
friendly
and
introduced
themselves.
So
we
were
grateful
to
field
that
meeting
with
them.
G
A
I'd
love
to
get
like,
maybe
a
demo
walkthrough
of
at
this
big
meeting
periodically
of
some
of
the
sub
projects
to
basically
highlight
they
know
some
of
the
cool
things
that
we
talked
about,
because
we've
never
really
done
that.
I
think
that
would
be
a
good
thing
and
maybe
like
we
can
rotate
through
the
sub-projects
for
demos.
Cuz
I'd
be
very
curious
to
see
where
they
add
on
some
project
is
import
component.
A
A
A
A
C
A
About
most
people
does
everyone
know
or
been
involved
in
cluster
API.
That
is
in
this
call,
because
we
could
you
know
they're,
they're,
always
doing
demos
all
the
time,
so
I
think
maybe
doing
a
highlight
here
of
some
of
the
key
things
that
they've
done,
not
necessarily
from
the
like
here's.
What
cluster
API
does?
That's
that's
interesting.
A
lot
of
people
know
those
things
but
I
think
one
of
the.
H
A
H
D
So
we
did
two
villages
last
week
and
which
fix
bug
fixes
like
in
Austria
I
would
have
0
to
9,
which
them
like
machines
at
like
a
now.
We
need
to
do
a
cluster
one
orphan
resources.
Speaking
off,
like
the
template
resources
that
you
attached
on
machines,
that
are
machine
deployment
and
for
kappa
I
instead
like
we
have
a
new
feature
that
passes
like
if
they
eat
since
tax
to
the
volumes
that
you
have
attached
to
the
instance
and
they
also
fixes
nullpointerexception
when
we
can't
get
them
device
size
from
a
double.
Yes.
D
So
for
one
fixes
there
on
any
questions
for
be
150,
we
have
good
progress
across
all
fronts:
control
plane
that
we
have
scale
up
and
create
capabilities
there,
and
they
new
has
appear
out
for
delete,
which
I
think
issue
should
be
almost
ready.
So
there
has
been
like
a
great
progress
there
in
closer
cuddle
he's
here.
Do
you
want
to
give
like
an
overview?
I
think?
Well,
the
TLDR
is
like
we're
trying
to
use
closer
cuddle,
also
in
our
dev
environment,
right
now,
like
maybe
alongside
tilts.
D
For
me,
one
for
three
and
we'll
probably
discuss
this
in
tomorrow's
meeting,
but
we're
probably
aiming
and
like
mid
well
like
a
month
from
now
for
an
RC
one,
and
this
will
give
time
not
so
for
infrastructure
providers
to
update
and
I'm
also
checking
with
the
Asha
folks
to
make
sure
they
look
at
their
like
up
to
date,
and
they
have
all
the
support
they
need
to
push
it
to
be
150.
We
also
do
kind
of
let
it
dry
in
a
plan
of
action
to
automate
all
the
things
for
animation.
Even
that
release.
D
D
One
thing
that
I
wanted
to
highlight
here
is
like
we
might
need
to
work
on
like
some
more
like
a
better
policy
on
like
when
it's
time
to
like
do
proposals
and
when
it's
time
to
do
a
freeze
to
do
code.
So
if
anybody
has
like
examples
where
ideas
are
like,
we
can
bring
the
culture
API,
we
send
them
to
me
and
I'll
make
sure
to
like
keep
it
discussion
tomorrow
or
next
week.
I.
A
Think
one
of
the
things
we
talked
about
to
request
API
was
to
have
this
long-standing
backlog
of
high-level
features.
I
think
a
problem
that
close
cherry
Icee
I
think
suffers
from
is
that
we
go
through
this
longer
phase
in
the
beginning
of
a
cycle
where
people
draft
new
proposals
I
would
expend
the
way
the
main
KK
works.
His
proposals
get
drafted
all
the
time
anymore
and
things
they
synchronously
so,
like
you
just
accept
certain
proposals
which
they
feel
it's
reached.
A
certain
point
maturity.
A
So
that's
the
typical
modus
operandi
for
the
rest
of
KK
and
you
get
in
what
you
can
get
in
for
cycles,
provided
people
are
going
to
devote
time,
energy
and
resources
towards
getting
those
things
done.
So
if
those
proposals
are
ready
to
go,
then
you
have
a
X
week
period
where
you
know
you
can
create
new
proposals
and
then
you
lock
down
and
then
that's
what's
gonna
be
in
a
release
and
that's
just
the
way
it's
gonna
go
yeah.
