►
Description
Jenkins users, are you ready to containerize your continuous delivery pipeline? In this video Per Bergman, Sr. Architect at Zilker Technology reviews using Jenkins X. If you have been hearing about Jenkins X and considering a shift from traditional Jenkins, this meetup will help you start moving in the right direction
A
B
A
C
B
Okay,
so
I
have,
I
have
like
eight
seven,
eight
slides
first,
but
most
of
the
time
will
be
a
hands-on
demo.
So,
but
I
have
some
slides
just
to
introduce
some
things.
You
said
there
me
and
some
background
materials
so
see.
We
can
start
this.
C
C
B
Here
we
go:
let's
try
to
play
this,
how
we
can
do
it
like
this,
so
yeah
welcome.
So
I'm
working
at
a
company
called
technology,
we
do
general
consulting
and
devops
is
one
piece
and
this
presentation
here
will
have.
I
will
have
some
refreshing
refreshers
about
what
what's
the
historical
lineage
behind
all
of
this,
so
we
have
hardware
virtualization
of
operating
system,
virtualization
and
kubernetes,
which
is
orchestration,
and
then
we
have
devops.
B
On
top
of
this,
it's
like
four
layers
I'll
go
through
that
and
then
what
is
jx
and
then
I
will
finally
run
a
demo,
actually
build
an
app
and
build
and
deploy
this
app
using
the
google
kubernetes
engine.
But
this
would
work
as
a
jx.
I
call
it
jx
instead
of
jenkins
x,
so
jx
would
work
on
all
the
aks
eks
gke,
on-premise
rack
space
and
some
other
kubernetes
clusters.
I've
been
using
gke
because
that's
what
they
develop
and
test
first,
but
eks
worked
fine
as
well,
so
my
background,
so
my
first
computer.
B
I
wanted
to
show
that,
because
it's
kind
of
cool,
so
it
was
at
cx81
this
one
here,
it's
one
kilobyte
ram.
I
started
to
code
1982
when
you
have
1k,
like
my
my
mac
here,
has
64
gigabytes
of
ram.
This
computer
had
one
kilobyte
of
ram,
so
it's
kind
of
kind
of
astonishing
how
the
development
evolution
have
gone
from
1982,
maybe
so
that's
yeah
and
somehow
I
worked
I'm
from
sweden.
I
worked
in
in
sweden,
9192
until
2000
moved
to
the
u.s.
B
I
did
a
telecom
first
massively
disputed
system
to
manage
radio
and
subscribers
in
southern
china.
B
They
could
not
run
in
a
big
machine,
not
even
the
biggest
some,
so
we
had
to
dispute
it
in
many
small
machines,
and
this
was
98.
Over
20
years
ago
I
worked
at
amazon,
clickstream
processings,
all
the
clicks
for
all
the
web
servers
back
then
2000
web
servers
sending
log
straight
log
data
every
minute
I
did
work
for
ibm
for
a
while
business
process
management.
B
One
startup
company
with
psychometrics,
so
collecting
reviews,
user
reviews
and
product
reviews
from
ordinary
people,
not
experts.
So,
like
you
call
people
look
up
opinions
about
core
models
and
we
could
run
natural
language,
processing
and
machine
learning
on
that
to
get
the
bias.
What
kind
of
opinion
do
people
have
about
certain
products?
B
This
was
used
by
the
marketing
departments
of
these
companies
and
albuquerque.
Does
my
new
mexico
connection?
I
live
12
years
2006
to
18,
and
I
worked
the
last
two
years
at
expansive
in
albuquerque,
which
is
now
moved
away
to
seattle
and
there,
so
that
company
did
data
collection
from
oil
and
gas
and
other
producers,
and
we
did
a
pipeline
all
the
way
to
market.
So
it's
so
expensive
actually
owns
a
trading
system
where
you
can
buy,
sell
financial
instruments,
and
here
we
used
both
these
last
companies.
B
We
used
ecs
and
also
far
gate,
which
is
like
node-less
containers.
So
you
don't
have
to
pre-allocate
virtual
machines,
which
is
the
way
to
go
if
you
can
so
for
gate
without
pre-building
servers
and
now
forget
actually
works
with
eks,
as
well
as
the
big
disadvantage
with
forget
was
always
that
you
cannot
connect
to
nfs,
because
you
have
big
models
and
big
data
sitting
on
disk,
and
you
cannot
connect
from
far
gate
instance
to
like
nfs
or
efs,
but
you
can
now
so
now.
B
B
So
that's
a
big
we
use
terraform
ansible
and
also
actually
some
azure
there
and
in
general
devops
have
been
involved
since
2006.
