►
Description
KCD Italy 2021* - The video of the talks are now available!
Anche lo sviluppo del software ha un impatto importante sul clima e sull’ambiente.
Qualche esempio? I Bitcoin producono tra 22 e 22.9 milioni di tonnellate di anidride carbonica all’anno. Minecraft, il videogioco più giocato al mondo, ha prodotto nella sua vita 600 milioni di kg di CO2.
Gli sviluppatori che hanno realizzato questi software non pensavano di avere un impatto così importante sul clima mondiale. Ora, però, è il momento di pensarci per il futuro.
Per questo ho realizzato kube-green, un progetto OSS su GitHub che ha l’obiettivo di ridurre la carbon footprint di applicativi Cloud Native.
In questo talk vedremo come funziona, una live demo e la sua roadmap per avere un mondo sempre più pulito e un software funzionante e scalabile.
🤓 SPEAKER: Davide Bianchi, Senior Tech Lead @ Mia Platform
👉🏼 Here the slidedeck: https://www.slideshare.net/KCDItaly/kubegreen-davide-bianchi
______________________
* KCD Italy 2021 is the first edition of a community-organized non-profit event that will take place on November 17th and 18th and will gather adopters and technologists from open source and cloud native communities for education, collaboration, and networking.
📯 Follow us to be notified about the upcoming MeetUps: https://community.cncf.io/kubernetes-cloud-native-milano/
A
Good
morning
Good
morning,
everyone
welcome
to
this
first
event
that
takes
place
at
Eataly.
I
am
Edoardo
Dusi
taqlid,
In
sparkfabrik,
who
is
also
one
of
the
organizers
of
today's
event,
and
I
will
accompany
you,
Or
rather
by
ferry
But.
Until
lunchtime
Before
starting
the
event
in
all
respects,
however,
I'll
tell
you
something
about
case
cds
and
leave
you
some
technical
indications
So
as
you
know,,
it's
an
event
supported
by
silence
F,
which
of
course,
we
thank
immediately
and
is
organized
by
the
Italian
Kubernetes
Community
But.
A
Why
not?
The
international
events
would
be
enough,.
The
various
cubes
are
there
right
away,
but
in
reality
the
sien
chef
has
the
promotion
of
the
clown
as
its
Mission,
and
you
see
all
the
Open
Source
themes
in
the
most
capillary
way
possible
and
therefore
relies
on
the
local
communities
and
local
meetups
that
they
know
more
closely
the
characteristics
of
the
market
and
the
problems
that
exist
locally.
They
can
therefore
activate
discussions
in
the
local
language,
and
this
is
why
today's
event
is
partly
in
Italian
and
partly
in
English.
A
So
don't
worry
if
you
are
there
following
listen
only
to
the
English
language,
because
in
any
case
you
will
have
part
of
the
Talks,
for
you
too.
So
I
have
to
tell
this
place
for
a
moment.
As.
You
may
have
noticed
this
logo
Remembers
a
pizza
and
because
a
pizza
for
one
thing
we
must
find
an
identity.
Therefore,
the
colors
and
signs
that
represent
us
that
they
represented
Italy
in
front
of
the
whole
global
community,.
A
You
have
no
idea
how
many
discussions
we
had,,
how
many
debates
we
carried
on,
but
in
the
end,,
as
happens
on
Saturday
nights,,
the
pizza
won.
But
why?
What,
is
more
Italian
than
pizza,
perhaps?,
pasta
with
tomato,
however,
pizza
means
many
things.
It
means
Italian.
Taste
in
the
world
means
sociability
sharing
simplicity,
and
then
it
resolves
domestic
issues.
A
By
bringing
north,
center
and
south
into
agreement,
we
could
have
chosen
the
Colosseum,
perhaps,
but
then
why
not
the
Duomo
of
Milan
the
tower
of
Pisa
The
Valley
of
the
Temples
or
the
Dolomites,
as
we
could
have
chosen
a
part
of
the
whole
and
therefore
the
pizza
recognized
international
symbol
recognizable,
but
with
a
few
more
details,
Claudine,
as
you
have
seen,
the
pizza
of
del
Casey
of
Italy
is
arranged
in
slices
inside
of
the
helm
formed
by
rolling
pins.
