►
From YouTube: Kubernetes Community Meeting 20180201
Description
We have PUBLIC and RECORDED weekly video meetings every Thursday at 10am US Pacific Time.
Notes: https://docs.google.com/document/d/1VQDIAB0OqiSjIHI8AWMvSdceWhnz56jNpZrLs6o7NJY
A
A
C
Hi
everyone.
Thank
you
so
much
for
her
having
me
today
this
this
meeting
so
I'm
going
to
present
these
cubes
cubes
en
configuration
tool.
So
first,
a
quick
introduction,
I'm
inserted
systemness
time
manager
for
the
digital
architecture
area
in
Avery's,
resist
Spanish
consultancy
company,
and
we
are
part
of
the
entity
data
group
and
during
our
experience
government
projects.
Lately.
What
we
realize
is
that
we
would
really
love
to
burn
aids,
or
at
least
me
and
my
team.
C
C
We
love
the
idea
to
handle
all
this
infrastructure
as
code,
but
sometimes
when
we
have
to
quick,
create
a
new
service
or
it
gets
rid
of
these
conflation
files.
We
found
that
we
have
to
do
a
lot
of
manual
tasks
but
engineer
that
in
the
end
needs
some
some
automation.
So
we
start
thinking.
How
can
how
can
we
do
automate
this?
So
we
have
the
end
zone
EMM
files,
that
we
should
write
the
disk
based
on
some
user
preferences
and
user
preferences
that
we
can
collect
through
some
sort
of
questionary.
C
So
thinking
about
what
the
other
other
communities
have
done
with
with
yeoman,
we
thought
okay.
Maybe
we
can
make
some
tool
that
in
some
sort
of
way,
can
make
those
questions
and
put
these
questions
in
a
JavaScript
object
and
then
simply
parse
it
to
to
the
XML
file.
So
this
is
in
the
entity.
Small
idea
is
behind
huge
and
so
simply
create
some
questions.
Okay,
ask
some
questions,
transform
it
to
an
object.
That
much
is
the
big
demo
files
that
your
net
is
needs
to
to
deploy
something
and
simply
write
it
to
disk.
C
Can
simply
if
you
work
with
before
with
human,
it's
a
simple
last
call
this
new
generator.
Okay,
and
let's
say
that
we
want
to
create
a
service
that
is
our
block
for
for
cube,
gen,
okay
and
the
system
starts
asking
us
all
the
questions
that
we
need
to
to
do
in
order
to
deploy
the
service,
okay,
in
which
name
we
want
to
deploy
and.
A
C
Thank
you
so
much
appreciate
it.
It
would
be
distracting,
so
this
I
will
say
in
Eugene
start
asking
you
all
the
questions
it
needs
to
to
to
create
the
files
we
want
to
create
a
replication
controller.
I
want
to
to
create
WordPress
I
want
to
deploy
WordPress
image.
I
want
one
replica
I
want
to
listen.
The
port
80
I
want
to
listen.
The
port,
80
and
I
want
to
expose
this
out
of
the
cluster.
C
Yes,
of
course,
it's
a
block
so
and
I
want
to
expose
this
as
an
engines
proxy
so
and
I
want
to
expose
this
at
the
hostname
blue
dot
cube,
gen
dot
says
plot
net
okay.
So
once
the
system
have
completed
all
the
questionary,
it
simply
creates
the
fields
of
the
file.
Sorry,
with
all
this
the
structure
as
kubernetes
expected,
and
then
you
need
only
to
make
an
apply
F
to
the
folder
that
is
called
block.
Okay,
I.
C
A
E
C
G
C
Want
to
deploy
it
to
the
cube
namespace
I
want
a
replication
controller
with
the
WordPress
image.
I
want
one
replica:
listen
in
the
port
80
service
port
80
I
want
to
expose
to
the
world
with
an
engines
proxy
in
blog
dot,
cube
gen
dots,
which
is
black
dot
net
and
ingress
path
route
in
greece,
port
80.
So
it
creates
all
the
files
and
the
only
need
to
really
huge
CTL
lie.
C
C
Okay,
so
I
have
already
created
a
wild-card
DNS,
so
we
can
go
right
to
the
browser
and
when
the
service
is
ready,
okay,
it's
running
right
now.
Look
for
these
seconds.
We
automatically
get
everything
up
and
running
without
having
to
type
a
single
line
of
code.
With
a
with
a
code,
editor,
okay,
we
can
make
some
you
want
to
see,
show
you
the
code,
but
it's
actually
doing.
C
C
C
Help
you
get
rid
of
the
boilerplate
code
on
copy
paste
that
we
found
really
painful
during
our
experience.
Have
you
Eve
to
focus
on
the
important
things
anyway,
you
need
to
know
what
what
you
are
doing,
and
it's
not
intended
as
I
was
saying
to
replace
your
text
editor,
but
to
complement
it
actually.
