►
From YouTube: Spinnaker Success Stories - Fi... AbdulBasit Kabir & Grace Akinsola, Balaji Siva & Marc Andre Vanier
Description
Spinnaker Success Stories - Fireside Chat with Interswitch & a large Telecom Company - AbdulBasit Kabir & Grace Akinsola, Interswitch; Balaji Siva, OpsMx & Marc Andre Vanier, Telecom Company
Enterprises are looking to deliver releases faster and safer. Spinnaker has emerged as a leading tool to provide scalable multi-cloud deployments among these Enterprises. Spinnaker allows the ability to enable self-service, control (through policy integration), and deploy safely with automated verification. In this session, we will talk to InterSwitch and a Large Telecom Company in North America that use Spinnaker in production, and learn about their use cases, best practices, and lessons learned.
For more Continuous Delivery Foundation content, check out our blog: https://cd.foundation/blog/
A
Hi
everyone
welcome
to
the
cdcon
spinnaker
success
story
session,
fire
chat
with
interswitch
and
the
large
telecom
provider.
My
name
is
balaja
siva.
I
am
a
vpf
product
at
opsa
max
for
people
not
familiar
with
ops
max.
We
provide
enterprise,
spinnaker
solutions
for
customers
and
I'm
I'm
happy
to
have
a
great
panel.
Today
we
have
grace
akinsola
and
abdul
bassett
kabir
from
interswitch
and
mark
andre
vanir,
who
is
freelancing
at
a
large
telecom
company.
B
Hello:
everyone,
my
name
is
grace,
I'm
speaking
to
you
from
lagos,
nigeria.
I
work
with
insurance
as
a
program
manager
on
the
infrastructure
transformation
program.
I've
had
some
experiences
doing
software
testing,
submission
engineering
and
as
well
as
solution
mastery
in
the
past,
but
I'm
here
now
today
for
the
program
manager.
Thanks
for
having
me.
C
A
D
A
A
B
B
So
that's
what
we
are
currently
doing
and
also
we
are
trying
to
also
implement
the
canary
analysis
out
of
the
pipeline.
So
we've
been
able
to
do
different
integrations
from
stash
deep
pocket
jprogs
on
the
cube
jenkins.
So
we
have
moved
past
that
initial
stage
and
the
other
stage
we're
trying
to
build
a
staging
environment
into
the
pipeline
and
also
put
in
the
canary
analysis
and
deployment
path.
So
that's
where
we
have.
A
Got
it
if
you
have
a
follow-up
question
like
what
is
there?
I
don't
know
you
didn't
get
too
much
into
interest
switches,
it's
a
payment
company
out
of
nigeria.
What
what
is
there?
What
is
your
you
know
how
many
software
engineers
are
there?
You
know
what
what
sort
of
driving
the
the
spinnaker
usage
you
know
at
in
interest
rates.
B
Our
software
team,
we
have
about
android
developers,
qas
and
like
10,
develop
engineers,
so
we
have
like
a
very
large
scene,
and
so
I'm
using
spinnaker
to
orchestrate
our
deployment
is
something
that
I
believe
would
be
beneficial
to
enlarge
the
audience
and
help
us
a
little
bit
faster.
So
that's
kind.
A
D
D
A
C
Okay,
to
answer
the
question
of
why,
let
me
take
a
few
steps
back,
so
indoor
switch
has
been
existing
for
since
the
early
2000s,
and
you
know
in
the
early
2000s
there's
nothing
like
all
the
hype,
other
micro
services,
containerization
kubernetes
and
all
that.
So,
of
course,
we
started
our
software
engineering
journey.
The
traditional
way
the
monolithic
way
copy
into
server,
then
ssh
into
servers.
Then
ci
became
ci
on
automation
became
popular
that
was
before
the
terms
devops
srs
infrastructure,
engineers,
platform
engineers
before
all
those
became
popular.
C
So
we
have
our
clusters
in
several
places.
For
some
compliance
reasons,
there
could
be
some
some
clusters
that
don't
can't
be,
or
maybe
on
the
cloud
or
can't
be
in
a
particular
region
or
whatnot
and
having
having
such
a
system.