Also
in
deference
to
the
outgoing
president,
because
covernet
es
is
passion
even
before
work.
A
Naturally,
he
wanted
to
take
himself
a
little
seriously
Come
on
also
to
lighten
this
year's
Mood
a
bit
which
is
heavy
for
everyone.
So
here's
the
pizza,
it's
an
invitation
from
Italy.
The
whole
community
has
reached
all
over
the
world
and
we
can't
wait
to
meet
again
shake
hands,
get
to
know
each
other
in
person
and
eat
a
pizza
together
and
wait.
And
you'll
see
what
we
thought
for
the
2022
edition.
A
Well,
How.
You
know
this
edition
of
kubernetes
Community
Days
Italy
is
completely
free
virtual.
It
is
structured
in
two
half
days
live
today
and
tomorrow
we
have
10
speakers
who
will
take
turns
on
stage
an
amazing
lineup
that
we
will
present
to
you
gradually,
which
ones
I
will
show
a
bit
of
the
agenda,
but
remember
that
the
event
is
over.
They
will
leave
invited
by
the
kubernetes
Cloud
knative
Milano
group.
So
it
is
as
if
this
event
will
continue
over
time.
A
A
We
want
to
improve
two
charitable
initiatives
carried
out
by
our
partners.
Are
new
life
project
has
decided
to
dedicate
hours
and
commitment
to
the
recovery
and
regeneration
of
donated
PCs
a
family
in
difficulty
and
more
store-age
actively
participated
in
the
fundraising
project
for
a
Congolese
orphanage
and
to
find
out
more,
follow
our
Twitter
and
LinkedIn
channels
and
speaking
of
partners,
we
inform
you
that
ethical
binary
has
a
surprise
for
you.
So
stay
alert.
We
will
share
more
information
later
in
the
chat.
We.
A
Hydrated
is
made
available
free
of
charge
by
Siemens
CF,
so
you
have
two
Tabs
1
for
chat
and
one
for
more
and
more
questions
that
please
use
also
to
facilitate
my
work
as
friends
who
I
will
then
collect
more
questions,
and
neither
Ele
I
will
turn
to
the
speakers
at
the
end
of
each
Talk.
If
there
will
be
any
if
there
will
be
time
available.
A
A
Last
thing,
Of
course:
you
see
this
magnificent,
magnificent
t-shirt
that
I
wear
very
well.
We
will
give
away
100
during
these
two
days
as
super
easy
in
chat
comes
a
link
to
a
form
in
which
you
must
leave
us
data
in
order
to
organize
shipments
in
the
next
few
days
in
the
form
It
will
be
published
in
Random
during
Tomorrow
morning.
It
is
intended
exclusively
I
recommend
to
those
who
live
in
Italy
So.
A
B
A
C
C
It's
a
pleasure
if
I'm
up,
here.,
look
Today
I'm
here
as
a
contribution
to
the
work
on
how,
for
about
three
years,
three
and
a
half
years
and
I
have
the
pleasure
of
driving
towards
the
first
Talk
first,
but
in
the
meantime,
I
would
like
to
thank
all
the
Organizers,
because
this
journey
today
is
the
result
of
a
2
year
and
tiring
journey
in
in
the
midst
of
a
pandemic,
etc.
and
never
even
the
beginning
of
something
new.
C
C
For
those
who
don't
know
her
Thursday
from
one
of
the
founders,
Viber
Netis
E1,
who
immediately
sensed
the
power,
the
power
of
the
community
and
what
can
be
done
by
working
together
at
this
point,,
however,
I,
would
say.
No
more
talking
and
it's
time
to
call
the
first
speaker
on
stage.
Davide
Bianchi.