I've
used
most
of
these
raf
w's
ibm
bill,
forge,
ibm,
sold
puppet
chef,
ansible
jenkins,
aws
tools,
travis
and
also
azure
pipeline.
B
So
now
the
historical
looking
backs
this
started,
so-called
virtual
machines
actually
in
the
60s
ibm,
was
early
under
some
british
system.
Two
atlas.
This
is
mid-60s.
B
The
idea
is-
and
I'm
sure
you
know
about
this-
but
idea
is
that
you
have
hardware
which
is
the
gray
box
here
and
then
you
have
a
layer
called
hypervisor.
You
can
do
hyper
vxn,
which
is
kvm
esx,
that's
vmware,
so
some
some
kind
of
software
that
makes
makes
you
install
different
operating
system
each
thinking
to
have
the
machine
on
their
own,
but
the
hypervisor,
which
is
software
and
some
hardware,
makes
network
disk
cpus
kind
of
shared.
But
you
don't
really
know
that.
B
He
has
gotten
cheaper
and
one
interesting
take
on
this,
which
you
should
check
out
unless
you
know
about
this,
is
so
called
the
so-called
nitro
suite
of
projects
from
amazon.
So
basically
2013
aws
said:
let's
look
at
ec2
how
that
works,
and
they
spent
six
seven
years
or
maybe
six
years,
redoing
or
changing
ec2
to
use
so
called
nitro,
which
is
hardware
for
networking,
storage,
security,
ship
and
also
hypervisor,
which
is
much
thinner
than
any
other
hypervisor.
So
it's
almost
bare
metal.
B
So
that's.
If
you
use
a
newer
ec2
instance
like
t3,
it
will
use
nitro,
but
only
one
or
the
older
ones.
So
amazon
has
done
a
lot
of
work
here,
yeah.
So
that's
the
hardware
hardware
virtualization.
Then
we
come
to
docker
here,
which
is
just
one
example.
There
are
other
container
technologies.
Doctor
is
just
probably
the
most
well
known.
So
the
idea
here
it's
really
coming
from
google,
I
guess
2004,
5
or
even
older,
like
solaris
zones,
and
things
can
do-
could
do
this
back
then,
but
no
one
really
used
this
mainstream.
B
And
then
add
they
added
some
features:
control
groups,
namespaces
and
some
special
file
systems.
So
the
idea
is,
you
start
a
process
container
that
thinks
it
has
the
whole
linux
kernel.
So
if
you
you
can
log
in
and
see
it
like
process
id
one
which
is
the
root
or
zero
or
one
root
process
id.
But
it's
not
really.
B
It's
mapped
to
a
process
in
the
linux
kernel,
so
this
makes
the
illusion
of
having
a
process
operating
system
process
like
java
process
has
a
whole
linux
kernel,
but
it
doesn't
that's
the
whole
idea
with
operating
system
virtualization,
which
is
different
from
hardware
virtualization,
and
the
disadvantage
is,
of
course,
that
you
can
rip
down
the
whole
machine
and
then
kill
everything
on
it.
This
is
your
software.
It's
nothing
hardware
in
this,
but
hardware,
virtualization
has
instructions
in
the
cpu
that
can
protect
memory
and
things
like
that.
B
B
Assets
like
libraries
and
things
loaded
in
once
it
goes
very
fast
to
start
a
new
container.
It's
already
there
so
to
speak
and
docker,
then.
So,
if
you
run
docker
command
line
docker,
it's
really
a
little
process.
Talking
to
the
docker
agent
really
and
that
uses
container
the
run
c
run
container.
B
That's
the
thing
that
takes
the
files
and
make
it
run.
This
run
c.
So
at
the
bottom
of
kubernetes,
and
all
of
these
you
have
docker
really
so
the
fourth
layer,
then,
is
the
container
orchestration.
So
now
you
have
the
ability
ability
to
run
a
process,
but
what,
if
you
have
5000
of
them,
like
kubernetes,
did
have
a
restriction?
5000
containers
in
a
cluster
I
think
is
maybe
lifted.
That's
not
unreasonable.
B
Actually,
so
how
do
you
orchestrate
now,
multiple
containers
or
many
containers?
That's
why
these
technologies
came
to
came
to
be
like
kubernetes
and
others
mesa
and
whatever
lots
of
them-
and
this
is
a
picture
of
how
kubernetes
cluster
looks
like
so
to
the
left.
You
have
the
so-called
control
control,
the
master
which
implements
the
control
plane
in
kubernetes,
which
is
that
you,
you
send
a
command
like
apply
some
jaml
file
to
start
a
container.
B
It
will
talk
to
these
nodes
or
the
machines
running
the
workload
and
the
control
plane
idea
is
that
you
tell
it
the
future
state
and
there's
a
loop
here.