If
you
don't
have,
if
you
don't
have
one
tool
that
can
streamline
your
process
end
to
end,
then
it
will
be
there'll,
be
a
lot
of
you
know,
tribal
knowledge.
A
A
A
D
A
No,
it
makes
sense,
I
think,
a
lot
of
the
sort
of
the
previous
stories
like
you
know
you
fire
and
forget.
Right
I
mean
you
can
start
a
script
or
you
can
do
something,
but
you
don't
know
what
happened
to
it.
The
dev
doesn't
know
what
happened
to
the
production,
whether
it's
running
in
production
or
there's,
no
end-to-end
visibility.
A
A
lot
of
the
things
you
know,
spinnaker
really
provides,
and
I
see
I
see
other
customers
also,
like
you
know,
using
it.
For
some
similar
reasons.
That's
great,
maybe
go
back
to
a
little
bit.
You
know
dive
deeper.
What
are
the
some
of
the
use
cases
abdul?
Maybe
you
can
talk
on
for
the
interswitch
that
you
I
think
grace
mentioned
about
kubernetes
and
maybe
monoliths.
C
What
we
had
was
the
change
management
process
and
a
lot
of
different
ways
of
taking
the
checkbox
or
for
the
exchange
management
process.
The
our
like,
for
example,
the
every
deployment
needs
to
go
through
some
uat
environment,
some
staging
environment
before
you
get
into
production
environment
so
which
means
that
for
juventus
application
we
had
multiple
kubernetes
clusters
and
each
cluster
has
its
own
way
of
deploying
some
urine
cube
ctrl
from
a
command
line.
Song
jenkins
does
cube,
ctl
runs
a
cube,
ctrl
command
and
the
likes
for
others.
C
You
needed
to
you
know,
deploy
so
like,
for
example,
production
deployments
shouldn't
happen.
Ideally
I
mean
every
environment
production
shouldn't
deploy
directly.
There
should
be
at
least
some
canary
deployments,
and
you
know
putting
all
those
together
together
with
different
scripts
and
different
manual
stages,
so
part
of
the
use
case
we
had
for
spinnaker.
Was
you
know
getting
all
that
serialized?
C
This
stage
happens
before
the
next
stage.
This
environment
needs
to
be
up
and
all
that
another
use
case,
which
for
me
is
one
of
the
most
exciting
use
cases
is
automating.
Our
change
request,
ticketing
process,
of
course,
having
the
change
process
having
a
change
management
process
means
that
you
have
to
have
some
ticketing
records
and
whatnot,
and
if,
if
you
have
the
ticketing
system,
then
there
would
most
likely
be
approval
on
it.
C
B
C
A
Yeah,
I
think
I
think
you
know
you
are
hitting
up
on
some
things
that
you
know
that
we
adopt
some
actually
care
about,
which
is
one
is
obviously
this
as
a
cicd
tool
right,
you,
you
have
automation
of
the
workflow
right.
You
know
our
construction
workflow,
you
know,
take
an
artifact,
you
know
deploy
to
kubernetes
or
something
like
that:
roll
back
roll
forward,
but
there's
also
a
lot
of
process
elements
that
are
actually
stopping
the
pipeline
to
go
faster.
A
Like
approvaling,
you
know
somebody
manually
approving
something
checking
all
the
things
that
they
need
to
approve
the
check
before
they
approve
or
the
verification
you
talked
about,
like
you
know,
you're
deploying
a
kennedy
but
who's
doing
the
actual
evaluation
right.
If
you,
if
you
don't
automate
it,
then
you're
gonna
make
mistakes
and
there's
so
there's
also
policy
check.
I
think
you
know
some
other
things.
I
think
you
have
another
session
in
the
cdcon,
where
you
talk
about
a
policy
check
which
is
really
another
big
sort
of
hidden
things.
A
It's
not
part
of
the
ci
cd
tool.
Typically,
people
do
it
offline,
but
why
not
part
of
being
make
it
part
of
cicd
tool,
and
I
think
you
into
switch,
I
think,
is
a
great
example
of
leveraging.
You
know
really
getting
that
thing
out
using
screen
recorder.
