►
From YouTube: Kubernetes Community Meeting 20190425
Description
We have PUBLIC and RECORDED weekly meeting every Thursday at 10am PT
See: https://github.com/kubernetes/community/blob/master/events/community-meeting.md for more details.
A
Hello,
everybody,
it
is
April
25th
2019.
This
is
the
kubernetes
community,
meaning
thanks
for
joining
us
before
we
get
started
just
a
few
ground
rules
as
a
community
meeting
and
we'll
be
posting
this
and
streaming.
This
live
on
YouTube.
So
please
be
mindful
of
what
you
say
is
being
recorded.
If
you
are
not
speaking,
please
mute
yourself
so
that
we
can
make
the
environment
as
noise
free
as
possible.
Today
we
will
be-
and
please
keep
in
mind,
that
the
code
of
conduct
is
always
an
effects
of
be
excellent
to
each
other.
A
I
am
still
looking
for
a
note-taker.
If
you
need
the
notes,
I'm
gonna,
post
them
on
chat
in
the
sidebar.
That
would
really
help
us
out
we're
gonna
have
a
demo
today
from
Daniel
who's
gonna
go
in
a
bit,
then
some
release
updates
from
Claire
I'm,
a
contributor
tip
of
the
week
from
me
and
then
to
cigs
are
going
today.
Cig
p.m.
stephen
augustus
and
say
testing,
who
I
don't
know
the
person
who
the
person
is,
but
hopefully
they'll,
show
up
on
time
and
with
that
daniel
you
have
ten
minutes.
Take
it
away.
B
So
basically
just
the
levels
that
an
operator
is
a
piece
of
software
that
extends
kubernetes
to
understand
a
specific
specific
applications,
operational
requirements
so
basically
how
its
installed
I
would
scaled,
how
it
fails
over
and
how
its
upgraded.
So
this
is
like
an
evolution
from
all
the
aspects
of
packaging
and
getting
your
application
initially
installed.
B
The
advantage
of
this
is
that
it
uses
communities
under
the
hood
right,
so
you
have
very
flexible
and
architectures,
while
not
having
to
reinvent
the
secret
at
the
wheel
of
things
like
secret,
ending
service
discovery,
another
core
concepts.
What
you're
really
end
up
using
is
the
same
queue
control
tuning
that
you
have,
and
that
allows
your
abs
to
be
no
much
smarter
and
truly
hybrid,
because
the
operator
ones
on
kubernetes
cluster,
which
could
run
in
your
notebook,
as
well
as
the
data
center
or
the
public
cloud.
B
So
in
order
to
kind
of
shape
how
operators
come
to
life
on
the
cluster
and
how
you
build
operators,
there
is
an
open-source
project
called
the
operative
framework,
which
we
announced
the
queue
con
last
year.
This
is
a
set
of
tools
that
help
you
quickly
get
started,
providing
an
operator
and
also
helps
you
run
multiple
operators
on
your
cluster.
B
So
it's
something
for
the
communities
in
terms
of
builders
to
easily
create
applications,
kubernetes
needed
applications
as
we
like
to
call
them
using
common
methods
and
provide
standard
set
of
tuning
and
also
for
the
consumers
of
these
applications
on
cluster.
It
keeps
apps
apps
to
date.
This
is
primarily
responsibility
of
the
operator
that
second
manager,
and
it
also
provides
a
lot
of
security
around
the
permissions
that
will
operate
the
house,
which
is
usually
quite
privileged
and
long
that
application
on
the
cluster.
B
If
you
want
to
try
it
out,
we
have
several
links
here
on
the
slides,
as
well
as
in
the
notes
section.
What
I
really
recommend
is
the
kind
of
connect
courses
where
you
get
a
browser-based
interactive
tutorial
on
running
operators,
writing
operators
and
using
them
in
production.
And
last
but
not
least,
we
have
launched
a
couple
of
weeks
back
a
community
initiative
to
gather
operators
for
the
kubernetes
community
on
operated
I/o
with
that
I'm
going
to
switch
to
the
demo.
B
So
here,
I
have
a
local
kubernetes
cluster
I'm
in
a
namespace
that
I've
set
up
previously
for
this,
and
what
I'm
going
to
do
is
I'm,
going
to
start
deploying
an
operator
and
use
it
to
install
and
manage
a
distributed,
key-value
store,
which
you
all
are
very
familiar
with
its
HCP.
So
let's
go
ahead
and
start
it.
It's
not
an
operator.
