►
From YouTube: Kubernetes Office Hours 20200916 (EU Edition)
Description
Office Hours is a live stream where we answer live questions about Kubernetes from users on the YouTube channel. Office hours are a regularly scheduled meeting where people can bring topics to discuss with the greater community. They are great for answering questions, getting feedback on how you’re using Kubernetes, or to just passively learn by following along.
For more info: https://k8s.dev/events/office-hours
A
I
hope
everyone
has
had
a
good
day
and
a
good
week
it
looks
like
we
are
live,
welcome
everybody.
It
is
the
third
wednesday
of
the
month.
That
means
it's
time
for
the
kubernetes
office
hours.
This
is
our
monthly
live
stream.
Where
we
sit
in
the
kubernetes
slack.
We
answer
all
of
your
questions
about
kubernetes
as
many
as
we
can
in
the
next
hour,
and
then
we
hope
you
have
a
bunch
of
fun
and
learn
a
bunch
of
things
while
doing
it
before
we
start.
B
Oh
hi,
I'm
here
I'm
senior
software
architect
at
spectrum,
I'm
currently
mostly
focusing
on
bringing
workloads
onto
communities
doing
distributed
services
like
designing
api
interfaces
and
so
on.
So
that's
my
day-to-day
little
bit
of
operation,
but
a
little
bit
less
from
that
in
the
last
couple
of
weeks.
C
Cool:
hey,
hey
everyone,
I'm
pablo
swarzotkos,
I'm
from
lithuania.
People
typically
call
me
pav
for
simplicity,
I'm
certified
kubernetes
administrator
and
I
know
a
lot
about
monitoring,
observability,
sre
and
the
website
of
things.
I'm
also
a
maintainer
of
project
anas,
which
is
a
long-term
storage
solution
for
prometheus.
So
yeah.
A
Awesome
and
rachel
is
new
here-
welcome
rachel.
Everyone
welcome
rachel,
introduce
yourself.
D
Thanks
for
having
me,
I'm
rachel
leaking,
I'm
a
kubernetes
field
engineer
at
vmware,
I'm
mainly
focused
on
oppa,
that's
kind
of
what
I
do
most
of
the
time
and
then
as
well
as
now
recently
working
on
some
of
the
tons
of
stuff
for
vmware.
A
A
And
I
am
your
host
george
castro
and
I
also
work
at
vmware
and
we've
been
running
this
project
for
about
two
years
two
years,
something
like
that
and
we
are
working
on
a
blog
post
on
all
the
things
we've
learned
with
doing
office
hours.
So,
if
you're
in
chat
right
now,
if
you're
joining
us
on
youtube,
you'll
see
the
slack
channel
that
we
have
there
on
the
side
so
feel
free
to
whack
in
your
question
as
the
show
goes
and
we'll
get
to
them.
A
So
first,
let's
talk
a
little
bit
about
the
rules.
Here's
how
we
go
like
I
said
before
the
kubernetes
code
of
conduct
is
in
effect,
so
please
be
excellent
to
each
other.
This
is
also
a
judgment-free
zone,
so
everyone
had
to
start
from
somewhere.
So
please
help
out
your
buddy
by
having
a
supportive
environment
in
the
channel
for
us.
A
We'd
appreciate
that
also,
if
you're
listening
in
the
channel,
let
us
know
how
the
audio
sounds
and
also
let
us
know
where
you
are
in
the
world,
what
you're
working
on
what
your
title
is
and
how
you're
using
kubernetes.
We
love
to
hear
all
that
kind
of
stuff
so
that,
while
we
are
talking,
we
have
a
nice
cool
sidebar
on
the
side
of
people
having
a
good
time,
and
while
we
will
do
our
best
to
answer
your
questions,
the
panel
doesn't
have
access
to
your
cluster,
so
live
debugging
is
off
topic.
A
We
can't
ssh
to
your
stuff.
We
have
no
idea
how
your
networking
works,
but
what
we
will
do
is
do
our
best
and
try
to
help
you
get
you
moving
towards
the
next
step
of
solving
whatever
your
problem
is
so
panelists
you're
encouraged
to
expand
on
your
answers
with
your
experiences
and
pro
tips
audience
you
can
help
by
piecing
in
urls
to
the
official
thoughts,
blogs
or
anything
that
may
be
relevant
to
the
question
at
hand.
