►
From YouTube: CNCF Cooperative Delivery WG 2021-09-30
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
A
B
B
A
Saw
it
I
didn't
have
a
chance
to
review,
but
I
haven't
really
gotten
into
the
go
code
there
honestly.
So
I
this
will
be
a
chance
if
I
review
that.
B
Yes,
so
it
was
for
smaller
for
smaller
exercises,
because
I
simply
created
one
series
for
the
arms
and
one
one
for
the
legs
and
made
a
bit
more
configurable
via
my
environment
variables,
because
I
had
to.
I
had
to
adopt
it
for
a
course
for
the
unifor
university
and
therefore.
B
But
it's
funny
because
this
this
example
I
will
use
outside
of
kubernetes
and
we'll
try
to
demonstrate
kind
of
a
service
mesh
without
giving
it
this.
So
this
can
could
be
very
funny
with
console.
A
So
yeah
I've
been
working
on
I
mean
I,
I
got
catch
customized
helm
all
those
done,
but
I
like
there's
one
last
bug,
but
I'm
gonna
move
on.
I
really
want
to
do
argo
and
flux.
So
that's
and
captain-
I
guess
I
know
you
all
are,
I
think,
captain's
pretty
cool.
So
I
want
to
get
those
practiced.
A
A
B
A
But
yeah
that's
enough
trouble
shooting
github
actions.
A
It's
me,
you
and
robert
rob
earth.
C
It's
actually
just
pronounced
the
way.
You'd
think
it
would
be
with
no
wish
got
it.
B
A
B
A
This
is
the
theme
that
I've
been
pursuing
for
a
little
while
now,
like
I
kind
of
see
kubernetes
and
the
operator
framework
and
all
that
kind
of
stuff
as
the
foundation
for
the
standard
api
for
infrastructure
management,
so
that,
like
app
devs
can
come
in
and
you
know,
use
the
same
paradigms
to
deploy
a
database
or
a
cassandra
database
or
a
kafka
topic,
or
you
know
dynatrace
observability
space.
Just
it
should
all
be
kind
of
kind
of
similar.
That.
B
A
I
kind
of
sensed
that,
in
your
like,
you
were
kind
of
you
mentioned
that
a
couple
times
in
there
yeah.
I
love
that.
Actually,
I
have
a
diagram
and
I
can
show
it
with
you
and
that
one
of
the
things
that
has
made
me
think
of-
and
I
wanted
to
bring
up,
maybe
in
this
meeting,
is
just
as
a
challenge
is
like
is
serverless
a
mode
of
cooperative
delay.
Is
that,
like
the
ultimate
like?
Is
there
like
a?
A
A
One
thing
in
the
article
that
I
was
a
little
disagreeing
with
I'm
not
sure
is
that
you
kind
of
say
that
kubernetes
itself
or
the
cluster
itself
is
static,
like
should
be
pre-configured
and
just
share
the
contracts
which
I
guess
it's
okay
like.
I
would
definitely
want
to
see
that
all
encoded
and
the
contracts
or
whatever
encoded
somewhere,
you
know,
and
then
so
then
it's
not
quite
as
static,
so
that
that
I
might
be
being
pejorative
by
saying
static.
But
you
know
maybe
maybe
a
little
more
flexibility.
A
The
infrastructure,
stuff
was
or
a
little
more
dynamism
there
at
the
cluster
level
might
might
fit.
Is.
B
This
I
I
agree
with
this
at
some
point,
so
the
cube
need
this.
Caster
itself
is
not
very
static,
so
the
nodes
change
come
go
and
whatever,
but
the
object
itself
is
more
or
less
the
all
the
time,
especially
when
you,
when
you're
working
with
production
systems
or
whatever
at
least
you
know,
at
least
in
my
landscape.
B
B
B
There
is
such
thing
as
an
infrastructure
team.
There
is
an
application
development
team
and,
as
we
all
know,
often
the
the
common
sense
is
we
put
something
over
the
fence,
and
I
want
to
do
something
against
this.
So
the
the
idea
I
had
with
this
blog
post
was
to
give
people
an
idea
how
how
such
things
could
look
like.
So
it's
for
me.
B
It's
obvious
that
then
kubernetes
caster
needs
maintenance
and
instead
that
it
is
a
separate
component
and
also
the
platform
itself,
which
is
kubernetes
and
platform
components
need
some
kind
of
need,
some
kind
of
of
care.