B
You
could
basically
say:
okay,
you
go
and
deploy
it
as
a
mobile
application.
So
it's
a
deployment
with
a
card
that
has
a
certain
set
of
rules
attached
to
ro
bindings,
and
that's
it
that's
really
what
an
operator
is
on
the
cluster.
However,
an
operator
needs
a
little
bit
of
special
care
because
it
is
so
privileged.
It's
usually
not
something
you
want
to
have
in
the
same
manses
as
your
application,
and
it's
also
going
to
be
something
that
you
want
to
install
from
like
a
repository
of
operators.
B
So
that's
also
like
vacation
packaging
aspect
and
play
here,
and
you
also
want
to
be
easily
able
to
do
to
get
updates
to
those
because
updates
certifications
means
fewer
security,
vulnerabilities
and
more
functionality
and
more
abilities
to
run
new
versions
of
your
application
so
to
create
an
operator.
In
this
sense,
we
call
we
create,
what's
known
as
a
subscription.
This
is
an
object
that
comes
with
the
operative
rainbow
and
it's
managed
by
a
component
and
the
opéra,
the
lifecycle
manager
on
my
cluster.
B
The
subscription
is
really
an
intent
to
subscribe
to
an
operator,
install
it
and
keep
it
updated
as
it's
pouring
into
the
catalog.
So
here
I'm
referencing
an
application
named
@cb
which
comes
from
a
catalog
named
operator,
hot
io
catalog,
which
is
in
an
engine
a
sole
M.
So
this
is
all
managed
by
all
M
and
when
I
create
this
subscription.
B
What
will
happen
the
background
is
all
M
will
basically
pick
the
application
from
a
catalog
installed
it
in
the
namespace
that
I'm
currently
in,
and
therefore
deploy
the
operator
and
set
up
two
provisions
correctly
to
work
in
this
particular
namespace.
So
if
I
watch
here
in
the
background,
I
get
a
thought
it's
running
up,
which
is
named
as
the
operator.
B
So
to
use
the
operator,
we
use
a
concept
called
CR
DS,
so
this
operator,
shipped
with
a
couple
of
CDs,
which
you
can
see
here,
one
of
them
is
called
s
CD
clusters,
and
this
is
to
do
a
nice
thing
about
operators
versus
other
aspects
and
approaches
where
you
package
upon
manifest
and
template
them.
You
really
get
a
nice
interface
to
deploy
your
application,
so
what
you're
going
to
do
is
you're
going
to
create
an
LCD
cluster
by
creating
an
object
named
at
CD
cluster.
In
this
case,
it's
called
community
demo.
B
It
has
a
couple
of
parameters
that
make
sense
where
it's
the
cluster
in
this
case,
which
is
like
a
size.
How
many
nodes
I
want
to
have
my
CD
cluster
and
what
version
they
should
be
running,
so
that
could
be
all
of
it
just
to
get
one
up
and
running
I
edit.
An
additional
section
here
that
attaches
persistent
volume
frames
to
this
to
this
at
CD
pods.
So
to
create
this
I'm
going
to
instantiate
the
object.
B
The
operator
is
a
custom
resource
controller
in
the
background
that
watches
for
these
resources
to
appear,
and
that
can
now
basically
a
scoop,
and
that
is
what's
the
status
of
my
a
CD
cluster.
So
there
is
one
that's
currently
being
built
up
and
I
can
describe
it
and
see
what
what
these
statuses,
so
the
status
is
currently
scaling
up,
because
I've
just
started
deploying
it's
currently
its
size.
B
One
designs
as
it's
read,
so
you
can
see
it's
the
same
gold
seeking,
eventually
consistent
nature
of
the
traditional
standard,
kubernetes
operators,
but
this
time
contain
a
specific
project
for
ED
CD,
and
you
can
basically
see
this.
These
pulse
coming
to
life
in
the
background
I
already
have
one
of
them
up
and
running
it's
a
second
one
is
coming
up,
and
this
is
already
executing
a
CD
specific
logic
right.
B
So
here
the
controller
the
sed
operator
makes
sure
that
the
new
parts
really
join
the
class
in
the
clean
fashion
that
they
reach
consensus
on
reach
quorum,
which
is
really
important
for
a
city
to
work
as
a
highly
available
distributed
value
store.
Now,
when
this
is
finished,
I
could
basically
start
using
Ed
CD
at
no
point
that
I
need
to
know
about
how
a
city
comes
to
life
or
how
it's
kept
highly
available.
