►
From YouTube: Keynote: Extending Jenkins with Spinnaker for a Multi-Target World - Isaac Mosquera, Armory
Description
Keynote Session Sponsored by Armory - Extending Jenkins with Spinnaker for a Multi-Target World - Isaac Mosquera, Armory
With the 3 major cloud providers battling it out for cloud dominance, we're left to reason through a dizzying array of new infrastructure and tools. What this means to your organization is that you're no longer deploying to 1 target (e.g. EC2, Bare Metal, PCF, Kubernetes, Lambda, Serverless), but now deploying to many targets. Jenkins is at the core of many CI/CD pipelines today, so we'll discuss how to pair Jenkins with Spinnaker for a multi-target world.
0:00 Intro
0:50 Armory's Journey
7:55 What is a target?
8:52 How Jenkins works
10:45 Deploying with Spinnaker
13:20 What does a pipeline look like?
14:46 Using Spinnaker and Jenkins
A
B
B
B
B
Now,
for
those
who
don't
know,
virtual
private
server
was
kind
of
like
the
predecessor
to
what
we
know
as
aws,
except
there
was
a
lot
of
manual
clicking
and
interaction.
So
you
would
go
onto
this
website
click
a
whole
bunch
of
different
buttons,
and
then
you,
you
know
you
hit
that
last
button.
It
would
provision
for
you
a
vm
and
then
I
would
ssh
into
that
vm
and
because
I
was
a
lot
younger
in
2005,
I
had
a
lot
more
hair,
but
I
was
just
very
naive
about
automation.
B
It
wasn't
a
thing
that
I
really
cared
about
or
even
knew
about.
I
would
manually
do
everything
I
would
manually
set
up
the
apache
web
server.
Yes,
the
apache
web
server
was
also
the
most
dominant
web
server
at
the
time
we
would
also
configure
tomcat
and
we
would
have
ourselves
an
inter
service
that
people
could
consume
and
then
in
2006.
B
So
we
had
image
storage,
which
was
important
to
our
social
startup
and
sqs,
and
the
queuing
systems,
which
was
also
really
important
for
us,
and
so
we
jumped
over
to
east
to
to
aws,
and
we
made
that
bet,
because
at
the
time
we
didn't
you
know
it
was
still
a
new
service
and
looking
back
now
that
was
the
greatest
move,
because
what
we
didn't
know
was
that
there
was
going
to
be
another
15
years
of
innovation,
in
that
that
aws
was
going
to
be
the
cloud
technology
to
really
kind
of
lead.
The
innovation.
B
Now.
The
other
thing
that
I
think
is
more
important
than
just
kind
of
that
innovation
is
the
behavioral
changes
that
happen
to
me,
because
aws
was
not
about
giving
you
something
that
was
like:
hey
here's,
a
vm
aws
really
ushered
in
this
idea
of
automation,
and
so
I
saved
an
enormous
amount
of
time
by
moving
to
ec2,
not
because
of
the
vm
was
so
much
different
was
because
everything
that
it
was
about
was
about
automation
and
reducing
the
overhead
of
operational.
You
know
scripting
and
things
that
you
had
to
do.
That
was.
B
And
they
were
ignoring
the
frick
train
that
was
called
kubernetes
at
the
time.
So
that's
why
they
came
out
with
ecs.
They
didn't
want
you
to
move
workloads
off
of
their
service.
Of
course
they
would
want
you
to
stay
with
amazon,
and
so
this
was
a
another
great
move
in
innovation,
and
now
I
don't
even
have
to
worry
about
servers.
I
can
just
purely
focus
on
my
code
with
respect
to
lambda.
Again
I
just
have
to
figure
out
how
to
get
my
code
up
there
safely
and
then
in
20
oops.
B
B
Now,
in
between,
there
was
a
lot
of
other
innovation.
Things
like
rds
came
out
around
here.
There
was
around
here.
There
was
also,
let's
see
emr
and
a
lot
of
other
places
to
run
compute
right
in
different
styles.