So,
let's
think
about
operators,
let's
think
about
easter,
let's
think
about
whatever.
So
you
can't
say
that
easter
itself
belongs
to
an
application
or
belongs
to
a
microservice.
B
There
are,
there
are
components
in
your
in
your
landscape
which
might
come
which
might
belong
to
many
services,
but
they
have
to.
They
have
to
be
maintained
somewhere,
and
this
is
this
was
my
idea
of
the
whole
infrastructure
context
and
so
on,
where
we
said.
Yes,
we
have
infrastructure,
this
is
shared
by
different
teams
and
so
on,
and
we
agree
that
we
have
istio
inner
version,
1.15
or
whatever
in
in
the
control
panel.
B
I'm
not
really
aware
of
the
of
these
two
versioning
at
the
moment,
so
the
developers
at
ball
above
are
are
using
this
infrastructure,
so
they
are
using
the
communities
cars
that
they
might
use
the
service
mesh.
They
might
use
an
opa
and
whatever,
and
they
might
have
they
might
can
rely
on
that.
B
There
is
still
larger
than
0.16
larger
than
0.14
running
on
this
customer,
and
this
and
the
rest
is
totally
obvious
for
them,
so
they
can
create,
let's
say
lambda
functions
or
whatever
they
need
as
long
as
their
policies
on
the
infrastructure
structure.
Layer
for
that,
because
currently
I
see,
I
often
see
that
this
is
managed
by
pull
requests
and
managed
by.
Yes,
I
put
something
in
a
repository
for
infrastructure
repository,
and
this
is
this
is
one
of
the
things
we
we
should
deal
with
in
this
working
group.
B
So
for
me,
the
perfect
time
to
deploy
infrastructure
is
the
time
for
which
is
very
short
before
you
apply
before
you
deploy
the
application.
A
I
will
I
want
to
I
mean
if
it's
okay
with
you
I'll
just
comment
on,
I
think
what
you
put
the
dmarc
here
kind
of
at
the
cluster
you
said,
or
maybe
higher.
Actually
you
said
service
mesh
opa,
you
even
said
developers
could
deploy
lambdas.
I
just
I'll
just
tell
you
for
for
sure
where
I
am
they
can't,
but
it
should
be.
It
should
be.
A
I
do
agree
that
it's
probably
an
infrastructure
capability,
but
I
do
wonder
if
there
are
other
places,
because
I
see
this
like
in
flux
and
argo
like
they
emphasize
deploying
clusters
and
I'm
like
who
are
you
trying
to
talk
to
who's?
The
audience
like
are:
is
the
application
teams
also
deploying
a
cluster?
I
don't
even
know
like
no
application
teams.
Are
you?
Don't
we
don't
find
application
teams
managing
their
own
clusters?
That's
I
think
you're
shaking
your
head,
thomas.
Yes,.
B
So
when
you,
when
you
read
about
microservices
and
so
on,
you
often
see
that
that
services,
such
as
databases
and
so
on,
are
part
of
the
application
itself.
B
So
if
someone
writes
an
application
within
database,
the
most
of
the
authors
think
that
or
say
that
the
application
belongs
to
the
the
database
belongs
to
this,
and
this
was
one
thing
I
wanted
to
point
out
in
this
blog
post
that
that
the
database
is
not
part
of
the
infrastructure
below
it's
part
of
the
application.
A
B
A
That's
interesting.
I've
been
thinking
about
the
same
thing
like
let's
say
pipelines.
Let's
say
you
want
to
run
pi
tecton
pipelines
out
whatever
kind
of
pipelines
there's
so
people.
Think
of
that
as
an
app
delivery
concern.
Of
course
it
is,
but
somebody
has
to
manage
the
deployment
of
tekton
well,
you
could
have
a
managed
instance,
I
guess-
or
you
could
run
it
yourself.
Somebody
that's
infrastructure
that
that's
a
new
thing
that
I've
been
realizing.
A
You
know
the
past
few
weeks
is
like
because
I'm
like,
what's
devops
versus
infrastructure
and
they're
really
different,
because
because
the
devops
people
are
using
infrastructure
still
just
like
the
application
people
you
it's
not
like
the
infrastructure.
People
are
the
devops
people,
they
run
jenkins,
the
actual
servers
and
they
run.