A
Every
single
episode
that
we've
had
so
far
there's
always
a
new
tool
or
something
that
people
are
pasting
into
the
chat
and
that
really
helps
us
out.
You
can
also
help
us
out
by
tweeting
spreading
the
word
and
helping
pay
it
forward.
This
panel
is
also
made
up
entirely
of
volunteers.
So
if
you
want
to
rotate
in
please
let
us
know
we'd
love
to
have
new
people
like
rachel
wrote
in
and
help
out
in
fact,
today
we're
a
little
light
on
panelists,
so
this
might
be
a
little
challenging
for
us.
A
So
if
there's
anybody
in
the
audience
who
wants
to
take
a
stab
at
coming
on
the
panel,
let
us
know-
and
you
can
hop
on
the
zoom
chat
right
now
with
us.
That's
how
pierre
got
started.
That's
how
chris
got
started
and
I
pablos
you
were
in
the
audience
first
right
and
then
volunteered.
If
I
remember
correctly,.
C
C
An
amazing
kubernetes
t-shirt.
A
And
you
won
the
t-shirt
that
is
excellent
and,
like
I
said,
if
we
do
address
your
question
on
the
air
in
the
next
50
minutes,
or
so
I
like
to
give
away
t-shirts
away
at
the
end,
but
you
have
to
stay
till
the
end
to
win
your
t-shirts
also
before
we
start
we'd
like
to
thank
the
following
companies
for
supporting
the
community
with
their
developer
volunteers,
giant,
swarm,
stockx,
pivotal
pusher.com,
weaveworks,
vmware,
university
of
michigan,
red
hat
and
utility
warehouse,
and
I
forgot
yours
in
there-
spectrum
thanks
for
pierre,
that's,
fantastic
and
always
a
special
thanks
to
the
cncf
for
sponsoring
our
t-shirt,
giveaway
all
right
with
that
everyone.
A
Please
keep
pacing
in
your
questions
in
chat
and
we
will
get
to
them
in
the
notes
here.
The
first
question
comes
from
doogie.
Welcome
to
the
show
says:
I've
got
a
question
for
you.
My
requirement
is
dependable
egress
static
ips
for
my
cluster,
so
that
my
vendors
can
whitelist
our
access.
Ideally
we'd,
be
able
to
segment
the
namespace
or
pods
to
specific
node
groups
so
that
we
could
have
blocks
of
ips
for
different
purposes
running
on
gke.
A
B
If
you're
running
the
managed
source
of
kubernetes
on
google
cloud,
I
think
it
will
be
even
harder
because
you
have
like
very
little
control.
I
think
you
cannot
like
provide
your
own
ip
space
to
it,
draw
something
so
not
really
sure
how
really
to
address
this.
C
Yeah
I'm
looking
at
at
this,
I
think
like
cloud.
Not
this
is
the
way
to
go
to
be
honest
or
like
deploy
some
kind
of
a
vm
with
the
proxy,
which
then
would
you
that
then
you
would
have
like
a
static
ip,
which
would
be
that
vmc
like
public,
ip
or
yeah
cloud,
not
not
ip
and
that's
it
and
then
basically,
your
earpods
actually
talk
to
that
proxy
and
trucks
actually
talks
to
blender,
because
in
kubernetes
I
don't
think
we
have
any
kind
of
ways
of
yeah
controlling
kps.
A
A
I
know
I
can
overwrite
etsy
resolve.conf,
but
I
can
only
use
one
dns
server,
so
I
either
either
use
the
ms
active
directory
one
or
the
kubernetes
one,
but
not
both.
So
my
pod
will
lose
ability
to
resolve
the
kubernetes
internal
urls,
or
can
I
set
it
globally,
but
all
pods
will
have
the
same
dns
server.
Can
I
have
pods
using
a
specific
one,
interesting.
B
B
B
I
sent
the
configuration
like
the
the
docs
and
the
stack
channel.
That's
right,
yeah.
B
A
Great,
I
I
also
I
have
a
dumb
follow-up
question.
If
I
may,
why,
why
would
you
do
couldn't
you
set
the
the
upstream
on
the
internal
dns
to
the
active
directory,
one
and
kind
of
get
the
same
effect.
B
So
for
us
we
actually,
we
had
a
similar
issue
where
we
had
two
rds
instances
or
like
database
instances
with
the
same
host
name
that
were
external,
like
we
had
two
dns
servers:
okay,
similar
host
names
and
one
of
the
parts
should
access
this
one
but
the
other
one
there
and
we
wanted
to
have
dns
resolution
there.
