►
Description
Want to view more sessions and keep the conversations going? Join us for KubeCon + CloudNativeCon North America in Seattle, December 11 - 13, 2018 (http://bit.ly/KCCNCNA18) or in Shanghai, November 14-15 (http://bit.ly/kccncchina18).
Building Event-Driven Pipelines with Brigade - Brian Redmond, Microsoft (Intermediate Skill Level)
Building complex or even simple event-driven pipelines on Kubernetes has always been somewhat of an elusive task, until now. Enter Brigade, a lightweight opensource event-driven tool that accepts a Javascript expression of a pipeline that gets seamlessly converted into the associated Kubernetes runtime objects. In this session we will build a pipeline from the ground up, hands on, to demonstrate just how simple to solve sometimes complex tasks with Brigade.
About Brian
Brian Redmond is an Cloud Architect on the Azure Global Black Belt team at Microsoft. Brian focuses on containers, microservices, DevOps, and cloud native applications in the Azure cloud platform. Brian has been working in technology for over 20 years and has a mixed background across application development to infrastructure. Brian is based in Pittsburgh, Pennsylvania, USA and enjoys running, biking, and tinkering with new technology. I spoke at KubeCon 2017 in Austin to a large room and got some really positive feedback. Looking forward to an opportunity to speak at KubeCon in Copenhagen and possibly events in Japan and China in 2018.
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.
A
Welcome
everyone,
I'm
gonna,
go
ahead
and
get
started
thanks
thanks
everyone
for
coming
I
know,
people
are
filling
in
I'm
gonna
go
ahead
and
start.
We
have
a
lot
to
get
to
you
today
and
in
a
short
period
of
time.
This
session
is
building
event-driven
pipelines
with
Brigade.
So
we're
gonna
talk
all
about
Brigade.
If
you're
new
to
Brigade,
you
don't
know
a
lot
about
Brigade.
A
This
session
is
perfect
for
you
we're
gonna
kind
of
start
from
the
beginning
and
we're
actually
gonna
build
the
entire
thing,
including
deploying
an
entire
cluster
and
and
kind
of
pipeline,
all
at
once.
Just
a
little
bit
about
myself.
My
name
is
Brian
Redmond
I
work
for
Microsoft
I'm,
a
cloud
architect
on
her
Azure
global
black
belt
team.
That
is
my
last
name.
That's
not
where
I
work,
that's
not
where
I
live.
A
Get
this
a
lot
I've
been
there
17
years
so
I've
been
talking
through
that
little
thing
for
much
of
my
my
my
career
I'm
based
out
of
Pittsburgh
in
Pennsylvania,
it's
kind
of
the
Northeast
and
the
state's
I
spend
my
time
running
marathons
and
doing
things
outside.
In
fact,
I
always
have
this
picture
on
my
intro
slide,
but
in
my
time
hop
today,
I
don't
know
if
you
use
time
hop.
This
was
actually
the
the
exact
date
of
that
race
three
years
ago
today.
So
this
is
the
pre-race
photo.
A
A
To
kind
of
keep
you
awake
we're
going
to
decide
the
best
programming
language
today,
whatever
whatever
we
decide
here
will
be
binding,
so
make
sure
that
you
vote
correctly,
you
can
vote
for
any
of
the
languages
as
much
as
you
want,
but
the
way
you're
going
to
vote
is
you're,
going
to
send
a
text
message
to
one
or
one
of
these
four
numbers,
and
your
vote
is
actually
going
to
be
your
message.
I'm
actually
going
to
do
a
sentiment.
A
Analysis,
I'm
gonna,
send
this
off
to
an
AI
service
and
Asherah
called
cognitive
services
and
we're
gonna.
Do
a
sentiment.
Analysis
of
your
text
message
and
I'm
going
to
convert
that
into
my
1
through
5
rating
scale
and
then
we'll
actually
collect
those
in
a
voting
database.
All
right.
It
actually
supports
multiple
languages.
Apparently
we
tried
it
with
Danish
earlier
it
doesn't
support
a
few
sort
of
languages
like
Japanese
and
so
forth,
but
in
theory
you
can
send
whatever
you
want.
Who
knows
what
kind
of
votes
and
messages
you'll
get
back
for
this?