D
I
guess
look
at
the
Persians
more
like
how
do
we
find
they
find
those
time
frames
and
like
at
which
point
in
time
like
the
proposal
was
late
for
thanks,
iteration
right
like
because
we
want
to
make
sure
that
were
inclusive,
but
at
the
same
time
we
need
to
make
sure
that,
like
I'd
release
it
kind
of
like
scope
down
to
what
we
can
do
right
then
like
what
we
wanted.
I.
D
A
This
should
be
public
facing
like
this
should
be
best
practices
like
we
should
put
this
in
general
in
the
mains
Sigma
a
lexical
community
channel
because
you're,
despite
what
cluster
API,
you
know,
cluster
API
you're
special,
but
you're,
not
special.
Now
right
so
like
this
problem
is
generic,
it
happens
everybody
right,
so
we
should
I
drift
something
out.
C
Cool
yeah
so-
and
there
was
quite
a
large
thread
that
started
up
on
Twitter
and
Ian
cold
water
like
sounded
the
Klaxons,
and
everybody
showed
up
and
started
talking
about
config
usability.
Basically,
a
user
was
trying
and
unsuccessfully
like
finding
the
non-existent
documentation
for
component
configs
with
Covidien,
and
that
got
us
into
several
threads.
A
conversation
I
brought
that
up
as
a
follow
up
in
our
working
group
and
chatted
about
that.
C
For
a
second,
it's
important
that
we
prioritize
config
generation,
OB
tech
who's
worked
on
a
lot
of
stuff
already
has
an
interest
in
picking
this
up.
Somebody
else
should
come
and
help
him
and
then
that
same
person,
Alex
snipping,
he
is
working
on
the
controller
manager
config.
You
can
see
some
of
his
progress
updates
in
this
Google
Doc
and
yeah
he's
working
on
making
sure
that
the
currents
trucks
that
are
in
the
code
are
serializable
might
wrote
an
excellent
cap
on
instant,
specific
config
that
you
should
go
check
out.
C
Seems
like
you're
taking
a
look
at
the
cup,
then
we
also
had
a
small
discussion
on
conversions
of
many
to
one
and
one
to
many
kinds
in
API
version,
upgrades
and
group
moves,
and
things
like
that.
This
is
something
that
has
happened
before
with
config
and
prior
art
is
mostly
like
this
hacky
stuff
that
we
did
in
kuba
DM
when
we
wanted
to
split
out
the
API,
and
you
can
see
there
recording
of
that.
A
very
interesting
work,
very
heavy
in
component
config
right
now.
So.
J
C
J
Consume
only
the
configuration
so
to
vent
around
the
configuration
of
our
own
project,
but
then
there
is
the
program
that
the
user
want
to
compare
from
ownership.
The
other
and
all
the
conversion
of
code
is
still
internal,
and
so
we
are
kind
of
blocking
in
moving
the
the
component
coming
from
one
version
to
another,
because
we
are
not
successfully
providing
a
conversion
part
to
the
users.
My.
C
Honest
intuition
here
is
that
we
should
try
to
convince
component
maintainer
x'
to
have
their
internal
types,
be
part
of
an
external
repository,
but
that
we
don't
have
to
import
all
of
so
we're
not
the
vendor
all
with
kicking
because
yeah,
it's
a
significant
burden
to
be
writing
the
external
to
external
conversions
for
every
API,
but
yeah
I
can
bring
that
up
for
next
week
or
the
week
after
with
working
group
that
won't
be
there
next
week.
I'll
put
that
on
the
agenda.
A
Want
to
make
sure
that
this
publicly
recorded
meeting
has
the
right
for
netkiller
here
it
is
very
stressful
or
difficult
for
us
to
be
able
to
export
api's
versus
hosting
them
for
things
that
we
care
about,
I,
think
we'd
better
for
us
to
maybe
link
to
the
main
documentation.
So
that
way
we
can
actually
have
like
a
link
to
the
comedian
could
think
that
we
would,
you
know
neckla
fire
do
whatever
we
want
to
do
so.
We
can
maintain
that
API
documentation,
but
it's
it's.
A
C
H
Yeah
first
I
just
want
to
say
that
we
are
definitely
looking
for
four
new
contributors,
especially
having
additional
reviewers
would
be
helpful,
even
if
you're
not
not
specially
fit
in
in
contributing
code,
but
with
what
you
know
are
interested
in
at
TD
and
sort
of
that
that
side
of
things-
yes,
please
see
me
or
directly
or
on
the
TDM
channel,
because
there
are
there,
are
like
over
the
holidays.
