►
From YouTube: DevOps For ASP.NET Developers Pt.6 - Release Pipelines
Description
After your application has gone through a build and has been tested by QA, the next step is to deploy it. This may be as simple as copying your build artifact to a web server. This could also mean publishing your build to multiple targets such a server cluster or a mobile application marketplace.
In part 6 of o ur series, Abel and Jeremy show us how to get our application builds deployed using Azure DevOps. They discuss incorporating approvals, creating custom gates, and even how to deploy your build behind your company firewall
[T01:24] - Reviewing the release pipeline
[04:12] - Available pipeline tasks
[06:37] - Working with build artifacts
[10:28] - Incorporating database changes
[14:39] - Using automated approval gates
[15:57] - How the Azure DevOps teams uses gates
[18:30] - Creating custom gates using HTTP endpoints
[22:24] - Setting up parallel deployments in a pipeline
[23:50] - Deploying behind your company firewall
DevOps for ASP.NET Core ebook https://docs.microsoft.com/aspnet/core/azure/devops?WT.mc_id=devops-channel9-cephilli
Use approvals and gates to controls your deployment
https://docs.microsoft.com/en-us/azure/devops/pipelines/release/deploy-using-approvals?view=azure-devops&WT.mc_id=TWC9-c9-cxa
Create a multiple platform pipeline
https://docs.microsoft.com/en-us/azure/devops/pipelines/get-started-multiplatform?view=azure-devops&WT.mc_id=TWC9-c9-cxa
Start with Azure DevOps for Free: http://dev.azure.com
Create a Free Azure Account: https://aka.ms/c9-azurefree
A
Hey
everyone
Donovan
Brown,
here
welcome
to
the
DevOps
for
asp.net
developer
series.
In
this
episode,
we're
going
to
focus
on
advance
to
release
topics
with
release.
You
can
do
more
than
just
automate
the
deployment
of
your
code.
You
can
also
run
integration
tests,
security
scans
or
deploy
your
containers.
Let's
see
how
Abel
and
Jeremy
add
these
advanced
features
for
their
pipeline.
B
Alright
Abel
I'm
super
excited.
After
our
last
talk
about
unit
tests,
I
integrated
some
unit
tests
updated
them
actually
found
some
defects
that
wouldn't
have
been
caught
till
later,
and
so
the
codes
built
gets
pristine.
It's
beautiful,
but
it's
not
running
anywhere.
I
know
it's
good
I
know
it
compiles.
I
know
my
test
run.
What's
the
next
step,
you.
C
Need
a
deployment
pipeline
right,
you
need
to
be
able
to
deploy
your
code
into
production
somewhere
or
into
someplace
sure.
So,
luckily,
for
you,
as
your
DevOps,
is
here
to
help
you,
you
can
use
that
your
job
ops,
to
deploy
your
code
anywhere
anywhere
into
Azure,
of
course,
and
you
can
deploy
your
code
behind
your
firewall
in
front
of
your
firewall,
someone
else's
cloud,
don't
use
someone
else,
use
our
cloud,
but
the
tooling
literally
doesn't
care
right,
we'll
let
you
deploy
anywhere.
However,
you
want
for
any
language
as
well,
not
just
for
dotnet,
okay
yeah.
C
So
if
we
jump
to
my
screen,
what
you
see
here
is
my
deployment
pipeline.
So
when
you
set
these
things
up,
what
you
need
to
do
is
you
first,
you
need
to
decide
what
stages
do
I
have
right.
So
in
this
case
here
I
have
a
dub
stage,
a
QA
stage
and
a
prod
stage,
pretty
simple.
Okay.
So
after
you
figure
out
what
stages
you
have,
then
you
can
go
into
each
stage
and
from
here
you
start
adding
the
tasks
that
you
want
to
run
to
deploy
your
application
into
that
particular
stage.
So.