You
know
techton
the
the
controllers
and
operators,
but
they
are
not
the
user.
That's
a
different
class.
B
So
there's
one
one
one
kind
of
team
or
a
part
of
a
company
which
provides
the
the
the
infrastructure,
the
infrastructure-
let's
say:
yes,
the
the
possibility
to
deploy
infrastructure.
This
could
also
be
operators.
This
could
also
be
cross
plane
and
whatever
we
know,
but
there's
also
the
other.
The
other
point,
the
other,
the
other
side,
which
consumes
exactly
this.
A
So
what
do
you
mean
by
that?
Like
top
down
buy-in?
Sorry,
what
are
you
getting
at?
You
said
executives
like
what
do
you
mean
by
that
like,
so,
we
need
to
explain
to
executives
that
infrastructure
needs
to
be
considered
like
what
is
what
what
is
it
that
with
executives?
What
are
you
getting
at?
I.
B
B
Is
this,
this
might
be
a
problem
you
you
could
just
describe
to
infrastructure
guys,
and
I
am
an
infrastructure
guy,
so
retrieved
it
to
the
application.
Well
today
to
the
application,
ops
world
so
and
I
think
some
I
think
some
people
have
to
should
learn
this
or
should
at
least
consider
consider
such
a
way
of
working
and
giving
up
giving
up
some
control
but
doing
control
via
policies
and
not
viable
regrets,
because
if,
as
an
application
developer,
I
would
need
some
kind
of
infrastructure.
B
Well,
it
was
very
funny
funny
to
wrap
this
okay.
Yes,
as
I
think
the
other
twos
will
not
join
us,
this
should
be
more
or
less
a
kick
of
meeting.
Unfortunately,
there
are
not
so
many
so
many
people
people
there
right,
but
that's
that
yes,
hopefully
this
will
get
more
in
the
future.
B
So
at
first
thank
you
for
coming
and
one
positive
thing
and
in
the
beginning
we
are
a
working
group
now,
so
we
have
all
we
have
all
necessary
worlds
and
we
can
act
as
a
working
group
of
the
undertaker
deliverance.
This
was
the
first
thing.
The
second
thing
is
we,
and-
and
I
will
will
send
out
this
after
the
meeting.
B
One
of
the
first
things
we
have
to
do
is
such
a
such
a
working
group
is.
We
have
to
we
have
to
vote
on
on
chairs
and
so
on,
because
someone
has
to
drive
the
progress
further
and
alex
and
me
will
join
the
will
join
the
working
group
for
the
state
leads
we
can
all
we
can
only
join
and
cannot
cannot
lead
to
the
working
group.
B
So
if
someone
of
you
is
interested
in
sharing
this,
please
I
will
send
out
something
for
some
nominations,
something
for
nomination
and
please
comment
on
this
and
and
then
we
can
vote
on
this.
B
B
Yes,
furthermore,
to
the
to
the
working
mode
of
this
working
group,
so
this
was
more
more
kick
off
meeting
now
when
alex
and
me
started,
writing
writing
the
chart
and
so
on.
We
thought
about
making
this
working
group
a
bit
more
interactive.
B
So
that's
that
we
invite
some
team
some
some
external
end
users,
or
also
also
vendors,
to
show
their
their
approaches
for
such
things
and
that
we
can,
it
might
might
get
a
nice
community
also
for
end
users
to
get
there
and
try
to
find
out
what's
up
with
app
delivery
and
infrastructure
delivery.
So,
as
I
wrote
the
same
thing
as
I
wrote
the
blog
post
before
this
was
a
nice
discussion.
I
think
the
last
20
minutes
and
I
think
we
can
have
much
more
of
such
discussions.
B
So
I
think
this
will
will
not
be
the
only
way
or
how
one
could
deal
with
this
it.
I
also
think
there
are.
There
might
be
people
who
are
doing
this
in
their
ci
pipelines
or
whatever.
I
think
there's
nothing
nothing
wrong
with
it
so,
but
but
I
think
it
would
be
cool
if
some
some
people
would
choke
their
approaches
and
that
we
could
define
some
kind
of
patterns
or
whatever
to
get
to
get
this.
