►
From YouTube: Buenos Aires 2019 | OpenShift @ Banco Galicia Gonzalo Gomez (Banco Galicia) | OpenShift Commons
Description
OpenShift @ Banco Galicia
Gonzalo Gomez (Banco Galicia)
Buenos Aires 2019
A
B
To
be
more
agile,
in
order
to
improve
our
customer
experience,
the
truth
is
that
customers
are
demanding
more
and
more
from
us.
We
understood
that
we
had
to
work
differently
with
a
different
technology
that,
at
the
same
time,
allows
a
culture
of
empowerment
and
agility
to
flow,
and
we
found
the
red
hat
these
types
of
characteristics,
and
that
is
why
we
chose
it
as
partners.
The.
D
B
Was
really
a
challenge
for
people
of
culture
and
eventually,
technology
a
bit?
What
we
were
looking
for
is
to
refound
the
way
of
doing
things
in
the
company,
with
a
a
more
collaborative
and
agile
culture,
of
being
able
to
fail
and
try
different
things,
and
the
truth
is
that
we
found
in
the
opensource
culture
of
redhat
an
excellent
complement
to
make
these
things
happen.
What
was.
A
B
C
E
B
Many
companies
tie
applications
to
particular
cloud
providers.
We.
What
What
we
wanted
to
do
is
take
a
flexibility
strategy,,
so
that
's,
where
we
found
that
openshift
allows
us
to
have
this
flexibility
and
that,
on
the
other.
Hand,
allows
us
to
be
more
efficient
in
costs,,
also
taking
this
path,,
but
we
see
that
we
are
going
to
reduce
up
to
20%
in
licenses
and
hardware.
E
D
The
financial
sector
are
beginning
to
understand
the
benefits
of
developing
a
digital
transformation.
Strategy.
Galicia
has
been
one
of
the
pioneers
in
Argentina.
At
Redhat,.
We
believe
that
open
banking
not
only
allows
services
to
be
offered
through
non-traditional
channels,,
but
also
fosters
technological
disruption.
By,
promoting
an
open
leadership
model
and
giving
employees
a
lot
of
prominence,.
We.
B
Are
convinced
that
this
new
innovative
ecosystem
with
an
improved
experience
or
my
channel
is
allowing
us
to
get
closer
to
younger
audiences
and
make
day-to-day
operations
easier.,
but
this
is
only
the
beginning,
because
the
truth
is
that
in
the
near
future,
we
hope
that
100%
of
our
transactions
and
operations
and
acquisition
of
new
products
can
be
done
completely.
Digitally.
F
F
The
video
summarizes
a
large
part
of
what
we
want
to
tell
you
today.
Let's
go
to
tell
our
last
year
and
a
half
of
history
in
an
hour,
and
we
compete
strongly
with
lunch.
So
I
don't
know
if
we
will
put
it
into
a
probe,
but
we
are
going
to
try
to
at
least
set
the
pace.
We
are
going
to
start
with
an
agenda
that
we
have
these
four
important
topics.
F
Well,
everything
that
was
this
year
and
a
half
since
we
started
with
openshift
all
the
time
we
are
going
to
be
coming
with
the
culture,
because
the
culture
really
is
something
that
everything
goes
through
this
transformation
and
that
basically
has
to
do
with
90%
of
whether
this
is
a
success
or
not,
well,.
What
we
say
is
that
we
are
struggling
with
that
aspect
due
to
image
quality,,
but
it
seemed
to
me
quite
good
to
represent
how
we
worked
before
in
Galicia
Before,
starting
this
transformation
a
little
bit,.
F
It
was
this
two-year
cascading
project,
where
the
analysis
and
design-construction
stages
were
planned,
which
were
not
fulfilled.
Either,
alaska
tables,
where
I
had
two
years
building
a
project
that
the
client
had
not
yet
seen
where
there
were
many
meetings,
In
different
areas.
There
was
no
collaborative
work,
everything
delayed
everything,,
a
lot
of
bureaucracy
and
soon
you
didn't
get
to
production,.
Everything
ended
up
exploding
in
your
face
and
the
client
really
hadn't,
seen
anything
with
which
the
client
had
changed.
The
context
in
those
two
years.
F
changed
and
really
what
had
been
helped
at
the
beginning
was
no
longer
useful
and
we
ended
up
more
hours
in
this
well-
known
image
that
the
client
did
not
know
what
he
wanted.
In
reality.