I
was
out,
and
there
are
some
outstanding
PRS
that
I've
yet
to
get
to-
and
you
know
great
says
they
have
more
reviewers
I'm
sure.
H
That's
true,
for
you
know
for
all
the
sub-project
and
then
yeah
the
next
next
sort
of
up
on
my
my
task
list
is
working
on
the
API
types,
but
I'm
currently
doing
some
things
with
cluster
API.
So
for
that,
that's
next
on
the
queue
and
something
interesting
that
I've
seen
we've
gotten
a
ping
about
someone
who
was
interested
in
you
know,
cog,
is
a
company
and
is
interested
in
using
cluster
API,
but
also
having
separate
machines
for
itd
and-
and
you
know,
wanting
to
figure
out
how
to
how
to
use
that
TDM.
H
For
that,
and
that's
something
that
just
and
I
have
talked
about
also-
and
you
know
in
recent
meetings-
I'm
sure
you
know
soon,
I
think
once
the
control
plane
management,
piece
lens
and
cluster
API.
You
know
we
can
start
thinking
about.
Okay,
how
our
or
whether
you
know
where
are
these
sort
of
these
and
separate
ICD
clusters
can
can
fit
in
cluster
API
and
we're
a
TD
again
can
fit
in
as
well
I.
A
Think
I
think
the
separate
managed
SCD
is
very
analogous
to
like
the
load.
Balancer
we're
like
you
could
POC
some
types
and
management
layer
independently,
and
we
could
always
see
if
it
makes
sense
and
then
maybe
have
a
separate
contribs
section
where
you
could
have
these
concepts
as
part
of
cluster.
If
you
have
it
not
like
fully
part
of
the
main
API,
because.
A
J
So
we
are
starting
to
get
ready
for
the
announcement
freeze,
so
we're
working
on
the
open
announcement
and
keeping
up
to
date
and
so
forth.
They
I
want
to
point
out
all
the
audience
here
to
the
cap
that
do
poem,
I
wrote
about
movie,
but
I
mean
out
of
cake.
A
feedback
are
welcome.
This
is
a
really
complex
and
interesting
operation
that
we
are
going
to
tackle
this
cookie
cycle
and
also
to
point
out
to
you
to
cover
the
main.
J
K
J
To
all
the
scaffolding
in
place,
we
have
to
start
implemented.
Currently
is
lagging
a
little
bit
of
contributor
onto
to
the
two
other
priorities
and
there
is
no
plan
because
they
didn't
order
well-defined
plan,
because
basically
it
is
a
a
prototype.
But
if
you
are
interested
in
something
special
research
I'll
reach
out
to
me
so.
K
K
D
F
Next
up,
video,
yes
from
one
of
the
exciting
things
that
we
emerged
this
week
was
an
experimental
version
of
VM
free
driver
doctor
driver.
So
now
you
can
run
mini
cube
in
docker
as
opposed
to
in
a
VM
and
the
exciting
thing
about
it
is
we
do
treat
the
docker
container
just
like
the
VM?
We
would
so
we
have
most
of
the
mini
cubes
functionalities
like
such
as
add-ons
or
whatever
cool
things
that
we
had
tunnels
and
mounds
and
we're
working
on
mounts.
F
F
We
like
to
also
increase
the
community
engagement
of
triage,
so
the
first
step
for
that
would
be
measuring
how
much
people
help
in
reviewing
pr's
or
closing
issues
or
responding
to
issues,
and
that
would
be
great
way
to
measure
and
include
that
in
our
release
notes
as
a
thank
you
to
the
people
who
participated
in
the
community
trash.
This
is
just
something
we
want
to
do,
but
we
haven't
done
it
yet.
So
if
you
guys
also
have
done
something
similar
I
would
love
to
hear
your
ideas
as
well.
A
F
A
A
So
I'd
be
very
interested
in
evaluating
it
and
kicking
the
tires
on
it.
We
have
as
a
community.
We
have
not
done
the.
We
have
not
done
the
best
of
effort
to
highlight
those
who
do
the
glorious,
scut
work
of
making
sure
everything
is
in
a
milestone,
making
sure
that
it's
actionable,
etc,
etc.
But
that's
a
lot
of
work
as
a
sig
lead
is
a
sig
chair.
