►
From YouTube: Day Two Kubernetes: Tools for Operability - Bridget Kromhout & Zachary Deptawa, Microsoft
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Day Two Kubernetes: Tools for Operability - Bridget Kromhout & Zachary Deptawa, Microsoft
Artisanally hand-crafting our own container hosting solutions can be a fun learning experience, but for repeatable production use, we want to deploy and manage Kubernetes clusters in a reproducible fashion. Using open source tools like Helm, Draft, Brigade, and Terraform, we can deploy and update our Kubernetes clusters via a trusted, versioned, repeatable process. We’ll discuss what containers and Kubernetes clusters are at a high level, look into the practical application of open source tools to simplify cluster management, and show you how to deploy Kubernetes clusters in a repeatable and portable fashion.
To Learn More: https://sched.co/GrY7
A
A
So
if
you
find
yourself
sitting
somewhere
where
you
can't
see
a
screen,
probably
move,
if
you
want
to
watch
zach-
and
I
laugh
and
talk
to
each
other-
and
you
cannot
see
our
faces
right
now-
probably
move
will
be
in
approximately
these
locations
and
if
the
obstructed
view
is
causing
a
problem
for
you,
now
is
the
time
to
fix
that
also
I
sort
of
imagined
since
we're
already
sort
of
filling
up-
and
this
is
difficult
to
find
location
far
away
from
where
all
the
snacks
were.
More
people
will
probably
filter
in.
A
If
you
find
yourself
doing
the
traditional
nerd
thing
of
one
empty
seat
between
every
person,
if
you
feel
the
need
to
move
a
little
in
towards
the
sides
that
might
make
it
easier
for
people
who
come
after
you,
particularly
if
you're
sitting
in
the
back
and
if
you're
saying
to
yourself
but
I,
need
a
lot
of
room.
That
is
ok,
but
maybe
right
now
in
the
minute
before
we
start
move
a
little
further
forward.
A
A
A
A
Basically,
hi
will
enter
ourselves
in
a
minute
but
I'm
just
gonna.
Let
you
know
right
now.
If
you
were
not
looking
for
kubernetes,
you
are
in
the
wrong
place,
possibly
at
the
wrong
conference.
But
yes,
you,
probably
if
you're
interested
in
day
two.
If
you
want
to
have
arguments
about
what
date
one
versus
day
two
means
we
can
do
that
at
the
booth
later.
But
if
you're
interested
in
day
two
kubernetes,
you
are
in
the
right
place,
we're
gonna
we're
gonna
talk
about
awesome,
you're,
actually
running
it.
Now.
A
What
okay
and
let's,
let's
intro
our
speakers,
real,
quick,
I'm,
Bridget
Crum?
How,
if
I,
don't
look
familiar
but
I
sound
familiar,
you
may
have
listed,
listen
to
the
arrested
DevOps
podcast
I
live
in
Minneapolis
I
work
for
Microsoft
and
I.
Am
the
chief
cook
and
bottle
washer
of
the
international
conspiracy
known
as
DevOps
days.
The
one
in
Seattle
is
gonna,
be
I.
Believe
it's
April,
23rd
and
24th,
and
their
CFP
closes
January
6.
A
B
I'm
Zack
I
live
in
South
Texas
I
work
for
Microsoft,
as
a
cloud
advocate
and
I
focused
pretty
heavily
on
DevOps
and
partner,
really
closely
with
tashi
Corp
and
that's
why
you
see
the
Hashi
Corp
logo.
There
and
I
also
raised
these
cute
little
goats
in
South,
Texas
and
I.
Seen
for
several
bands
down
there.
So.
B
A
Got
to
find
Michael
tasty
and
have
a
goats
in
the
silo
discussion,
all
right,
that's
deep
cut
from
old
stalks
all
right!
So,
let's,
let's
start
with
the
stage
setting
like
what
even
is
well.
If
you
want
to
know
what
even
is
kubernetes
you're,
definitely
in
the
wrong
talk,
we're
not
gonna,
explain
that,
but
let's
kind
of
scaffold
to
where
we
are
and
why
we're
talking
about
the
stuff
we're
talking
about
starting
with
clout.