C
Qgn
final
comment
comes
with
a
big
generator
and
then
some
sub
generators
and
help
you
to
create
only
a
deployment
file
or
a
replication
controller.
Okay,
you
have
some
some
comments
to
help
you
get
rid
of
of
all
the
fight
you
need
to
create
during
your
Europe
project,
okay
and
on
that's
all.
Thank
you.
B
Jam
so
this
is
a
this
is
Joe.
So
sorry,
oh,
very
interesting,
I'd
like
to
invite
you,
if
you
haven't
already
involved
or
anybody
else,
who's
interested
in
this
space
with
the
application
definition
working
group
where
we're
exploring
a
lot
of
these
ideas,
trying
to
kind
of
find
commonality
about.
How
do
you
describe
applications
where
the
common
patterns
and
how
do
we
start
sharing
techniques
and
tools
so
so
yeah,
very
cool?
Thank
you.
So
much.
H
F
C
C
So
I
was
taking
this
very
simple
idea:
we're
not
using
template
at
all,
but
simply
only
using
a
JavaScript
object
that
we
dynamically
feel
with
the
required
fields
from
the
kubernetes
configuration
file
and
then
simply
pass
it
through
a
yeah
mol,
parser
and
right
to
do
this.
If
you
see
the
source
code
is
really
really
really
really
simple.
All.
E
The
curve
Jesus
here
yeah,
let's
see
her
speaking
I'm
I'm
a
feature
slate
for
all
water
generally.
Cyst
asked
me
to
cover
this
part
today
for
him,
since
he
is
absent,
so
we
have
several
news
since
there
was
community
meeting
last
week,
forget
in
Wanda
general
ease.
First
of
all,
we
we
have
already
completed
the
features,
features.
E
Freeze
things
feature
freeze
successfully.
They
incurred
on
Monday,
and
today
we
have
thirty
five
features
that
again
it's
really.
They
find
them
the
link
to
to
the
future
threatened
spreadsheet.
In
the
MIDI
notes,
we
also
have
successfully
cut
several
alpha
sinks
to
Caleb,
who
did
this
job
for
us,
and
we
also
expect
in
their
alpha
in
the
in
the
nearest
time
and
since
the
features
phrase
has
successful
past
few
days
ago,
where
all
the
started
out
one
marketing
activities.
Well,
we
started
a
discussion
forum,
marketing
planning.
E
We
would
like
to
to
make
the
marketing
strategy
for
this
release
much
more
consistent
and
successful,
comparing
to
the
press
releases.
So
so
we
all
day
we
always
started
working
on
that
and
sensitive
mutation
Gatling
who
are
living
at
for
us.
That's
that's
all
for
me.
For
now
any
questions
for
that
and
what
return
release.
H
Yep
root,
so
you
may
find
its
no
I
I
have
one
I.
Have
one
comment
one
time,
so
the
ordinary
number
of
issues
that
have
no
labels
on
them
I'm
going
to
be
pestering
all
the
signals
about
those,
but
if
people
could
go
in
on
their
own
and
actually
start
seeing
that
things
that
are
in
the
milestone
actually
have
kind
and
priority
and
area,
etc,
labels
that
would
be
really
nice.
I
E
Mean
I
guess:
there's
I
get
it
yeah,
yeah
I'm!
Sorry,
I
did
yes,
so
I'll
put
the
link
right
there.
I
E
F
I
I
Okay,
hi,
it's
Erin
again
I
keep.
It
keeps
signing
me
in
to
seek
testing
even
though
I
square,
my
zoom
window,
says
Erin
creaking,
Berger,
alright,
so
anyway,
I'm
here
to
give
you
this
week's
graph
of
the
week
I
lied.
Actually
it's
going
to
be
a
table,
but
it's
still
Deb
stats,
so
it
counts
right.
This
is
linked
in
the
meeting
notes.
I
It's
what's
called
the
developer
summary,
so
I'm
gonna,
preface
all
of
this
by
we
all
know
like
measuring
purely
metric
like
assigning
or
distilling
somebody's
value
down
to
a
single
numbers
of
bad
faith.
Okay,
gaming
metrics
is
not
good
right,
but
there's
also
this
thing
like.
If
you
don't
measure
it,
how
do
you
know
if
you're
succeeding
at
it,
and
so
today,
I
just
wanted
to
take
a
quick
tour
of
the
developer
summary
which
collects
all
of
the
various
github
events
and
associate
them
with
each
of
the
github
users
that
created
them?
I
So
if
I
look
at
the
github
events
for
1.9
now,
I,
don't
know
how
real
this
is
so
I'm
going
to
just
try
blowing
this
off
real
fast.
You
can
see
that
Jordan
Liggett
is
the
most
active
github
user
for
the
current
ongoing
or
leased
by
github
events.
Now
what
our
github
events
they
could
be
any
number
of
things.