So
the
control
plane
master
will
go
and
ship
did
it
start
five
containers?
No,
it
hasn't
okay,
so
it
waits
a
while
and
then
it
shakes
again.
So
there
is
like
the
future
state
and
the
current
state
it
tries
to
make
them
the
same.
If
you
see
what
I
mean,
so
if
there
is
a
problem
with
an
image,
so
it
can't
really
start
start
it.
B
You
will
get
an
error,
it
will
stop
the
loop.
Then
you
have
to
fix
something.
So
it's
look
like
this
or
like
back
and
forth
between
the
master,
node
master
nodes
and
here's
the
difference.
If
you
run
like
amazon's
kubernetes
eks,
you
have
to
kind
of
build
your
cluster,
including
the
master,
but
that's
google's
engine.
You
don't
see
the
left
side.
You
only
need
to
worry
about
the
nodes,
that's
it
makes
it
easier
and
google
obviously
were
the
ones
that
started
this
whole
thing
yep.
So
now
we
have
hardware
virtualization
operating
system,
virtualization.
B
B
B
I
have
the
link
after
the
last
slide
accelerated
the
book
looking
at
the
state
of
the
art
of
devops
and
automation
and
architecture,
and
they
these
guys
jx,
took
that
book
and
said:
okay,
let's
implement
it,
and
this
is
what
they
came
up
with
so
so
it
has.
It
depends
on
kubernetes
clusters.
It
won't
work
with
anything
else.
It's
mostly
command
line
interfaces.
Only
thing
I
will
show,
but
you
can
run
some
user
interface.
It
depends
on
github
and
web
hooks
and
images,
of
course,
and
it
uses
helm.
B
So
it
kind
of
uses
all
these
things
out.
There
lots
of
components-
and
the
idea
is
to
this-
is
really
like
git
ops,
you,
you
would
have
a
repository
match
in
your
environment,
for
example,
for
example,
I
have
staging
and
production
I
will
use
staging,
so
staging
is
living
actually
in
a
repository
in
github
and
there
are
web
hooks
to
tell
it
to
do
stuff.
So
you
don't
do
like
we
did
at
my
last
company.
B
We
write
jenkins
scripts
and
things
to
do
service
deployment,
etc.
This
is
actually
driven
by
commits
to
git,
which
this
is
what
they
call
githubs.
So
you
commit
a
change
to
something
that
go
and
modify
the
environment.
So
that's
how
you
push
an
application.
You
don't
run
some
other
scripts.
It
will
do
that.
B
B
D
A
So
if
you
have
questions
post
it
to
the
chat,
yeah
and.
D
B
Sure
yeah,
yes,
and
that
last
slide
I
will
come
back
to
here
before
this
is
just
the
references
I
referenced
to
in
in
one
slide
here.
As
you
can
see
so
the
book
I
use,
I
will
show
the
infra
app,
which
is
a
really
good,
app
to
manage
kubernetes
pods
on
your
laptop
devops26
toolkit.
This
is
the
book
which
covers
so
I
will
cover
about
20
of
this
book,
maybe
15,
and
it's
already
out
of
date.
So
it's
a
bit
tricky.
B
You
can't
just
use
this
book
anymore,
as
is
things
will
not
work,
but
it
does
work
to
level.
So
this
is
the
definite
guide
to
yankee
sexes,
this
book
and
it's
a
series
of
devops
books,
the
same
authors.
It's
like
volume,
one,
two,
three,
four,
five,
six
seven
covering
different
things
so
highly
recommended
for
this
accelerate
is
the
book
which
is
more
of
a.
B
Interviews
and
statistics
about
how
people
build
and
scale
businesses
with
devops
and
architecture,
so
nitro
is
also
mentioned.
The
amazon
kind
of
infrastructure
they
build
everything
on
and
also
a
little
command
called
cube
ctex.
I
can
recommend
it's
not
so
important,
but
I
wanted
to
have
the
references
here
so
now
I'm
going
to
build
the
demo
and
the
first
thing
I
have
to
do
ahead
of
time.
B
Since
I
need
a
kubernetes
cluster,
it
takes
a
while.
It
took
25
minutes
to
build.
So
I
cannot.
I
cannot
build
it
during
the
demo,
so
I
needed
to
build
it
ahead
of
time
and
this
is
a
command.
So
this
jx
is
something
you
install
on
your
laptop
and
it
has
comma
lots
of
commands
and
one
of
them,
if
you
can
see
here,
cjx
create
cluster.
I
run
this
command
as
it
is,
and
you
give
it
a
name
project
id
which
is
the
gte
or
rather
google
cloud
platform
project.