B
B
But
I
want
the
white
paper
to
be
honest,
not
as
not
to
be
an
initial
goal
of
all
of
this,
so
it
should
be
more
more,
be
getting
a
community
getting
getting
ideas
and
getting
cool
solutions
for
this
and
then
and
in
the
end,
if,
when
we,
when
we
think
we
have
enough
information,
then
just
write
the
white
paper.
I
think
this
could
be.
B
So
I
think
we
did
the
first
thing
in
the
potato
head
that
we
let's
say
we
spin
up
a
handcast
and
deploy
the
application
there.
This
is
a
bit.
This
is
more
small
thing
of
of
cooperative
delivery,
but
we
could
also
make
this
a
bit
a
bit
larger
and
as
far
as
I've
heard
yesterday,
microsoft
also
provides
azure
credits
for
open
source
projects.
B
Or
whatever,
so
wherever
we
have
the
possibility
to
deploy
and
find
that
get
ideas
how
to
how
to
deal
with
this
topic.
C
There
are
certain
platform-specific
ways
of
doing
some
of
these
things
that
might
be
valuable
to
document.
For
instance,
you
know
aks,
I
will
have
the
the
get
ops
add-on
tool,
which
is
basically
just
spinning
up
flux
and
everything
like
that.
But
at
least
you
know
having
that
as
a
as
a
way
of
showing
off
how
to
enable,
for
instance,
get
ups.
You
know
and
stuff
like
that.
C
So
there
are
certain
platform,
specific
things
where
we
could
use
azure
or
whatever
to
at
least
you
know,
get
things
written
down
and
have
screenshots
and
stuff
like
that,
and
and
obviously
everything
asher.
You
know
that
I
can
probably
do
I
have.
I
have
a
lot
of
credit.
I
don't
need
anyone
else's
credits.
A
C
Oh
aks
in
itself
just
cost
the
the
amount
of
money,
whatever
you
type
of
machine
you
put
it
in
and
and
obviously
you
know,
gain
to
connect
out
of
everything,
but
I
I
I
was
reading
up
on
some
of
the
goals
that
we
kind
of
said
before
this
meeting
just
to
have
something
to
think
about
and
contribute
with,
and
things
like
the
like.
C
The
second
to
last
point
that
give
end
users
id
as
an
example
how
they
could
integrate
application
infrastructure
deployments
there
are,
there
are
stuff
one
can
show
off
with
using,
for
instance,
only
terraform
for
the
entire
run,
both
the
infrastructure
and
actual
app
delivery
and
you're.
Not
having
examples
like
that,
but
also
having
you
know,
yeah
you
could,
you
could
use,
you
could
use
the
flux
provider
for
terraform
and
every
time
you
run
up
a
cluster,
it
would
automatically
point
towards
a
get
repository,
and
then
there
you
have
your
app
delivery.
You.
B
C
There's
a
lot
of
those
kind
of
examples
that
can
get
written
down
and
you
know,
conveyed
in
a
ways
that
people
actually
understand
if
it
fits
their
needs.
So
there's
a
lot
of
things
like
that.
C
One
can
start
doing
already
now,
but
having
a
place
where
this
ends
up
would
be
great
like
having
you
know
some
of
these
deliverables
or
whatever
I'll
call
it.
If
we
can
kind
of
like
establish
some
sort
of
cadence
for
how
where
and
how
and
who
should
look
at
it
before
you
know
it
gets
accepted
and
all
those
kind
of
things.
If
we
have
that,
I
can.
I
can
write
a
lot.
B
So
I
think
we
will
need
some
kind
of
kid
repository.
I
think
we
should
use
that
keep
proposing
for
that,
because
there
were
a
lot
of
lots
of
discussions
regarding
such
things
and
then,
if
we
find
out
that
this
doesn't
fit
anymore
or
that
that
it
gets
too
much
to
to
be
managed
here,
then
we
can
put
it
out
in
a
separate
repository
organization
or
whatever
we
can
define
code
owners
for
for
our
cooperative
delivery,
stuff
yeah
and
therefore
this
should
be
no
problem.
C
Yeah,
if
we,
if
we
can
yeah,
even
just,
have
like
deployment
examples
as
a
sub
folder
directory
and
and
have
markdown
per
dip,
you
know
example
or
something
that's.
B
I
think,
in
the
first
step,
if
we
have
some
general
information
or
whatever
we
want
to
document
somewhere,
the
app
delivery
repository
would
be
the
correct
thing
for
that.