That's
all
part
of
the
operator,
so
the
operator
basically
tell
me
at
some
point
here
when
it's
done
provisioning.
B
Where
can
I
connect
to
discuss
sir,
what's
to
state
how
many
members
do
I
have
and
so
on?
I
can
basically
shortcut
there's
a
little
bit
for
the
sake
of
time
and
go
into
one
of
these
parts
here
that
make
up
the
SV
cluster
and
in
the
HDD
part
I
obviously
have
all
the
SC
detuning
and
they're
like
HDD
control
and
I
could
ask
it.
What's
the
cluster
health
and
you
see
it's
a
free
node
healthy
as
the
d
cluster,
so
this
is
the
real
deal.
B
It
also
runs
at
a
specific
version
version
3
to
13.
So
now
I
could
already
start
using
this
with
an
application
that
depends
on
this
particular
service.
What
I
can
also
do,
and
what
is
a
really
interesting
terms
of
today
to
operations,
is
basically
managing
this
sed
cluster
from
the
perspective
of
the
customer
resource,
so
it
can
also
say
loop
control,
edit
and
then
my
lcd
cluster
and
could
start
modifying
it.
So,
let's
say
I
want
to
have
a
new
version
of
this,
because
I
depend
on
a
certain
feature.
B
I
could
basically
change
this
version
tag
here.
Let's
change
it
to
3,
3
12
and
then
watch
the
cards
getting
updated.
So
this
is
the
work
of
the
SED
operator
as
well.
It's
in
the
background,
looking
at
the
help
of
the
cluster
and
starting
to
one
by
one
with
the
chrome
in
mind,
we
start
them
and
have
them
read
your
the
cluster
in
a
clean
state
and
reach
consensus
before
continuing
to
the
next
one
and
therefore
very
carefully
updating
arm
updating
the
whole
deployment.
So
this
is
already
done.
That's
all
of
it!
B
So
I
can
go
back
into
my
pot
and
see
what
has
changed
so
again
at
CD
control,
cluster
help
and
I
still
helping
cluster.
So
that's
good
and
that's
easy
control.
Minus
V
I
have
attitude
now
running
in
version
3
feet,
12
like
exactly
like
I
specified,
so
the
reconciliation
nature.
Nature
is
a
core
aspect
of
T
of
the
operator,
so
what
I
can
also
do
is
basically
scale
the
cluster.
This
way
so
I
can
again
hit
edit
and
just
from
the
size
here
to
let's
say
5
and
in
this
reconciliation
dupe.
B
The
operator
would
pick
this
up
and
start
spawning
new
parts
in
the
sector.
So
I
can
completely
control
the
application
from
a
very
narrow
and
user
oriented
surface,
which
is
really
the
API
of
this
operator,
which
is
shipped
in
the
form
of
a
custom
resource
definition.
So
here
you
can
see
it
will
spawn
additional
parts
which
is
really
cool
and
then
I
have
a
5
node
cluster.
Where
did
this
operator
come
from?
Well,
I.
B
Have
all
M
operator
lifecycle
manager
running
in
my
cluster?
It
is
itself
running
as
a
set
of
operators,
as
you
might
imagine,
and
one
of
them
is
actually
an
API
server,
which
reads
the
catalogs,
which
are
basically
cards
running
with
the
database
inside
serving
similar
purposes
like
yum
repository
serve
for
rpms
and
introducing
the
package
manifest
API
I
can
query
this
and
see
all
the
operators
that
I
have
available
on
my
cluster.
These
are
basically
all
the
operators
that
I
get
from
offer
adapter
Dale.
B
This
is
a
very
interesting
operator
because
the
cockroach
operator
is
basically
provided
by
the
home
community.
Now
the
home
community
did
not
write
that
operator,
but
it
turns
out
everybody
can
write
an
operator
with
the
Opera
SDK.
Your
parade
STK
is
one
other
part
of
the
operative
framework
which
allows
you
to
convert
an
existing
help
chart
into
an
operator.
So
when
you
see
my
prompt
here,
you
can
basically
say.
B
So
if
you
look
at
this
generator
scaffolding,
here,
that's
a
deploy
section
which
generated
all
the
memo
manifests
that
are
needed
to
deploy
this
operator.
The
only
thing
that
I
need
to
do
is
build
the
actual
docker
file
that
makes
up
the
operator.
Also,
this
has
been
scaffolding
for
me
from
the
SDK
I
need
to
run
it
from
the
project
directory
and
that's
it
I've
just
built
an
operator,
zero
code
approach.