B
You
have
to
have
an
account
there,
a
region,
what
type
of
machine
the
nodes
should
use.
How
many
nodes
I
do
have
two,
I
think
yeah
and
some
other
things.
So
this
will
run
for
20-25
minutes
and
I
can
show
you
I
have
to
probably
scale
this
up
a
bit.
B
C
B
Oh,
my
screen
is
smaller
than
yeah
and
and
here's
the
pool
here
down
with
a
this
type
of
machine.
It's
not
super
big,
but
this
is
kind
of
the
basis.
Now,
of
course,
you
can
create
other
pools
more
more
nodes.
This
is
for
the
demo.
I
will
rip
this
down.
This
costs
some
money
after
this
demo
but
yeah
and
then
what
it
does,
the
jx
crate
it
actually
creates.
All
of
these.
B
I
don't
know
how
many
this
is
18
pods,
so
just
imagine
doing
this
manually,
so
the
jx
crate
command
actually
created
all
these
parts
and
they
not
going
to
cover
all
of
this,
but
you
have
the
tech
tone
down
here.
That's
that
that's
like
the
build
system.
You
have
all
the
things
with
gc,
since
it
creates
lots
of
parts,
it
has
crown
jobs
that
goes
and
kills
them
if
they're
not
needed.
So
you
can.
A
Yeah
we
have
a
couple
questions.
One
is
related
to
this
screen
here.
The
first
question
is:
is:
is
jenkins
x,
running
off
of
the
custom
resource
definitions.
B
B
C
B
A
The
question
was,
you
know
just
from
a
broader
perspective,
why
is
jenkins
not
part
of
jenkins
x
and
what
then
makes
jinx's
ex
so
great
sort.
B
No
yeah,
so
so,
basically,
this
started
a
couple
years
ago
and
they
they
used
headless
jenkins
inside
and
you
could
maybe
two
months
three
months
ago,
you
can
actually
they
call
it
classic
yanking
sex
you
can.
Actually
I
mean
jenkins,
is
just
a
build
system
that
has
the
notion
of
steps
and
dependencies
and
upstream
downstream.
B
It
doesn't
really
do
anything
like
if
you
don't
use
the
user
interface,
it's
just
a
sequencer
and
they
found
that
the
tecton,
sequencer
or
step
engine
was
faster
and
better
and
easier,
so
they
changed,
but
they
kept
the
name
jenkins
x.
B
So
so
that's
why
the
name
it
should
be
called
something
else.
Maybe
they
change
that
one
day
it
started
with
a
jenkins
engine,
but
they
replaced
it
simply.
A
B
Okay,
I
will
show
when
I
show
the
demo
you
recover,
you
don't
really
see.
I
mean
I
use
jenkins
and
even
the
one
before
what's
called
hudson
or
something,
but
this
you
don't
really
need
to
log
in
to
see
any
of
that
you
don't
have
to
do
the
user
interface
or
anything.
You
know,
but
you
will
see
better
when
I
and
I
show
the
demo-
I
think
yeah
okay.
So
what
also
came
out
from
this
jx
command
is
capture
repositories
here.
B
So
this
is
my
public
github,
so
this
one
environment
called
environment,
jx
rock
staging
which
what
I
will
use
and
then
a
production
which
I
will
not
show
today,
but
this
one
is
created
automatically
by
this
command
line,
so
it
will
have
actually
jenkins
file.
So
so
the
text
on
engine
can
actually
read
the
jenkins
file.
This
is
probably
quite
yeah.
It's
just
some
steps
here.
So
so
it
automatically
generates
a
bunch
of
files
and
you
never
really
modify
these
unless
there's
some
specific
stuff
going
on.
B
I
think
that's
it
for
that
part.
So
now
it
also
created.
Can
you
see
this
yeah?
I
can
do
it
make
it
bigger.
So
this
is
a
command
line.
B
This
is
jx,
it
has
lots
of
options.
I
will
not
cover
all
of
this,
but
it's
ready
to
now
create
they
call.
The
have
quick
starts,
which
is
the
quick
start,
is
like
a
recipe
to
build
a
demo
application
and
they
are
growing
a
growing
list.
I
don't
know
how
many
we
can
see.
So
when
I
create
quick,
quick
start
here,
it
will.
Let
me
I
think
it
downloads
this
from
jit
somewhere
the
list
it
takes
a
bit
of
time.
B
D
B
So
now
it
will
create
a
repository
in
the
top
and
also
create
a
local
file
system
and
shake
in
the
master.
So
so
it
will
yeah
it's
actually.
If
I
go
now
to
the
see
here,
this
refresh,
my
repository
should
have
a
yeah,
the
export
just
created
15
seconds
ago,
so
that
came
from
that
command,
and
I
will
also
run
this
here.