A
I'm
actually
using
open
fast,
so
I'm
running
a
kubernetes
cluster
in
Azure
I
work,
I
work
with
Azure
and
are
in
our
Azure
cloud,
so
I'm
running
and
aks
cluster,
and
on
top
of
that,
I
have
open
fast
installed
and
I
have
just
a
simple
function
there
that
function
receives
a
web
hook
from
Twilio.
That's
where
your
text
messages
are
actually
being
sent
and
from
that
I
actually
call
into
this
Azure
cognitive
services
and
then
I,
deploy
or
insert
the
result
into
Azure
cosmos,
DB.
So
very
simple.
A
A
So
often
we're
deploying
these
web
apps.
This
is
kind
of
what
we
do.
We
deploy
it
we
say
apply
and
off
it
goes.
We
cross
our
fingers,
even
if
we
have
some
kind
of
deployment
pipeline,
we're
often
in
that
kind
of
scenario,
where
we're
just
kind
of
hoping
it
works.
Well
we're
going
to
end
up
doing
a
deployment
pipeline
with
our
web
app
I,
don't
know
that
everybody
hates
pipelines,
certainly
if
they're
working
well,
you
might
kind
of
love
them,
but
in
general,
we're
building
pipelines
to
do
see.
Icd
and
the
pipeline.
A
We're
going
to
build
today
is
very
simplistic.
It's
by
no
means
a
description
of
a
good
pipeline.
It's
really
meant
to
be
showing
you
kind
of
how
you
would
do
automation
and
then,
obviously
from
there
you
might
build
a
much
more
complex
pipeline.
Certainly
in
my
case
I
work
at
Microsoft.
It
might
be
great
to
actually
say
you
know,
file
new
pipeline
and
have
clipping
pop
up
and
actually
build
out
the
proper
kind
of
pipeline.
For
me,
we're
not
going
to
do
that.
A
We're
not
going
to
see
Clippy
today,
besides
besides
right
here,
so
we're
gonna
use
Brigade
to
do
this
and
again
what
is
what
is
brigade
equals,
event-driven
scripting
for
kubernetes
in
some
ways
it's
actually
similar
to
some
of
the
surrealist
frameworks
like
like
a
like
an
open
faucet
trigger
some
kind
of
some
kind
of
workflow,
and
my
various
functions
in
this
case
are
going
to
be
containers
in
kubernetes.
I
may
have
a
pretty
complex
workflow
a
series
of
containers,
some
might
run
in
some
steps
might
run
in
parallel.
Others
might
just
run
in
series.
A
I
want
to
describe
that
pipeline
in
this
case,
with
JavaScript
and
so
I
used
javascript
to
define
that
workflow
and
the
difference
here
is
I
could
have
a
very
long
running
kind
of
pipeline.
It's
not
a
simple,
quick
function
that
runs
and
then
returns
a
result.
I
actually
want
to
build
something
that
can
foresee
ICD,
for
example,
to
integrate
my
code
run.
A
series
of
tests
actually
deploy
canary
builds
I
want
all
that
kind
of
complexity
and
Brigade
can
help
me
with
that.
It
comes
with
some
basic
gateways,
things
like
github
and
docker
registry.
A
So
you
can
actually
trigger
this
kind
of
workflow
off
of
those
kind
of
things,
but,
frankly
you
can
trigger
it
off
anything.
Absolutely
anything.
You
can
write
your
own
gateway
to
trigger
it
off
of
data
being
written
into
written
into
a
queue
into
a
storage
container.
Whatever
that
may
be,
it
is
well-suited
for
C
ICD,
certainly
helpful
when
we
want
to
have
things
like
secrets
in
our
pipeline,
like
passwords
to
things
like
a
container
registry
or
things
like
that
Brigade
kind
of
helps.
A
It
make
it
very
easy
for
me
to
put
those
into
a
project
file
and
inject
those
into
my
pipeline.
So
I'm
not
putting
passwords
right
in
my
script,
which
would
be,
would
be
a
big
mistake.
So
here's
what
the
architecture
looks
like
the
components
that
will
install
initially
in
Brigade
are
a
gateway
in
this
case
by
default.
I'll
get
this
github
gateway,
I
also
get
a
controller
which
really
does
the
work.