So
that's
why
we
use
this,
but
I'm
not
sure
if
we
actually
had
to
do
any
workarounds.
A
Okay
and
pop,
I
see
you've
pasted
in
the
link
to
doogie's
question.
Do
you
want
to
go
over
that,
or
is
that
just
the.
A
Awesome
yeah
all
right
simone.
I
hope
that
helps
you
out.
Let
us
know
how
that
works,
and
also,
if,
if
you
have
a
follow-up,
question,
feel
free
to
just
go
ahead
and
ask
again,
we
have
no
problems
going
back
to
another
question.
All
right,
evcs,
welcome.
I've
seen
you
on
the
show
before
says
we're
using
an
init
container
centos
7
to
do
some
iptables
rules
to
redirect
traffic
to
localhost
6379
to
remote
ip6379,
which
works
fine.
We've
been
moving
to
centos
eight
containers
for
our
base
now,
which
replaces
ip
tables
with
nf
tables.
Uh-Oh.
A
I've
got
the
equivalent
nft
commands.
However,
I'm
always
getting
error
cannot
process.
Rule
operation
not
supported
the
init
containers
is
running
with
net
underscore
admin
as
it
was
before,
and
I've
also
tried
adding
net
underscore
raw.
It
works
successfully
just
running
via
docker,
on
my
local
machine,
with
the
same
capabilities
using
dash
cap
dash
ad,
but
I
can't
get
it
working
in
kubernetes.
Could
this
be
down
to
the
underlying
os
on
the
kubernetes
nodes?
Gke
cos,
I
think
that's
cloud
os
right,
which
is
like
choreos
or
some
other
factor.
I
haven't
considered.
C
There's
the
issue
is
that
by
default,
kubernetes
users
be
all
the
way
like
the
old
ip
tables
yep.
A
F
C
A
I
found
a
very
long
thread
from
one
of
the
sig
from
tim
hawking
going
wow
this
ip
tables,
none
of
tables
transition
is
bungled.
Anyone
using
kubernetes
nissa's
anyone
using
kubernetes
on
sufficiently
new
distros
that
have
iptables
1.8.x
need
to
switch
to
iptable's
legacy
mode.
The
kernel,
implementation
change
and
there's
no
stable
api
to
hold
on
there's
a
huge
long
threat
on
this.
Let
me.
A
F
B
A
A
A
A
B
A
Find
I
found
this
so
I
found
this
one
yeah
container
optimized
os
yeah.
I
believe
that's
that's
google's
gke
core
os
like
so.
This
is.
A
A
D
So
in
that
ticket
george,
I
was
looking
at
that
one
as
well.
Yeah.
E
D
A
A
Oh
interesting,
so
on
debian
systems
you
can
do
an
update
alternatives
to
set
ip
tables
to
be
legacy,
but
you
said
that's
not
available.
In
centos
seven
guy
champton
says
I
started
to
play
with
centos
eight
and
it
comes
with
ip
tables.
I
have
tables,
but
without
legacy
right-
and
you
said,
seven
doesn't
come
that
with
that
either
then
they
mentioned
that
the
ip
tables
user
space
tools
are
provided
in
a
container.
Let
me
check
that
out.
C
B
It
santos
eight
that
he
has
trouble
with.
He
said
he
works
with
centos,
seven
like
we're,
using
it
in
container
center
s7
to
do
some
ip2,
which
works
fine
and
then
he
says
we
have
been
moving
to
centos
eight
containers
for
our
base.
Now,
and
it
does
so,
I
mean,
doesn't
kubernetes
not
use
nf
tables
at
all
right
now
like?
Is
it
not
still
mostly
running
ip
tables,
and
might
that
be
the
issue.
A
A
A
A
A
If
you
keep
scrolling
down,
then
you
see
in
in
71
305.
If
you
scroll
down,
you
see
all
the
hyper-connected
links
where
everyone's
like
new
operating
systems
are
busted
right,
cluster
ip
udp
port
not
available
on
centos
8.
Let's
see
if
that's
related.
F
B
So
did
anyone
on
the
panel
ever
migrate
from
gke
to
eks,
so
that's
google
offering
to
amazon
offering
a
managed
offering
and
what
were
the
things
that
didn't
quite
work
the
same
way
or
what?