So
I
really
like
your
use
case
and
mark.
I
think
you're
also
looking
to
do
something
similar,
but
let's
talk
about
before
you
get
there.
Let's
talk
about
your
specific
use
case.
D
Yeah,
so
mainly,
we
want
to
do
canary
deployment
in
order
to
to
measure
the
new
version
of
the
of
the
deployment
with
the
previous
one,
and
we
had
to
adapt
our
services
in
order
to
expose
those
those
matrix
and
find
which
matrix
we
want.
We
want
to
track
for,
for
those
canaries
and
also
the
notification
part
of
it
is
really
better
than
when
we
used
to
have
previously.
So
when
a
deployment
fails,
we
have
a
notification
right
away.
D
D
We,
when
we
deploy
we're
confident
that
it
is
not
really
worse
than
the
previous
version
and
that
we
can
find
that
the
different
bugs
are
or
issues
really
fast,
instead
of
it
being
someone
who
used
the
application
or
the
portal
that
raised
an
incident.
And
then
you
have
to
work
under
pressure
to
react
to
what
you've
deployed
a
few
days
ago.
A
You
got
it
that
makes
sense,
I
mean
definitely
the
safe
deployments
ability
to
see
what's
happening.
You
know
the
auditability
of
it
traceability
of
the
issues
and
those
things
are
very,
very
definitely
possible
with
spinnaker.
Maybe
grace.
Do
you
have
anything
else
to
add
around
the
current
use
cases
we
could
talk
about
future
use
cases,
but
do
you
want
to
have
anything
to
do
about
the
current
use
cases
from
your
site.
B
A
Okay,
all
right
so
maybe
maybe
we
can
switch
to
sort
of
the
next
topic
or
not.
So
what
are
the
benefits
you
have
seen?
I
know
you've
talked
about
some
other
things
already.
You
know
in
terms
of
some
of
the
process.
Is
there
a
tangible
benefit?
I
mean,
have
you
guys
looked
at
what
was
the
time
it
took
to
before
now?
It's
it's
better
or
say
you
know,
like
you
know,
2x3
xyx
better
or
is
it
cost
saving?
A
So
how
do
you
and
somebody's
looking
at
the
audience,
obviously
is
interested
in
in
sort
of
knowing
hey?
You
know
what
is
this?
What
what
is
the
significance,
saving
or
significant
advantage?
I
have
and
there's
a
tangible
way
to
measure
it.
Do
you
guys
look
at
something
like
that
or
abdul
or
face
from
your
side.
B
I
was
able
to
learn
questions
on
canary
environment,
and
so
with
all
this
in
place,
it's
quite
obvious
that
we'll
be
able
to
really
cut
our
department
time
time
between
when
the
code
is
committed
and
the
developer
partnership
is
work.
So
when
we're
able
to
take
it
to
production,
we're
able
to
maybe
do
it
by
like
70
percent.
E
B
So
yeah,
so
that's
what
I
believe
we're
going
to
be
able
to
achieve
and,
of
course,
the
change
request
process
as
well
right
now,
it's
already
improved
before
now.
You
could
take
like
two
days
before
your
change
gets
approved
because
of
the
different
undergraduate
process,
but
now,
as
as
we
have
moved
into
the
pipeline,
I'm
saying
that
we'll
be
able
to
have
shorter
approval
life
cycles
within
the
pipeline.
So
yes,
so
everybody's
the
same.
A
C
Yeah,
so
so,
basically,
because
our
adoption
is
still
new,
that
is
not
something
that
you
know.
We've
focused
on
on
isometric
to
measure
see
you
what
happens
there,
but
one
thing
I
think
it's
a
bit
related
to
the
question.
One
one
thing
that
I
know
for
that
really
we're
already
seeing
is:
we
can
always
pinpoint
where
the
process
failed
and
since,
since
the
process
intent
has
been
serialized,
we
we
we
know,
we
know
where
our
bottlenecks
nexa,
we
know
what
is
failing.
Is
they
fail
on
uhc
is?
C
A
Got
it
that
yeah?
That's
that's
huge
right.