The
Gateway
just
says
to
the
controller,
now
take
this
action.
A
It
actually
spawns
a
worker
that
will
actually
run
our
actual
pipeline
and
that
spawns
a
series
of
jobs,
those
jobs.
All
of
these
components
are
just
kubernetes
pots.
We
do
have
this
project
file
that
feeds
into
and
to
that,
as
the
event
is
fired,
we're
going
to
collect
all
that
data
out
of
our
project
and
actually
run
this
kind
of
thing
into
end
and,
of
course,
there's
also
an
API
pod
for
access
to
the
data
inside
out
of
reg
aid
in
general.
A
That's
a
quick
picture
of
the
architecture
I'm
actually
going
through
the
slides
pretty
fast,
because
in
general
I
think
people
learn
best
by
demo,
so
we're
gonna
try
to
get
to
the
demo
pretty
quickly
effectively
we're
going
to
install
a
brigade,
we're
gonna
push
code
into
github,
which
will
fire
a
web
hook.
I'll
do
two
things
in
that
in
that
Brigade
pipeline
I'll
actually
send
my
code
in
my
docker
file
up
to
my
Asha
container
registry
and
I'll,
have
it
built
in
the
cloud.
That's
where
I'm
gonna
store
it
anyway.
A
We're
announcing
some
new
features
in
that
in
that
service.
That
will
actually
allow
you
to
build
containers
in
the
cloud
they're
announced
to
get
next
week,
but
they're
letting
me
talk
about
it
today,
which
is
kind
of
fun
and
then
we'll
actually
do
a
helm,
upgrade
and
deploy
it
in
the
cluster,
and
in
this
case
it
isn't
there
already.
A
So
it
actually
will
be
a
brand
new
install,
but
this
pipeline
will
also
allow
us
to
do
upgrades
as
well,
so
with
that,
let's
do
some
live
coding
and
I
and
I
hate
to
tell
you
this,
but
we're
we're
gonna
actually
do
pair
programming.
Today,
it's
gonna
be
me
programming
with
you.
If
anything
goes
wrong,
where
it's
collectively,
all
of
our
fault,
I've
been
pretty
successful
on
my
own
I've
tried
this
out
a
few
times.
So
if
we
fail
here
today,
it
might
be
a
little
bit
more.
Your
fault
I
hate
to
say
so.
A
So,
let's
get
started.
Actually,
let's
go
over
and
see
what
kind
of
votes
we
have
here,
good,
everybody's
participating.
We
don't
know
what
the
results
are.
We're
gonna
work
hard
to
get
that
here
in
a
moment.
So,
let's
take
a
look
at
our
cluster,
and
so
you
can
see.
I
have
I,
have
a
few
pods
actually
in
the
cluster.
Already
I
do
have
these
api's
that
they're
actually
working
in
conjunction
with
the
stuff
that
I
installed
with
open
fast.
A
So
you
can
see
I
installed,
open,
Foss
I
have
a
function
in
another
namespace,
but
these
api's
are
actually
the
way
that
the
data
is
getting
into
the
database
again.
We're
gonna
need
to
deploy
this
web
app
in
this
extra
lap
deployment.
I'm
going
to
use
some
snippets
over
here
in
a
gist
to
do
some
of
these
some
of
these
steps
so
we're
gonna,
install
Brigade
first.
A
And
we'll
see
that
kind
of
pop
up
pretty
straightforward,
we'll
see
those
three
components
come
up:
the
API
the
controller,
and,
in
this
case
the
github
gateway
I,
had
some
other
parameters
and
things
we
can
look
through
that
chart.
We
could
do
a
lot
of
different
things
for
that
installation.
I,
don't
have
to
use
the
Gateway
I
can
configure
it
with
our
back
all
those
kinds
of
things,
but
we
really
just
skipped
right
past
that
looks
like
Brigade
is
already
running
I'll.
Let
you
take
a
look
at
the
services
in
this
cluster
as
well.
A
You
can
see
I'll
end
up
getting
IP
addresses
for
the
API
and
the
github
gateway,
we're
gonna
end
up
using
that
github
gateway
for
our
web
hook
that
we're
gonna
plug
into
and
to
github
here
in
a
moment.