What.
A
A
Hopefully
ours,
but
the
important
thing
there
is
you're
picking
the
degree
of
abstraction,
that's
right
for
the
workloads
and
the
workflow
patterns
that
you're
using,
which
is
to
say
when
somebody
tells
you
that
your
kubernetes
is
wrong
because
you're
not
doing
it
with,
like
you
know,
a
managed
service
in
a
public
cloud.
I
mean
you
know
what
it's
actually
okay,
to
get
to
your
level
of
abstraction.
That's
right
for
your
organization's
needs.
Okay.
What
about
what
about
containers?
What
is
a
container?
Yes.
B
A
Pretty
sure
that
if
we
announced
an
exciting
C
group
startup,
we
would
not
get
nearly
as
much
funding
as
people
do
for
container
initiatives.
So,
okay
and
finally
kubernetes
I
think
the
biggest
shocker
for
me,
especially
if
you're
new
to
the
kubernetes
space,
is
to
look
at
this
and
think
wait
a
minute,
2014
really
but
yeah.
The
project's
actually
been
around
for
a
little
while
yeah.
A
Day
and
I
think
that
the
valuable
thing
there
is,
if
you
are
feeling
like
the
tech
FOMO
and
they
I'm
already
behind
on
kubernetes
the
date
on
the
slide,
is
giving
me
anxiety.
Don't
worry
about
it,
because
it's
a
work
in
progress
and
we're.
Finally,
getting
to
the
point
where
you
know
actual
responsible
adults
would
probably
recommend
to
their
professional
colleagues.
A
Yes,
we
can
in
fact
do
this
in
production,
so
has
Jake,
probably
fine,
okay,
so
I
use
this
slide
a
lot
and
the
main
point
that
I
try
to
make
here
is
people
get
really
excited
about
the
shiniest
new
tech
that
they
think
will
definitely
make
everything
easier?
That
is
not
true.
Sorry,
that
is
a
lie.
You
know
yeah.
B
A
A
boss,
mind:
Tim
girl,
Suzette
instacart,
now
likes
to
call
it
conservation
of
complexity,
you're
going
to
have
complexity,
whether
it's
IPC
or
a
network
latency
like
you're,
going
to
have
it.
So,
if
you're
trying
to
decide
if
kubernetes
is
right
for
you,
we're
gonna
go
through
a
bunch
of
cool
open
source
projects.
That
kind
of
help
with
that.
But
if
you're
deciding
whether
or
not
it's
right
for
you,
it
might
not
be,
and
that
is
okay.
You
do
not
have
to
kubernetes
everything
I.
A
Should
if
that
will
solve
problems,
you
actually
have
like
wanting
to
be
able
to
do
independent
deployments
at
different
times
of
those
pieces
but
yeah?
Okay,
so
what
you
all
came
for
tools,
yay
tools
in
the
kubernetes
ecosystem,
so
there
there
are
a
lot
of
tools
here.
You're
you
have
an
entire
three
days
of
conference.
There's
no
way
we're
gonna
cover
all
of
them
in
this
session,
so
we're
basically
just
cherry
picking
some
yeah.
B
A
A
A
A
And
I
think
that's
one
of
the
things
that
really
saves
people
a
lot
of
time
right
and
I.
Think
that's
why
a
lot
of
you
are
probably
using
terraform
is.
If
you
are
multi
cloud
sure
you
write
a
whole
bunch
of
Confirmation
and
you're
like
that,
was
great
and
then
you're
like.
Oh
now,
I
need
to
write
arm
templates,
that's
even
more
and
it's
nice
to
not
have
to
to
have
another
layer
of
abstraction
that
saves
you
from
having
to
do
that.
Right.
B
Yeah,
it
makes
it
very
simple:
it's
it's
terraform
plan
terraform
apply,
that's
how
you
use
it
and
also
I'll
mention
the
module
registry.