Where
what
possible
issues
would
you
say
I
should
look
out
for
that's
from
christian
roy,
and
so
does
someone
want
to
dig
at
this,
or
is
everyone
still
busy
with
the
ap
cables.
A
B
So
I
mean
one
of
the
major
differences
I
would
say
is
a
recycle
between
gke
eks,
like
eks,
is
usually
lacking
a
little
bit
behind
it's
currently
on
1.17
and
we
are
currently
on
the
official
release
of
1.19,
I'm
not
sure
where
gke
is
at
so
like.
I
would
make
sure
that
your
version
is
at
least
supported
by
eks
that
you're
currently
running
and
yeah
I
mean
some
settings
are
different.
Like
load
balancers
are
differently
initialized,
so
there's
like
provider
like
club
provider,
specific
annotations.
B
So
if
you
ever
set
something
like
an
increased
timeout
or
something,
you
would
need
to
change
that.
But
these
are
the
things
that
I
would
look
out
for
so
gke
is
1.18
yeah.
B
B
A
C
Yeah
just
add
to
that
usually
like
storage
in
different
providers.
D
D
C
Differently
so
make
sure
to
look
at
that
and
another
big
thing
is,
I
think,
load
balancers.
So
I
think
google
has
like.
I
think
google
has
like
a
little
bit
better
loadable
answers.
We.
It
was
like
just
easier
for
us
to
manage
them
and
like
basically,
for
example,
if
you
want
to
get
the
client
ap
back
proxy
back
to
you,
it
was
like
way
easier
to
do
than
in
amazon.
C
Yeah
so
little
things
like
that.
But
I
think
if
you
are
doing
the
transition
anyway,
you
will
learn
them
as
you
go.
F
B
Yeah,
that's
true.
Actually
it's
one
thing
that
I
have
not
taken
the
micro
in
these
clusters
right
now
that
I
get
the
host
ap
like
the
request,
ip
from
the
right,
like
the
forwarded
header
or
whatever.
A
Look
at
this
so
pop
pop
left.
Thanks
christian,
we
have
room
for
follow-up
questions
here,
while
we
finish
up
this
iptables.
One
pop
welcome
to
the
show
sent
me
sent
us
a
link
here
to
the
kubernetes
website.
A
It
says
if
your
system's
iptables
tooling
uses
the
nf
nf
tables
backend.
You
will
need
to
switch
the
ip
tables
to
link
to
legacy
mode
to
avoid
these
problems.
Okay,
I
think
we
we
determined
that
this
is
the
case
on
at
least
debian
10,
ubuntu,
1904,
fedora,
29
and
newer
releases
of
these
distributions
by
default.
B
C
Just
I'm
just
I
mean
the
whole
issue.
Is
it's
really
hard
with
like
switching
ip
tables,
because
I
mean
that
I
remember
reading
that
issue
a
while
back.
Maybe
my
knowledge
is
out
of
date,
a.
C
F
A
D
A
E
A
To
me
interesting,
but
let's
still,
let's
still
show,
let's
show
them
how
to
find
the
network
office
hours,
because
you
know
sure
the
people
that
work
on
this
stuff
every
day
probably
are
doing
a
better
job
than
us
digging
in
github
the
best
that
we
can
I'll
find
that
information
and
paste
it
in
slack.
I
got
that
pierre.
Meanwhile
we're
about
we're
about
halfway
through.
So
if
anybody
has
more
questions,
please
feel
free
to
whack
them
in
the
channel.
A
B
Just
posted
it,
oh
okay,
all
right,
it's
like
so
also
the
link
is
the
mailing
list
link,
I
think,
and
it
might
be
worth
to
send
a
link
there
like,
like
your
question
there
as
well
right,
there's
a
lot
of
people
there
very
knowledgeable.
A
Ip
table
source
rpm
includes
iptables
legacy,
however,
the
rpm
spec
drops
the
legacy
ones.
So
if
you
want
yeah,
it's
yeah
that
might
be
an
option
for
you,
grab,
grab
the
spec
file
and
just
build
a
legacy.
Rpm.
A
Yeah,
that's
it
it's
usually
a
bluetooth,
that's
dropping
all
the
old
legacy
stuff
and
then
the
centos
stuff
gets
to
it
like
two
years
later.
I
think
that's
really
interesting
and
while
lead
there's
leaving
you
a
link
there
on
on
the
ip
table
saying
so.