I
mean
that's
what
I
was
talking
earlier
about
end-to-end
visibility
for
dev
and
off,
so
you
can
really
see
where,
where
things
are
stuck
or
where
it's
failing
and
what's
what's
the
bottlenecks,
are
that's
useful
mark?
What
about
you,
what
are
the
tangible
benefits
you
have
seen
from
process
improvement
time,
cost
anything
that
you
can
sort
of
share.
D
Yeah,
it's
centralized
a
bit
of
the
view
of
the
the
clusters.
So
if
you
have
multiple
clusters,
you
you
will
see
which
deployment
is
deployed
in
which
cluster
with
you
will
see
live,
and
you
will
also
see
the
different
version
of
those
of
those
deployments.
You
can
see
logs
pod
logs
and
things
like
that.
So
it's
all
centralized
in
like
your
cd
slash
platform
application,
so
so
that
that
can
be
useful,
not
trying
to
switch
switch
application
for
application
in
order
to
diagnostic
to
do
a
diagnostic
and
what
goes
wrong.
D
Also,
the
templates
can
can
help
us
really
streamline
the
various
deployment
or
various
type
of
applications,
so
we
create
we
create
a
large
pipeline.
So
our
pipeline
has
like
30
steps,
something
like
that
with
approved
wall
and
with
with
some
conditions,
and
you
can
chain
them
also.
But
it's
really.
D
I
could
not
manage
hundreds
of
deployment
with
with
that
kind
of
step.
So
it's
really
it's
really
fine
and
we
also
use
a
configuration
as
code
also
with
the
json
and
with
the
the
spin
klee
also
in
order
to
to
manage
our
various
various
platform
or
environments,
because
we
have
two
different
installation,
one
for
patching
and
one
really
in
prod
to
to
to
do
the
real
work
and
yep.
It's
instant
feedback.
Also
that
when
it
fails,
it
sends
you
a
notification
when
any
validation
or
manual
judgment
is
required.
D
D
D
Maybe
if
you
are
leveraging
the
the
spring
expression
language
in
order
to
to
enhance
the
pipeline,
it
could
go
really
wrong.
So
it's
that's.
That's
one
thing.
Also.
Some
tasks
are
a
bit
harder
to
to
execute.
For
example,
we
are
doing
some
security
testing
on
on
some
on
some
components
to
extract
these
reports
and
try
to
make
it
available
to
into
the
pipeline
is
a
little
bit
more
difficult
than
some
other
tools
can
provide
and
yeah.
D
The
the
platform
also
need
need
the
maintenance,
something
there's
some
some
new
feature
that
you
want
to
have
so
it's
important
to
to
to
get
it
up
to
date.
Sometimes,
and
those
kind
of
ci
cd
tools
is
platform,
don't
really
get
updated
or
upgraded.
Often
so
I've
seen
that
there's
a
lot
of
improvement
from
version
to
version
on
spanish.
A
Got
it,
I
think
you
know
we
work
with
you
against
that
ops
max
with
you,
so
you
help
you
along
the
journey
with
your
with
your
challenges
for
sure
and
then
help
you
with
you
know
new
features
as
well
abdul.
You
want
to
share
quickly
any
any
any
couple
of
best
practices
or
suggestions
to
the
audience.
C
My
first,
my
first
suggestion
is:
if
you're
a
large
enterprise
and
you're
really
serious
about
your
spinach
adoption,
get
a
vendor
you.
You
can't
underestimate
the
value
you
get
from
from
having
a
vendor,
because
and
if
you're,
not
a
large
enterpriser,
and
maybe
they
have
a
project
for
that,
they
need
to
dedicate
some
time
to
sit
down
and
actually
learn
and
go
through
what
spinnaker
can
offer
without
a
vendor.
We've
been
using
spinnaker
for,
I
would
say
months
and
we
barely
scratched
the
surface
of
what
we
could
do
with
the
vendor.
C
We've
gone
way
ahead
of
what
what
you
can
do
and
yeah
just
just
like
what
mark
said.
Put
it
in
a
sandbox
environment,
put
it
in
a
test
environment
play
around
with
it
and
and
see
what
you
what
you
have
to
offer.