Everybody
can
do
this
usually
requires
not
a
lot
of
adjustments
to
what's
generated.
B
A
C
What
I'd
like
to
ask
is
if
folks
could
take
a
test
run,
especially
with
the
helm,
stuff,
we'd
love
to
have
more
operators
contributed
to
operator
hub
do
and
if
you
have
any
questions,
I
add
them.
The
link
to
the
Google
group
for
the
sig
on
operator
framework
is
in
the
chat
and
I'll
add
it
into
the
doc,
but
we're
really
looking
for
people's
feedback.
There
is
one
question
now.
A
B
What
I
have
done
is
basically
I
have
installed
over
a
lifecycle
manager
from
the
github
repository
it
ships
with
a
default
catalog,
which
is
exactly
what
you
see
here
and
you're
free
to
basically
carve
those
up
into
stuff
that
you
need,
or
just
manually,
create
like
these
subscription
files
that
you
need
in
order
to
tell
or
what
I'm
to
install
something.
So
the
web
part
is
really
just
a
nice
shortcut.
You
would
not
necessarily
do
this
like
in
other
community
solutions
like
overture.
B
If
you
have
like
you
said
that
provide
you
install
buttons
and
things,
but
really
what
it
comes
down
to
are,
are
basically
subscriptions,
which
is
an
attempt
to
install
an
operator
and
keep
it
up
to
date.
It
over
time
similar
to
your
subscription
and
the
repository.
So
that's
how
you
instantiate
it
and
there's
also
that
construct
of
a
cluster
service
version,
which
is
the
actual
operator
object.
It's
just
a
call
that
way,
which
is
which
tells
all,
and
what
to
do.
So.
If
I
look
at
my
series
here,
I
have
this
one
and.
B
If
I
look
at
it,
I
see
like
information
that
is
used
to
feed
user
interfaces,
to
tell
what
needs
to
be
deployed.
What
custom
resource
definitions
needs
to
be
set
up?
What
Commission's
this
operator
needs
in
order
to
run
correctly
a
textual
description.
This
is
a
bit
of
a
bloated
file
because
it
contains
also
UI
xi
information,
but
for
the
sake
of
deploying
it,
what
it
needs.
Sorry.
A
We
have
the
hard
cutoff
now
agenda
worth
sorry,
but
you
have
the
links
where
people
can
find
you
and
continue
the
discussion.
So
sorry
about
that
also,
thanks
for
being
flexible,
we
did
snafu
this
demo
because
of
the
time
zone
change
from
a
little
while
ago.
So
thank
you,
Daniel
for
coming
on
with
that
we're
moving
on
to
the
release
updates,
Claire.
D
I
will
try
to
keep
this
super
brief,
so
we're
nearing
the
end
of
week
three
for
the
115
released.
We
have
our
enhancements
free
scheduled
for
next
Thursday,
April,
30th
friendly
reminder
to
be
included
in
enhancements.
Freeze,
your
enhancement
must
have
an
open
issue
in
kubernetes
enhancements
and
it
must
have
a
come
in
an
implementable
state
by
EOD
on
the
30th.
If
it
doesn't
have
either
of
those
two
and
you
still
need
it
included
in
the
115
milestone,
you
will
have
to
submit
an
exception
after
the
30th
and
right
now
we
are
tracking
46
enhancements.
D
24
of
those
are
alpha.
14
of
those
are
beta
and
seven
are
stable,
so
lots
of
alpha
features
it
sounds
like,
but
we
will
see
how
many
of
those
fall
off
the
milestone
after
enhancements
freeze
next
week
and
then
after
enhancements,
freeze,
I,
think
the
next
well
actually
technically
before
enhancements.
Freeze
the
next
notable
date
is
on
Monday.
We
should
be
trying
to
cut
our
alpha
2,
and
that
is
all
that
is
happening
in
week.
4
of
the
115
release
any
questions
for
Claire.
A
A
A
This
is
something
that
I
started
a
long
time
ago
and
haven't
had
a
time
to
really
revise
and
our
outreach,
intern
édouard
went
back
and
updated
it
a
bit.
So
what
I
really
like
is
it's
a
list
of
links
of
things
that
you
can
find
in
the
project
that
are
useful
for
you
for
contributing
specifically
the
second
half
of
this
sheet
is
all
brand
new
and
it's
really
useful
and
it
really
kind
of
ties
in
a
bunch
of
pages
from
around
the
community
repo
into
one
kind
of
cheat
sheet.