D
B
B
Yeah,
so
they
are
yeah,
it's
it's.
You
know
initializing
some
work.
So
it's
running
all
this
stuff
in
the
cluster
here
now
for
my
new
component.
B
Yeah
yeah,
you
see
yeah,
you
see
the
steps
here.
Meta
pipeline,
credential
working
directory
place
tools
it
source,
so
it
kind
of
put
it
does
merge
pulling
the
pipeline
here.
So
this
is
the
text,
and
this
is
kind
of
what
jenkins
would
do
and.
B
B
That's
already
in
there
I
said
that
that's
the
beauty,
that's
that's
the
beauty
and
the
point,
and
you
of
course
have
to
modify
some
things
in
the
real
scenario,
but
I've
been
building
the
last
five
years
from
scratch.
You
know
ansible
cloud
formation,
jenkins,
git
web
hooks.
All
of
that
is
done
by
these
guys.
You
don't
have
to
do
anything
unless
you
want
to
modify
some
some
helm
special
version.
I
mean
it's
already
doing.
I
I
don't
do
I
don't
do
more,
so
I
can
open
my.
B
C
B
Yeah
yeah,
I
I
used,
I
ran
it
on
eks
and
google.
I
know
the
support
aks.
So
that's
I
sure,
and
then
on
premise
like
on
my
you
can
run
probably
on
mini
cube
or
something
on
your
laptop
or
and
then
there
are
some
other
ones
like
rackspace.
I
think
your
line
nodes
there
are
some
other
ones.
It's
more
matter
of
value
certification.
A
A
Add
to
some
to
your
pipeline,
then
some
kind
of
custom
like
calling
testing
or
changing
something
yeah.
Then
you
customize
the
gene
concepts
itself.
B
A
D
A
What
I
think,
what
the
what
is
hard
about
understanding
jx
compared
to
jenkins
is
we're,
so
we've
been
taught
for
quite
a
long
time
now
about
this
kind
of
static
pipeline
that
you
define
the
different
steps
in
the
process.
Now.
D
A
But
it's
built
it
on
based
on
the
tools
that
it's
just
going
to
assume
that
you're
going
to
use
like,
like
I
see
gradle
in
here
so
there's.
B
B
You
can
import
your
own
code,
but
you
have
to
add
stuff,
like
the
yankees
x,
which
bill
bill
pack
means
this
is
a
gradle
type
project,
so
it
might
say
rust
here
or
go
or
node.js
or
something
so
you
don't
have
to
follow
there.
This
is
just
hard
to
demo.
If
you
have
to
build
this
from
scratch,
you
can
import
an
existing
code
base
and
you
have
to
add
stuff
like
the
shorts
here,
which
is
for
helm.
B
B
Yeah
yeah,
the
idea
is
you
don't
really
need
to
have
a
person
100
percent.
We
had
like
two
people
hundred
percent
doing
devops
you,
you
maybe
need
a
50.
Let
the
developer
have
their
cluster
and
go
from
code
to
at
least
staging
without
devops
person.
That's
kind
of
the
idea,
I
think,
to
speed
up
that.
So
this
will
this
demo
will
actually
will
build
and
execute
in
the
cluster.
A
B
Yes,
yes,
yes,
I
just
want
to
run
because
it
complained
about
something:
no,
it
passed
okay.
So
this
is
the
code
and
if
you
look
now
at
my
command
window,
it
actually
created
a
pull
request
automatically
in
jita,
but
it's
set
to
automatically
approve,
but
you
can
change
that
flag
saying
it
will
stop
here.
You
have
to
go.
Some
approver
have
to
go
in
and
approve
this,
but
it
succeeded.
So
I
can
stop
this
and
do.
B
So
now
you
have
a
public
dns
name
actually,
with
my
application
deployed
out
there,
so
I
should
be
able
to
do
curl
and
it
should
actually
are
not
yet
it
takes
a
while
forward.
B
So
it
is
running,
it's
not
just
come
up
yet
so
this
is
both
built
and
deployed
to
a
cluster
and
assigned
an
external,
in
this
case
dns
name.
So
this
is
not
just
my
local
environment.
It
takes
a
while
now
because
it
runs
out
there
in
gke,
so
you
normally
have
to
go
and
just
shake.
What's
going
on
here,
oh,
I
can
actually
show
the
other
tool.
Meanwhile,
it's
called
infra.
B
B
B
Yes,
so
it
distorted,
so
this
tool
runs
on
a
laptop
is
smart.
I
feel
it's
much
faster
than
to
use
the
google
console
or
the
eks
console.
Sometimes
this
is
quicker,
but
it
shows
the
same
thing,
but
it's
not
dependent
on
a
vendor
like
aws
or
google
or
microsoft.