Right,
like
emr,
was
all
about
hadoop
right,
and
so
this
is
the
rough
timeline
plus
thousands
of
other
services,
but
these
are
the
ones
that
really
kind
of
stick
out
in
my
head
and
then
from
there.
Also.
B
B
B
B
B
Now
you
may
say:
isaac,
I'm
not
a
multi-cloud
operation
and
and
there's
probably
a
lot
of
companies
that
are
not
multi-cloud,
but
you
are
absolutely
multi-target
you
deploy
to
many
of
these
different
locations.
Even
within
a
single
cloud,
you
don't
deploy
to
aws
what
used
to
be
just
ec2,
you
could
just
deploy
to
ec2,
and
that
meant
you
were
deploying
to
aws
is
now
a
multitude
of
services.
So
let's
define
what
a
target
is.
B
That's
where
configuration
database
schema
will
be
a
target
now
why?
It's
important
to
define
even
the
word
target
is
because
anything
that
is
a
target
then
needs
to
be
audited,
needs
to
be
part
of
compliance
needs
to
be
rolled
out
safely
and
needs
to
be
rolled
back
safely.
It
goes
through
a
different
process
when
it's
about
actually
a
deploying
your
code
to
a
target
right.
Even
phones
applications
could
also
be
a
target
right,
and
so
there
are
many
different
things
that
we
can
consider
target.
B
C
B
B
B
They
had
thousands
of
microservices,
as
many
of
you
are
moving
towards,
and
there
increases
the
complexity,
and
not
only
does
it
increase
the
complexity.
You
also
want
to
package
up
the
best
practices
of
deploying
into
these
servers,
and
so
in
2016
netflix,
google,
microsoft,
pivotal
all
decided
to
open
source
this
software
delivery
tool
called
spinnaker.
So
what
is
spinnaker
and
how
can
it
help.
C
B
C
B
Continuous
integration
right
these
aren't
these
are
concepts,
these
aren't
targets
right,
and
so
what
we're
doing
is
we're
separating
out
the
logic
of
deployments
from
the
actual
targets,
and
what
this
allows
us
to
use
is
reuse
these
these
great
best
best
practices
and
concepts.
So
then,
what
would
happen
is.
B
B
We
can
do
canary
for
kubernetes
right
and
by
abstracting
this
out.
It
saves
us
a
lot
of
time
from
having
to
rewrite
a
lot
of
the
logic
that
we
had
to
do
in
the
jenkins
world.
When
we
heard
that
there
was
a
new
service
coming
out
from
aws
or
google,
which
is
every
year
we
get
hundreds
of
new
services.
B
B
Now
the
way
that
spinnaker
will
help
you
in
this
multi-target
world
is,
if
you
notice,
we
only
have
two
deploy
stages
right.
What
we
were
have
been
doing
with
jenkins
is
rewriting
all
of
this.
All
of
this
all
of
this.
Every
time
we
had
a
new
piece
of
infrastructure.
So
if
you
want
to
save
yourself
time
using
spinnaker,
you
can
still
call
out
into
jenkins.
B
Eks
becomes
trivial
and
you
get
it
out
of
the
box
declaratively,
and
so,
when
you
think
about
spinnaker,
it's
not
it's
not
about,
in
fact
the
deployment
targets
it's
about
codifying
the
best
practices
that
are
out
there
in
the
world
for
software
delivery.
Software
delivery
is
a
lot
more
than
just
deploying
code.
It's
about
shipping
ideas
and
getting
them
into
production.
B
It's
about
all
of
the
hard
work
that
all
of
us
do
together
and
combine
up
work
and
effort
from
a
lot
of
different
teams.
It's
an
endeavor
in
human
collaboration.
It's
how
we
work
together
is
how
security,
qa
operations
developers
all
work
together,
because
delivery
isn't
ever
done.
We
may
never
be
perfect,
but
we're
always
trying
to
iterate
towards
that
perfection,
and
when
we
focus
on
best
practices,
we
focus
on
our
practice.