B
C
B
C
A
B
C
B
C
B
C
Infrastructure
as
code,
so
in
my
arm,
template
I
basically
describe
exactly
how
I
want
my
infrastructure.
You
look
like
and
then
in
this
particular
stage.
That
first
step
is
it
takes
that
arm
templates
goes
to
Azure
and
creates
for
me
that
exact
environment
that's
described
in
my
arm,
template
or
if
it
already
exists,
it
doesn't
do
it
back
or
if
it
exists,
but
it
looks
a
little
bit
different.
It
will
tweak
it.
So
it
looks
exactly
like
how
my
arm
template
describes
it.
Okay,.
C
Into
it,
okay
right,
so
it's
super
simple
to
do.
Out-Of-The-Box
you
get
hundreds
of
tasks
that
you
can
just
start
using
right
and
and
you'll
start
seeing
man.
This
is
for,
like
all
languages,
all
different
platforms,
whether
it's
the
Apple
App,
Store,
Amazon,
Cloud.
Of
course,
a
sure
stuff.
There's
gonna
be
a
ton
of
Azure
stuff.
So,
basically
anything
you
want
to
do
with.
Azure
is
already
pre-built
and
you're
ready
just
to
rock
and
roll
okay.
C
If
you
want
to
do
something
that
doesn't
exist,
out-of-the-box
jump
into
our
marketplace,
so
our
marketplace
now
has
I
think
over
nine
hundred
tasks
like
build
and
release
tasks
that
our
apartment's
of
creating
so
those
you
can
just
download
and
start
using
and
they're
for
you.
If
you
scroll
through
this
list,
it's
massive
right
and
it's
just
for
all
sorts
of
stuff,
whatever
you
can
think
of,
there's
it
probably
it's
out
there.
If.
C
Way,
you
can
write
your
own
custom
tasks
as
well.
Okay,
so
a
custom
task
is
PowerShell
or
no
all
right,
which
means
whether
you're
on
a
Windows
system
or
a
Linux
or
Mac
OS
system.
It
doesn't
matter
right.
You
you'll
be
able
to
write
a
task
that
does
whatever
you
need
it
to
do
and
if
you
really
think
about
it,
if
it's
node
or
PowerShell
I
mean
basically
anything
from
the
command
line,
I
can
write.
C
It
becomes
it's
really
easy
to
get
what
you
want
done
done
so,
but
I
always
tell
people
whether
it's
for
your
build
or
whether
it's
for
your
deployment,
if
you
don't
know
how
to
set
up
the
pipeline,
ask
yourself:
how
would
I
do
this
from
the
command
line
and
what
are
the
steps
that
I'm
going
to
do
because
most
likely
you'll
find
a
corresponding
task?
That
does
those
steps,
and
if
you
don't
have
that
task,
you
can
write
one
okay,
so
easy
enough
to
customize
this
and
make
this
do
whatever
you
need
it
to
do.
A
C
C
C
We've
created
our
environments,
we've
created
our
tasks
that
will
run
in
each
of
the
environments.
The
other
thing
that
you
need
to
decide
is:
where
are
your
artifacts
that
you
wanted
to
poi?
Where
are
those
coming
from
all
right?
So
these
are
gonna,
be
your
build
artifacts.
If
we
go
ahead
and
click
on
add
you'll
be
able
to
see
so.
B
C
C
B
C
The
output
of
your
build
is
stored
back
into
Azure
or
at
your
dev
ops.
It's
actually
blob
storage,
but
it's
it's
all
we
need
to
know
is
it's
it's
held
inside
of
at
your
dev
ops
and
from
there.
If
I
select
a
build,
I'm
able
to
pick
which
project
and
where
which
build
I
want
to
use
right
and
then
it
will
go
ahead
and
grab
that
particular
asset
has
my
build.
B
C
C
B
C
C
B
C
Cares
it'll
be
fine
right,
alright,
so
after
it's
done,
I
sit
here,
I
do
have
a
post
appointment
or
over
I'm
waiting
for
someone
to
actually
say
you
know
what
dev
looks
good,
let's
go
ahead
and
dump
it
to
QA.