If
we
have
concrete
examples
so
where
we,
where
we
want
to
deploy
an
application
cooperatively,
then
we
should
use
the
potato
head
and
should
also
also
define
it
there,
because
the
potato
head
is
also
part
of
the
of
that
delivery.
C
Yeah
well
everything's,
okay
with
me,
but
I
also
think
there
there
might
be
some
value
to
you
know
having
practical
suggestions,
for
instance,
if
you,
if
you,
if
you
every
time
you
want
to
you,
know,
deploy
a
new
cluster,
there's
certain
base
services
that
needs
to
get
deployed,
because
that's
what
you
deliver
to
your
people,
you
know
like
prometheus
and
stuff,
and
you
know
it
would
kind
of
be
cool
to
actually
use
that
as
a
an
actual
practical
example.
C
But
but
obviously,
if
you
point
flux
or
argo
to
a
to
a
git
repository,
it
doesn't
really
matter.
What's
there,
it's
gonna
get
deployed,
so
you
know
getting
to
at
least
that
point.
B
If
we,
if
we
are
planning
to
do
some
delivery
examples
or
whatever,
where
we
need
specialty
repositories
and
so
on,
we
have
now
organization
for
the
potato
head
and
we
can
create
the
repositories
as
far
as
we
as
we
want.
So,
if
you,
if
we
need
to
point
to
some
repository,
this
is
also
possible.
I
think
yes-
and
I
think
you
you
spoke
about
one
very,
very
important
thing.
B
B
We
want
to
make
their
lives
easier
and
we
want
to
to
get
them
in
a
position
that
they
say.
Yes,
I
need
a
solution
and
I
can
take
a
look
in
some
repositories
of
the
depth
of
the
gap,
delivery
and
find
out
how
to
deal
with
this
topic
and,
in
my
opinion,
it's
pretty
it's.
It's
not
really
relevant.
B
If,
if
this
is
at
some
cloud
provider,
if
this
is
a
cncf
project
or
not
it's
in
the
project,
it
should
also
it
should
only
bring
some
value
for
the
further
for
the
end
users,
because
that's
what
we.
What
we're
here
for
to
be
honest,
and
so,
if
we,
if,
if
we
have
some
kind
of
terraform
example
where
only
terraform
deploys
the
whole
on
the
whole
infrastructure
and
and
the
applications,
it's
also
perfectly
fine,
because
it's
a
it's
a.
B
So
I
think
there
will
be
many
solutions
which
will
be
based
on
terraform.
Obviously,
because
it's
one
of
the
tools
which
is
mostly
used.
C
Yeah-
and
you
know
it's
one
of
the
tools
that
people
are-
you
know,
want
to
use
or
are
looking
into
using,
and
you
know
it
makes
sense
to
me
and
then
again
that's
kind
of
what
I
work
with.
So
maybe
that's
why
it
makes
sense
to
me,
but
but
but
there
are
obviously
other
things
or
you
know,
there
are
obviously
other
ways
of
kind
of
getting
the
entire
cooperative
delivery.
C
You
know
point
across
and
showing
how
to
kind
of
you
know,
work
with
both
infrastructure
and
app
delivery
in
a
way
that
makes
it
easy
for
developers
and-
and
that
doesn't
mean
it
has
to
be
infrastructure.
As
code
they're,
you
know,
certain
providers
have
cloud
providers,
have
certain
services
that
is
aimed
to
make
it
easier
for
people
in
in
this.
You
know
integrate
into
things
like
kubernetes,
so
there
might
be
many
different
kind
of
things
that
we
can.
C
You
know
show
as
practical
examples
and
they
might
even
be
short
things
that
don't
necessarily
have
to
be.
You
know
extreme
long
document
and
complete
how
to
step
by
step.
So
there's
a
lot
of
things
we
can
do
relatively
easily,
but
but
but
yeah.
If
we,
if
you
get
a
a
directory
for
this
and
maybe
set
up
the
code
owners
on
on
that
directory
or
the
corporate,
the
working
group
directory
at
least.
C
Yeah,
I
would
at
least
help
be
able
to
contribute-
and
I
think
things
like
the
other
things
that
are
kind
of
been
set
as
goals
are
things
like
landscape
radar
and
the
white
paper
yeah.
That
does
a
lot
of
work.
C
A
I
mean
if
we
contribute
a
bunch
of
examples.