The
registry,
that's
our
form
that
io
and
that's
basically,
essentially,
packages
of
resources
that
are
put
together
in
a
meaningful
way.
Think
of
it
like
a
function
with
inputs
and
outputs.
It's
reusable
code.
Instead
of
pasting
over
and
over
all
these
resources,
you
can
use
modules.
It's
also
a
good
way
to
get
started.
If
you
haven't
played
with
terraformable.
A
Oh
and
one
of
the
things
we
should
talk
about
that,
we
that
we
discussed
that
we
were
talk
about
is
that
it
might
surprise
people
if
they're
gonna,
use
terraform
with
azure
say
with
aks.
Usually
you
think
of
terraform
is
just
something
that
used
to
launch
infrastructure,
but
what's
the
but
wait
surprise,
but.
B
Wait,
there's
more
so
terraform
can
actually
speak
to
anything
with
an
api,
so
any
of
the
providers
you
can
use.
If
there
is
not
a
provider
for
something
that
you
have
say
it's
proprietary.
You
can
actually
write
your
own
provider,
which
is
super
awesome,
but
you
can
deploy
straight
to
kubernetes,
so
I
can
and
I'll
show
you
in
a
second
you
can
deploy
in
a
KS
cluster
on
Azure
and
then
can
deploy
into
that
kubernetes
cluster,
using
a
kubernetes
provider
in
the
same
codebase
right.
A
Exactly
so,
we've
mentioned
aks
a
couple
of
times,
as
I
mentioned
before
we
work
for
Microsoft,
and
so
one
thing
that
we
like
to
show
people
is
that
of
all
the
acronym
soup
out
there
and
there's
you
know,
there's
your
GK
ease
and
your
EK
SS
and
whatnot
aks
is
the
managed
kubernetes
service
from
asher.
Yes,.
B
A
There's
there's
a
couple
of
things
about
this
that
you
know
again
like
well,
we'll
publish
the
slides
later,
so
you
don't
need
to
take
pictures
of
every
command
and
then
try
to
type
it
in
laborious
ly.
But
it's
valuable
to
think
about
how.
Of
course,
you
want
to
be
able
to
instantiate
clusters
like,
of
course
you
need
to
do
that
right
and.
B
This
is
actually
just
showing
the
AZ
CLI
tool.
Just
a
za
KS
create
a
za
KS
install
CLI.
If
you
don't
already
have
coop
CTL
installed,
it
will
install
it
for
you,
a
Z,
a
KS
get
credentials
will
actually
pull
the
credentials
and
set
your
context
for
your
kubernetes
cluster,
and
then
you
can
use
coop
CTL
and
it's
literally
those
commands.
So.
A
If
you've
been
to
one
of
the
workshops
where
you
do
everything
from
like,
you
know,
Kelsey
Hightower's
kubernetes
the
hard
way,
this
is
obviously
faster,
also
I.
Think
one
of
the
other
things
that
we
don't
always
think
about
is
cool.
You
stood
up
a
kubernetes
cluster.
That's
awesome!
If
you're
going
to,
actually
you
know
use
it,
you're
gonna
need
to
upgrade
it
and
you
might
not
want
that
to
be
a
terrible
disaster.
Yeah.
B
For
sure
so,
there's
tools
like
the
of
course
you
can
list
your
clusters,
that's
pretty
straightforward,
but
also
AZ
aks
upgrade,
which
will
upgrade
your
cluster
in
place
to
the
version
that
you
specify
and
a
really
cool
thing.
It
looks
like
a
harmless,
simple
command
and
it
is
a
a
short
command
here
to
type,
but
it's
doing
some
magic
on
the
back
and
it's
following
best
practices.
It
will
actually
upgrade
the
control
plane
it'll,
make
sure
that
the
nodes
are
then
unschedulable
and
pulled
out
of
rotation.
B
B
Let's
see
if
I
and
this
so
this
is
just
gonna,
be
a
quick
demo
of
terraform
and
you
can
see.
I
have
a
Cates
demo
environment,
it's
empty
right
now,
and
what
I'm
going
to
do
is
show
really
quickly.