The
last
thing
is:
spanaka
is
a
bit
opinionated
on
how
things
should
be
does
so
for
a
good
reason.
The
same
way
cuba
entrance
is
opinionated
on
how
applications
should
be
deployed
and
should
come
up,
and
you
know
that
traffic
should
get
into
the
application
and
all
that
is
there
for
good
reason.
C
It's
basically
the
best
practices
from
the
different
organizations.
I've
been
doing
just
for
decades,
now
condensed
into
a
platform.
So
don't
you
could
if
you
want
to
choose
spanaka,
to
do
what
you
want
regarding
your
your
your
own,
your
own
process,
your
own
practices
that
are
not
in
line
with
the
good
practices,
but
I
would
say,
go
with
the
opinionated
with
the
opinionated
spiritual
best
practices.
A
Got
it
thank
you
in
the
last
minute,
or
so
maybe
a
quick
thing
of
what
are
you?
What
are
you
looking
forward
in
the
future
in
your
spinnaker
deployments
like?
What
are
you
maybe
like
a
very
quick
30
seconds
each
or
you
know
like
what?
What
you're
looking
forward
to
the
next,
let's
say
six
months
mark,
you
can
start.
D
Yeah
yeah:
let's
really
try
to
involve
more
more
people
in
in
the
organization
in
order
to
to
use
it
to
try
to
administrate
it
and
train
them
really
to
the
best.
What
can
spinnaker
do
for
them
that
they
maybe
don't
know
right
now,
also
add
more
validation,
so
is.
Is
it?
Is
there
any
vulnerabilities
or
things
like
that
that
can
be
spotted
or
contract
tests,
also
that
we
don't
have
actually
and
yeah
it's
really
to
complete
the
migration
move,
more
services
into
it
and
then
really
to
complete
the
full
continuous
delivery.
A
C
Yeah,
so
it's
very
similar
to
what
mark
said
and
spanaka
as
it
is,
is
a
great
tool
and
if
you
add
the
enterprise
features
from
a
vendor,
then
you
get
a
lot
more.
That
you
could
do
like
the
possibilities
are
a
lot
more
and
for
for
us
we
see
that
we
we
see
an
opportunity
where
our
I.t
governance
can
come
in
our
infrastructure
infrastructure
team
can
come
in
our
data
center
team
can
commit
our
infosec
and
security
team
can
come
in
and
a
lot
of
other
teams.
C
A
That's
right,
yeah,
the
talk
is
really
good.
I
had
a
review
of
that,
so
it's
really
good.
You
should,
if
you
haven't,
if
you
haven't
watched
it,
you
should
watch
it
again,
watch
it
at
a
later
time.
Thank
you.
Thank
you
for
the
panel.
I
think
we
are
out
of
time.
Jack
marks
thanks
absolute
thanks
grace
for
for
your
time
today,
and
hopefully
the
audience
got
something
out
of
it.
We
have
shared
some
information
about
about
the
contacts
so
before
you
reach
out
to
us.
B
E
F
E
E
I
think
thanks
it's
like
almost
eleven
o'clock,
for
you.
Thank
you
for
being
here
to
answer
the
q,
a
and
and
folks
who
want
to
reach
out
feel
free
to
reach
out
to
us.
I
think
the
contacts
were
provided
in
the
in
the
video
if
you
want
to
go
back
and
listen
to
it,
definitely
reach
out
those
and
happy
to
give
you
our
thoughts
on
things
that
we
talked
about
all
right.
Thank
you.
Thank
you.
Thank
you.
E
Yeah,
I
think
the
the
primary
challenge
we
see
with
spinnaker
is
essentially
sort
of
the
day
zero
day.
One
initially
like
the
ability
to
you
know,
bring
up
spinnaker
and
have
it
work
for
that
all
the
environments
requirements
around.
You
know
the
connections,
the
tooling,
all
the
you
know,
maybe
to
make
it
work
for
them,
like
that's
sort
of
the
day
zero
day,
one
problems
to
bring
it
process
all
the
process
that
you
need
to
have,
and
next
would
be,
I
think,
I
would
say,
is
sort
of
the
cell
service
enablement.