B
E
Ok
Even
today,
I'm
talking
about
how
Green
is
an
Open
Source
project
that
aims
to
reduce
resource
consumption
on
Kubernetes
Clusters
Pollution
is
a
problem
even
in
digital.
Even
in
the
cloud
is
everything
we
do
on
the
Cloud
to
a
involves.
A
consumption
of
resources
on
some
servers
are
physically
located
somewhere,
even
if
far
from
us,
after
sending
an
email,
Blood
downloading
images
on
a
Cloud,
it
produces
CO2.
I.
Take
some
examples
that
Minecraft,
for
example,
in
its
entire
life,
has
produced
a
product.
E
So
looking
at
these
software
made
in
the
past
and
looking
to
the
future,
We
have
to
imagine
that,
always
when
we
are
coding
Anyway
things
we
write,
we
don't
know,
we
don't
know,
wait
between
5-10
years,
what
impact
it
could
have
on
the
environment,
what
birth
it
could
have
on
the
ecosystem,
But
energy
costs
and
consumption
above
all,
are
due
to
what
inside
a
Data.
Center
are
mainly
due
to
computation
to
storage.
E
Networking
and
then
to
the
equipment
used
to
cool
the
servers,
to
light
up
the
Data
Centers
to
the
management
and
obviously
for
the
production
of
the
servers
But.
How
much
CO2
does
a
server
produce?
This
question
which
I
asked
myself
but
on
which
it
is
difficult
to
find
answers
on
the
Go
climent
website?
There
is
an
approximation
of
when
a
server
produces
CO2
equivalent.
If
the
server
is
installed
on
a
Cloud
that
uses
100%
green
electricity,
this
produces
160
kg
of
CO2
equivalent
per
year
for
each
server.
E
Even
then,
if
the
provider
does
not
use
Green
electricity,
we
reach
up
to
458
kg
of
CO2
equivalent
on
the
Home
Data
Center
press.
This.
This
number
even
doubled
the
number
of
readings
double
because
the
large
providers
optimize
resources
to
the
maximum
consumed.
Therefore
they
have
a
consumption
of
resources
is
much
lower
than
that
of
a
Home
Premium
Data
Center
But.
Therefore,
starting
from
these
data,
how
can
we
tell
how
much
a
Kubernetes
cluster
can
produce
if
we
take
a
cluster
of
4
nodes,
each
node
composed
of
two
Core?
E
E
However,
let's
take
this
approximation
and
see
that
with
640
kg
of
CO2
equivalent
in
reality,
the
consumption
of
a
car
that
travels
15,000
km
in
a
year
therefore
has
a
consumption
that
has
a
certain
significance
in
terms
of
impact
pollution,
and
if
we
assume
that
it
cannot
be
of
this
size
more
or
less,
we
insert
about
fifty
Pods.
This
leads
us
to
say
that
each
Pod
in
Tria
produces
12
kg
of
CO2
equivalent
per
year.
This
is
still
an
approximation,
however,
We
are
waiting
for
this
number
and
we
can.
E
Go
produce
absorb
the
same
amount
of
CO2
But
because
with
the
Cloud
there
is
the
risk
of
going
out
of
control
because
once,,
when
Center
on
premises
got
there,
to
scale
up,
was
a
purchase
of
new
machines,,
a
purchase
of
something,
and
it
was
a
very
high
step.
So
the
cost
was
very
much
under
control
because
it
meant
adding
many
machines
or
number
of
machines
inside
your
own
Data
Center.
This
greatly
increased
the
waste
but
kept
everything
under
more
control.
However,.
It
doesn't
have
anything
at
all.
E
E
Therefore,
it
very
easily
continues
to
increase,
even
if
perhaps
our
business
complains,
but
our
customs
Our
costs
for
these
others
of
ours
continue
to
increase
But.
Why
does
it
continue
to
rise
We?
They
are
various
in
style.
When
we
develop
solutions,
Claudio
knative
can
increase.
First
of
all,
Because
we
write
code
that
is
deficient.
Then
we
use
needlessly
complex
algorithms
or
we
go
to
create
too
many
microservices
for
what
is
the
purpose
of
our
project
and
come
on.