This
is
vs
code,
that's
just
as
very
straightforward,
terraform,
environment.
It's
got
a
main
TF
file
where
I'm
setting
the
provider
I'm
setting
the
back
end
for
terraform
I've
also
got
a
variable,
z,
TF
file,
I'm,
setting
some
variables
that
I'll
be
using.
You
can
see
I'm
setting
a
tag
there
as
well
on
the
outputs.
B
These
are
meaningful
outputs
for
me.
You
can
see
that
I
actually
set
a
coop,
config
output
right
there
that
I'll
be
using
later
and
I
know.
This
is
going
a
little
fast.
So
if
there's
anything
I
want
to
see
you
later
at
the
booth
just
come
and
let
me
know
I
mean
this
Kate's
TF
file.
This
is
actually
setting
up
an
a
KS
cluster.
It's
everything
that
you
need
to
set
up
the
AKS
cluster
and
and.
B
B
Now
you
can
see
the
provider
kubernetes
right
there
that
I
just
kind
of
what
I
talked
about
earlier.
This
Kate
spod
TF
is
actually
going
to
deploy
an
engine,
X
pod
and
an
engine
X
service,
and
it's
just
a
sample
service.
You
can
see
it's
using
engine
x1
7-9,
if
I'm
not
mistaken,
and
port
80
here
as
well,
and
so
all
I'm
going
to
do-
and
this
is
vs
code-
well
actually
really
quickly.
B
This
is
a
sample
environment
variable
file
and
there
are
sensitive
things
that
you
do
not
want
to
store
in
your
codebase
ever
like
your
client,
ID
and
client
secret
and
subscription
ID,
so
I
actually
export
that,
and
that's
just
an
example
of
that
file.
So
here
I
am
at
the
terminal.
I'm
just
gonna
run
a
terraform
version
to
make
sure
terraform
is
installed
and
see
the
version
output
really
quick
and
at
terraform
an
it
will
actually
inspect
all
those
terraform
files
and
look
what
you
know:
pull
what
plugins
I
need
what
providers
I
need.
B
Well,
so
terraform
plan
is
basically
inspecting
all
of
those
files
figuring
out
what
the
state
is
now
and
where
the
state
wants
to
be,
and
it's
figuring
out
a
dependency
graph
on
how
to
get
there.
The
really
cool
thing
about
the
plan
is
that
it
is
going
to
see
what
it
can
build
in
parallel
and
it's
also
going
to
see
what
depends
on
what
resource
depends
on
another
resource,
and
it
will
do
that
all
intelligently,
you
can
see.
B
I
ran
a
terraform
apply,
and
now
you
can
see
in
the
control
panel
that
things
are
starting
to
populate.
My
demo.
Cluster
is
creating
and
through
the
magic
of
video.
This
won't
take
the
normal
sixish,
seven
ish
minutes,
but
it
will
finish-
and
you
can
see
that
I've
got
my
terraform
output
here
in
green
and
you
can
see
six
were
added
six
resources.
B
These
are
all
the
outputs
that
I
can
figured
I
know
this
is
going
a
little
fast,
but
what
I'm
going
to
do
is
take
this
output
that
I
have
configured
the
terraform
output,
cube,
config
and
basically
I'm
going
to
export
my
kubernetes
credentials
and
I'm
going
to
use
coop
CTL,
just
the
standard
cube,
CTL
tooling,
to
interact
with
this
cluster
right
away,
and
now
you
can
see
I'm
checking
my
cluster
info
to
make
sure
I'm.
Looking
at
the
right
cluster
I
can
do
get
pods.
You
can
see
my
engine
next
demo
pod
get
services.
B
A
A
Cool,
so
that
is
an
example
of
using
terraform
with
aks
in
this
specific
case,
but
again,
Terra
forms
open
source
and
you
can
use
it
with
your
cloud
provider
of
choice,
but
we
also
want
to
see
a
bunch
more
exciting,
open
source
by
the
way,
spoiler
alert.