So
the
next
thing
we
need
is
a
brigade
project,
and
this
is
kind
of
the
structure
of
what
a
project
file
would
look
like.
This
is
where
I
put
all
those
that
information
that
I
consider
secret,
or
even
it's
just
configuration
details.
A
The
project
is
also
deployed
via
a
helm,
chart
I'll.
Let
you
take
a
look
at
that
here
and
so
we
said
helm
install
I,
pointed
to
the
Brigade
project,
chart
and
I'm
gonna
pass
in
that
animal
file,
which
again
is
the
same
as
vile
as
I
just
showed
you,
except
all
the
values
are
actually
filled
in
and
we
have
a
little
CLI
with
Brigade.
That
can.
A
Let
me
do
a
few
different
commands
and
it's
going
to
show
me
the
list
of
my
projects
and
I
only
have
this
project,
of
course,
because
Brigade
was
just
installed
freshly
here
now
and
so
any
kind
of
project
we
could
have
a
bunch
of
different
projects
and
tie
them
to
various
events
that
come
in
via
our
gateway.
Okay,
everybody
good
so
far,
it
seems
like
a
bell
might
have
been
appropriate
there.
Okay,
so
now
we
need
a
pipeline,
and
this
is
it
the
pipe.
A
This
is
empty
pipeline
and
effectively
we're
going
to
have
various
event
handlers
in
our
case
we're
going
to
handle
the
push
event
and
we're
going
to
do
sort
of
an
after
event.
So
when
we're
done,
we're
gonna
actually
perform
some
action
if
we're
successful,
and
so
we
need
to
add
some
things
here
and
so.
I'm
gonna
first
just
set
up
a
series
of
variables
and
most
of
the
time,
I'm
going
to
cut
and
paste
this
code.
To
be
honest
because
having
you
watch
me
type
might
be
pretty
painful.
A
But
we're
gonna
pull
something
from
that
that
Brigade
project,
so
this
is
where
I'm
not
going
to
type
my
actual
password
or
information
from
from
that
projects
I'm
pulling
out
of
that
project
file.
This
is
a
great
way
for
me
in
my
pipeline
to
not
have
secrets
in
the
actual
pipeline
file,
so
I
pulled
that
in
and
set
that
as
my
ACR
server
again
we're
going
to
set
a
bunch
of
variables
I'm
just
going
to
paste
all
of
these
in.
A
We're
setting
things
like
my
Azure
credentials,
I'm,
building
a
dynamic
image
tag
based
on
the
commit
ID,
just
some
basics,
like
that,
this
just
kind
of
sets
the
stage
I
can
also
write
things
out
to
the
to
the
log,
and
so
I
can
say
you
know
started
obviously
do
whatever
I
want
in
there
now
I
needed
to
find
my
various
jobs,
and
so
I'm
gonna,
say
first
I'm
gonna
do
a
CR
and
it's
equal
to
a
new
job
and
we'll
call
it
AC
or
we'll
call
it
job
runner.
You
see
our
builder.
A
A
And
so
what
it
says
for
that
job,
I'm
not
going
to
use
storage,
but
but
multiple
jobs
may
need
to
pass
information
back
and
forth.
So
we
support
storage
between
the
different
jobs,
I'm
telling
it
the
image
I
want
to
use,
and
so
I've
got
an
image
in
my
my
container
registry
that
has
the
address
CLI
built
in
for
it
and
that'll.
A
And
we'll
do
a
similar
thing
here:
I'll
paste
this
in
and
briefly
explain
it
so
again,
I'm
using
a
container
that
has
the
helm
CLI
in
it
and
I'm
doing
a
helm,
upgrade
install
that'll
take
care
of
since
the
the
app
isn't
already
there
it'll
do
a
fresh,
install
pointing
to
the
chart.
That's
in
the
same
repo
and
again,
all
I'm
doing
is
dynamically
setting
the
image
tag
so
that
we'll
get
this
new
to
point
or
this
new
image
that
we
just
built,
and
so
that
part
will
take
care
of
for
us.
Now.
A
At
that
point,
our
pipeline
is
done
run
each
says,
run
them
in
order.
If
anyone
any
one
step
fails,
then
the
whole
thing
is
kind
of
gone
wrong
and
stop
an
error
out,
and
so
we're
done
from
a
standpoint
of
our
defining
our
pipeline.