Dev
stats
breaks
this
down
by
a
variety
of
different
types.
I
Comments,
I,
believe
covers
both
commits
comments,
review
comments
and
issue
comments,
so
this
would
be
somebody
whose
vocal
who's
actively
participating
in
the
project,
but
name
may
not
actually
be
pushing
any
code,
and
there
again
you
can
see
Jordan,
look
it's
doing
a
great
job,
stone,
asleep
Steven
I
can't
pronounce
your
last
name,
I'm,
so
sorry
and
then,
and
the
other
from
sleep
testing.
If
I
take
a
look
at
review
comments
specifically
because
we
all
know
like
reviewing
is
really
the
best
way
to
get
code
forward
right.
I
It
helped
the
funnel
sort
of
like
people
who
push
the
code.
People
who
review
the
code
people
can
actually
get
into
it
again,
Jordan
great
job
in
St
TTS.
Thank
you
guys.
So
these
are
the
people
that
I
would
definitely
want
to
pat
on
the
back,
because
they're
doing
a
great
job
of
reviewing
the
pull
requests.
I
Unfortunately,
this
isn't
broken
down
by
repository,
so
I
can't
tell
if
you
like,
which
repos
these
are
happening
again.
Specifically
I
know
that
a
lot
of
the
sig
testing
folks
have
pretty
large
commits,
because
we
do
a
lot
of
fast
frequent
commits
into
the
testing
for
repo,
especially
when
it
comes
to
changing
jobs.
I
So
if
I
wanted
to
take
a
look
at
who
was
who
has
submitted
the
most
poll
requests
to
all
of
the
kubernetes
projects
in
the
past
during
the
current
release
cycle?
It's
me
honestly.
This
is
the
entire
reason.
I
wanted
to
show
you
this
graph,
it's
just
a
humble
brag
about
it,
but
also
you
can
see
that
the
other,
the
top
two
top
three
are
folks
from
state
testing.
So
we
do
a
lot
of
frequent
pull
requests.
I
It
doesn't
necessarily
mean
we're
generating
a
lot
of
code
that
contributes
directly
to
Kurban
Eddie's,
so
these
are
definitely
active
participants
in
the
project,
but
I
just
I'm.
Also
using
this
as
an
example
appointed,
this
doesn't
necessarily
mean
you're,
getting
huge
substance
of
contributions.
The
majority
of
Michael
requests,
for
example,
were
to
change
basically
either.
I
The
word
from
approve
reviewers
to
recruiters
or
to
update
the
code
of
conduct
across
every
single
one
of
our
repositories,
and
so
the
way
you
could
see
that,
for
example,
is
looking
at
the
active
repositories
event,
so
that
just
shows
who
has
touched
how
many
different
repositories
or
forts
I
forget
what
the
threshold
for
this
is
and
again
you
can
see.
I've
got
80
I
touched
all
83
repos
that
are
part
of
the
huger
Nettie's
project.
I
So
I
encourage
you
to
use
this
to
go
humble
brag
to
your
aim,
teammates
or
maybe
to
create
a
wall
of
shame
or
something
I.
Don't
know!
You
know
this
range
allows
you
to
view
by
releases,
so
you
could
maybe
sort
of
dig
through
by
hand
and
sort
of
compare
how
things
have
changed
across
releases
or
you
could
take
a
look
at
things
in
sort
of
the
last
quarter
week,
a
month
decade,
boy
I,
hope
you
get
to
that
point
on
anyway.
That
has
been
this
week's
graph
of
the
week.
It's
you
any
questions.
A
I
A
J
First
thing
is
that
there
is
a
new
time
for
civic
cluster
ops,
so
we
moved
the
meetings
up
an
hour
earlier
to
try
and
accommodate
European
schedules
a
little
bit
and
the
second
major
one
is:
we've
been
trying
to
figure
out
cig
working
group.
What
sort
of
what
we're
doing
sig
up
sig
cluster
ops
is
not
a
code
committing
sig.
So
our
goal
is
not
the
writing
code.
J
It
didn't
didn't
feel
to
us
like
we
were
exactly
a
working
group
or
you
know,
and
we're
not
exactly
a
sig,
because
we're
not
responsible
for
a
code
commit
place,
and
if
we
were
going
to
do
Docs
we
would
participate
in
the
docks
Zig
from
that
perspective,
so
we're
switching
the
format
to
the
more
meetup
style
and
we'd.
Like
to
invite
people
who
are
interested
in
talking
about
today's
operations
to
share
their
experiences,
talk
about
what
they,
what
works?
J
What's
not
working,
what
they're
trying
to
do
as
a
meet-up,
so
a
lot
of
sharing
are
very
conversational
in
our
meetups.
We
do
want
to
have
scheduled.