Those
will
help
too,
like
even
you
describing
that
you
like
to
use
terraform
like
you
could
probably
do
similar
things
with
crossplane
and
we're
all.
A
There
at
my
environment,
I
had
to
use
a
lot
of
imperative,
which
is
less
mature.
I
would
say
so
yeah
as
we
like
put
these
things
together,
we'll
probably
come
up
with
a
maturity
model
like
imperative,
with
bash
scripts,
declarative
with
terraform
declarative
with
kubernetes
resources
and.
A
B
Yeah
and
those
would
be
cool
so
at
some
point,
if
we
see
some
kind
of
some
kind
of
patterns
or
how
how
such
things
could
work,
we
could
all
also
write
some
write.
One
write
a
short
blog
post
or
whatever
regarding
this
as
red
hat
did
with
operate
the
maturity
model
or
whatever.
So
it's
all.
Things
are
opened
to
get
to
the
point
of
robot
before
I've
throwed
one
white
paper
soda
operate
the
white
paper.
A
B
C
And
and
and
to
that
point,
when
the
the
first
thing
in
the
list
there
is
vendor
and
then
user
into
use,
do
we
want
to
do
the
same?
There
have
a
a
should.
I
I
kind
of
look
at
this
as
kind
of
just
gathering
a
lot
of
data,
so
we
have-
and
I
you
know
a
directory
for
for
those
interviews
with
you
know
we
could
start
doing
that,
but
I
I'm
not
really
sure
how
to
go
about
that.
Should
we
have,
you
know,
set
questions.
C
C
Yeah
that
one
is
not
totally
clear
to
me.
What's
like
the
the
best
approach,
I
see
the
value
of
it,
but
I
I
I
it's
kind
of
open-ended.
I
think.
B
I
think
this
is
also
a
thing
we
can
discuss
a
bit
later,
especially
when
alex
is
here,
because
I
think
he
brought
brought
up
this
topic
yeah
and
but
I
don't
think
that
this
is
one
of
the
things
we
should
we
have
to
tackle
in
the
beginning.
B
So
in
the
beginning
I
would
for
me
it
would
would
be
really
cool
if
you
get
some
projects
and
users
in
there
and
get
some
demonstrations
or
some
some
discussions
on
how
this
works
in
their
environment.
Well,
where
they
face
problems.
A
So
we've
said
three
kind
of
streams:
we've
got
practical
gathering
up
and
loosely
gathering
loosely
coupling
bunches
of
examples.
We've
got
gathering
feedback
from
end
users
on
their
biggest
problems
and
jobs,
and
then
we
have
writing
an
opinionated
white
paper
and
we're
saying
focus
on
gathering
up
examples
to
start
and
then
the
next
two
will
will
come.
B
A
And
databases
and
queues-
and
it
would
like
it's
a
different,
so
we
talked
about
imperative,
declarative
flux.
This
is
a
whole
different
paradigm,
where
the
application
developer
just
declares
the
capabilities
they
need
and
the
infrastructure
injects
it.
So
that's
almost
like.
A
Everything
on
its
head,
if
that's
something
that
I'm
just
bringing
up
something,
I
hope
to
to
try
in
our
in
the
potato
repos
at
some
point-
and
I
mentioned
to
them
that
we're
doing
this-
and
maybe
they
want
to
you-
know,
talk
to
us
at
some
point
or
visualize.
The.
C
Thing
that
would
be
great
and
that
actually
kind
of
solves
a
lot
for
developers.
I
think
it's
one
of
those
things
where,
where
we're
gonna
probably
create
something
in
the
future,
and
I
have
to
talk
to
developers
who
aren't
really
that
cloud
native
of
you
know
they
kind
of
done
a
lot
of
stuff
in
the
past,
but
not
in
the
real
cloud
in
any
way
and
having
something
like
that,
would
make
it
a
lot
more
easy
for
them.
So
I
could
just
say:
here's
a
kind
of
framework
you
can
work
out
of
it.
C
You
know
write
whatever
you
want,
but
here's
how
you
get
to
that
and
that
helps
them,
and
obviously
that
makes
like
the
application
better
suited
for
these
kind
of
environments.
But
at
the
same
time
I
I
I
met
clients
looking
over.