A
That's
really
useful
and
I've
also
put
in
a
link
to
a
epic
bug,
like
a
tracking
issue,
which
I
will
paste
in
chat
again
for
people
to
translate
this
into
their
native
language.
So
they've
got
a
bunch
of
Chinese
Korean,
Japanese,
German,
Spanish,
French,
Indonesian,
Portuguese,
Italian
and
Hindi.
Each
of
them
had
their
own
issue
filed
in
github.
So
if
you
speak
any
of
those
languages,
I
want
to
help
contribute
to
translating
that
we
would
appreciate
it
and
with
that
we're
gonna
move
on
to
sing
updates.
A
B
E
Cool
all
right,
so
I'm
gonna
actually
keep
it
short
and
sweet.
If
I
like
to
do
that's
the
wrong
button,
this
is
the
right
all
right,
hey
everyone.
I
am
steven
augustus,
I'm
one
of
the
sig
PM
chairs
we're
going
to
go
through
just
a
few
slides
today,
Zig
p.m.
so.
What
have
we
been
up
to
some
of
our
most
recent
updates?
Is
we
recently
did
a
chair
rotation
ish,
so
Aparna
it
has
been
a
chair
for
quite
some
time.
E
We
were
graduating
her
two
emeritus
status
and
myself
and
Jace
are
stepping
up
as
SiC
p.m.
chairs
and
then
also
we
made
a
miter
miter
monofin
minor
modification
to
the
cig
p.m.
mailing
list
to
kind
of
bring
it
in
line
with
all
of
the
all
the
other
sig
mailing
lists.
So
that
happened
yesterday
just
in
time
for
this
meeting,
so
you
can
reach
us
at
kubernetes,
cig
p.m.
at
Google,
Groups
comm.
E
E
So
everything
that's
involved
with
making
keps
successful
and
and
kind
of
spreading
the
word
throughout
SIG's,
the
second
one
being
project
management,
so
there
has
been
a
it
seems
like
there's,
been
a
need
overall
across
SIG's
and
working
groups,
and
any
any
group
within
the
kubernetes
community
to
have
project
management
to
have
some
sort
of
grooming
process
for
four
org
level
and
project
level
boards.
So
what
we'd
like
to
do
is
essentially
staff
project
managers
for
forsakes,
so
expect
to
hear
more
of
that
I
think
I
know.
E
E
architecture
right,
I
believe
a
has
been
covering
some
of
the
stuff
for
four
contrib
x,
so
that
will
probably
continue
but
yeah.
That
program
will
be
developing
over
the
next
few
quarters
if
you
have
interest
as
a
project
manager
or
interest
as
a
cig.
Who
wants
a
project
manager?
Please
allow
them,
please
let
us
know
the
the
next
bullet
is
the
CNCs
end-user
community
so
end-user,
as
in
specifically
the
CNC
F
end-user
community,
which
is
a
which
is
a
group.
E
It's
a
membership
level
for
CTF
and
what
we
usually
do,
every
quarter
quarter
or
two
is
do
a
presentation
to
the
end-user
community
right,
so
these
are
actual
consumers
of
kubernetes
and
what
we've
been
realizing
in
that
kind
of
in
that
communication.
Is
that
it's
kind
of
one-sided,
so
we
want
to
be
able
to
kind
of
open
up
a
feedback
loop
between
that
end
user
community
get
there.
E
You
know,
get
their
notes
on
what
we
could
be
doing
better
as
a
community
and
bring
that
back
in
and
kind
of
so
we're
working
with
Cheryl
hyung,
who
is
the
director
of
ecosystem
at
Linux
Foundation,
and
we
want
to
so.
We
are
first
I
think
our
first
meeting
is
going
to
be
June
11th.
So
if
you're
part
of,
if
you're
part
of
the
CTF
end
user
community,
we'll
be
doing
a
presentation
on
both
sig
p.m.
E
E
How
do
we
message
what
happens
during
not
just
the
release
cycle,
but
how
do
we
track
things
that
are
out
of
tree
for
kubernetes
right
and
what
we're
realizing
is
that
being
able
to
just
being
able
to
essentially
kind
of
enumerate
all
the
things
that
have
happened
across
a
release
cycle?
It
basically
happens
within
the
change
log
of
kubernetes,
kubernetes
I.