We
did
not
know
what
the
client
wanted,
because
we
did
not
listen
a
little
bit
then
in
January
In
2018.
The
bank
decided
to
seriously
face
this
transformation
that
comes
from
the
executive
committee
to
the
last
developer,
or
to
whom
to
try
to
change
this
a
bit
and
perhaps
also
not
to
be
left
out
of
the
market.
So
in
January
I.
F
Remember
that
they
brought
us
together
in
a
room
of
what
we
had
chosen
for
this
process,
and
they
told
me
that
we
had
the
responsibility
to
change
the
way
in
which
we
are
going
to
work
from
here
to
the
future,.
The
president
comes
and
told
her
that
they
will
be
able
to
innovate,.
They
will
not
have
to
ask
permission
from
anyone
and
they
will
be
able
to
materialize
the
ideas
that
occur
to
them
for
the
destination
that
they
had
chosen
for
us
in
great.
Without.
F
It
is,
then,
we
started
with
what
is
travel
optimization.
That
was
the
first
team
that
was
put
together
for
this,
and
we
had
to
transform
a
complete
experience
of
what
the
bonding
process
was
from
a
client
with
the
bank.
That
was
like
the
reference
they
wanted
to
use
because
it
was
the
one
that
had
the
most
impact
now
that
it
is
a
trip
for
us
according
to
gerry,
as
we
also
call
it
internally.
A
trip
speaks
of
a
set
of
interactions
that
are
relevant
to
the
client
for
a
purpose
and
a
very
clear
beginning.
F
F
F
We
had
the
possibility
of
removing
the
paper
signature
to
replace
it
with
a
digital
signature,
and
with
that
we
spent
a
process
of
half
an
hour
in
only
5
minutes,
kaka
I'm
going
to
mix
it
a
bit
with
the
culture
because
When
we
started
climbing
this,
at
the
time,
people
in
their
course
told
them
che
and
now,.
What
are
we
going
to
do
with
the
time
we
have
left?
Over?
We
are
going
to
lose
the
work
that
took
us
half
an
hour,.
It
takes
us
four,
more
minutes,
about
half
the
day,,
so
well,
too..
F
There
is
a
very
important
change
management
process
that
is
necessary
to
accompany
these
in
vip
and
that
they
come
to
digitize
these
processes,
let's
say
so
a
little
now
how
we
did
it
or
how
we
have
been
wanting
to
live.
We
put
together
these
multidisciplinary
teams
where
there
was
a
product,
a
list
of
products,
an
ibex
designer
business
order
was
going,
and
that
was
a
role
that
did
not
exist
in
the
bank.
At
that
time,
we
accompanied
him
by
a
scrum
master.
F
With
this,
what
we
did
was
cluster
hoist
them
insight
to
put
together
the
profiles
of
those
clients
for
whom
we
are
going
to
design
this
process
and
with
that,
what
we
did
was
take
it
to
have
a
general
consensus
and
that
everything
the
world
can
contribute
new
ideas
to
a
vision
workshop
where
all
the
ideas
that
had
been
revealed
and
the
science
of
those
clients
that
we
had
been
studying
can
be
put
together.
The
daily
path,
the
ideal
trip
for
this.
F
My
foot
two
weeks
that
I
demonstrated
here
before
a
difference
as
we
had
been
working
before
what
would
be
the
static
pyramid
and
the
origin
of
the
changes
is
that
cascade
project
of
two
years
and
when
we
started
working
with
envipe
I
think
it
allows
us
to
disintegrate
and
revolutionize
the
product
as
the
client
is
giving
feedback
not
before
we
arrived
with
the
finished
product,
and
if
someone
did
not
like
it,
there
was
no
chance.
We
had
to
start
everything
is
zero.
In
this
case
we
had
a
process
that
obviously
was
not
digital
at
all.
F
F
We
had
at
that
time
with
a
vision,
comes
to
my
channel
and
then
enter
the
part
of
enchanting
that
we
are
there
now
after
a
year
and
a
half
where,
if
we
are
doing
things
that
put
some
more
relevant
emotion
and
more
experiences
as
a
summary,
we
basically
had
these
56
steps
where
there
is
an
offer.
Well,
in
my
channel
where
we
lowered
those
times
of
half
an
hour
and
46
minutes
end
to
end,
we
eliminated
a
lot
of
paper,
almost
225
tons
per
month,
which
is
a
lot
of
money.
F
It
is
like
the
final
photo,
and
here
is
a
small
proof
of
what
this
scaling
was
of
wine
once
we
are
sure
that
it
works,
that
it
is
resolved.