We
do
also
need
to
setup
the
web
hook
and
so
I've
got
some
quick
commands
to
make
this
easier
for
us
again.
You'll
see
that
we
we
had
this
service
earlier
and
you
can
see.
A
Github
and
we're
gonna
get
JSON
and
I
need
a
secret,
and
this
is
actually
something
that
is
important
in
that
project
file.
We
passed
a
password,
that's
the
password
that
the
get
that
the
github
hook
is
going
to
say
to
my
to
the
Gateway
to
actually
execute,
and
so
we
need
to
actually
paste
that
in
here
and
we're
gonna
get
the
the
just
the
push
event.
You
can
see
it's
active
and
seems
to
be
working.
So
let's
go
ahead
and
push
our
code
ship.
It
right.
A
We
do
a
push
right
into
production
right
into
master,
that's
normal
right,
so
luckily
things
have
kicked
off
correctly
and
we
have
this
worker.
That's
that
started.
The
worker
is
actually
the
one
that's
going
through
my
JavaScript
file
that
we
just
wrote
together
and
executing
those
steps,
and
hopefully
what
it's
going
to
do
is
actually
kick
off
these
these
job
runners
and
again,
the
reason
it's
called
job
Runner
you
saw
where
we
actually
put
in
that
name.
We
actually
see
that
this
thing
is
already
running.
A
If
we
take
a
look
at
the
logs,
it's
logged
me
into
my
subscription.
You
know
actually
go
out
and
start
performing
the
different
steps
and
going
through
and
and
running
through
that
pipeline.
In
this
this
is
probably
the
one
that
takes
the
longest.
This
is
where
I'll
start
to
do.
Do
things
to
distract
you
while
you're
waiting
for
that
to
finish,
but
we
could
look
at
these
logs
and
that's
one
way
to
see
what's
going
on,
it's
probably
not
the
best
way
to
see
what's
going
on.
A
A
A
And
it
is
seven
seven,
four
five,
so
if
I
did
that
right
and
we've
added
cosh
T
and
once
that's
up
and
running
we'll
be
able
to
see
the
service
for
it
and
connect
to
it
and
I'll
give
it
a
minute.
You
can
see
I'm
waiting
for
a
public
IP
address
from
the
cloud
and,
of
course
cosh
T
is
now
up
and
running.
Well,
we
can
actually
hit
that
with
the
port
forward.
If
we're
not
willing
to
wait
and
I'm
not.
A
A
Since
we
just
did
this,
we
only
have
one
build
and
you
can
see
we're
still
in
the
middle
of
the
AC,
our
builder
job,
so
they
actually
go
in
and
take
a
look
at
the
logs
that
are
it's
going
on
here
and
we
could
scroll
down
and
you
can
see
we're
pretty
far
along
we've
actually
completed
the
the
container
registry
build,
so
the
build
is
actually
complete
since
we
built
it
in
the
cloud.
That's
where
we're
gonna
store
it
odds
are,
by
the
time
I
go
back
and
take
a
look
at
this.
A
We've
actually
completed
the
helm
step,
because
that
goes
pretty
fast.
Let's
go
over
and
take
a
look.
You
can
see.
We
now
have
a
webapp
application
in
container,
creating
and
again
I'm
going
to
pop
out
of
this
and
see
if
we
have
a
service
for
that,
we're
waiting
for
the
IP
address
for
that
one.
So
cosh
T
is
a
cool
way
to
see
what's
happening
with
with
builds.
I
can
have
lots
of
different
projects
in
that
interface.
In
fact,
someone
out
in
the
community
actually
built
another
tool
for
me
to
see.
A
What's
happening,
I
love,
this
flock
I,
don't
know
who
slack
is,
but
he
made
a
brigade
term.
So
let's
actually
see
how
this
does
it's
kind
of
my
command
line,
version
of
cache
D.
So
look
at
this.
We
have
this
sort
of
terminal
interface
to
that
and
we
can
see
the
same
information
we
just
saw
over
in
cache.
D
I
can
actually
go
in
I,
see
a
successful
push
event.
We
can
see
the
two
jobs
that
are
run.
We
can
see
the
logs
from
the
helm
standpoint.