E
Think
that
that's
not
entirely
in
accurate
place
for
that
to
be
put,
but
it
has
classically
been
the
place
for
for
these
updates
to
go
strictly
for
the
sake
of
being
on
that
kind
of
PR
and
communications
train
for
for
each
quarter.
So
what
we'd
like
to
do
is
is
figure
out
a
way
for
SIG's
to
still
present
the
infirm
what's
been
happening
for
them
day-to-day
weather.
This
is
vlog
posts,
podcasts
different
things
like
that,
right,
so
figuring
out.
E
So
some
things
that
we've
been
noodling
on
things
that
you
can
expect
to
see
within
the
next
quarter
or
so
a
kept
one
sheet
so
lucky
had.
The
awesome
idea
of
this
will
essentially
be
the
the
FAQ
of
how
to
be
successful.
When
writing
a
cap,
we
think
that
that's
a
reasonable,
reasonable
output
for
for
the
process.
You
know
over
over
the
last
few
quarters,
especially
as
we
use
made
caps
a
requirement
for
the
release
process.
There's
been
confusion,
there
are
kept
set
lag
and
we
want
to.
E
These
go
hand-in-hand,
one
cheater,
as
well
as
the
kept
retrospective
I'm
thinking,
based
on
timelines,
that
we
probably
won't
be
able
to
do
something
like
that
until
after
keep
gum,
but
essentially
what
we'll
do
is
go
from
we'll
pick,
maybe
the
top
six
SIG's
or
so
and
have
a
have
a
sit
down
in
one
of
their
Sigma
T's
to
talk
about
their
experience
with
caps
and
what
we
can
be
doing
better
to
that
end
as
after
the
retrospective
and
I
think
I.
Think
after
cube,
con
will
plan
to
do
you
a
kept
office
hours.
E
So
this
is
this
will
be
an
additional
meeting,
probably
bi-weekly
so
it'll
it'll
it'll
be
the
the
week
right
after
the
the
standard
say
p.m.
meeting,
and
what
this
is
is
you
can
essentially
say:
hey
I'm,
getting
ready
to
write
a
cap.
I,
don't
know
where
to
begin
or
hey
I've
got
a
Kemp
that
I've
already
submitted.
How
can
I
move
it
forward?
You
know
it's.
It's
marked
as
provisional
and
I
haven't
been
able
to
get
traction
on
or
I've
got
a
Kemp
that
I
want
to
move
from
provisional
to
implementable
and
again
I.
E
E
The
additionally
I
think
it
I
think
part
of
it
is
the
people
who
are
listed
in
cap
in
the
cap
owner
files
right
now
are
essentially
the
sig
chairs
and
sig
technical
leads.
I
want
to
make
it
known
that
if
you
have
people
in
your
sig
that
you
trust
to
review
and
approve
of
review
and
approve
this
content,
you
should
add
them
as
as
owners
for
your
individual
cap
subdirectory.
So
you
know
again-
and
this
extends
this
extends
both
in
the
on
the
on
the
cig
level,
as
well
as
kubernetes
wide
level
right.
E
So
these
are
top
level
caps
that
are
reviewed
by
architecture
and
maybe
a
few
other
groups.
If
you
again,
if
you
have
reviewers
and
approvers,
that
you
trust
that
are
not
just
the
the
sig
chairs
and
technical
leads
feel
free
to
add
them
to
the
owners
files
for
your
for
your
individual
individual
sub
directories,
and
then
you
know.
Finally,
we
want
to
really
move
forward
on
closing
out
the
the
non
KK
targeted
kubernetes
enhancements
is
like
writing
in
hieroglyphics
there,
but
but
the
everything
that
is
not
in
tree
gruber
Nettie's.
E
We
want
to
start
shutting
down
me.
We
currently
have
a
list
of
I
want
to
say,
136
enhancement,
issues
that
we
track
and
and
someone
and
it's
usually
a
singular
person,
and
occasionally
some
shadows
across
the
across
the
cycle,
have
to
read
each
of
those
and
try
to
and
try
to
project
what
the
the
current
status
is
for
each
of
those
enhancement
issues,
so
I
think
I
think
removing
some
of
that
burden
by
by
kind
of
starting
to
thin
out
the
the
non-urban.
E
A
discriminatees
targeted
issues
will
be
helpful,
so
we're
going
to
try
to
do
that
over
the
next
quarter
as
well.
So
how
can
you
reach
us
who
we
are
Kaleb
miles?
He
heard
of
wretzky
Jay,
singer
DeMars
and
myself.