Let's
say
if
we
start
to
increase
volume
channels
and
different
new
functionalities.
We
end
up
in
what
was
November
in
this
case
last
year,
with
seven
thousand
sales
per
day
in
all
those
channels,
with
the
same
solution
at
each
An.
Also
interesting
part
that
also
has
to
do
with
culture,.
The
operations
model
was
transformed
quite
a
bit
into
an
operations
model,.
F
It
was
a
fairly
automatic
role
in
terms
of
data
loading
and
nothing
else,.
Today,
operations
went
from
32
to
11
people
in
what
this
process
was,
nothing
more
and
they
stopped
loading
data
to
have
a
much
more
relevant
control
where
they
control
fraud.
They
ensure
delivery
data,
they
control
data
quality
and
operate
the
product
with
the
product.
In
some
way,
all
the
time
they
are
giving
feedback
to
say.
Milk
This
is
happening
at
the
branch,
pay
attention
to
prioritizing
it
for
the
next
sprint,
because
it
is
important
in
terms
of
numbers,.
F
It
may
seem
that
good,
those
people
were
left
without
a
job,
on
the
contrary,.
They
went
to
other
parts
of
the
organization
where
they
were
trained
and
trained
to
have
an
opportunity
to
work
in
another,
very
different
role
than
they
were
used
to,
and
the
truth
is
that
it
worked
very
well.
What
was
done
was
to
make
an
area
more
efficient
and
take
advantage
of
the
potential
of
those
people
who
are
hidden
in
other
positions
in
the
organization
that
continues
to
grow
advantages
and
challenges.
F
This
one
is
not
going
to
work
in
this
new
way
of
working.
The
mba
wenn
is
one
of
them
radically
changes
from
day
to
day.
When
you
can
decide
what
you
want
to
do,
and
there
is
no
one
to
tell
you
no,
it
is
not
good
that
they
always
say
what
you
have
to
do,
and
that
is
what
we
were
used
to
at
that
time.
F
F
So
when
what
challenges
is
it
also,
if
decision
making
and
thus
prove
to
be
wrong
quickly
to
capitalize?
On
that
experience
and
learn,
let's
say
well
speed
in
decision
making
learning
of
other
roles,
we
had
a
multidisciplinary
team
with
which,
after
a
year
and
a
half,
we
had
to
the
producer,
who
perfectly
understands
technology
and
now
developers
who
are
conducting
interviews
with
users
with
the
practice
of
uex,
good,
that
the
entire
team
responds
according
to
the
needs
of
the
moment,
and
everyone
learns
from
everyone.
F
Flexible
hours
was
also
an
advantage
because,
when
working
for
objectives
We
began
to
have
that
advantage
from
there,
it
did
not
comply
in
the
rajatabla
area,
but
also
in
that
party
that,
as
always
happens,
some
go
too
far
from
mambo
and
around
there.
At
some
point
they
arrived
at
the
3rd
afternoon.
There
is
always
a
control
station.
Let's
go
back
to
the
above.
That
It
is
the
challenge
and
finding
the
balance
obviously
had
an
impact
on
the
motivation
of
the
team,
and
if
this
team
is
motivated,,
better
results
are
obtained,.
F
F
Ours
could
be
that
of
technology
at
the
beginning.
Well,
obviously,
as
Santi
said,
and
the
openshift.
Video
is
a
fundamental
part
of
our
transformation
because
it
is
what
gave
us
the
basis
to,,
let's
say,,
streamline,
everything.
The
first
process
would
be
in
VIP
and
we
had
a
server
at
the
bank
at
that
time
for
a
minimum
of
three
weeks
so
as
not
to
put
a
maximum
and
what
that
did
was
that
when
they
told
us
in
12
weeks,
they
have
to
have
VIP
and
in
production
and
well.
F
I
already
have
three
weeks,
try
the
first
one
in
a
row,.
The
development
will
be
impossible,
so
well,,
since
we
had
freedom
to
transform,,
also
the
technological
one,
and
we
were
able
to
open
the
game
to
new
things,.
We
began
to
test
openshift
and
it
ends
up
being
a
fundamental
part
until
today,
and
it
continues
to
continue
being
a
colored
fact
here
at
the
time
as
we
were
so
enthusiastic
about
new
things.
The
first
thing
they
did
was
set
up
openshift
in
asur
and
well,
nothing
later
with
a
little
rationality.