This
is
a
generic
kubernetes
manager
tool,
which
is,
I
use
a
lot,
but
let's
see
if
this
one
got
further
yeah.
So
now
we
see.
B
The
app
actually
running
out
on
the
internet,
you
could
actually
hit
this
endpoint
yourself,
so
it's
good
to
know
that
it
says:
nip
dot
io
there
at
the
end,
it's
a
free
service
that
makes
an
ip
address
a
dns
name.
So
this
way
you
can
publish
a
temporary
dns
name
to
your
app
without
having
just
the
app
addresses,
because
some
of
these
services
want
to
have
dns
and
not
just
ip
addresses.
B
So
that's
what
this
is
yankees
x
creates
that
automatically,
but
you
can
put
your
own
domain
custom
domain,
dns
name
there
for
the
real
production
system.
This
is
kind
of
temporary.
So
now
I'm
going
to
modify
this
code
and
show
what
happens
when
it's
processing
through
the
change.
So
I
will
go
in
just
do
some
dumb
changes
that
so
we
can
see
like
here
is
the
and
two
endpoints.
C
This
one
doesn't
have
creating
yeah.
B
This
one,
which
shows
anything
that
happened
to
my
project-
it
will
you
see
now
it
doesn't
say
before
it
says
something
about
master
number
one.
Now,
actually
I've
commit
or
push
two.
So
now
it
starts
automatically
to
pick
up
my
code
changes
and
it
starts
the
new
pipeline
to
build
my
new
version.
So
we
have
to
wait
a
little
bit.
So
that's
how
it
would
that's
it.
B
This
will
take
a
bit
of
time.
I
can't
do
much
about
that,
so
so
here
it
will
do
in
the
in
the
old
world.
It
will
do
so
in
jenkins.
It
would
be
listening
for
a
commit
to
some
branch
around
the
pipeline
and,
at
the
end,
try
to
redeploy
the
service.
We
used
ecs
python
command
line
that
that
could
tell
amazon's
docker
to
redeploy
a
service,
for
example,
but
I
don't
have
to
do
any
of
that
now
it
just
does
it
so
repair.
D
B
I
I
don't
really
have
that
covered
in
this
demo,
because
I
only
have
four
minutes
left
of
the
actions
so,
but
so
the
book
I
cover
15
20
percent
of
this
book
there.
Maybe
40
of
the
book,
is
actually
how
to
modify
these
pipelines.
It
just
gets
very
complicated
because
you
have
to
change
their
examples.
Are
a
bit
complicated,
it's
hard
to
show
in
a
short
demo,
but
you
can
modify
the
pipelines
like
you
want
to
use.
B
D
B
No,
I
would,
I
would
just
modify
what
you
need:
helm,
shorts,
scaffold
definitions
and
you
inject
snippets
there.
You
don't
really
need
to
modify
so
one
thing
you
can
typically
do
is
you
you
build
towards
staging,
but
there
is
a
promote
step
I
haven't
set
this
up,
but
promote
will
do
like
push
to
prod
but
but
from
the
command
line.
So
you
don't
just
automatically
push
the
prod.
So
now
it's
waiting
on
the
pull
request.
So
it's
soon
down
here.
B
B
B
B
In
the
cloud
yeah
jx
is
just
the
command
line
to
you.
You
install
on
your
system
to
drive
the
command
line,
but
it
communicates
only
with
the
cluster
running
somewhere.
We
could
run
mini
cube
on
your
laptop.
It
could
run
aws,
google,
I
sure
whatever
it
just
needs
a
connection,
so
you
need
to
authenticate
from
your
system
to
it.
B
So
this,
in
this
case
infra
here,
has
one
two,
three,
four,
five,
six
seven
eight
different
target
clusters,
I'm
using
the
one
I
built,
but
you
can
just
switch
and
then
it
will
show
what's
going
on
inside
that
cluster,
so
it
doesn't
really
care
where
it
runs.
B
A
So
my
question
would
be:
is
in
a
future,
would
you
be
interested
in
doing
a
future
meetup
presentation
on
how
to
modify
a
pipeline
in
jenkins
x.
B
A
Sure,
the
last
time
I
spoke
with
koskay,
he
you
know
basically
was
had
the
thoughts
that
being
able
to
have
pipelines
already
built
out
as
a
starting
point
was
sort
of
the
the
the
the
starting
point
of
jenkins
x
and
getting
getting
rid
of
plugins.
B
B
I
want
to
do
a
side
car
with
envoy,
and
how
do
you
add
that,
yes,
you
need
to
update
your
helm
and
scaffold,
so
it's
kind
of
gets
very
specific,
so
so
the
use
case,
the
demo
use
case-
has
to
be
something
useful.