We
ferment
too
much
responsibility
and
go
to
having
just
this.
E
Let's
say
the
microservices
partner,
a
bit
exasperated
having
instantiated
primordia,
but
not
using
almost
any
of
the
functions
Daniel
tremor,
which
is
a
con
cremor.
They
consume
resources
to
keep
themselves
inactive
put
How
to
set
also
some
rickers
I
have
oversized
limits
And.
This
leads
us
to
have
Then
to
allocate
with
Enrico
and
is
within
the
classic
pin
of
resources,
which
then
cannot
be
used
by
other
Pods.
E
This
leads
us
to
having
to
scale
without
Pro
effectively
that
those
resources
increase
the
static
checks
so
much
to
have
in
one
high
reliability,
which
may
not
always
be
necessary
or
for
badly
configured
CPAs.
For
example,
if
we
have
spies
configured
with
a
minimum
Maximum,
but
on
C
more,
but
we
don't
care
about
Lario
15
anymore,
at
which
point
she
spies
always
puts
the
pods
at
maximum
skating
and
therefore
it
is
not
useless
or
having
many
environments
of
Real
Time
replicas
replicated
a
lot
for
the
development
for
you
on
your
own.
E
Therefore,
when
the
requests
increase,
the
dog
names
the
number
of
the
problem
when
the
number
of
requests
decreases
or
it
consumes
more
or
whatever
the
No
does
too.
But
so
how
do
we
reduce
the
screen
is
done
badly.
Here
is
a
solution
for
Green
as
Green
is
an
Open
Source
project
I
have
been
working
on
in
recent
months,
which
aims
to
reduce
the
consumption
of
our
resources,
kratter
covernet,
how
everything
is
based
on
an
idea
about
the
idea
that
today
software
development
takes
place
through
projects.
E
These
projects
are
then
divided
into
N
runtime
environments,
where
usually
one
is
production.
Therefore,
the
one
used
training
finished,
the
end
users,
while
the
others
are
production
environments
lament.
Are
you
complaining
or
something
else?
If
we
consider
that
these
that
we
see
as
in
Medina
Time
Verdi
are
used
more
or
less
only
during
working
hours,
which
are
40
on
average,
This
means
that
for
more
than
75%
of
the
time
these
environments,
these
put
Real
Time,
they
are
running,
but
they
are
never
used.
If.
We
bring
these
runtime
environments.
E
This
way
we
see
that
each
environment
is
a
The
Space
on
the
space
contains
n
Pods
and
each
time
it
has
a
consumption.
You
say
more
is
a
defined
RAM
consumption,
even
if
used
in
idle,
because
in
any
case,
I
have
to
keep
the
server
or
the
instance
active,
or
what
is
ours
What
it
contains
and,
in
addition,
also
an
allocation.
You
say
more
eram
or
one
reserves
a
little
more
and
RAM
inside
the
Bernese
that
can
only
be
used
from
that
bridge.
So
the
idea
is
simply
going
to
turn
off
the
pods
How.
E
Do
we
turn
them
off?
We
are
going
to
turn
them
off
simply
by
putting
in
the
tests,
while
the
zero
replicas
that
land
we
imagine
that
they
are
released
via
the
first
ones,
so
that
we
can
configure
how
they
have
to
replicate.
It
is
possible
to
go
put
p30
so
that
they
are
actually
turned
off
and
the
resources
remain
on
the
classic
Venice
without,
however,
that
the
Pod
is
running
But.
How
is
it
possible
to
achieve
the
goal?
I
have
various
ideas
come
up
with.
E
A
possible
idea
is
to
turn
off
manually
then
every
evening
to
each
project
for
each
runtime,
environment,
necessary,
This,
obviously,
a
solution
that
I
took
as
it
is
not
maintainable
over
time.
It
is
very
boring
for
those
who
have
to
manage
an
environment.
Another
possibility
was
to
make
a
con
job,
which
is
released
in
each
environment
in
Udine
Space,
which
must
be
turned
off,
and
it
goes
automatically
turn
off.