Okay,
I've
also
added
a
pre-deployment
improver
for
my
QA
environment,
because
maybe
they're
in
the
middle
of
testing
that
are
not
gonna
be
happy.
If
I
just
blast
new
code
right
in
the
middle
of
their.
B
C
I'm
giving
them
the
option
saying
now
we're
ready
and
then
it'll
go.
The
bits
will
immediately
flow.
Okay
and
I
also
have
a
post
deployment
approver
as
well.
So
the
post
deployment
approval.
It
is
kind
of
cool,
because
when
your
deployment
runs,
you
get
the
choice
of
saying
this
is
great,
go
ahead
and
pass
it
on
through
and
it
flows
on
to
the
next
environment
right
and
then
the
next
environment.
If
you
have
a
pre
deployment
approver,
you
have
the
ability
to
approve
the
deployment
but
schedule
when
you
want
the
deployment
to
happen
as
well.
C
Okay,
so
you
can
say
yes,
I
approve
it
to
do
be
deployed
into
QA
and
let's
go
ahead
and
do
it
immediately
right
or
you
can
say,
I
do
approve
it,
but
let's
wait
until
2:00
in
the
morning
or
something
like
that
sure.
So
you
have
the
ability.
Did
you
do
that
as
well?
So
basically
that
right
there
in
a
nutshell,
is
your
deployment
pipeline
define
what
your
stages
are
define?
What
task
will
run
inside
of
each
stage
and
then
go
ahead
and
define
your
approvers
before
and
after
each
after
each
stage
as
well.
B
C
C
B
C
C
Basically,
you
have
a
tool
that
encapsulate
your
schema,
the
changes
that
you
make
to
your
schema,
so
you
can
use
SS,
DT,
sequel,
server,
data
tools,
I
think
that
is
Microsoft's
all
of
that,
and
that
literally
encapsulate
the
scheme
of
your
database.
You
make
a
change
to
it
and
and
it
changes
the
SS
DT
project,
part
of
the
build
steps.
It's
really
cool,
because
an
SS
DT
project
is
just
part
of
your.
Your
visual
studio
solution,
part
of
your
build
steps
when
it
builds
an
SS
DT
project.
C
B
B
C
B
C
Can
say
what
is
the
difference
between
the
schema
that's
described
in
my
backpack
versus
the
schema
that's
described
in
in
my
def
database,
and
it
will
say
these
things
are
different,
and
these
are
the
scripts
that
you
need
to
run
to
change
your
def
database.
So
it
looks
exactly
like
your
schema.
Got
your
deck
back
and
then
there's
another
task
that
will
just
go
ahead
and
run
that
script.
So
it
changes
your
database
schema
to
match
what
you
think
it
is
so.
B
B
C
Has
some
database
tools
that
are
absolutely
fantastic
as
well
so
SSD
T
is
basically
state
based
right.
So
this
is
the
state
of
your
data.
This
is
the
state
of
your
database.
What
it
should
be-
and
here
are
the
auto
Jennifer
you
the
scripts
that
will
change
it
from
this
particular
state
to
the
state
that
it
needs
to
be
at
okay,
so
a
Red
Gate.
C
It
does
have
state
based
as
well,
but
they
also
do
versioning
right
from
one
version
to
the
next
to
the
next
and
I
really
like
that,
because
you
have
super
fine
game
control
over
exactly
what
happens.
So,
basically,
it's
like
okay,
I,
add
a
column
so
from
version
zero
there'll
be
no
column
from
version.
One
now
of
my
database,
I'm
going
to
add
a
column,
and
so
these
are
the
scripts.
I
have
to
run.
That
goes
from
version,
one
to
version
two.
C
B
C
B
A
sequel
aspect
for
what
makes
sense
for
that
and
then
a
no
sequel
aspect:
yeah
yep,
but
it's
good
to
know
that
that
can
all
go
as
part
of
this
pipeline.