Most
of
them
are
doing
very
specific
things.
This
is
already
doing
from
code
to
production
without
any
change.
So
question
is
what
what
makes
sense
to
demo
for
for
people.
That's
not
just
a
very
super
special
case.
I
want
to
run
some
special
node
version.
B
A
So
maybe
a
good
case
study,
for
it
would
be
how
to
add
in
a
security
scan.
We
did
one
of
our
last
meetups.
We.
D
C
C
A
Would
you
add
a
security
scan
or
how
would
you
know
you
do
a
run,
something
like
deploy
hub
to
do
automatic
the
configuration
management
around
what
was
just
deployed
yeah.
A
A
good
follow-up,
yes
on
jinkx
yeah.
A
B
The
problem,
often
with
this
stuff,
is
all
asynchronous.
Just
because
this
thing
says
it's
done.
It
doesn't
mean
that
google
cluster
is
done.
So
there's
always
a
wait,
see
yeah.
So
this
one
has
version
002
now,
so
it's
me
it
went
from
one
to
two,
so
we
can
see.
B
If
I
do
that
it
doesn't
say
yeah.
I
did
too,
but
infra
would
show
an
actual
log
file,
yeah
yeah
yeah.
So
now
we
see
there
are
actually
two
different
pods
and
old
and
the
new
see
if
I
look
in
the
log
here.
B
C
B
B
B
Yeah,
you
see
so
I
have
five
so
now,
I'm
just
running
my
new
version
of
this,
so
it's
fully
commit
push
build
redeploy
because
there
was
an
old
service
without
me
doing
really
anything.
That's
that's
the
kind
of
the
idea
with
this.
B
What
you
and
there
is
this-
maybe
how
to
increase
this
a
bit.
So
these
end
points
are
automatically
created,
so
there
is.
B
C
D
B
There
was
a
mat
here:
it
was
a
master
branch,
commit
master
branch,
promotion,
etc.
So
you
can,
you
can
kind
of
see
a
bit
what's
going
on?
I'm
not.
I
don't
really
need
to
use
any
of
this,
but
if
something
fails
normally
okay,
so
it
jumps
into
the
repository,
and
here
you
can
see,
pull
requests.
C
B
B
Nothing
yeah
because
the
thing
is
with
this:
you
would
use
terraform,
maybe
to
build
your
njx
and
terraform
to
build
your
ingress
and
security
settings
for
your
provider
like
aws.
You
might
have
load
balancers,
so
kubernetes
would
do
some
without
and
you
you
would
build
that
with
terraform
and
then
this
would
just
create
repositories
and
push
code.
You
don't
really
need
to
do
so.
Much
changes
if
you
have
your
cluster
up
and
running
and
it
does
work
with
existing
clusters
as
well,
not
just
from
scratch
like
I
did
here.
So
they
try
to
separate.
B
Like
someone
runs
terraform
whatever
drum
to
build
clusters,
you
would
do
maybe
ansible
if
there's
some
special
thing
going
on
and
then
this
would
handle
just
code,
changes
and
builds,
and
you
don't
really
need
to
go
and
tweak
so
much.
It
will
do
a
service
upgrade
automatically
replace
the
like.
You
saw
pod
version.
One
plays
with
two:
you
don't
have
to
do
anything
of
that.
A
And
then
they're
they're
one
there's
another
question
about
using
jenkins
x
to
create
on-demand
automated
tasks
like
prod,
like
dever
testing
environments,.
B
Yeah,
you
have
to
build
kubernetes
clusters.
First,
you
have
to
build
github
repositories.
You
have
to
build
web
hooks.
You
have
to
sync
all
this
together,
you
have
to
build
helm
package
management
access
security,
so
just
techton
is
not
doing
only
that.
B
Right,
you're
right
and
it's
not
just
the
pipelines,
it's
all
the
other
components.
It's
docker
registries.
It's
helm,
shorts.
It's
scaffold
for
builder,
deploy
jit
git,
ops
web
hooks,
your
code,
quick
starts.
I
mean
there's
lots
of
stuff
behind
this.
It's
not
just
a
little
component
like
that,
and
also,
as
I
said
this
book,
maybe
I
can
pick
up
the
pull
request.
Yeah
go
back
to
my
last
slide
here,
so
this
book
here
you
can.
I
have
a
kindle
version,
but
it
shows
all
answers
to
most
of
these
questions
in
this
book.
B
I
just
don't
have
time
so
I
covered
20,
the
first
20
but
there's
another
80.,
so
they
do
things
like
create
a
dev
environment
in
kubernetes
and
give
you
an
endpoint
to
run
visual
code
web,
which
is
electrons.