What
I
was
joking
also
this
idea,
because
at
that
point,
I.e.
E
E
Another
possibility
is
to
create
a
kubernetes
operator
that
goes
to
turn
off
the
pods
adore
the
fixated
through
a
configuration,
the
Sur
definition
card.
Therefore,
something
integrated
within
the
Bernese
from
here,
which
has
the
power
to
modify
the
diploma
in
mind
of
all
the
cast
or
another
alternative
which
mentally
was
being
able
to
turn
off
the
pods
AND
reactivate
them.
When
a
new
request
arrives
at
the
moment,
I
preferred
to
focus
on
the
kubernetes
operator
that
works
with
a
chrono
So
with
fixed
times,
as
it
was
much
simpler
than
to
reactivate.
E
You
can
make
a
request,
and
in
any
case
maybe
we
could
already
get.
What
is
the?
Let's
say
the
for
the
purpose,
therefore
Reduce
resource
consumption,
but
let's
see
what
a
kubernetes
operator
is,
what
kubernet
is
in
reality.
It
is
a
loop
controller
that
takes
care
of
verifying
that
the
desired
State
is
always.
B
E
E
To
change
the
current
state
more
to
nothing
if
we
go
to
take
this
architecture
and
see
in
concept
like
Green,
we
see
it's
called
sleep
info
and
from
this
structure
that
we
see
at
the
top
right,
where
the
specification
defines
the
weekend
in
which
it
must
come
into
play,
which
is
week.
15
means
Monday-Friday
sleep
at
20
in
the
evening.
Therefore,
at
20
in
the
evening,
all
the
deployments
and
replies
will
be
set
to
zero
at
8
in
the
morning.
They
will
be
awakened
in
the
Europe
Rome
time
zone.
E
So
once
this
is
done,
what
happens?
It
is
freed
In,
the
return
of
modification
to
the
controller
like
Green,
who
goes
to
modify
the
deployment
and
modifying
the
number
of
replicas.
But
let's
see
now
A
real
case,
the
first
time
that
he
finished
creating
as
Green
the
first
we
installed
In
a
first
Cluster
this
Cluster.
E
We
see
that
there
is
a
square
wave
that
has
5
glasses,
are
5
working
days,
5
narrow
valleys,
like
work
working
days,
which
are
nights
and
a
slightly
longer
shirt,
which
is
the
weekend
as
we
are
busy
shutting
down
for
the
non-working
hours
and
port
of
these
special
trains,
and
we
see
that
the
difference
in
the
request
example
of
these
in
this
graph
from
day
to
night,
is
about
twenty
of
GB,
where
also
the
actually
consumed
memory.
Therefore,
not
only
the
leased
one
but
the
consumed
one
is
about
fifteen
GB,
less
consumption.
E
We
see
in
addition
to
the
memory
always
leaving
the
same
graph,
Cellole
analogous
graph
for
the
ACI
more
for
the
same
Cluster
designed
in
space,
and
here
too
we
see
a
considerable
reduction
in
the
consumption
of
resources.
Where,
however,.
We
see
that
even
before
it
stood
at
idle
on
the
two
c's
plus
and
when
it
is
used,
it
went
further.
While
we
see
that
at
a
certain
point
onwards,
it
begins
to
have
this
square
wave
with
a
reduction
in
food
consumption.
E
E
See
now
some
numbers,
it's
been
a
while
and
let's
see
that
this
same
cast.
Now
we
have
75
in
heart.
Space
of
48,
with
which
the
total
Poldo
number
of
the
whole
cast
is
activated
is
1050
while
when,
as
Green
active,
they
become
450.
Therefore
600/600
Pods
are
turned
off
and
the
consumed
memory
is
reduced
by
33.
Gb
more
consumed
is
reduced.
3.5
there
is
more
This
is
the
effective
consumption
that
is
reduced
Anita.
It's
not
for
nothing.