F
F
Take
it
to
any
good
cloud,
our
micro-
service
layer,,
which
is
also
something
new
for
the
bank.
He,
talked
to
us
a
lot
about
this,.
He
started
with
network
with
spring
in
some
things
with
an
out
and
then
when
we
started
to
add
a
new
one
yesterday
they
are
not
even
new
trips
with
these
teams
and
we
did
not
have
much
time
to
get
strict
in
metrology.
We
started
to
add
python.
Well,
I
think
that
some
beyond
and
frei
turning
around
that,
obviously
adri
and
team
did
not
like
much
power.
F
Each
of
technology,
9
implied
a
new
pipeline
and
one
to
evolve
to
maintain
to
learn
with
which.
Well
now
we
are
more
than
most
of
us
have
it
in
spring
and
in
this
We
try
not
to
add
new
new
technologies.
At
that
time,,
we
prefer
to
hire
talent
for
whatever
the
tool
and
masters
are,,
and
we
believed
that
we
could
have
that
flexibility.
A.
F
F
That
what
we
valued
the
most
in
all
this
time
for
putting
5,
because
we
did
not
fit
more
and
if
the
car
is
going
to
be
very
long
provisionally,
the
infrastructure
is
again
mixed
with
the
cultural
that
I
told
you
before
and
this
it
gave
us
a
lot
of
speed
and
a
lot
of
flexibility.
Also
for
what
is
the
scaling
demand
of
each
application?
What
is
in
the
fourth
point,
the
cultural
elimination
of
my
machine
works.
This
was
a
tremendous
change.
F
It
still
happens
to
us
today
that
we
have
many
homework
and
always
My
machine
works
for
me
and
well,,
but
you
have
to
program
the
standard
and
it
has
to
work
in
open.
If
we
go
towards
that,,
it
eliminated
a
lot
of
problems
that
were
there
before.
When
I
developed
a
request,,
it
deployed
technology
and
energy,,
it
operated
an
application
that
the
developer
did
not
know.
About.
F
Automation
is
another
One
of
the
advantages
when
joining
the
mein
becomes
automation,,
which
is
to
automate
everything,
and
that
also
begins
to
generate
some
fear
and
then
the
culture
with
I'm
going
to
lose
my
job
I
continue
to
automate
access
to
what
I
do
tomorrow
and
well.
Nothing
is
also
a
is
a
an
enormous
challenge
to
convince
people
that
what
automation
is
going
to
do
is
free
up
their
time
to
be
able
to
do
other
things
later,
this
historically,
since
the
industrial
revolution
until
now,
at
least
any
automation
or
improvement
of
the
tools,.
F
What
they
end
up
seeing
is
a
multiplication
of
new
specialties
with
which
much
more
work
was
generated
than
the
one
that
was
eliminated
so
being
able
to
have
that
manises
of
automation
and
reinventing
oneself
and
being
able
to
try
new
things
or
learn.
New
things
also
motivates
people
to
continue
growing
and
never
stop
learning,
Let's,
say
and
well,.
Another
key
thing
that
at
least
I
like
a
lot
is
what
is
blue,
green
or
ibi
testing
of
each
new
microservice
version?
That
is
also
a
very
cultural
issue.
F
All
the
time
we
are
seeing
the
vision
of
the
guys'
development
team
We
have
the
possibility
of
being
a
plugin,
and
it
will
begin
to
spread
traffic
little
by
little
and
increasing
traffic
for
each
functionality,.
Lowering
the
impact
of
incidents
in
production
and
they
don't
use
it.
Everything
is
ready
and
they
don't
use
it.
Suddenly.
You
find
that
something
exploded
and
you
gave
up,
but
you
didn't
do
it
because
of
the
green,,
not
good.
F
So
that's
also
against
the
culture
of
being
able
to
generate
that
change
and
that
incorporation
of
technology
also
one
thing
that
happens
to
us
a
lot
is
you
have
a
two-week
sprint
or
we
have
a
two-week
sprint
and
on
Monday
First,
it
already
ends
in
a
vague
or
a
new
function,
and
we
wait
until
the
end
of
the
sprint
to
move
on
to
production.
Itself.
We
have
a
tool
to
connect
for
Biblián.
F
F
Agility
scale
here
we
also
have,
within
these
pillars
of
transformation
in
the
part
of
methodology
and
culture
and
talent
that
was
happening.