You
actually
give
you
actually
you're,
given
a
dev
environment
in
the
cluster,
you
can
develop
you
push
and
disrupts.
So
that's
one
example
of
an
extension.
Some
people
use.
You
can
also
integrate
your
dev
environment
with
your
tools
like
intellij,
on
your
laptop,
but
the
updates
will
trigger
the
build
environment
as
well.
B
A
A
Yeah
add
jenkins,
you
know
jenkins
x,
to
the
jenkins
pipeline
to
go,
create
to
execute
those
the
build
and
the
creation
of
the
container.
B
Yeah,
you
could,
I
guess
I
mean
that's,
how
they
started,
but
they
they
have
found
that
takedown
pipelines
was
easier
and
faster.
But
that's
I
don't
know
exactly
why
that
happened,
but
you
could
yeah,
it's
just
command
line
right
tool
has
millions
of
options,
and
I
only
showed
like
three
here
so
you
know
yeah.
What
was
I
thinking
about?
I
forgot.
B
No,
no,
no!
That's
fine
yeah,
but
this
book
shows
much
more.
I'm
just
warning:
it's
already
a
bit
obsolete.
I
know
that
some
of
the
new
scaffold
so
scaffold
is
a
tool
that
is
a
a
mini
program
that
can
take
both
the
build
and
deploy
it's
like
a
jamal
and
it
works
together
with
all
of
this.
But
scaffold
has
been
upgraded
so
much
so.
The
examples
won't
really
work
anymore
out
of
the
book
so
to
speak,
yeah.
But
you
would
see
the
idea
like
how
it's
supposed
to
work.
A
B
B
A
D
A
Done
a
great
job
of
giving
us
the
you
know
a
a
peek
under
the
hood.
I
had
really
no
idea
that
this
is
what
jenkins
x
was
doing
and
I
find
it
fascinating.
It
is
a
shift,
though,
in
terms
of
how
we
think
about
pipelines.
Yes,
and
now
I
understand
what
cross
k
was
talking
about
when
he
was
like
we're,
trying
to
standardize
as
much
as
we
can
in
lincoln's
x,
containers.
B
Yeah,
so
she
has
to
yeah
I
reply
to
that.
I
think
if
you
see
the
full
jenkins
x,
you
can't
really
show
in
an
hour,
but
it's
the
book
give
so
they
have
like
commands.
Here's
a
whole
cluster
and
a
whole
dev
environment
for
code
change.
There
is
nothing
to
install
even
on
your
laptop
except
a
web
browser,
so
it
will
automatically
create
pods
for
development.
B
It
gives
you
an
end
point
for
you
only
you
you
work,
you
commit
to
push
so
so,
basically
it
can
remove
the
whole
laptop
from
the
equation,
but
I
wouldn't
like
to
work
with
that,
but
this
kind
of
cool.
So
so
they
do.
You
have
both
the
development
environment
under
kubernetes,
as
well
as
the
build
and
actual
service
deployment.
That's
kind
of
new
to
me.
Normally,
you
would
sit
with
your
laptop
code
and
then
commit.
B
This
is
not
even
that
it's
not
even
running
development
on
your
laptop
anymore
and
some
people
would
I
I
mean,
I'm
not
sure,
I'm
ready
to
go
there,
but
it
looks
kind
of
cool,
because
visual
code
can
run
in
a
web
browser
as
well
as
a
gke
has
what's
called
a
cloud.
Shell
and
azure
has
a
cloud
shell,
so
this
is.
This
is
not
so
new
that
you
run
actually
the
whole
thing
somewhere
else:
development
as
well,
not
just
deployment
and
build.
B
A
Your
opinion,
this
is
a
production
grade
product.
It's
not
just.
B
Development,
absolutely
yeah,
but
of
course
you
would
in
a
real
world
you
you
would
use
your
own
dns
and
you
would
of
course
have
to
have
security
and
all
that,
like
I'm
more
aws
than
google,
but
all
the
security.
All
that
stuff
you
can't
just
I
mean
you
have
to
do
work
for
your
production
like
security.
B
D
A
All
right
well,
thank
you
so
very
much
for
doing
this
presentation
for
us
today.
I
appreciate
the
effort.
I
know
that
none
of
this
is
ever
easy
to
put
together.
A
So
we
will
have
you
back
to
do
a
follow
up
on
jenkins
x,
there's
so
much
to
learn
about
this.
If
you're,
a
jean
consider.
D
A
100,
absolutely
all
right:
everybody
stay
safe
out,
there
stay
healthy
and
we
will
see
you
in
a
month
and
we'll
we
will
bring
pear
back
and
we'll
give
them
a
couple
months.