E
We
see
that
for
1050
Pod,
the
consumption
of
CO2
in
one
week
in
kg
out
of
two
without
Cube
Green
would
be
242.
However,
Active
Green
So,
considering
1050
Pods
during
working
days
and
450
Pods
during
nights
and
weekends,
the
production
of
CO2
equivalent
is
reduced
by
91
kg
per
week,
therefore,
in
any
case,
a
considerable
reduction
in
consumption
for
our
coffees
for
nothing
beyond
this.
E
E
Yes,
what
we
talk
about,
then
at
the
end,
but
let's
see
now
A
demo
of
how
Cube
Green
actually
works,
So
here
or
Kind
the
place
can
afford
in
docker,
where
made
the
Ghetto
of
the
Pods
and
let's
see,
and
let's
see
that
I
said
one
that
I
won't
go
to
sleep,
because
it
is
possible
to
exclude
some
probably
from
the
slip.
An
Eco
Service
with
a
replica
is
a
service
with
4
epic
Now
I'm
going
to
get
this,
which
is
a
configuration
of
slip
info
for
every
day
a
week
at
9:53.
E
And
let's
see
that
I
don't
know
9:52
But
if
I
go
to
see
the
pods,
they
will
come.
Stars
are
already
in
terminating
because
because
in
any
case,
it's
there
when
he
goes
to
check
if
he
has
to
fall
asleep
or
not
a
little.
There
is
a
confidence
interval
of
plus
or
minus
a
minute,
so
I
brought
953
plus
or
minus
a
minute.
So
53
went
to
sleep
ipod.
We
can
go
and
see
that
the
number
of
iPods
are
probably
set
to
0.
E
E
Maintenance
is
done
through
a
Secret
which
is
installed
inside
each
spent,
which
defines
how
the
pods
were
at
the
beginning.
So
let's
go
and
see
for
a
moment
the
sleeping
bit
we
come
to
the
specification
we
saw
earlier,
the
coffee
you
want
and
the
timezone
Europe
Rome
had
to
be
specified.
In
addition
there
is,
it
is
the
former
itch
where
we
went
to
exclude
the
deployment
of
sleep
from
the
slip
here.
It
is
also
possible
to
remove
the
one
you
want
to
open,
so
simply
go
to
switch
off
the
namespace
without
going
to
reactivate
it.
E
This
can
be
useful,
for
example,
for
projects
that
are
used
very
much
rarely,
and
so
when
I
come
back,
Friday,
maybe
I
want
I
want
that's
enough
for
me.
It's
perfectly
fine
to
deploy
my
resources,
wait
for
Palma
not
running,
and
every
night
I
want
it
to
be
turned
off,
because
I
don't
care
if
it
stays
on.
Even
now,
it's
9:54
So.
If
I
go
to
see
The
Pod,
we
see
that
they
have
been
reset
correctly.
E
E
B
E
They
can
be
save
a
history
of
the
changes
made
by
the
controller,
for
example,
I
promise
you
so
that
you
can
manage
see,
verify
monitor
then,
who
were
the
slips
of
which
in
space
of
which,
just
how
many
to
suspend
the
crowjobs,
not
only
the
Flow
But
when
I
go
to
sleep
all
the
weight
I
would
also
like
to
suspend
the
crowjob.
This
is
a
work
in
progress
and
in
the
sense
that
there
is
a
specific,
is
already
open
on
which
I
am
working
and
should
soon.
B
E
To
manage
resources
and
then
also
view
the
history
see
how
many
Spaces
they
have
in
fact
goes
And
how
long
they
are
asleep
when
they
are
asleep.
In
addition
to
view
the
history
of
when
they
were
tormented
and
a
bar
to
be
able
to
view
energy
savings.
Therefore,
I'm
a
little
on
the
CO2
saved
this
that
this
is
being
worked
on
shortly.
We
should
have
the
first.
The
first
then
be
able
to
suspend
the
operation
of
everything
as
Green
and
therefore
be
able
to
say
not
on
the
whole
crankcase.
Do
not
go.