A
It
looks
like
it
actually
worked
and
we
can
actually
see
everything
that's
happening,
so
I
love,
seeing
that
brigade
term
again
we
can
look
through
the
logs
I,
have
lots
of
different
things
to
be
able
to
get
sort
of
visibility
into
what's
going
on
in
the
pipeline.
Now,
let's
see
if
I
have
an
IP
address
waiting
for
an
IP
address
and
we
will
continue
waiting
now,
while
we're
waiting
I
want
to
do
something
else
with
our
pipeline,
and
so
we
had
this
after
event.
A
So
my
Twitter
code
says
just
like
before
we
had
these
jobs
that
we're
doing
our
build
process
now,
I
have
a
have
a
container
that
that
will
do
Twitter
options
or
functions
so
I'm
gonna
actually
feed
into
it.
A
series
of
environment
variables
that
have
access
to
a
Twitter
application
and
then
I'm
going
to
tweet.
Then
a
live
tweet
from
Brigade
a
coop
con
and
we're
gonna
see
this
actually
work.
Now
notice,
I
have
a
bunch
of
project
details
here,
and
so,
when
we
made
the
project
at
the
beginning,
we
didn't
supply
these.
A
Go
here
at
the
end
of
this
project,
file
and
we'd
fill
in
these
values
and
again
I've
already
done
this
kind
of
for
us,
but
we'd
put
we'd
go
over
to
Twitter
and
collect
those
from
our
Twitter
application
and
from
there
we
have
to
do
a
helm,
a
helm
upgrade
for
our
project.
So
this
is
actually
a
nice
thing
because
I,
because
I
have
this
ability
to
have
these
values
in
a
project
file.
A
It's
very
easy
for
me
to
add
things
to
the
project:
I,
don't
really
changed
anything
of
the
current
functionality,
I'm,
just
adding
the
capability
to
Twitter,
and
so
now
we
plug
those
things
into
the
into
the
project.
When
we
execute
this
pipeline,
we're
now
going
to
be
able
to
tweet
live
and
we're
hopefully
going
to
see
that
actually
happen
here
on
my
Twitter
when
we
do
this
and
so
I've
saved
that
code,
I
do
need
to
show
you
the
webpage.
Let's
see
if
we
have
an
IP
address,
we
do.
A
A
We
have
a
good
bit
of
votes
coming
in,
but
the
reason
the
color
is
a
bit
terrible
is
I,
don't
know
if
you
know
this
I'd
actually
did
this
a
little
bit
of
a
similar
thing
at
coop
Conant
in
in
North
America
in
December,
you
can
actually
put
HTML
colors,
you
can
put
funny
names
in
there
and
so
let's
actually
see
what
the
name
is
representing
this.
Let's
go
into
our
app
into
the
web
app
here
and
take
a
look
at
our
HTML
and
it
actually
looks
like
that.
A
Color
is
steven
seagal,
sorry
for
the
steven
seagal
fans,
but
that
is
a
bad
color
and
what
would
be
the
color?
We
would
replace
Steven
Seagal
with
Chuck
Norris.
Of
course
we
would
so
we've
not
only
changed
the
color
we've
updated
our
pipeline
to
tweets
I.
Think
we're
ready
to
to
push
this
and
again.
We
will
ship
it
and
push
this
out
and
we'll
see
this
actually
execute.
Go
back
over
here
and
we're
gonna
see
the
same
steps
execute
here
again.
A
We
will
actually
end
up
kicking
off
another
job
to
do
the
Twitter
part
of
this,
and
if
we
go
over
to
my
Twitter
page,
let's
just
make
sure
we
don't
have
any
tweets
there
already.
You
can
imagine.
I
probably
tested
this
a
couple
times
so
I
tweeted
about
Jeremy's
session
on
service
broker
earlier,
and
that's
pretty
cool
to
see
that
there,
but
I
don't
have
any
recent
tweets.
So
when
this
thing
finishes,
we're
gonna
see
that
there
and
and
we'll
wait
for
that
to
actually
finish
in
the
meantime.
A
I
do
want
to
make
one
other
point.
The
the
the
pipeline
we're
doing
here
with
brigade
is
all
about
github.
The
github
is
our
is
our
gateway
again
I
mentioned
at
the
beginning.