We
started
to
have
these
teams
scattered
all
over
the
place
and
we
needed
to
somehow
give
it
an
organic
framework
to
be
able
to
scale
so
that
it
wouldn't
be
out
of
control,
not
because
each
one
was
going
to
decide
what
agility
was
like
for
that
team,
how
to
develop
seno
pencil
applications
for
that
team.
F
F
So
at
some
point
we
had
this,
which
was,,
let's
say,
a
barbaric
mess,
because
anyone
already
did
anything,.
Everyone
said
it
was
one
product
without
knowing
what
it
was,.
We
had,
like,
many
of
these
cells
or
bubbles
everywhere
for
branches
in
an
inorganic
way.
So
that's
where
we
made
the
decision
to
put
together
a
team
that
we
called
the
rocket
that
was
the
center
of
excellence
of
agility
to
scale
this
in
an
organic
way.
The
rest
of
the
bank,
then
it
was
The
idea-
was
to
somewhat
train
these
scrum
masters
that
we
did.
F
N't
have
so
many
at
the
uex
to
download
an
alignment
of
common
practices
among
all
of
us
how
we
were
going
to
break
down
in
open
and
how
we
were
going
to
get
to
production.
What
our
deployment
model
was
going
to
be
good
With,
a
lot
of
things
that
fell
within
this
scale,
agility
framework
and
then
start
those
teams
that
we
have
been
multiplying,
put
them
in
what
we
call
tribes.
I,
don't
know.
If
anyone
knows
the
Spotify
model,
well.
F
Well,
our
model
is
the
Spotify
model,,
but
more
or
less.
It
tends
to
always
do
in
some
places,.
They
listened,
don't
copy
spotify,
because
you're
not
spotify,.
So
what
we
try
to
do
is
something
that
really
follows
that
line,
but
that
adapts
to
the
culture
we
have
at
the
moment
in
Galicia,
no,.
So
what
we
started
to
do
is
define
two
tribes
to
also
in
vip
mode
and
learn
from
what
is
happening
in
them.
F
We
abruptly
broke
off
the
product
segment
in
one
day
and
we
put
together
everything
that
was
a
digital
channel
in
what
was
a
product
tribe,
and
we
stopped
talking
about
channels
in
the
bank.
From
that
we
began
to
talk
about
products
if
there
is
an
accounts
school
that
looks
at
accounts
in
all
channels.
Before
we
had
the
branch
area
that
looked
at
accounts,
the
online
banking
area
that
looked
at
accounts.
That
was
quite
a
big
change.
That
It
is
also
a
movement
that
mobilizes
culturally.
F
Today
we
have
two,
one
for
daily
banking
and
another
for
collections
and
payments,,
and
we
were
learning
to
later
continue
scaling
the
same
model
to
the
rest
of
the
bank
and
below.
We
have
what
we
call
the
chapter
waste
center,
that
we
have
cybersecurity
the
design
practices
of
the
Galician
system,,
just
as
it
has
to
do
with
everything
that
is
visual
components
that
what
they
do
is
also
have
a
cross
look
and
coherence
in
what
the
entire
organization
is
and
the
one
that
is
in
the
middle.
F
That
I
am
going
to
talk
about
in
the
next
step,
to
see
with
the
tribes
that
we
call
enablers,
because
there
comes
a
time
when
the
product
tribes
run
out
of
things
to
do.
If
they
do
not
follow
the
capabilities
behind
what
our
package
is,
and
today
they
have
inherited
applications
where
the
We
put
these
tribes
together
on
the
first
day,.
F
F
To
me,
a
bit,
as
I
was
telling
you,,
so
we
had
what
was
bank
and
not
in
the
image
at
the
time,
100%
monolithic
applications
where
to
make
a
deployment
you
had
to
regress
at
least
three
weeks,.
Even
if
you
make
a
minimum
change
to
ensure
that
you
do
not
break
anything
from
the
channel,
we
had
different
business
applications
scattered
around
some
zenet
others
in
visual
6.
There
was
a
bit
of
everything
and
we
did
not
have
something
that
was
really
single
channel.
F
F
F
These
tribes
are
in
their
squads,
having
all
these
spd
in
Riyadh,,
with
which
we
disarmed
the
bullet
and
martin
monolith
to
have
them
oriented
to
products
in
that
piece.
Nothing
without
application
flu
that
are
already
independent
of
each
cell.
We
are
forming
the
apis
layer.
Today,
we
already
have
250
microservices
only
of
what
is
the
back
of
channels
later
in
open
chief.
There
are
many
more
in
the
rest
of
the
bank.