The
the
secret
agenda
to
this
particular
session
is
to
get
you
excited
about
and
hopefully
contributing
to
and
using
all
of
and
make
the
opening
issues
in
all
of
these
open
source
projects.
Yes,
so
starting
with
helm.
What,
in
fact
is
home,
helm.
B
A
Think
that
one
place
where
this
is
key
is
because
we're
all
we
none
of
us
when
we're
doing
anything
in
production
again
we're
talking
about
the
day
two
realities
of
operability.
None
of
us
have
like
one
set
of
charts
and
one
version
of
all
the
things
and
one
cluster
that
it
runs
on
our
laptop
like
that.
A
B
A
A
So
many
issues
with
that
term
we'll
talk
about
it.
Okay,
so
I'd
say
starting
at
least
with
the
with
the
complexity.
We
have
an
awful
lot
of
llamo
when
we
talk
to
people
about
kubernetes
and
we
talk
about
how
I
feel
like
we
need
to
have
a
Content
warning
for
gamal
like
I'm.
So
sorry,
if
Yama
love
sets
you
but
there's
going
to
be
lots
of
it
and
you
want
to
keep
track
of
it
all
and
that
like
helm
and
charts,
are
a
really
good
way
to
do
that.
Yeah.
A
A
I
think
sometimes,
when
I'm
doing
a
workshop,
it's
like
and
we're
gonna
install
prometheus,
we're
not
going
to
take
all
the
time
to
like
read
all
the
things
and
type
all
the
things
we're
just
gonna
use
the
Prometheus
chart
and
hey
installation
right
on
our
kubernetes
cluster.
Okay.
So
let's,
let's
talk
about
the
possibly
semi
controversial
thing
about
roll
backs.
Roll
backs
are
damned
dirty
lie
because
a
share
can't
turn
back
time.
And
neither
can
you
so
like
what
you
can
do
is
you
can
hopefully
roll
forward?
Yes,.
A
B
Let's
take
a
look
at
how
I'm
really
quickly
so
this
is
just
gonna,
be
another
quick
demo
here
of
helm
and
you'll,
see
if
I've
started
it
did
I
start
it.
Oh
there
we
go
yeah,
so
you
can
see
that
I'm
just
doing
a
witch
helm
and
by
the
way
I've
installed
helm
and
drafting
these
demos
using
homebrew,
so
I'm
doing
coop,
cuddle,
config,
current
context
and
cluster
info
just
to
make
sure
I'm
working
on
the
right.
The
right
cluster
before
I
move
forward
Helmand.
B
B
I
should
have
highlighted
it,
but
that's
the
tiller
for
helm,
which
we
can
talk
about
later,
as
well,
because
of
helm
version,
3
I'm,
doing
a
hell,
Murray
Pro
update
here
to
update
the
helm,
repository
and
I'm,
just
gonna
search
for
MySQL
to
install
a
simple
helm,
chart
and
I
picked
percona
MySQL,
because
I've
used
it
a
ton.
In
my
past
days,
at
Rackspace
and
yeah,
you
can
see
it
built
here
via
that
how
I'm
trying
to
deployed-
and
it
also
spit
out
some
notes
for
us
as
well-
that
are
helpful
from
this
helmet
right.
B
So
I'm
gonna
do
get
pods.
I
can
see
that
it's
not
ready.
There's
your
old
one
ready
get
services.
You
can
see
my
dandy
noot
is
already
there,
but
it's
it's
not
ready
yet
so
I
believe.
If
I
recall,
yeah
I
get
pods
again
there.
It
is
it's
ready
and
now
that
that's
ready,
I'm
gonna
use
those
helpful
notes
and
hints
and
I'm
gonna
get
my
root
password,
which
is
super
not
safe,
but
I'm
going
to
use
this
just
for
demo
purposes
and
I'm
going
to
actually
create
a
pro
kkona
client
inside
my
cluster.
B
B
So
there
you
go,
you
can
see
the
MySQL
databases,
I
log
out
I'm,
going
to
delete
that
Percona.