Now
you
showed
me
some,
you
can
just
keep
kind
of
the
end
of
the
gate
open
or
you
can
have
like
this
manual
step.
Is
there
any
way
to
automate
approvals?
Actually.
C
And
after
each
environment,
you
can
use
these
automated
approval
gates,
which
uses
automation
and
AI
to
help
you
decide,
should
this
gate
pass
or
fail?
Okay.
So
to
give
you
a
quick
example:
let's
go
ahead
and
jump
in
and
look
and
see
what
that
looks
like
out
of
the
box.
You
get
I
think
five
different
gates
that
you
can
just
start
using.
Okay,
so
we'll
work
from
the
ground
up
from
the
bottom
up.
The
first
one
is
the
security
compliance
assessment
gate,
and
this
is
super
cool
right.
C
B
C
C
B
C
C
B
A
B
C
We
literally
use
this
every
single
day
to
not
only
build
as
your
DevOps
but
to
deploy
actually
DevOps
as
well.
What's
what's
kind
of
nice,
is
we
get
to
use
our
software
engineering
team
as
our
first
line
of
testers
right?
So
the
first
thing
that
we
do
when
we,
when
we
were
ready
for
a
new
release,
is
we
release
it
to
internally
to
us
right.
B
C
Then
we
sit
there
for
like
48
hours
and
we're
all
using
it
in
our
day-to-day
life,
and
this
is
what
we
use
to
manage
all
of
our
work
and
if
we
run
into
bugs
or
if
you
know
the
performance
is
down
we're
using
telemetry
for
everything,
we're
gathering
telemetry.
So
if
the
telemetry
is
off
or
whatever
you
BAM,
it
will
just
automatically
stop
as
well
nice.
C
B
C
You
can
say
you
know
what,
if
the
response
time
is
more
than
XYZ,
then
throw
alerts,
and
if
there
are
enough
alerts,
you
know
we're
gonna
block
this,
because
the
performance
is
just
too
bad
at
it
right,
very,
very
powerful.
So
the
last
gate
is
the
invoke
REST
API
and
the
Evoque
as
your
functions.
So
this
is
the
extensibility
point.
You
can
write
an
azure
function
or
a
REST
API
that
will
get
called
by
the
gate,
which
means
this
is
just
custom
code
right.
So.
A
B
C
So
two
gates
that
I've
written
right-
it's
right
off
the
bat
two
gates
that
I've
written
I,
was
at
a
hospital
that
had
a
horrible
role
well,
for
whatever
reasons
they
have
interesting
rules
as
its
healthcare,
you
literally
could
not
deploy
into
production
without
a
document
that
was
signed
uploaded
to
their
DocuSign
server.
Okay,
so
this
became
a
bottleneck
in
the
deployment
was
first
we
have
to
track
down
the
doctor
and
be
like
dear.
C
Do
you
need
to
sign
this
paper
and
then
you
need
to
upload
it
and
then,
when
upload,
it
needs
to
be
verified
by
somebody
and
then
it
can
finally
move
on
and
then
sometimes
people
didn't
check
right.
They
just
said
eff
that
I'm
sick
of
waiting,
just
you
know,
pump
it
on
through,
and
then
you
get
a
huge
amount
of
trouble
once
I
hit
right
over
else,
it
becomes
involved,
but
the
big
thing
was:
it
was
just
a
bottleneck.
So
what
I
did
was
I
created
a
natural
function
that
checked
DocuSign
and
said:
hey?
C
A
C
So
when
you
create
a
gate,
you
set
up
the
polling
frequency
for
your
gate
and
you
also
set.
How
long
do
you
want
there
to
run
until
the
times
out
so
I
basically
said
this
is
demo
code,
so
I
said
run
for
every
five
minutes
and
then
let's
run
for
24
hours,
okay,
but
but
you
can
really
tweak
that
to
whatever
you
need
it
to
be
so
deployed
into
QA,
QA
tested.