So
if
you
know
it's
helpful
for
everybody,
if
you
can
say
I'd
like
to
have
the
mic
today
and
talk
for
a
little
while
about
what
what's
going
on
so
we're
you've
been
doing
this
already
in
our
our
cigs,
but
more
formally
announce
it
and
I'll
start
doing
to
do
a
blog
post
for
the
community
and
explain
some
of
how
things
are
going
on
along
those
lines.
J
That
means
we
definitely
want
to
do
demos
we're
very
welcoming
for
demos.
So
if
you,
if
you
felt
like
10
minutes,
was
too
short
in
the
community
session-
and
you
had
a
demo
that
you,
you
think
would
be
more
interesting
if
you
could
do
a
30
minute
or
a
40
minute
review
and
get
some
specific
feedback,
especially
about
the
operational
components
of
it.
J
J
A
All
right
on
to
more
of
me
on
behalf
of
Sagada
scaling,
our
update
work
continues
on
the
vertical
pod
autoscaler.
You
can
follow
it
at
the
kubernetes,
slash
autoscaler
repository,
so
the
official
mobile
was
was
merged
and
accepted,
and
so
work
is
well
underway
on
that
we
are
currently
investigating
a
couple
minor
additions
to
HP
a
version
two
to
improve
flexibility
with
regards
to
standalone
or
kind
of
unassociated
metrics
before
we
graduate
so
we're
investigating
whether
uncle
and
make
changes
those.
A
So
we
may
do
another
data
there
and
we're
also
continuing
to
get
feedback
on
the
metrics
api's
adapters,
and
you
know
how
easy
or
difficult
it
is
to
integrate
them
with
APA
trying
to
get
real
user
feedback.
So
if
you
have
that,
we
would
encourage
you
to
either
join
us
in
our
meetings
or
hop
on
pouncing
auto-scaling
rant
angrily
at
us.
I'll
drop
a
link
to
our
meeting
information
meeting.
Those
any.
A
D
I
ask
you
questions
before
we
move
on
sure,
go
ahead,
think
related,
hey
yeah,
so
this
bob
was
so
sick
scalability.
One
of
the
things
we've
been
working
on
is
trying
to
revise
our
charter
and
the
question
came
up
at
the
meeting.
What
we
need
to
do
I'm
asking
this
question
because
it
may
be
of
interest
across
things
generally.
D
I
Moving
along,
thank
you:
okay,
so
hi
I'm,
Aaron,
trickin,
Berger
I'm,
taking
off
my
state
testing
hat
I'm,
taking
off
my
graph
of
the
week
hat
I'm.
Putting
on
my
steering
committee
hat,
welcome
we're
gonna,
try
and
make
sure
that,
as
you
may
know,
the
steering
committee
and
week
meets
on
a
bi-weekly
basis.
That
tends
to
be
when
a
lot
of
decisions
actually
happen,
or
we
have
some
stuff
to
report.
So
we're
gonna,
try
and
carry
that
forward
to
repeating
reporting
here
to
the
community
on
a
bi-weekly
basis.
I
So
you
know
kind
of
what's
going
on,
although
I
will
take
this
time
to
remind
you
that
we
do
publicly
record
our
meetings
and
they
are
available
on
YouTube
and
our
meeting
notes
are
also
public.
So
today,
I
wanted
to
talk
to
you
about
the
concept
of
formalizing
sub-projects,
so
I'm
going
to
share
my
screen
again
real
quickly.
All
these
links
are
available
in
the
meeting
notes
as
well
and
that
plan
on
Wow.
I
Why
is
the
window
down
there
cool
and
I
plan
on
having
a
thread
out
to
kubernetes
that
a
little
bit
later,
so
this
work
is
based
off
of
a
proposal
that
was
started
in
Civic
architecture
back
in
November
30th,
and
so
the
idea
is
a
sub
project.
Is
we
want
to
try
and
align
the
ownership
structure
with
the
code
structure
that
we
have
today,
we'd
like
to
get
that
defined
in
kind
of
machine
conceivable
form,
so
that
our
automation
can
help
improve
things
for
us,
rather
than
increase
the
amount
of
friction
we
have
here?
I
So
we
think
that,
for
example,
a
sub
project
is
probably
either
a
single
repo
or
a
collection
of
multiple
sub
directories
by
the
idea
being
the
siga
actually
be
a
great
example.
Today,
where
say
gap
staff
already
has
like
the
chart
sub
project,
which
would
be
defined
by
the
charts
repo.
They
have
an
ongoing
weekly
meeting
to
discuss
the
care
and
feeding
of
charts
and
best
practices,
there's
also
the
Hellmann's
sub
projects,
and
you
know
today
that
I
think
is
defined
as
the
owners
file
that
sits
inside
of
the
helm
repository
now.
I
These
sub
projects
can
have
like
their
own
leads
completely
independent
of
the
cichlids.