Client
I,
don't
need
it
anymore.
Goodbye,
percona,
client
and
now
I
can
actually
use
helm
to
list
that
deployment
and
I
can
use
home
to
actually
delete
it
from
my
kubernetes
cluster.
So
it's
it's
full
scope
management
there
and
you
can
see
that
it's
gone
now.
A
Awesome
alright,
so
that
we've
already
now
we
looked
at
home-
that's
cool!
Let's!
Let's
talk
about
draft
a
little
bit,
because
if
people
want
to
iterate
quickly
like
that
that
inner
loop
of
development,
it
can
seem
like
a
lot
of
work
to
set
up
a
whole
chart-
and
you
know,
do
you
really
want
to
sit
there
and
learn
everything
about
home
and
write
docker
files
so
enter
draft
yeah.
B
A
B
A
B
B
Let's
look
at
it
really
quickly.
This
is
just
a
draft
demo,
so
it
needs
helm
so
I'm
just
doing
a
home
version
to
make
sure
that
it's
they're
drafting
it
will
install
draft
for
me
and
draft
version
just
to
show
which
version
I'm
on
make
sure
it's
there,
coop
cuddle
I'm
again
I'm
checking
my
cluster
because
I
always
want
to
double-check
which
cluster
I'm
working
with
before
I
work
with
it,
you
can
see,
I've
got
an
app
dot,
py
file
and
a
requirement
text.
B
That's
all
I
have
it
is
a
straightforward
flask
app
draft
create
detected
that
it's
got.
Python
says
it's
ready
to
say
all
drops
me.
Docker
files
drops
me
a
draft
AMA
file
and
these
are
all
sort
of
best
practices
or
what
draft
thinks
you
might
want
to
might
want
to
use.
So
you
can
go
in
and
edit
those
if
you
want
to
and
from
here
I've
actually
already
configured
ACR
the
container
registry
on
azor
and
so
draft
up
is
building
container.
B
It's
deploying
it
pushing
it
to
the
registry
and
actually
deploying
it
into
my
cluster
and
what's
really
cool
and
we'll
see
it
in
just
a
second
now
that
it's
there
I
can
do
get
pods
I
can
see
my
example
Python
app
right
there
and
I
can
do
draft
Connect
and
what
that's
doing
is
creating
a
proxy
or
a
tunnel
to
where
I
can
hit
a
localhost
port
and
actually
get
response
from
that
container.
Immediately
so
say
that
I
want
to
make
a
change
to
that
and
see
that
change
quickly,
reflected.
B
Let's
make
a
change
I'm
going
to
change
that
to
hello,
cube
con
and
once
I
do
that
all
I
have
to
do
is
a
run
draft
up
and
it's
going
to
push
those
changes.
Push
the
changes
to
that
container
redeploy
it
I
can
do
draft
Connect
right
away
again
after
I
see
you
can
see
the
old
pod
terminating
the
new
one
running
draft
connect
again
and
then
I'm
going
to
curl
the
new
localhost
port,
and
you
can
see
that
I
have
hello,
coop
con
and
so
draft
will
also
allow
you
to.
B
A
Excellent
okay,
so
helm
draft
these
by
the
way
all
I
have
websites,
helm,
dot,
SH,
draft
at
SH,
etc
will
have
links,
I
was
about
to
say,
show
notes,
but
not
podcasting
will
have
links
in
the
last
slide.
Okay,
let's,
let's
talk
about
Brigade
a
little
bit
and
we
don't
that
we're
done
with
the
demos
now.
So,
if
you're,
like
demo,
heavy
we're
done
with
the
demos
now,
but
we're
gonna
just
show
you
a
couple
more
things:
Brigade
I
think
is
when
people
are
like.
A
B
A
If
you're
sick,
you're
writing
yeah
mo,
you
could
write
some
JavaScript,
okay
and
I
guess
also,
whenever
you're
introducing
tooling,
like
this
I
want
to
acknowledge,
there's
other
tooling
in
the
ecosystem,
but
it
bears
some
similarity,
definitely
parallel:
evolution
with
github
actions,
though
those
are
only
available
for
public
repos
right
now,
and
so
some
people
I
think
I've
already
started
putting
open-source
out
there
about
knitting
them
together
with
Brigade.