We
are
trying
to
finish
it
leave
full
production,
tries
kyl,
and
then
we
are
using
a
lot.
F
G
Thanks
I
am
adrià
secas,
as
the
presenter,
roberto,
I,
say
santiago,
who
is
there
and
well
everything
that
gonza
told
us
is
really
oriented
to
this
new,,
very
dynamic,
agile
work
methodology
and
how
it
impacted
within
the
systems
area
Really,
when
we
started
working
with
the
cells
and
we
began
to
receive
all
these
requirements,.
We
realized
that,
in
order
to
keep
up
with
the
agility
of
these
cells,,
we
had
to
make
a
change,
not
only
methodologically,
but
also
in
tools,
so
that
we
could
have
an
idea
of
.
A
cell.
G
often
develops
between
5
or
10
microservices
per
day,,
which
have
to
be
deployed
in
different
environments,,
including
production,,
and
when
the
cells
arrive
with
these
requirements
in
the
systems
area,,
they
collide
a
bit
with
the
reality
of
the
methodology
we
use
at
that
time
and
really
They
saw
us
a
bit
like
that,.
What
I
was
telling
Gonza
that
you
know
in
case
we
had
a
procedure
that
took
weeks
to
provide
some
provisioning
of
a
server
of
a
base.
G
After
all,
the
very
large
number
of
these,
your
live
environments
that
are
these,
these
very
short-lived
environments,
that
it
was
also
very
difficult
to
determine
when
they
stopped
using
them
or
not
to
be
able
to
recycle
those
resources,
well,.
Everything
related
to
deployment
orchestration
was
handled
through
a
group
of
deployments,
which
ended
up
creating
a
significant
bottleneck,,
often
causing
downtime
in
deployments,
because
the
implementation
team
really
didn't
know
exactly
how
to
automate
a
deployment
in
production.
I
didn't
know
what
were
the
specifics
of
what
was
going
to
crash
in
a
good
environment.
G
Everything
was
really
fragmented
configuration
how
to
ensure
the
equality
of
all
environments
and
all
servers
and
all
components.
In.
An
ecosystem,
feedback
between
developers
and
platform
administrators
is
also
very
long,
from
when
something
collapses
in
production
until
it
arrives,
and
it
is
really
validated
that
there
are
no
problems
when
there
are
no
problems
and
that
feedback
goes
back
to
the
implementation
team.
to
be
able
to
do
a
roll
back.
It
really
took
days
living
with
comebacks
in
production
and
finally,
well,.
G
The
traditional
silo
scheme,
in
which
everyone
does
their
part,
the
server
team,
has
their
time
to
enter
a
server.
The
network
team
has
their
time
to
deliver
a
virtualized
url.
The
platform
team
has
time
to
install
a
server
application
or
a
web
server,
and
all
this
really,
when
we
talked
to
the
cells,
was
to
hit
a
wall
and
really
the
unthinkable
a
bit
from
the
bank
to
reach
an
architecture
that
It
can
keep
up
with
a
cell.
So
the
wisest
decision
was
really
to
create
an
architecture
as
a
service
to
create
a
peace.
G
G
We
chose
the
stack
of
tools
that
openshift
gave
us,
and
this
was
a
really
great
value
that
helped
us
to
be
able
to
absorb
all
the
new
methodology
and
all
this
new
paradigm
of
containers.
Why?
Because
openshift
gives
us
all
the
tools
to
be
able
to
create
a
cedice
chain
of
really
almost
out
of
the
box
with
an
old
playbook,
you
can't
create
your
entire
tool
stack
of
6
cds.
We
use
jenkins,
we
use
an
internal
git.
G
G
G
That
was
the
best
in
the
last
50
years,
because
it
was
the
only
one
we
had
so
that
's
the
joke
that
we
always
told
you
and,
as
you
know,
it's
a
bit
difficult
to
find
this
profile
in
the
market
to
really
find
a
profile
that
understands
both
the
development
of
the
life
cycle,
of
an
application
and
infrastructure
operations.
But
well
we
managed
to
form
it
and
what
we
did
was
once
we
defined
Our
entire
plan,,
which
I
don't
know.
G
So
through
all
this
development
that
the
divos
did,,
the
self-managing
cell
of
100,
can
not
only
create
its
applications,
but
to
do
the
passages
100%
and
a
little
bit
to
show
you
what
developer
stevens
is
that
we
create
in
three
simple
steps.