These
are
the
people
who
are
responsible
for
you
know,
operating
in
that
corner
of
the
world
and
maybe
following
more
specific
conventions
that
are
appropriate
for
the
entirety
of
the
project
as
a
whole.
The
other
kind
of
sub
project
would
be
the
workloads
API.
So
that's
an
example
of
a
sub
project
that
spans
multiple
sub
directories
within
a
kubernetes
repository.
That's
the
API,
that's
responsible,
for
you
know,
daemon,
sacks
replicas
sets
so
on
and
so
forth.
I
Ultimately,
the
idea
here
is
since
I've
been
talking
about
owners
files,
we'd
like
sub-projects,
to
be
defined
by
learner's
files
across
a
variety
of
repositories
and
we'd
love
those
repositories
to
be
the
source
of
truth
and
have
that
fed
back
up
into
one
location.
Thanks
to
the
wonders
of
automation.
But
before
we
go
ahead
and
implement
that
automation,
we
want
to
make
sure
that
we're
on
the
right
path
here
and
we
want
to
make
sure
that
in
the
interim
we
have
the
content
sort
of
correct
for
us
as
humans
to
concern.
I
So
it
made
sense
to
add
fields
to
this
six
dot,
yamo
file
that
some
of
you
may
be
aware
of
it's
sort
of
like
our
database
in
the
animal
form
for
all
things
related
if
you're
updating
souq
meeting
times.
So
you
can
even
agendas.
You're
updating
state
leads
whatever
you're,
updating
this
file
and
then
you're
running
make,
and
it
generates
a
bunch
of
readings
which
is
loosely
speaking.
I
What
I've
done
here
and
I
just
took
like
a
total
pass
by
designing
at
least
all
of
the
repositories
to
six,
so
you
can
see,
for
example,
here
close
that
evacuated
window
to
your
cig,
apps
as
well.
So
API
machinery
would
be
an
example
of
we
probably
don't
need
all
of
these
single
repos
each
being
their
own
sub
project,
and
it
would
be
great
to
start
a
conversation
about
which
of
these
repos
maybe
belong
under
one
sub
project.
I
But
maybe
there's
one
sub
project
found
for
your
ladies
clients
that
has
all
of
the
kubernetes
client
repos,
further
you'll
notice,
if
I'm
clicking
through
on
some
of
these,
but
not
all
of
the
repos,
actually
have
owners.
So
the
next
step
would
be
to
go
through
and
make
sure
that
all
the
repos
do
actually
have
a
excuse
me
owner's
files
that
corresponds
that
people
who
currently
have
write
for
admin
access
to
those
repos,
an
example
of
a
sake
it
does
have
sub
projects,
would
be
cig
apps.
I
So
you
can
see
here
that
there's
a
regular
cig
meeting
and
if
we
look
down
and
sub
projects,
I
see
that
there's
a
chart
sub-project.
This
does
actually
have
an
owner's
file
and
there's
a
meeting
for
charts
that
happens
Tuesdays
at
9:00
a.m.
Pacific
infusion.
So
you
can
look
at
their
meeting
each
other
same
thing
for
help.
I
It's
almost
like
you're
reading
ahead
in
my
agenda.
Okay,
you
and
so
I
will
answer
that
real
quickly,
but
before
I
do
are
there
any
other
questions?
Okay,
sig
leads
I
linked
to
that
pull
request
in
the
meeting
notes.
I
will
be
sending
out
an
email
to
community
staff
later
today.
Please
take
a
look.
I
would
like
to
have
this
initial
implementation
merged
early
next
week.
I'll
put
an
actual
date
to
that.
Okay,
so
other
things
upcoming
from
the
steering
committee
will
start
with
Josh's
questions
since
I
believe
that
was
next.
I
On
my
thing,
Fredman
burns
has
been
working
on
what
we
have
been
calling
the
repositories
proposal,
which
I
think
we
used
to
sort
of
call
the
incubator
problem
and
we've
slowly
been
trying
to
pick
apart
like
how
we
don't
boil
the
ocean
here,
so
we're
roughly
coalescing
to
the
concept
of
three
different
tiers
or
classes
of
repositories,
I,
say
classes
in
the
sense
of
like,
what's
the
formalism
around
them,
how
much
process
is
involved?
How
much
consistency
is
expected?
I
This
will
all
be
spelled
out
in
a
doc,
so
I'm
getting
a
sort
of
ramble
off
the
top
of
my
head,
but
I
would
really
encourage
you
to
look
forward
to
a
doc
coming
from
the
steering
committee
and
encourage
you
to
dr.
things
seem
out
of
whack.
The
loose
concept
is
there
are
the
core
kubernetes
repost?
These
are
the
things
that
make
up
kubernetes.