It
looks
good,
they
approved
it,
I
immediately,
chucked
DocuSign
has
it's
been
signed
yet
nope,
another
five
minutes
later?
C
C
C
When
I
hit
my
gate
now,
I
start
asking
Donna
trace.
Are
we
given
these
parameters
that
I
want
tracked?
Are
we
good
and
they
told
me,
no
you're,
not
your
response.
Time
is
way
too
slow
bad
code
never
reaches
production,
so
you
can
make
these
gates
literally,
do
whatever
you
want.
If
you
go
to
the
marketplace,
you
can
see
where
our
partners
have
created
automatic
gates
as
well,
that
you
can
just
download
and
start
using
yourself
now.
B
C
But
yeah
it's
super
super
easy
to
do.
It's
literally
just
like
you
need
to
install
the
the
azure
pipeline's
extension
for
github
and
then
from
there
you
select
which
repo
you
want
and
you're
in
business
nice.
Then
you
drop
into
this
UI
and
the
way
you
go,
that's
great
yeah.
The
other
thing
that
I
wanted
to
show
you.
The
last
thing
that
I
wanted
to
show
you
is
you
can
make
these
pipelines
as
complicated
or
as
simple
as
you
want
them
to
be
right.
C
So
the
first
pipeline
that
we
showed
is
pretty
simple
that
we're
just
deploying
the
web
front-end
boom
boom
boom
yep,
but
remember
the
hotel,
360
app
is
a
lot
more
complex
than
just
a
web.
Front-End
its
you've
got
a
web
front-end,
that's
a
dotnet
core
app,
and
then
you
have
an
API
layer
as
well.
That
needs
to
run
inside
of
a
kubernetes
cluster
and
there's
also
a
mobile
portion
of
that
as
well
right.
C
C
That's
we've
already
seen
exactly
what
it
does
another
for
my
API
layer,
where
I'm
deploying
into
a
kubernetes
cluster
I'm
using
helm,
to
deploy
my
helm
application
into
your
kubernetes
cluster
and
then
finally,
I
also
have
my
mobile
portion,
where
I
use
App
Center
to
build,
deploy
to
my
alpha
testers,
beta
testers
and
eventually,
all
the
way
out,
even
into
the
store
itself
oops
wrong
test.
Here
we
go
right
all
the
way
into
the
Apple
App
Store
itself,
nice.
So
these
pipelines,
all
of
our
demos.
C
We
show
just
like
playing
simple
demos,
but
you
can
make
these
pipelines
literally,
do
anything
that
you
need
to
in
front
of
your
firewall
behind
your
firewall.
So
a
lot
of
times,
people
ask
me:
how
do
you
deploy
behind
your
Faro
behind
my
firewall
right?
You
need
to
install
our
agent
behind
your
firewall,
so
this
agent,
it
needs
to
be
able
to
see
this
agent
needs
to
be
able
to
reach
out
to
Azure
DevOps
the
product
and
it
needs
to
be
able
to
reach
whatever
machines.
C
B
C
B
C
That
you're
done,
that's
literally
it
well.
You
answer
a
couple
questions
and
then
you're
done
right
right.
So
it's
super
super
easy
to
do
in
front
of
your
firewall
behind
your
firewall.
If
you're,
if
you're
going
into
the
cloud,
you
don't
even
need
to
have
your
own
agent,
you
can
just
use
our
hosted
agents
as
well.
B
Right
so
we
have
hosted
agents,
but
in
the
case
I
was
at
a
previous
company
that
had
a
dedicated
build
box.
They
spent
the
money
on
it.
You
wanted
to
use
it,
so
we
orchestrated
everything
in
the
cloud,
but
the
actual
build
took
place
on
this
one
machine,
so
we
were
able
to
utilize
the
resource,
but
still
take
advantage
of
the
flexibility.
Absolutely.