A
cell,
a
developer
of
a
cell,
can
create
by
choosing
the
framework
that
is
going
to
use
all
the
objects
that
he
needs
inside
openshift,
also
using
an
archetype
to
be
able
to
have
a
standard
on
the
use
of
the
framework
and
through
templates
that
we
were
also
creating
in
openshift.
G
To
show
a
bit
the
current
architecture
and
how
many
They
are
the
number
of
projects
and
cells
that
we
currently
have.
Well,
as
I
was
saying,?
Progress
began
with
a
group
of
five
cells,,
but
today
there
are
32
in
production,,
which
obviously
did
not
begin
to
expand,.
All
the
resources
have
five
clusters
in
the
development,
laboratory.
production
integration
and
plateau
production.
We
have
23
nodes
for
each
of
these
plastics,
which
have
8,
corts
and
32
gigabytes
of
ram.
Show
you
here
a
little
bit.
What
a
screenshot
of
one
of
the
clusters
is.
G
This
is
the
development
caster
that
is
that
screenshot
this
week
it
has
868
potts.
It
has
180
corts
in
total,
640
gigabytes
of
ram.
We
have
to
multiply
this
by
five
for
the
five
clusters
that
we
have,
which
led
us
to
have
a
fairly
large
monster
and,
as
it
supports,
we
go
all
over
the
face
and
how
we
make
sure
that
all
these
pot
really
are
available
for
all
the
for
all
the
cells
that
meet
all
the
standards.
We
had
to
create
a
team
that
really
looks
after
the
scalability
and
the
relaxation
ability
the
cluster.
G
That
is
the
team
of
that
network,
since
those
of
box
saw
the
complete
application
life
cycle,
but
they
could
not
see
the
infrastructure,
and
we
created
this
team
thinking
about
automation
in
the
infrastructure
as
code
to
be
able
to
scale
the
cluster
in
an
automated
way.
Have
you
observe
the
ability
of
the
cluster
we
began
to
play
with
streets
in
engineering
break
a
node
see
how
we
recovered
to
see
where
we
recorded
those
spots
in
his
classes
and
immutability
and
resilience
also
automated,
so
that
the
cluster
always
remains
the
same
way?
G
Well,
these
are
a
bit
of
what
gonza
was
talking
about
this
new
work
culture
like
these
three
teams,
They
began
to
get
feedback
from
each
other
as
the
cells
passed
requirements
to
the
box,,
the
drawings,
requirements
and
infrastructures,,
and
that
really
created
a
work
culture
where
we
are
all
interconnected
and
we
all
know
where
we
are
going
and
on
the
same
page,
up
to
here,
yes..
We
see
it
really
seems
like
a
very
simple
story.
We
created
groups,
we
created
infrastructure,
everything
went
well
and
we
really
did
100
percent.
G
G
We
met
with
the
networking
team
and
one
of
the
main
things
that
one
has
to
define
when
you
are
going
to
create
your
cluster
is
the
range
of
ipes
you
are
going
to
use
inside
from
the
cluster
for
all
the
posts
that
are
going
to
create
that
range
of
ips
are
defined
by
segment
of
r,
a
bar
24,
a
bar
20.
Whatever
one
decides
that
the
cluster
is
going
to
grow
is
like
this,,
we
had
to
separate
them
from
ours
so
that
only
be
used
within
the
openshift
cluster.
G
We
got
together
with
the
networking
team
we
proposed.
Having
a
neighborhood
16,
which
is
about
65,000
ips,
they
told
us
it
was
crazy
that
we
were
never
going
to
use
it,
that
we
lower
it
a
bar
20,
which
is
about
four
thousand
pesos
four
thousand
pesos
each
openshift
node
by
default
since
q.
Ernest
really
by
default,
a
maximum
of
250
posts
is
what
is
recommended.
250
gives
us
16
nodes.
G
We
use
3
for
the
administration
of
the
masters,
3
for
the
infrastructure,
We
had
10
computing
applications
left
and
beyond
that
we
basically
couldn't
grow
in
three
months.
Four
of
us
ate
up
all
the
cluster
resources.
We
started
having
post
generation
problems
because
we
no
longer
had
space
for
anything
and
well,.
G
It
was
a
good
lesson
learned
from
the
infrastructure
as
code
that
we
were
able
to
build
another
cluster,
making
that
network
change
and
we
were
able
to
go
through
and
with
all
the
implementation
that
the
living
had
created
to
be
able
to
migrate
that
cluster
in
one
day
to
have
it
ready
and
have
recreated
all
the
dances
and
all
the
applications
in
another
cluster,
after
hundreds
from
pai
dance.