I
guess
then,
at
that
point,
though
you're
effectively
maintaining
your
own
iptables
package
for
your
infrastructure
right.
B
Is
it's
behind
a
paywall?
I
think
like
it's,
you
need
to
be
a
customer
of
red
hat,
I
mean
it
makes
sense.
Royal
8
is
enterprise
solution
right,
so.
A
C
Yeah
I
just
looked
at
santa's
seven
and
turns
out
like
center.
Seven
end
of
life
is
like
20
24.
so
like
there
are
like
you,
you
still
have
like
about
four
years
of
like
getting
security
updates
and
everything
yeah.
D
C
C
A
A
All
right,
we
got
more
questions.
I
hope
that
gets
you
who
was
this.
You
see,
I
hope
this
gets.
You
started
at
least
hopping
in
the
right
direction
for
sure,
stop
and
sig
network
and
give
them
that
feedback
there.
A
A
A
Over
to
firewall
d
does
that
anyone
keeping
track,
I'm
I'm
hoping
walid,
is
keeping
track
of
distro
distro
news
here
like
at
some
point,
do
we
all
switch
to
enough
tables
firewall
d?
I
don't
I
don't.
Have
you
ever
heard
of
firewalld?
Is
this
a
new
systemd.
A
A
All
right,
I
will
have
to
check
that
later
out.
I
don't
wanna,
I
don't
bog
us
down
going
into
every
every
firewall
thing
for
all
the
distros
of
what
they're
doing,
but
that
is
very
interesting.
Hopefully,
that'll
help
help
you
get
started
and
let
us
know-
and
that
would
be
awesome-
there's
some
discussion
now
about
the
next
question,
which
I'm
gonna
go
ahead
and
read
any
tool:
recommendations
for
kubernetes
security,
benchmarking,
everything
I've
seen
so
far
is
not
configurable
enough
to
cater
for
non-vanilla
deployments
of
kubernetes
non-vanilla
is
in
quotes.
A
I
don't
know
what
that
means,
but
hopefully
it's
kubernetes
enough.
We've
had
this
question
asked
before,
of
course,
cube
bench
always
gets
mentioned,
let's
see
panel.
What
are
you?
What
have
you
all
been
recommending
here
on
the
side
here
to
angelos.
D
Yeah
I
mean
someone
mentioned
cinnaboy.
I
know
you
can
create
custom
plugins
for
that.
So
if
you,
if
you
need
to
do
that,
you
can
do
that
as
well.
Of
course,
pop
is
on
and
I
love
cystic.
So
if
he's
he's
got
something
there
trying
to
see
if
there's
anything
else,
what.
E
C
E
B
Yes,
I
posted
a
couple
of
links
in
the
hackamd.
I
have
not
yet
had
the
time
to
post
on
the
slack.
C
A
But
wally
does
mention
he
wants
benchmark,
not
just
like
a
security
thing
pop.
Can
you
give
us
a
tl,
dr
on
the
systick
thing,
while
you're,
while
you're
in
there
and
which
one's
this
one
chekhov?
Let's
look
at
this
one
see
rachel.
This
is
why
I
started
the
show.
This
is
how
I
learn
all
the
cool
tools
all
the
kids
are
using.
A
A
Okay,
that
doesn't
answer
the
question
yeah
still
interesting,
though,
let's
something
I
do
do
everyone
who's
listening
in
is
we
take
all
the
urls
and
I
whack
them
in
the
show
notes
on
the
forums
afterwards,
so
you
can
always
just
subscribe
to
the
notes
and
get
all
the
urls
and
stuff
that
we're
talking
about.
So,
if
you
see
him
on
the
side
there
in
slack
and
don't
worry
about
writing
him
down
and
stuff,
we'll
we'll
give
you
a
list
of
links
there.
A
Agustin
holgrive
says
I've
seen
good
things
in
argo,
for
example,
argo's
custom
rollout,
which
is
the
deployment
but
nicer
the
fact
that
this
exists
could
mean
vanilla.
Kate's
deployment
may
be
getting
a
little
bit
incomplete.
Do
you
know
if
there
are
any
plans
on
improving
the
oh?
This
is
a
separate
question.
Sorry.
F
F
A
Pop
says
pretty
much:
the
systig
tool
allows
benchmark
customizations
and
provide
a
metric.
If
you
want
to
populate
this
data
to
something
like
rafana
cool
cool.