C
You
know
the
clusters
of
someone
that
created
an
application
for
them,
and
I
see
just
today
I
saw
not
only
the
application
they
created,
but
also
the
ingress
controller
and
search
manager,
and
everything
like
that
in
one
single
name,
space
everything
they
deployed
was
in
one
name
space
I
kind
of
like,
but
this
this
is.
This
is
not
correct,
but.
C
C
Why
it's
not
great
to
have
one
namespace
for
everything
that
you
do
and
things
like
that
and-
and
I
think,
there's
a
lot
of
stuff
like
that
we
can
kind
of
get
across
as
well
just
again
going
back
to
goals
where
we
can
be
written
about
kind
of
who
these
key
stakeholders
are,
and
specifically
the
ckid,
the
developer
part,
not
the
administrator
part,
but
the
you
know
developer
how
how
can
help
them
do
better
job
with
with
you
know,.
B
So
one
thing
rs
brought
up
when,
when
we
had
the
idea
for
this
working
group
was
also
kind
of
getting
delivery
ready
when
he
discussed
the
without
without
intention.
A
A
Do
we
need
is
one
of
the
themes
bringing
infrastructure
development
into
the
cloud
like,
of
course,
at
azure,
they're
building,
you
know
cloud
native
services,
but
like
out
at
the
customer.
Do
they
understand
what
modern
infrastructure
is
that
it's
all
software
defined
that
it's
dynamic,
that
it
should
be
agile
like
do
we
have
that
might
be
beyond
our
scope?
That's
what
I'm
kind
of
asking
like
do.
We
have
a
remit
to
educate
and
and
upskill
infrastructure,
folks
from
being
cis
admins.
To
being
you
know,
developers.
B
So
if
it,
if
it
fits
to
the
to
the
to
the
topic
area
of
the
of
the
working
group,
yes
because
we
are
cold
guys,
therefore
we
should
help
them
to
get
to
to
get
to
the
cloud.
A
A
Along
in
the
journey
but
yeah,
how
do
we?
How
do
we
do
that
nicely?.
C
I
I
think,
there's
a
lot
of
things
that
we
could
do.
That
would
help
a
lot
of
people,
and
I
I
you
know,
I'm
based
in
norway
and
and
I
work
with
a
lot
of
like
public
sector
governmental,
big
things
and
they're,
starting
to
get
into
like
their
own
private
cloud
idea
and
they're
and
they're
doing
a
fine
job
at
it.
You
know,
obviously,
a
lot
of
people
need
to
get.
C
You
know,
get
up
to
speed
on
certain
things,
but
things
like
what
does
what
does
a
cluster
like
a
cluster
look
like
when
it's
deployed
for
a
developer
if
a
developer
needs
a
cluster?
What
does
that
look
like?
What's
that?
Base
cluster
and
a
lot
of
them
are
just
thinking
yeah?
Well,
it's
a
cluster.
C
You
know
on
vmware
or
in
in
azure
or
whatever
they
use
as
the
backbone,
just
give
them
a
cluster,
and
it's
like!
Well,
you
know
what
about
monitoring?
What
about
policy?
What
about
the
network
policy
security,
things
there's
a
lot
of
things
there
that
kind
of
needs
to
be
present
and
I
think
getting
the
point
across
that
a
base
cluster
is
not
just
kubernetes.
For
instance,
it
could
be
a
thing
we
could
try
to
to
to
to
get
across
and
that
would
enable
developers
to
safely
deliver
the
application.
C
A
I
love
that
yeah
like
force
them
to
think
from
a
devs
viewpoint.
The
job
isn't
done
because
you
installed
a
server
with
microsoft
windows,
server
on
it
like
yeah,
the
dev
needs
more
than
that
to
action
yeah.
If
we
can
that's
what
I
was
actually
bringing
up
serverless
before
because,
like
maybe
instead
of
thinking
from
the
bottom
up
like,
I
need
to
provide
compute
and
storage
and
network,
and
all
these
things
we
should
think
from
the
top
down,
like
my
developers
want
serverless
or
something
like
that.
A
How
do
I,
you
know,
implement
to
meet
that
kind
of
thing,
but
and
that
ties
into
what
you
just
said-
that's
what
I
was
getting
at
like
if
we
can
help
infrastructure
folks
think
like
hey
the
app
devs,
what
do
they
really
want
and
that
will
transform
their
whole?
You
know
so.
Yeah
that'll
be
cool.
Let's
keep
that
in
mind.