I
So
it
would
include
those
things
it's
loosely
today,
everything
in
the
kubernetes
org
that
definition
may
not
completely
hold
true
cognitively,
but
from
a
technical
sense,
that's
sort
of
what
it
is
and
we're
going
to
move
to
the
expectation
that
you
know
things
have
a
certain
level
of
processor
on
there
as
far
as
making
sure
that
there's
good
stewardship,
but
we
don't
have
bit
rot
that
there's
adequate
levels
of
testing
that
the
consistency
is
followed
in
terms
of
process
automation.
Things
like
that.
I
If
you
take
a
step
up,
there's
I
think
what
we're
calling
cig
repos.
Then
the
idea
there
is.
We
want
to
empower
six
to
be
able
to
create
repos
as
they
see
fit,
whether
that's
they
just
need
a
vendor
neutral
place
to
experiment
on
some
stuff
or
they
want
to
start
a
project
as
a
sub
project
of
that
cig,
but
they're
not
really
ready
to
go
full
whole
hog
and
do
it
directly
in
core.
I
I
Your
experiments,
but
you're
also
may
be
doing
your
additional
tools
that,
for
whatever
reason,
like
your
sig,
thinks,
they're
super
stable
and
they're
super
usable,
but
they're,
not
part
of
core,
but
their
presence
in
an
organization
that
has
the
word
incubator
and
then
really
seems
to
be
holding
back
adoption
right.
So
what
if
we
had
a
single
word
named
Q
Burnett
e-cigs,
we
set
up
a
good
lightweight
process
for
anybody
to
create
their
repos,
so
on
and
so
forth.
So
Josh
to
answer
your
question,
you
would
go
talk
to
the
cig.
I
You
want
to
create
sub
project
under
they
create
the
repo
on
your
kubernetes
States
it'd,
be
slightly
less
process
formalism
under
that
third
tier
associated
repos,
basically,
I
think
just
make
sure
you
have
the
CL
turned
on
for
Yuriko
and
no
real
presence
in
any
kubernetes
org
required
Joe.
You
have
your
hand
code
of
conduct.
Awesome!
Oh
yes,
and
a
code
of
conduct
so
but,
like
I,
said,
if
I
get
the
details
wrong
here,
trust
the
doc.
Don't
trust
me
rambling
off
the
top
of
my
head?
I
Okay,
other
upcoming
thing,
so
save
charters
of
it
looked
to
see
sort
of
like
the
nerdy
analogy.
Here
is
look
for
the
interface
and
then
reference
implementations
of
sake
charters
right.
So
what
are
the
SIG's
expected
to
provide?
How
are
they?
What
are
the
stakes
expected
to
provide
looking
for
that
and
sort
of
the
phone
on
the
checklist
or
something,
and
then
we're
going
to
provide
some
really
well-thought-out
example,
templates
that
you
could
use
for
your
Charter
today.
I
You
know
deadlines
by
which
a
formal
policy
has
to
go
in
and
what
the
enforcement
mechanism
is
to
make
sure
that
things
are
are
checked
off,
but,
as
usual
with
the
steering
committee,
we
don't
want
to
do
a
top-down
approach
for
the
most
part,
and
it
could
be
that
there
are
some
SIG's
for
whom
a
strict
set
of
requirements
don't
necessarily
make
sense.
We're
open
to
an
exception
process
rather
than
a
top-down
mandate.
Again,
look
for
that!
D
I
D
I
A
D
Let
me
let
me
start
by
saying
that
I
think
the
the
tone.
Well,
maybe,
for
those
of
you
been
around
the
project
a
long
time,
but
maybe
I
should
start
with
that.
So
I've
been
involved
with
kubernetes
since
before
one
dot
Oh,
which
is
getting
to
be
a
lot
of
a
lot
of
turbidities
dog
years.
At
this
point,
and
I
am
now
at
Amazon
as
the
GM
for
open-source
container
services,
which
is
primarily
responsible
for
eks,
which
is
our
hosted
kubernetes
service
I,
think
the
tone
I
really
wanted
to
set
for
us.
D
You're
gonna
see
a
lot
more
involvement
from
us,
but
was
to
be
measured
more
on
what
we
do
and
less
than
what
we
say.
So
I
have
not
actually
asked
to
do
the
presentation,
but
I've
gotten
some
feedback
from
the
community
managers
that
there
was
enough
kind
of
curiosity
that
it
would
probably
be
a
good
idea.
So
so
doing
this,
just
by
way
of
showing
everybody
that
we
have
a
team
here,
we're
working
on
kubernetes
the
conference
room
that
you
see
here
is
in
Seattle
and
I
thought.
D
D
The
places
you
would
probably
expect
should
expect
us
to
see
start
showing
up
and
making
more
solid
contribution
are
around
testing
really
important
for
us
to
support
the
project
in
running
on
turbidities
should
be
awesome
on
AWS,
let
me
say
that
that
when
I
say
that
don't
don't
interpret
that
as
eks
should
be
awesome,
which
is
also
true,
but
we
really
want
to
help.