Steven
against
this
we
are
available
are
all
of
our
info
is
on
kubernetes
community
sig
p.m.
the
slack
channel
again
a
sig
pm
and
the
mailing
list
is
kerbin.
It
is
sig
p.m.
thank
you
all
for
your
talent.
F
Am
NOT
a
bot
just
right
now,
if
you
figured
out
how
to
have
my
bot
chat,
I'm,
so
I
don't
have
a
presentation,
unfortunately,
or
anything,
but
I
can
chat
about
various
things,
so
I'm
Eric
from
Sega
testing
and
guess
maybe
talk
a
little
bit
about
some
kind
work.
That's
been
going
on
some
prowl
work
and
some
changes
to
our
build
system
and
on
the
kind
side
of
things
kind.
If
you
don't
know,
is
our
system
for
deploying
a
kubernetes
cluster
into
a
pod
or
a
container
which
we
think
will
be
useful.
F
Eventually,
we
want
to
make
this
the
primary
way
of
doing.
Like
you
know,
ete
tests
on
your
local
workstation
without
having
a
cloud
provider
we're
working
on
getting
a
blocking,
pre
submits.
We
have
a
pre
submit,
that's
running,
that's
not
blocking
we're
working
on
making
that
blocking
long
term.
There's
been
some
discussion
with
cig
cluster
lifecycle
about
sort
of
making
that
the
primary
you
know
slash
only
supported
way
of
having
blocking
priests
limits
and
I
thought
recent.
F
You
know
this
sort
of
seems
to
be
having
a
lot
of
traction
around
the
community
people
like
to
use
it.
So
if
you
haven't,
please
try
it
out.
It's
faster,
obviously,
then
deploying
a
real
cluster.
We
recently
got
it
working
on
arm
64,
it's
pretty
sweet
on
the
pro
side
of
things,
probably
sort
of
our
key
immunity,
CI
CD
system.
So
you
know,
if
you
see
lots
of
labels,
shenanigans
and
whatnot.
That
is
what
is
doing
that
and
so
recently
we've
been
in
a
more
plugins
around
the
whole
organization.
F
To
start,
we
get
into
weird
situations
where
people
will
add
people
into
the
owners
file
who
aren't
community
members
and
so
that
winds
up
with
a
weird
experience
where
technically
I
think
they
can
approve-
or
at
least
they
are,
they
are
an
approver
in
the
owners
file.
But
github
won't
allow
us
to
assign
that
person
since
they're,
not
an
org
member
and
so
that
winds
up
causing
problems,
and
so
we
want
to
make
sure
that
people
who
are
owners
of
kubernetes
code
are
also
community
members.
F
F
When
we
merge
commits
to
prowl,
it
then
deploys
those
images
and
then
every
day
a
fade
about
like
well
prowl
will
create
a
pull
request
asking
for
us
to
give
it
permission
to
update
itself
and
when
we
merge
that
p.r
it'll
update
itself,
and
so
that's
all
been
working
pretty
well.
We
have
two
new
utilities
that
might
be
useful
and
we've
also
been
trying
to
improve
our
getting
started.
Docs
because
every
you
know
we
sort
of
get
a
trickle
of
people
who
are
interested
in.
F
F
There's
we're
also
working
on
this
tool
say
no,
which
will
essentially
run
a
proud
job.
Local
in
a
sort
of
convert,
a
proud
job
into
a
docker
run
command,
so
will
allow
you
to
rerun
jobs
locally,
and
so
as
long
as
you
yeah.
So
that
should
potentially
make
it
easier
for
people
to
debug
what
this
change
to
my
job
will
do
without
actually
having
to
merge.
It
there's
also
a
doc
out.
F
We're
sort
of
create
trying
to
create
a
road
map
for
a
prowl
about
things
we
want
to
do
a
lot
of.
It
has
to
do
with
sort
of
improving
some
of
our
monitoring
stuff,
starting
to
use
more
of
the
standard
kubernetes
like
api's,
like
the
things
for
working
with
CRTs,
and
we
also
want
to
make
it
eventually
there's
an
idea
of
moving
configs
for
jobs
into
the
repo
itself,
as
opposed
to
having
to
send
a
PRN
to
test
infra.
F
So
various
things
like
that
and
we're
still
collecting
info.
So
if
you
once
have
an
idea
of
things,
you
would
like
us
to
prioritize,
go
find
us
probably
on
cig
testing
first,
and
then
we
can
point
you
to
the
doc
on
the
building
side
of
things.