Brigade
is
a
lot
more
than
CI
CD.
We
can
kind
of
perform
any
kind
of
function
here.
As
far
as
a
series
of
steps
we
might
need
to
won't
want
to
create
some
kind
of
custom
gateway,
and
certainly
you
can
do
anything
you
want.
A
You
can
take
the
gateway
code
for
those,
it's
all
open
source
and
build
your
own
now,
but
we
did
want
to
make
it
easier
for
you
to
actually
build
a
brigade
pipeline
and
when
I
say
we
I
actually
mean
Matt
butcher
on
the
team
that
actually
did
a
lot
of
work.
Around
brigade
rarity
a
blog
post
around
this
and
I.
A
A
And
all
we
have
to
do
to
do
this,
it's
quite
simple:
I!
Do
a
draft
create
and
I
tell
it
this
sort
of
draft
pack
that
I
want,
and
so
that's
where
in
my
directory
here
is
empty,
we
obviously
just
created
it.
That's
where
the
intelligence
actually
comes
in
I've
actually
pointed
that
and
set
that
up
on
my
sheet,
my
machine
so
that
when
we
open
vs
code
here
we
see
a
skeleton
project
here
to
create
a
brigade
gateway.
A
So
this
is
a
great
place
to
get
started
this
one's
in
written
in
JavaScript.
They
can
be
written
in
any
language,
and
this
one
uses
a
sort
of
web
hook
kind
of
functionality.
You
don't
actually
have
to
do
that
yourself.
You
could
be
again
monitoring
things
like
a
queue
doing
whatever
you
need.
This
is
a
great
way
to
get
started.
I'm,
not
real
good
at
file,
new
empty
projects,
and
so
I
like
to
take
advantage
of
this.
This
might
give
me
a
chance
to
actually
create
a
gateway
myself.
A
All
I
need
to
do
to
actually
run.
This
thing
is
I,
say
draft
up
so
draft
us
and
things
for
me
and
I
and
I
actually
suspect
this
will
fail
and
it
doesn't
really
matter,
but
what
it
does
is
it
first
builds
the
image
for
me
and
then
it
pushes
it
into
a
container
registry
and
I.
Have
it
set
up
to
push
it
into
Azure
container
registry
and
I'm?
Actually,
I
may
not
be
logged
into
Azure
container
registry,
which
thankfully
it
doesn't
allow
me
to
push
that
there.
A
Perhaps
I
am
logged
in
or
not
it'll
be
fun
to
find
out,
but
what
it'll
also
do
is
push
this
this
into
my
kubernetes
cluster.
Whichever
one
is
my
current
context,
that's
super
simple
for
me
to
actually
get
this
thing
up
and
running,
get
it
get.
It
started.
Curl
the
endpoint
understand
how
it
works
and
then
start
coding,
my
gateway,
changing
it
building
it
and
then
you
all
you
do
is
that
same
process.
Draft
up,
run
it
again
and
it'll.
Take
care
of
this
part
for
me,
which
is
the
painful
part,
all
right
makes
sense.
A
A
A
That
looks
successful,
that's
exciting
bells
right.
We
live
tweeted
right
here
with
Brigade.
Now,
let's
just
go
see
if
Chuck
Norris
saved
the
day,
Chuck
Norris
did
save
the
day.
It
already
looks
a
bit
better,
that's
exciting!
So
we're
successful,
very
good
pair
programming.
We
did
it
alright,
so
again,
I
think
it's
best
to
learn
kind
of
building
into
n.
That
was
our
goal
here
today
we
didn't
have
any
catastrophic
failures.
I
will
take
questions.
It
may
be
easier
to
take
questions
kind
of
on
the
side.
A
There's
a
short
transition
time
here,
so
I'll,
probably
just
wrap
up.
You
can
find
me.
That's
pronounced
cheeseburger
71
again.
I
will
explain
that
later.
If
you
want
I'll
be
working
at
the
azure
booth.
A
bunch
of
us
are
there
to
talk
about
these
things
and
many
more
things
of
our
Azure
cloud.
You
can
email
me
as
sue.
That's
my
work,
email
address.
I
did
put
the
demo
code
out
in
github.
It's
mostly
explained
how
to
do
it
I.
You
know.