The
entire
community
users
on
AWS
have
a
great
experience
with
super
Nettie's
and
whether
you're
using
our
hosting
service
or
not.
D
So
so,
that's
one
area
that
we
think
is
really
important
and,
as
you
might
expect,
you
should
start
to
see
us.
In
fact,
we
have
been
showing
up
and
starting
to
contribute
more
with
sig
AWS
seems
kind
of
like
the
natural
place
for
us
to
show
up.
So,
if
you're
looking
for
folks
on
the
service
team,
including
me
sig
AWS,
would
be
a
great
place
to
go,
and
there
again
we're
very,
very
dedicated
to
the
notion
that
sag
AWS
should
be
a
place.
That
is,
you
know,
a
useful
welcoming
place
for
anyone.
D
That's
running
kubernetes
on
AWS.
However,
it
is
they're
doing
that
and
you
know
we're
gonna
spend
some
more
time
on
documentation
and
other
things.
One
thing
that
I
want
to
make
sure
everybody
knows
is
that
I'm
around
I'm
on
slack
I'm
available
the
team
here
is
starting
to
show
up
in
the
community
like
feel
free
to
reach
out
to
any
of
us
feel
free
to
reach
out
to
me.
We'd
love
feedback
ideas.
You
know,
especially
as
a
team,
we're
a
little
bit
newer
to
the
community.
D
We
need
as
much
kind
of
help
and
guidance
and
advice
as
we
can
get
and
we'd
love
to
have
it
I
think
just
to
emphasize
one
point,
although
we
made
a
bunch
of
statements
in
public
about
this,
but
I'll
just
emphasize
here
that
we
are
very
dedicated
to
an
upstream
experience.
We
are
we're
not
going
to
be
forking.
D
A
A
I
Okay,
I
guess:
I'll
go
since
I
Tim
ready
there
yeah
all
right
to
make
it
real.
Quick,
sig
testing
has
set
up
a
new
subject.
It's
called
the
testing
comments.
If
you
care
about
how
the
tests
are
written
and
what
they
are
actually
testing,
if
you
are
interested
in
contributing
more
tests,
if
you
are
interested
in
cleaning
up
the
existing
tests,
if
you
are
interested
in
fixing
the
flaky
tests,
please
come
to
the
testing
Commons
meeting.
It
meets
bi-weekly
Wednesdays
at
7
a.m.
Pacific
senator.
F
I
F
This
was
sprung
out
from
a
bunch
of
other
SIG's
who
are
starting
to
shift
repositories.
So
if
you
are
in
the
process
of
shifting
repositories-
and
you
want
a
common
location
for
putting
your
infrastructure
for
testing
as
well
as
the
conversation
that's
ongoing
with
regards
to
conformance
we're,
also
interested
in
that
area,
so
we
we
have
a
lot
of
common
ground
with
how
the
tests
are
written
and
commonality
in
the
infrastructure
for
running
the
tests,
not
the
actual
infrastructure
inside
of
the
environment,
but
the
the
code
infrastructure.
So
yeah.
A
G
All
right,
hi,
I'm,
Andrew,
Chen
I'm
part
of
sig
Docs
I'm
gonna,
show
like
a
really
quick
demo
of
our
MVP
that
we
launched
on
Tuesday,
but
before
that,
I
just
want
to
think
that
the
team
that
worked
on
this,
because
this
really
was
a
community
effort,
there's
Jennifer,
Rondo
and
Jessica
from
heavy.
Oh,
you
have
Joe
heck,
Luke,
hi
deca
and
then
Alex
Cantina,
who
is
our
graphic
designer
from
the
Linux
Foundation
and
Steve
Perry.
So
let
me
share
my
screen.
G
All
right,
let
me
know
if
you
can
see
this
now
yep
it
looks
good
cool,
so
we
replace
the
docks
homepage
with
sort
of
a
new
landing
page
that
shouldn't
that
incorporates
personas.
So
if
you
see
we
have
like
these
buckets
like
users
and
contributors
and
under
users,
you
can
you
know
if
you're
an
application
developer,
you
know
you
can
just
coming
in.
You
can
say:
oh
I
want
to
learn
kind
of
the
foundational
stuff
and
there
is
a
content
page
that
goes
over.
The
specifics
of
you
know,
setting
up
kind
of
a
development
environment.
G
You
know
and
then,
like
the
initial
things
like
developing
an
application
and
so
forth,
so
this
should
kind
of
streamline
the
onboarding
process,
for
you
know
specific
persona
so
that
they
know,
like
you
know
what
they
should
focus
on
then
for
cluster
operator,
there's
a
similar
thing
and
then
there's
different
levels
so
like
after
you
know
the
foundational
things
like
you.
Can
you
know
what
are
the
things
I
would
normally
work
with,
and
so
we
have
it.