I
can
attest
you
the
sheer
amount
of
github
finagling
that
I
do
is
it
has
to
go.
It's
pretty
much.
My
job.
A
Yeah
I
think
the
other
thing
I
wanted
to
talk
about
is
that
the
over
the
Venn
diagram
overlap
of
mini,
cube
and
kind.
We're
kind
of
like
this,
and
now
they're
they're
approaching
this
great
I've
always
said
that
I've
wanted
to
put
mini,
cube
and
kind
into
a
meat
grinder,
because
users
don't
want
to
choose.
They
just
want
to
have
like
a
knob
that
they
twist
right.
Yeah.
F
A
The
question
is:
have
you
been
working
with
with
Ben
and
James
to
try
an
approach,
a
steady
roadmap
that
you
that
everyone
agrees
that
this
is
a
good
plan
for
the
records,
because
I
think
the
big
win
that
kind
had
was
for
CI
test
automation.
That
was
the
huge
win,
because
we
had
the
doctor
and
doctor
approach
for
us
to
be
able
to
actually
get
CI
integration
for
kubernetes
for,
like
C,
RDS
or
whatever
right.
You
can
be
able
to
integrate
kind
directly
into
your
workflow,
and
that's
this
hugely
powerful
thing.
A
F
Yeah
I
would
love
to
see
that
I
mean
we
have
reached
out
to
Ben
a
few
times.
I
think
Thomas
also
reached
out
to
Ben
a
few
times
and
a
mini
CubeSat
has
been
a
feature
that
people
wanted
for
a
long
time
and
and
for
for
us
we
already
had
a
driver
mechanism,
and
doctor
was
just
another
driver
in
in
implementation.
F
F
When
I
first
tried
to
work
on
it,
I've
I
started
with
programming
it's
kind,
but
it
had
a
lot
of
implementations
that
minicam
had
already
solved
and
I
want.
Dean
want
those
limitations
now,
for
example,
running
and
different
kind
of
runtimes
you
want
to
do
you
want
to
do
either
pad
man
or
darker
not
be
bind
to
just
darker
or
be
able
to
have
our
add-on
system
or
our
tunnel
or
or
handling
the
mini
cubed
I.
F
Think
one
of
the
things
that
is
very
popular
is
because
it's
handles
users
in
different
countries,
very
well
like
in
China
or
India
or
countries
very,
very,
very
slow
internet
or
not
access
to
different
websites.
Oh
yeah,
if
you
detect
you're
in
a
country
that
doesn't
have
access,
for
example,
to
GCR,
we
automatically
detect,
for
example,
a
different
juicier
for
your
countries.
There's
a
lot
of
details
that
many
cube
does
and
I.
Think
this.
The
hard
part
of
it,
spinning
up
kubernetes
in
container
was
the
easiest
part
of
it.
F
A
I
think
I
think
maybe
I
can
prefer
conversation
between
everybody
to
sit
down
because
I
don't
have
a
real
stake
in
the
fight.
No
I
just
want
to
have
one
thing
that
I
tell
people
to
go
to
what
I.
What
I
care
about
is
consistency
and
reproducibility,
and
so
long
as
it's
convenient
based
I,
don't
care
yeah.
It
is
fast
right,
but
I.
What
I
care
about
is
reproducibility
speed
the
ability
to
integrate
test
automation
with
what.
F
I
think
they
also
like
to
mention
we.
Actually,
we
did
reuse
the
kind
space
image
at
a
contained
space
image
and
the
rest
of
it
is
just
running
a
starting
that
container
just
like
the
way
we
would
do
the
VM.
So
there
was
no
really
much
of
the
duplicate
work,
so
the
I
would
say
the
base
image.
That
kind
has
produced
that
we're
going
to
base
image
with
the
right
amount
of
tools
was
very
valuable
and
and
that
that
could
still
be
something
that
we
reuse
and
the
block
together.
I.
A
Will
leave
it
that
but
I'll
take
an
extra
to
maybe
set
up
a
conversation
with
Webber
on
your
deep
side,
you
think
is
appropriate
and
with
Bend
and
possibly
James
to
see
if
we
can
try
to
get
convergence
because
the
the
idea
of
having
two
tools
that
overlap
in
features
and
functionality
from
a
community
perspective
wholly
breeds
conflict.
So
I
would
ideally
like
to
have
a
single
set
of
tools
that
we
use
across
the
test:
automation
that
we're
creating.
So
that
way,
we
you
know
it's
a
common
building
block
or
set
of
building
blocks.