That
was
another
very
important
country
that
we
were
able
to
standardize
relatively
recently
when
the
whole
initiative
started,
and
we
had
so
many
types
of
technologies.
G
It
was
very
difficult
for
the
living
to
be
able
to
centralize
everything
in
a
single
jenkin
file
that
can
really
give
you
all
automation
to
all
the
developments,,
so
they
began
to
create
different
types
of
jenkins,.
There
are
different
types
of
plans
that,,
although
they
did
the
same,
the
same
steps,
compile,
analyze,
codes,
authorizations
were
tied
to
each
of
the
developments
of
those
microservices
that
had
been
created,.
G
This
caused
at
one
point,
we
have
250
jianqing
files,
totally
unmaintainable
the
bobs
or
the
off
that
we
had
at
that
time
were
overwhelmed
because
every
time
one
of
the
paisas
failed,
they
had
to
call
him.
They
had
to
be
playing
and
monitoring
these
these
files
all
the
time
what
we
ended
up
doing
after
a
major
development
and
after
many
many
deliberations
within
the
team
was
to
centralize
everything
in
a
single
pay
line,
file
which
is
used
across
all
applications.
Basically,
every.
G
Each
one
compiles
differently:
each
one
is
analyzed
differently,
each
one
has
its
own
characteristic,
and
that
was
something
that
took
us
a
long
time
and
it
was
a
long
road
too
how
good
we
were
able
to
overcome
it
and,
in
the
end,
too,
tell
you
a
little
about
a
problem.
We
had
with
the
file
system
that
we
used
back
when
we
did
the
first
implementation
of
openshift.
G
What
we
counted
on
was
an
nfs
as
the
decentralized
file
system,
although
network
that
has
a
fairly
large
sign
that
says,
do
not
use
nfs
in
production
because
it
can
cause
you
problems.
We
equip
ourselves
with
that.
It's
neither
good
nor
playing
and
we
use
it
the
same,
and
it
brought
us
problems.
It
brought
us
problems
from
ayo.
It
brought
us
problems
of
not
being
able
to
limit
the
amount
of
space
in
each
of
the
persistent
volumes
that
we
were
creating
there.
G
Well,
everything
that
to
the
automation
of
the
request
of
those
persistent
volumes
had
to
be
a
script.
It
has
given
you
the
voice
ones,
it
had
more
work
overload.
It
was
a
problem.
Finally,
we
went
to
a
solution
more
oriented
to
containers
and
for
a
card
that
that
last
name
container
storage,
that
It
is
basically
a
cluster
and
we
were
able
to
solve
it,.
It
changed
a
lot.
G
Good,,
if
not,
it
did
not
solve
life
for
us,
a
lot,
but
well,
like
this,.
We
had
10
15
and
its
more
important
than
it
was
learning,
and
we
went
also
seeing
the
skills
that
we
needed
for
each
of
the
teams
that
we
are
creating
for
the
divos
for
the
network
and
even
for
the
cells
that
were
educating
themselves
a
little
in
their
development
and
leaving
a
little.
What
I
told
gonza
de
me
and
my
machine
that
It
was
already
a
forbidden
word
on
the
floor
and
being
able
to
develop
for
containers
and.
F
Then
another
important
thing
that
was
a
change
and
a
lesson
that
we
had
each
the
box
per
cell
was
the
model
that
we
proposed
at
the
beginning
and
that,
in
addition
to
the
fact
that
it
has
a
dependency
on
all
the
developers
not
to
use
the
infrastructure
and
maintain
the
production
room
itself,,
there
is
almost
a
dependency
on
ticketing
with
the
the
box.
That
was
on
the
table,
with
which
we
returned
to
the
same
previous
scheme,.
So
what
we
did
was
separate
the
box
from
the
agile
cells.
F
G
Well,,
it
was
also
to
show
you
a
bit
with
this
that,
although
the
methodology
is
a
very
important
part,,
the
tools
are
also
very
important,.
Without
agile
tools,
it
is
impossible
to
achieve
the
agility
that
a
cell
requires
and
without
the
agility
of
the
methodology,
we
also
become
again
the
threads
as
I
told
gonza,
where
it
is
a
ticketing
system,
and
it
is
a
request,
was
to
show
you
a
little
how
we
really
adopt
that
from
the
tools
part
with
red
hat
and
the
methodological
part
with
the
agility.