You
know
broke
it
down,
so
you
can,
you
know,
see
quickly
like
and
with
links
to
the
documentation.
G
You
know
what
you
should
be
aware
of
and
in
fluent
with,
and
then,
if,
if
you
don't
like
this-
and
you
want
to
just
you
know,
continue
to
browse
the
docs
that
you
like
how
you
did
before
you
can.
Click
on
the
Browse,
Doc's
button
and
I'll
show
you
kind
of
a
high-level
overview
of
all
the
different
sections.
So
if
you
want
to
just
let
go
directly
to
install
tools,
you
can
do
that.
So
this
is
an
MVP,
so
that
means
you
know
we're
still
going
to
add
things
later.
G
So
we
would
really
appreciate
any
feedback
you
have
on.
You
know
usability.
If
it's
you
know
helping
like
onboarding
people
faster.
You
know
what
we
could
add.
What
we
can.
You
know
improve
what
you
know
what
you
like
about
it,
and
then
you
can
either
send
that
you
know
ping
us
on
the
slack
channel
for
sig
Docs
or
you
can
open
an
issue
on
the
in
the
repo.
A
G
A
B
So
so
we
talked
about
this
at
the
steering
committee
and
decided
that
you
know
it's
probably
a
sigle
architecture
thing,
because
this
is
about
naming
pieces
of
our
architecture
and
the
upshot
out
of
that
and
I'll
summarize
the
thinking
here,
there's
an
issue
that
I'm
putting
in
the
it's
in
the
nodes
and
it's
in
the
chat
here.
The
thinking
is
number
one
master
is
not
a
great
term
to
actually
describe
what's
actually
going
on.
B
The
quran
DS
control
plane
is
distributed
and
anytime,
where
you're
referring
to
master,
there's
a
certain
laziness
there
and
we
should
probably
find
ways
to
be
more
precise
about
it.
It
does
it
has
this
connotation
of
being
sort
of
a
single
thing
that
can
live
or
fail,
and
the
reality
is:
is
that
sometimes
folks
run
in
that
mode?
But
but
you
know,
especially
for
each
a
or
production-ready
environments,
the
control
plane
will
be
split
up
across
multiple
nodes.
B
So
it's
not
a
great
technical
term
for
what
we're
doing
and
then
then
obviously,
there's
there's
you
know:
divot
ease
around
the
historical
connotations
now
in
the
the
computer
science
literature.
This
is
a
common
name,
and
so
you
know
we
don't
want
to
like
I,
don't
think
there's
only
any
ill
intent
coming
up
here
but,
like
I,
think
you
know,
the
connotations
are
what
they
are,
and
so
we
have
to
be
sensitive
to
that
and
make
sure
that
that
we
don't
try
and
sort
of
fight
and
hold
on
to
something
for
for
silly
reasons.
B
That
being
said,
the
I
don't
think
this
is
anybody
sees
this
as
a
fire
drill,
type
of
thing
of
like
let's
rip
out
a
bunch
of
code
or
we'll
it's
rip
out
a
bunch
of
documentation
over
the
next
release,
or
something
like
this,
but
more
along
the
lines
of
let's
pick
better
terminology
and
agree
on
that
as
a
community.
Let's
try
and
start
using
that
in
our
day
to
day
sort
of
how
we
talk
about
things
and
then
make
sure
that
we
don't
make
the
problem
worse.
B
So,
as
we
see
stuff
as
we
update
stuff,
try
and
and
correct
for
this
and
and
you
know,
create
the
right
incentives
so
that
we
we
switch
to
this
new
terminology
over
time.
The
there's
obviously
parallels
to
when
we
you
know
this
is
historic,
like
ancient
history
right
now
for
kubernetes,
but
knows
used
to
be
called
minions
now,
they're
called
nodes,
it
took
I,
don't
know
something
like
you
know,
two
years
to
actually
complete.
B
All
of
that,
so
the
proposed
terminology
would
be
if
you're
talking
about
the
control
plane
as
a
whole
is
control,
plane
and
and
if
you're
talking
about
something
more
specific,
just
use
the
specific
term
about
that.
So
nothing
totally
like
locked
down
there.
I'd
love
to
hear
some
comments,
come
to
me
offline
or
go
ahead
and
put
some
stuff
in
that
issue
there.
But
but
this
is
what
we're
proposing
and
and
I'm
hoping
that
that
nobody
sees
it
as
as
hugely
intrusive
and
and
that
this
is
a
good
path
forward.
B
D
A
I
I'll
just
say
it
again:
Thank
You
Jordan
Liggett
for
reviewing
so
many
PRS.
That's
super
awesome.
Please
more
people
review
more
PRS
dethrone
him.
If
you
actually
click
through
release
by
release
the
dude's
been
at
the
top
of
the
reviews
for
almost
every
single
release
for
the
past
year,
so
go
for
it.
That's
a
good
metric,
ref
gaming.