E
Do
sleep
Wake
up
or
do
a
trigger
is
immediate
So.
I
want
to
put
to
sleep
immediately
or
wake
up
immediately:
That
determined
the
expenses
without
without
waiting
for
the
correct
time
to
come,
use
cards
or
m
to
decide
whether
or
not
to
slip-on,
and
then
what
was
the
original
idea
to
stop
in
case
of,
please
call
one
of
The
Spaces,
which
is
always
waiting
for
a
feature
that
I
have
in
mind
very.
E
Even
if
quite
complicated
and
then
deactivate
the
namespace
with
with
down
then
create
some,
let's
say:
membranes
Space,
where
we
are
going
to
define
the
creation
The
Space
time.
This
will
be,
will
be
kept
active
at
a
certain
point,
but
all
turned
off
Or
again
cluster-wide
management
in
which
you
can
put
a
CRD
that
acts
on
the
whole
cast.
Then
they
are
put
to
sleep
with
some
namespace
annotations,
the
various
unexpressed.
E
A
Hello
Davide
Excellent
excellent,
Talk
Among
other
things,.
A
lot
to
current
events
happening
right
now,
So
perfect.
Then
we
have
a
lot
of
questions.
Actually
I
saw
that
some
asked
the
questions
in
chat
instead,
who
is
coming
and
therefore
Mattarella
takila
asked
who
asked
in
chat
Please
use
yours,
but
since
there
are
so
many
and
we
probably
won't
be
able
to
actually
not
be
able
to
answer
them,
all,
I
ask
you
maybe
to
try
to
answer
them.
Directly.
A
E
Currently,
it
only
works
with
the
diploma
in
20
But.
Exactly
That
is
In
Progress
to
also
support
crowjob
And.
In
any
case
now,
the
new
support
suitable,
for
example,
of
step
Pool
rather
than
filling
boxes,
rather
than
anything
else.
It
is
possible
to
include
it
without
a
great
deal
of
work
and
therefore
that's
enough.
A
A
Ok
So:
this
is
a
bit
long.
Now,
I'll
read
it
to
you,
Pietro
piutti.
Have
you
considered
the
impact
of
the
hardware
on
the
environmental
impact?
First
part,
then,
even
before
filing
at
the
software
level,
the
choice
of
the
underlying
hardware
should
be
considered
and
favor
more
energy-efficient
systems
such
as
arm
compared
to
Intel,
AMD,
x86
architecture,
and
also
they
are
on
x86.
It
is
possible
to
choose
more
efficient
CPU
models.
That's
fine!
So,
basically,
if
there
is
also
a
consideration
to
be
made
on
the
hardware
and
not
only
on
the
software
above.
E
Yours
then,
certainly
also
a
modification,
for
example.
Changing
the
architecture
therefore
going
assoarma
could
lead
a
reduction
in
consumption
reports
a
fairly
considerable
reduction
in
consumption.
However,
one
thing
does
not
exclude
the
other
reality.
They
are
quite
in
parallel,
In.
The
sense
that
it
is
known
that
he
has
already
given
up
the
software
Shout
at
another
time,
I
have
also
in
parallel.
He
has
already
given
up
at
war.
We
have
neither
much
greater
than
consumption,
and
this
is
a
good
thing.
I,
don't
see
them
as
two
parallel
paths.
E
So
then,
on
the
hardware
side
of
action,
we
didn't
consider
the
only
one
in
the
first
instance,
because
in
any
case
we
are
linked
to
a
provider
that
supplies
its
hardware
So
either
internally.
We
are
also
evaluating
to
go
on
présentoir
architectures
But,
precisely
instead
of
on
the
classic
x86
internally.
My
step
and
I
take
it.
However,
this
is
one
in
progress
Also,
because
in
any
case
the
transition
can
involve
That
is,.
It
is
not.
It
is
not
Despicable.
E
A
E
B
E
Analogous
to
a
cronjob
in
which
every
certain
time
goes
to
sleep,
to
wake
up
our
in
space
for
in
the
future
point
1
in
the
road
map,,
I.e.
considering
the
possibility
of
going
to
turn
off
or
awaken
Space
itself,
is
something
that,.