So
it's
probably
this
is
again.
A
B
Actually
used
an
item
in
to
mention
that
we
actually
use
helm
as
well
on
a
lot
of
our
a
lot
of
our
projects
at
Microsoft,
I
mean
yeah.
We
use
our
own
things
that
we
contribute
to
because
we
love
them
and
we
use
them
because
they're
useful,
not
because
they're
shiny,
make
sure
you're
using
these
things,
because
they're
useful
and
not
adding
complexity.
For
you
absolutely.
A
A
A
A
Even
if
you
don't
know
you
have
it,
which
is
ok,
you
have
all
the
yeah
mole
from
you
know
all
of
our
helm
stuff,
and
you
have
your
your
docker
files
and
your
images
and
probably
all
of
the
backing
service
definitions
that
don't
really
go
in
any
of
those
places,
and
you
want
to
version
those,
and
you
want
to
tell
if
those
are.
You
know,
reproducible
and
verifiable,
and
this
is
what
C
NAB
does
it's
a
I
mean
I'm
just
gonna,
look
at
the
notes
to
make
sure
that
I
don't
skip
anything
I.
A
Definitely
want
to
talk
about.
I
think
that
it's
valuable
in
terms
of
delivering
an
entire
application
like
the
definition
for
an
entire
applique-
and
this
was
actually
on,
like
I,
said
on
stage
at
dr.
Khan
and
Matt
butcher
and
Gareth
Rush
Grove,
who
were
announcing
it,
are
actually
both
here
at
the
conference.
So
you'll
be
able
to
talk
to
you
know,
butcher
at
the
Microsoft
booth
and
presumably
Gareth
will
be
at
the
dr.
booth,
and
you
know
you'll
be
able
to
discuss
this
with
them
in
detail.
I.
A
Yes,
Gareth
is
tomorrow
morning,
he's
definitely
speaking
tomorrow,
so
you
can
go
look
for
his
talk
so
anyway,
if
you're
thinking
this
is
great.
So
there's
a
spec.
What
am
I
going
to
do
with
a
spec?
You
can
actually
look
at
the
reference
implementation
of
the
Specht
that
we
released
called
duffle,
and
so
this
is
again
like
it's
double
dot,
Sh,
it's
all
pretty
new
stuff,
but
it's
all
about
packaging
and
also
unpackaging
you're
distributed
apps.
Ok,
what's
next
so.
B
A
B
A
I
think
it's
about
20.
We
can
talk
about
how
we
shouldn't
have
siloed
organizations.
We
should
have
all
this
tooling
be
used
better
to
create
cooperation
across
our
organization.
We
have
green
fields,
we
have
blue
skies.
We
have
the
slow-burning
tire
fire
that
is
proud.
My
twitter
handle
like
we
can
talk
about
things
in
all
these
abstractions
right,
but
I
think
that
it's
really
important
right
now
for
us
to
get
hands
on
actually
try
this
stuff,
Jerome
Pettis,
Oni,
developed,
container
training
I
encourage
all
of
you
to
go
and
run
through
and
try
it
yeah.
A
Absolutely
so
like,
if
your
I
would
say
when
you
go
when
you
go
to
use
kubernetes
in
production,
you
should
strongly
consider
a
managed
service
and
we
work
at
a
cloud.
We
would
love
it
to
be
ours,
but
it
doesn't
have
to
be
ours.
But
you,
when
you
start
building
the
stuff
out,
you'll
see
that
you're
gonna
spend
an
awful
lot
of
time
on
undifferentiated
heavy
lifting
and
incidental
complexity
that
you
don't
actually
need.
A
B
A
And
then,
if
you've
been
looking
at
virtual
Kubla,
this
was
announced.