There
we've
been
experimenting
with
so
you
know,
most
of
our
CI
uses.
Basil
and
basil
has
a
mode
of
remote
doing
things
using
remote
execution,
and
so
we've
been
playing
around
a
little
bit
with
that
which
should.
F
Hopefully
we
have
a
weird
system
where
like
when
we
start
a
pull
request.
We
have
to.
We
have
a
big
spike
of
compiling
everything,
and-
and
so
we
want
to
give
that
a
lot
of
CPU,
so
it
goes
fast,
but
then
it
sort
of
spends
the
rest
of
the
hour
sitting
around
waiting
for
the
cluster
to
do
stuff.
So
we're
trying
to
smooth
out
that
spike
enos
by
sort
of
moving
all
of
the
build
into
a
remote
server
cluster
and
then
the
you
know
so
then
we
can
sort
of
just
get.
F
We
can
potentially
run
more
things
I'll
by
doing
that,
and
we've
also
been
testing
for
just
like
the
main
repo
has
switched
over
to
using
go
modules
and
we've
been
kind
of
experimenting
with
how
that
system
works.
We've
been
trying
to
sort
of
automatically
patch
and
do
minor
upgrades
of
our
modules
as
I'm
sure
everybody
will
be
shocked
to
hear
the
go.
F
Community
isn't
yet
super
great
at
doing
simper
correctly,
which
sort
of
conflicts
with
modules
assumption
that
people
are,
and
so
the
minor
upgrades
kind
of
are
hassle
right
now,
but
we'll
see
how
that
progresses
and
yeah
I
know.
If
anybody
has
any
questions
or
something
exciting,
I'm
sure
I
probably
forgot
something,
but
that
is
sort
of
the
few
things
I
could
rattle
off
off
of
the
top
of
my
head.
A
A
Okay,
moving
on
to
the
announcements,
the
LTS
working
group
still
running
a
survey
that
ends
on
April
26,
that
is
tomorrow,
I've
added
the
link
to
the
notes
and
we'll
put
them
into
chat.
So
if
you
can
check
that
out
and
give
them
as
much
data
as
possible,
we
would
really
appreciate
that
and
there's
a
link
to
the
working
group
itself
and
let
me
get
the
survey
there.
A
There
we
go
and
DIMMs
wanted
to
remind
everyone
that
we
now
have
a
hash,
PR
review
slack
channel
as
a
last
resort
if
your
PR
is
stuck
so.
The
idea
between
this
is,
if
you've
submitted
a
PR,
and
you
feel
like
it's
not
getting
any
traction
and
maybe
you've
tried
to
go
to
a
sig
meeting
and
it
was
too
busy
you
couldn't
get
on
the
agenda.
The
idea
for
this
channels
to
exist
is
to
kind
of
have
like
a
last
minute.
Hey
last
call
you
know:
I've
been
struggling
with
this
PR.
A
We
did
have
someone
join
contrived
acts
last
week
that
had
a
PR
they've
been
sitting
around
for
a
year.
So,
like
that's
the
kind
of
thing
we're
trying
to
avoid
with
that.
So
if
you
have
cycles
to
help
please
idle
in
there
or
I'm
sure,
if
you
ping
tain
them
he'd
be
happy
to
let
you
know
how
to
get
involved
shoutouts
for
this
week.
So
what
shoutouts
are?
We
have
hash,
shoutouts
and
slack
if
someone's
going
above
and
beyond
the
call
of
duty,
and
you
want
them
to
get
them
some
recognition
in
the
community.
A
We
read
these
out.
So
Nikita
would
like
to
shout
out
to
édouard
standing
the
contributor
cheat
sheet
and
following
up,
even
though
his
outreach
internship
had
ended,
tada
emoji
zachary.
Sarah
would
like
to
shout
out
to
brad
Topal
for
his
willingness
to
step
out
of
a
barcelona
speaking
slot
in
order
to
make
room
for
a
local
leader
to
step
in
that's
just
classy.
A
Well
done
there,
Brad
and
Lucas
would
like
to
thank
shout
out
to
Eugene
Glo
Tov
for
helping
me
with
AWS
e
lb
and
e
KS
stuff
thanks
and
with
that.
We
are
early
I'd
like
to
give
you
twenty
minutes
back.
Anything
else.
Last
minute
announcements,
anything
okay,
great
next
week
up
will
be
GCP,
storage
and
docks
will
have
their
status
updates
and
we
will
see
everyone
next
week.
Thank
you.