That's
definitely
interesting.
A
And
not
really
a
question,
but
constantinos
is
asking
hey
kos:
do
people
use
kubernetes
in
container
d
on
xfs,
butterfest
and
zfs?
How
is
the
support
there?
So
those
of
you
using
different
file
systems,
let
us
know
let
constantinis
know
in
in
slack
there.
Okay,
anything
anything
else
when
it
comes
to
benchmarking.
Here.
A
F
A
This
is
contributing
to
kubernetes.
Documentation
is
very
complicated.
Is
there
video
tutorial,
create
issues
create
pr's?
So
this
is
an
area
I
contribute
to
regularly.
I
don't
know
if
there
are
any
video
resources.
However,
I
use.
B
B
D
F
A
While
he
says
there's
a
kubecon
eu
2020
talk,
if
someone
can
find
that
that'd
be
great,
I
know
they're
on
youtube,
now
youtube
cncf.
I'm.
A
Let's
see
yeah
that'd
be
great
thanks
rachel,
so
looking
at
the
contributed,
kubernetes
docs
there's
a
guide
there.
The
I
just
fork
the
repository
and
create
a
branch
because
it's
in
markdown
and
then
there's
a
readme
in
there
for
building
it
locally.
With
hugo.
B
A
B
A
Go
there
yeah,
I
would
say,
if
you're
serious,
about
getting
like
if
you're
just
trying
to
get
one
thing
fixed
and
you
you
know,
you
want
to
do
an
one-off,
do
the
contributor
guide
and
do
that,
but
if
you're
interested
on
a
regular
basis
seriously
go
to
a
sick
docs
meeting,
it
was
the
first
sig
one
of
the
first
things
I
was
involved
in
they're
very
friendly
and
they
will
absolutely
help
you
get
a
workflow
down
for
contributing,
but
yeah
theoretically,
like
you,
can
hit
like
if
you
go
to
a
kubernetes,
docs
page
on
the
right,
it
says
edit
this
page,
you
can
click
on
that
right
and
it
will
take
you
to
hey
chris
is
here
and
it
will
take
you
to
a
it
will
take
you
right
to
github
and
you
can
edit
it.
A
The
problem
is,
is
if
you're
not
familiar
with
github
and
stuff
already,
that
might
be
too
complex
for
you.
You
know
right
off
the
bat
kind
of
thing,
but
I
would
definitely
get
involved
there.
If
not,
if
you
can
do,
can
one
of
you
record
on
your
next
contribution,
the
issue
template
is
huge.
You
know
that
is
a
good
idea.
A
Let
me
I'm
gonna
todoist
that
does
and
I'll
go
ahead
and
pass
that
along
to
someone
just
I've
I've
I've
debated
just
doing
things
like
just
recording
people,
fixing
issues
and
stuff
in
github,
you
know
or
like
on
twitch
or
something
record
contributing
a
fix
to
the
docs.
A
D
A
Right
and
yogi
says
I've
done
prs
for
other
projects
and
I'm
not
okay
with
github
yeah.
The
issue
I
found
with
contributing
to
docs
is
you
got
to
run
hugo
right
because
you
can
generate
the
page
to
make
sure
that
it
renders
right
and
then
sometimes
like
the
first
time
I
did
it.
My
version
of
hugo
was
different
on
my
pc
than
what
the
project
was
using,
so
it
rendered
different.
A
So
then,
there's
like
a
little
docker
thing
that
you
run
in
there
to
ensure
that
you're
running
the
same
hugo
that
everyone
does
and
then,
when
you
do
a
pr
we
have
netlify.
It
generates
an
entire
preview
for
you
and
it
looks
like
totally
sweet
and
that's
all
great,
but
you
wouldn't
know
that
if
you
weren't
contributing
already
so
I
I
hear
you
I
hear
you
we
could
probably
definitely
but.
B
Generally
speaking,
I
think
this
docs
meeting,
if
it's
any
like,
I
have
never
attended
it,
but
if
it's
anyone
like
similar
to
the
second
rubrics
meeting
like
you,
will
find
people
there
that
want.
A
A
You
know
anyone
in
the
community
can
submit
a
blog
to
that
kubernetes
website
and
blog
on
the
official
thing
you
just
got
to
do
a
pr
and
you
do
it
in
text
using
hugo,
it's
pretty
dope.
So
if
you
fancy
yourself
writing
a
blog
post
for
kubernetes,
that's
how
that
works
all
right.