However,
is
not
yet
on
which
I
still
have
no
clear
idea
on
how
to
do
this.
This
evolution
and
therefore
not
then
not
I,
know
how
it
will
be.
A
E
So
at
the
moment,
there
is
no
this
conflict
when
you
share
Pia.
If
we
put
the
replies,
Zero
nice
noticed
anyway
that
she
has
them
reactivate
So.
If
you
use
used
the
Cate
who
spies
there,
not
at
all
zero
should
let
her
sleep.
Let's
say
it's
not
true.
Let
her
reply
0,
then
in
the
sense
we
did
some
tests.
They
could
also
they
could
be
in
the
old
version
of
which
they
could
So
in
case.
Let
me
know
if
there
is
something
I'm
looking
at.
If
you
use
standard
plans,
I
can
think
of
I.
A
E
The
hour
is
Sila,
the
use
of
kinship
actually
has
a
very
low
influence
because
it
still
consumes
That
is,
at
a
location
of
a
few
hundred,
more
not
more
so.
Actually,
the
gain
is
much
more
than
who
knows
what
makes
you
lose
anyway.
This
Peretola
is
a
bit
installed
inside
the
character
To,
which
resources
must
be
allocated.
E
In
the
when
defined
the
various
possible
paths,
those
with
job
would
be
activated
only
when
necessary.
In
that
case,
there
would
not
have
been
less
consumption
of
resources
But.
In
fact,
it
would
have
been
more
inconvenient
for
the
user
So
from
one
point
of
view,
I
preferred
Finding
a
comfortable
way
for
the
end
user,
even
if,
in
any
case
that
means
having
a
little
more
about
our
Caste
tests,
a
podcast
turn
off
600
In
our
case.
It
has
been
a
month,
That
is,,
a
favorable
thing,
let's
say
numerically.
A
Ok
Since
I,
see
that
there
are
those
who
ask
Where
is
the
Q
and
in
the
chat
on
the
right
of
the
stream,
there
are
the
tabs
above
chat
participants.
There
are
sub
chats,
there
are
some
sub
Tabs,
there
is
the
general
stage
and
then
who
comes
So,
the
third
plus
and
there
you
can
ask
this.
Your
questions
I'll
go
on
to
the
next
ones.
A
We
still
have
a
few
minutes
and
your
exam
is
not
finished
yet
then
Carlo
Reggiani
says
or
a
super
Metal
openshift
cluster
and
I
was
just
looking
how
to
automate
shutdown
of
an
e
Space
Pods
at
night
to
free
up
resources
for
a
selenium
that
runs
at
night.
Is
it
possible
to
set
namespaces
to
shutdown
and
not
So
those
to
be.
E
Are
imported
In,
The
Space
to
be
spent
in
the
sense
that,
through
the
CRD,
the
wax
must
be
installed
inside
the
Crafter
to
be
extinguished,
so
it
is
already
in
time,
if
only
those
that
going
to
shut
down
you
don't
go
to
hatch
itself,
as
the
include
list
is
more
was
more
Safer
was
the
former
CUD.
Let's
imagine
we
have
only
one
Gast
prod
in
OffRoad,
doing
it
by
exclusion
was
a
little
more
dangerous
In
the
sense
that
they
risked
losing
a
production
and
from
the
Apennines
I
look
at
Unes
production,
Ok.
A
Then
I
would
say
that
we
can
close
this
session
here
more
and
in
as
I
said,
there
are
still
some
questions
and
maybe
you
will
have
more
Davide.
You
can
go
and
answer
them,
maybe
in
chat
or
in
or
in
dm.
They
are
direct
messages
directly
with
those
who
have
questions
on
social
media.
You
can
contact
Davide,
so
much
I.
Thank
you
for
your
Talk
and
for
this
additional
exam
session.
You
have
that
you
have
faced
and
nothing
I
wait
for
the
next
round.