It
was
worked
on
by
a
bunch
of
our
colleagues
and
announced
as
just
a
new
open
source
project
right
before
coop
con
this
time
last
year,
and
it
was,
you
know,
always
meant
to
be
kind
of
a
reference
implementation,
and
now
everything
is
very
exciting,
like
if
you
take
a
look
at
its
in
preview
now
right.
Yes,.
A
At
the
a
and
Brendan
burns
just
tweeted
about
this
I
can
blow
his
tweet
up,
so
you
can
see
it
a
little
bit
better.
It
is
happening.
The
virtual
nodes
are
available
in
preview
and
IKS
and
I.
Think
a
lot
of
other
providers
are
working
on
this
too.
This
is
basically
bringing
your
chocolate
and
peanut
butter
together,
bringing
your
server
list
to
your
kubernetes,
so
you
can
have
all
the
buzzword,
compliant
technology
that
you
could
possibly
want.
That's.
A
I
think
that
it's
when
we,
our
colleague
Eric
st.
Martin,
with
apologies
to
halt
and
catch
fire,
said
that
kubernetes
is
not
the
thing.
It's
what
gets
us
to
the
thing,
because
if
you
think
you're
gonna
go
back
to
your
organization
and
you're
gonna
say
excellent,
we
have
now
covered
all
the
Nettie's.
We
are
done
with
technology
we've.
We
have
a
you
know.
Capstone
project,
we're
done
like
no,
obviously
not
like
tech
is
gonna.
A
B
A
And
I
think
also
like
we're
both
standing
here
talking
to
you
from
Microsoft,
which,
if
you
haven't
heard
of
it
as
a
little
startup
across
the
you
know,
across
some
bridges
from
Seattle,
but
yeah
I
mean
seriously,
though
the
fact
that
Asscher
is
more
than
50%,
Linux
I
think
tells
us
something
about
where
the
industry
is
going.
I
think
it
tells
us
something
about
how,
when
you're
in
your
organization's
and
you're,
having
conversations
with
people
who
are
like
well,
we
have
the
way
we
do
things
yeah
I.
A
Like
if
you
have
you
know,
if
we
have
people
who
want
to
cling
to
the
past,
I
would
just
point
them
to
the
fact
that
Geoffrey
Snover
is
out
there
talking
about
a
sure
being
Linux.
So
maybe
it's
okay
to
change
and
it's
there's
not
also
a
place
where
you're
done
changing
and
now
you've
you
have
Devon
all
the
opsin.
You
have
continuously
improved
to
the
point
of
not
needing
to
improve
anymore.
That
doesn't
exist
either
right,
like
Microsoft.
A
A
So,
let's
moment
of
ridiculousness,
when
we
realized,
we
had
essentially
the
same
pin
tweet
and
it's
it's
hilarious,
because
we
we
both
are
really
excited
about
open
source
and
that's
why
we're
at
Microsoft,
which
is
amazing
so
and
one
of
the
one
of
the
things
that
I
think
is
so
amazing,
is
what
most
of
our
sessions
are
about.
Yeah.
B
A
You
know
time
slot
is
going
to
be
pretty
great
and
oh
Jessica
Dean
from
our
team
is
going
to
be
doing
some
stuff
about
all
the
things
we
just
talked
about,
but
with
Windows
because
hey
that
is
still
a
thing
that
exists
in
a
lot
of
people's
enterprises.
That's
all
today.
Tomorrow,
we
have-
let's
see
oh
there's
so
many
good
ones,
I
would
say
like
I
will
definitely
be
live-tweeting
like
at
least
half
of
these,
but
you
probably
want
to
make
it
to
him
yourself
too,
and
then
there's
even
more
on
Thursday.
A
B
A
And
that
there's
a
as
your
free
trial,
there's
also
credits
I
think
they
have
a
fifty
dollar
credit
at
the
Microsoft
booth.
So
you
can
go
and
get
all
your
selfies
with
you
know
fit
fit
be
and
when
absolutely
adorable,
we
didn't
even
talk
about
the
magic
heats
ball.
Oh
yeah,
it'll
answer
all
your
kubernetes
questions.