So
we
covered
security
thanks
yogi
for
the
question.
We
hope
that
gets.
You
started
a
gustin
hoolgrave
hope
I
got
that
right.
Welcome
to
the
show
we
got
about
10
minutes
left
says
I've
seen
good.
Oh
chris
carty.
H
Oh
yeah,
so
hi,
I'm
chris,
I'm
coming
broadcasting
from
ottawa
ontario,
I'm
currently
a
customer
engineer
with
google
cloud
canada.
Before
that
I
was
running
on
prem
kubernetes
for
a
municipality
up
here.
So
it's
kind
of
my
domain
of
expertise.
A
Yes,
welcome
to
the
show
augusta
says:
I've
seen
good
things
in
argo,
for
example,
argos
custom
rollout,
which
is
a
deployment
but
nicer
the
fact
that
this
exists
wait
hold
on.
I
think
my
text
got
cut
off.
Oh,
okay,
the
fact
that
this
exists,
I
could
mean
that
vanilla,
kubernetes
deployment
getting
a
little
bit
incomplete.
Do
you
know
if
there
are
any
plans
for
improving
the
deployment
resource
kind
of
unclear
on
this
one.
B
D
B
F
B
B
Right
so
I
mean
you
can
do
the
same
thing
still
with
like
nginx
ingress
and
normal
deployments.
It's
just
like
different
in
a
way
where
the
logic
is
handled.
So
I
would
think
that
I'm
not
sure
if
there's
any
advancements
planned,
but
I
would
say.
H
Yeah,
I
would
kind
of
lean
towards
that,
because
yeah
you
have
tools
like
that.
You
have
flagger
things
like
istio
linker
d
that
can
control
something
like
some
parts
of
that
aspect,
so
I
think
they're
leaving
it
as
the
building
block,
but
would
this
sig
be
for
that
api
machinery
to
check
there,
so
they
might
have
more
insight.
F
A
B
A
Okay,
we'll
give
you
a
second
there
any
other
questions,
while
while
pierre
is
digging.
A
H
H
H
Is
really
cool?
Okay,
so
augustine
just
provided
the
rollouts
there.
I
haven't
heard
of
anything
any
plans
to
change
vanilla
deployments
to
that
level
of
detail,
but
that
being
said,
someone
out
there
might
have
that
under
underway.
B
Yeah
and
looking
at
the
issues
that
are
related
to
deployment,
it
always
seems
to
be
api
machinery,
so
chris
was
right.
So
at
least
that's
what.
I
guess
right
now
like
considering
all
these
different
resources
here
like.
B
If
you
want
to
modify
your
rollout
strategy,
you
usually
have
to
at
least
somehow
change
the
way
you
serve
your
traffic
or
get
your
traffic
so
like
it's,
in
my
opinion,
a
little
bit
out
of
scope
for
receipt
deployment,
spec,
and
but
this
is
my
opinion,
like
I'm,
not
contributing
actively
to
that
part
of
human
leaders.
So
I
don't
know.
Maybe
they
have
plans
for
that,
but
yeah.
A
Sorry
our
answer
there
augustine
is
in
his
detailed
ours,
deep
as
you
probably
like,
but
I
don't
I
don't
think,
there's
any
major
changes
there.
B
I
can
quickly
go
to
the
proposals.
A
A
A
Yeah,
it's
okay!
If
there
was
any
plan
to
be
more
visible,
I
guess
that
answers
my
question.
Yeah
would
have
you
all
seen
the
new
the
new
feature,
the
deprecated
api
warning
thing
that
it
gives
you
now
that
was
on
the
blog,
that's
where
I'm
going
to
start
buying?
Yes!
So
if
you
oh,
I
definitely
want
to
get
this
out
there,
because
this
is
important
so
so
from
now.
Why.
A
H
A
B
A
I've
I've
rolled
I've
rolled
the
dice
and
we've
figured
out
the
t-shirts,
angelos
mimi's
and
us
augustine.
You've
won
the
kubernetes
t-shirt
because
we've
brought
your
question
on
the
air.
So
right
after
this,
I
will
pm
you
and
let
you
know
how
that
works.
I'll,
give
you
a
code
to
the
cncf
store
and
then
you
can
go,
get
a
kubernetes
shirt
which
we
always
forget
to
wear
on
the
actual
show.
But
it
looks
like
this,
except
for
the
kubernetes
logo.
A
A
While
he
brings
a
good
point,
maybe
because
it's
still
alpha
since
1.6
yep
today
on
the
podcast
drop
a
link
to
the
podcast,
because
I
think
you
all
should
be
listening
to
this.
If
you
like,
kubernetes
content,
you
definitely
want
to
check
this
out.
Oh
darren,
shepard
from
rancher
slash.
So
do
they
do
we
still
call
them
ranger?
Do
we
just
call
them
all
souza?
I
don't.
I
don't
know
what
we
call
them
so
follow.
A
B
Actually,
I
have
one
more
question.
I
think
I
really
read
something
about
kubernetes
now,
having
a
like
release
cycle
for
alpha
and
beta
apis
that
they
will
get
like
get
removed,
get
like
if
they
don't
graduate
to
fully
in,
like
four
release
seconds
or
something.
Oh.
A
B
I
read
about
it
somewhere.
It
might
have
been
on
moving
forward
from
better
with
this
blog
post.
This
is
actually
super
interesting
and
it
kind
of
puts
a
little
bit
of
pressure
on
the
kubernetes
yeah
six
like
yeah.
B
F
A
B
B
H
A
H
A
Went
over
this
at
tgik,
I
remember
going
through
all
the
release
notes
back
trying
to
find
out
when
ingress
was
started,
and
I
went
all
the
way
back
to
1.4.
I
just
forgot,
but
yeah,
that's
interesting,
all
right!
So
with
that
I've
been
talking
to
the
west
coast
panels,
there's
definitely
going
to
be
a
show
for
the
west
coast.
Americans
sometime
this
afternoon,
we're
gonna
shoot
for
well,
I
haven't
gotten
confirmation
yet
so
we're
gonna
shoot
for
5
p.m.
Pacific
time
is
going
to
what
we're
going
to
shoot
for
we
will.
A
I
will
sticky
that
in
the
office
hours
channel
you're,
more
than
welcome
to
hang
out
in
the
office
hours
feel
free
to
ask
questions.
So
if
you
get
stuck
throughout
the
month
and
you
post
questions
there
when
we
get
to
the
next
show,
we
just
go
through
them
again,
and
things
like
that.
So
with
that
thanks,
everybody,
augustin
and
angelo
stand
by,
I
will
pm
you
after
the
show
and
we
are
going
off
the
air
now
panel,
any
last
thing
to
say
after
this
after
I
say,
goodbye,
stick
around
and
we'll
wrap
it
up.
B
I
have
one
more
thing
to
say
sure:
go
if
you
ever
want
to
like
sit
on
this
panel.
We
are
actively
looking
for
people
to.
D
B
To
organize
this
session
and
like
contribute
back,
you
don't
need
to
contribute
to
actively
to
any
sick
or
something
if
you
have
knowledge
about
kubernetes
and
feel
like
you
can
answer
any
of
these
questions,
you're
more
than
welcome
to
hang
out
every
third
wednesday.
A
Yeah,
so
the
way
it
works
is,
I
have
a.
I
have
a
panelist
there's
about
10
of
them
and
a
few
days
before
I
said,
hey
who's,
going
to
show
up
for
this
one
and
as
long
as
we
have
enough
to
run
the
show
they
show
up
so
the
commitment's,
not
even
like
you
know,
I
say
it's
an
hour
and
a
month,
but
if
you
don't,
if
you
can't
show
up
that
thing,
it's
we
have
enough
people
where
we,
you
know
we
can
run
a
panel.
So
it's
an
excellent
way
to
go.
A
I
think
chris
started
off
with
office
hours
on
the
project
and
now
he's
been
speaking
at
kubecon
works
at
google.
Now,
look
at
that
all
started
because
of
office
hours.
A
Yeah
yeah,
so
like
don't
you
know
and
like
I
said
I,
I
started
the
show
to
learn
kubernetes,
I
don't
know
a
lot
about
kubernetes
still
to
this
day,
so
you
know
I
figure
it's
just
a
fun
place
to
participate
and
we
can
all
learn
from
each
other
and
have
a
good
time.
So
with
that
congratulations,
t-shirts
owners.
Thanks
panel.
We
are
hang
out
for
a
second
panel
and
for
everyone
else.
We
will
see
everyone
either
later
today
or
next
month.
Thank
you.
Everyone
wait.
Everyone
ate
all.