If
we
can,
if
we
can
facilitate
that
through
here,
maybe
like
educate,
you
know
how
to
write
bash
scripts
now,
do
it
in
go
or
something.
C
What
do
we
want
to
kind
of
set
off,
as
you
know,
first
order
of
business
like
what
what
do
we
want
to?
Where
do
we
want
to
start
with
this?
Do
we
just
you
know,
fly
off
to
see
those
bands
type
thing,
or
do
we
want
to
kind
of
set
up
some?
B
So
what
I
would
I'd
like
to
to
what
what
would
be
good
for
the
next
one
to
meetings?
What
would
be
that
we
get
some
kind
of
some
kind
of
examples
for
this,
so
that
someone
could
be
there
and
try
to
try
to
show
how
such
such
an
approach
could
look
like,
and
also
that
that
we
get
more
more
traction
into
the
working
group
that
we
get
more
more
people
working
in
the
working
group
so
that
we
we
have
something
to
show
and
something
which
which
enables
people
to
do
something.
C
I
didn't
have
edit
access
for
the
the
docs
all
right,
cool.
A
A
That's
all
I'm
going
to
do.
I'm
taking
away
contribute
some
examples,
I'm
going
to
have
some
time
next
week,
so
maybe,
when
I'm
working
on
potato
head
I'll,
try
to
integrate
dapper
or
something
in
one
of
them,
I'm
also
going
to
work
on
a
blog
just
to
kind
of
drum
up.
I
don't
know
I'm
not
that
I'm
not
that
popular
or
anything
but
like
just
just
start.
You
know
whatever
conversation.
B
C
If
yeah,
that
that's
probably
also
a
kind
of
better
way
because
you
kind
of
may
might
get
some
some
cncf
muscles
underneath
everything,
but
there
are
other
venues:
where
one
could
you
know
post
stuff
and
probably
get
some
people
in
just
you
know
as
being
part
of
the
gatehouse
working
group,
open,
open,
get
ops
has
a
web
page
where
we
can
put
a
blog
and
we
could
discuss
things
around,
get
ops
and
kind
of
get
into
this
topic
and
ask
if
people
want
to
contribute
their
ideas
to
you,
know
and
kind
of
get
people
in
that
way.
C
A
Yeah
it'll
be
interesting
if
there's
get
up
people
that
are
using
git
ops
for
like
a
combo
of
app
and
dev
like
is
there
somebody
sorry
about
infra?
Are
there
people
that
are,
you
know,
deploying
a
whole
cluster
or
deploying
a
cert
manager
or
whatever.
C
C
C
C
So
you
know,
obviously
I
try
to
get
them
into
the
getups
space
there
and
then
that
kind
of
just
you
know,
went
on
and
we're
talking
about
cross,
plane
and
other.
You
know
solutions
like
that,
and
that
might
be
a
really
cool
example
in
the
future,
but
right
now,
but
but
even
even
there,
just
the
fact
that
they're
gonna
go
from
doing
a
ci
cd
based
deployment
of
applications
of
their
application
to
probably
using
either
flux
or
ocd
to
to
actually
enable
get
ups
and
and
do
everything
there.
C
A
B
Yes,
because
it
would
be
cool
to
see
such
things
as
and
what
I
could
also
contribute
is
a
great.
As
I
said,
I
have
created
some
kind
of
showcase
for
capable
captain
terraform.
I
could
do
the.
I
could
show
this
in
one
of
the
next
meetings,
I'm
not
sure
when
the
next
one
will
be,
because
I
don't
know
how
it
is
on
location.
B
A
A
B
B
B
C
Yeah,
yeah,
and,
and
even
even
just
like
one
of
the
things
like
the
examples
that
were
talked
about
and
the
end
user
interviews,
but
also
just
the
the
entire
that
the
goal
of
capturing
the
current
practices
and
the
landscapes
like
what
what
are
people
actually
doing?
Should
we
should
we
create
a
a
a
questionnaire
that
we
kind
of
distribute
out
to
see
if
kind
of
results
we
are
getting
or
you
know
something
like
that,
we
can
do
and
then
do
a
write
up.
A
By
the
way
we
didn't
even
mention,
but
one
of
the
reasons
I'm
particularly
interested
service
binding,
like
that's
to
me,
one
of
the
issues
that
I've
had
and
I've
seen
it
where
I
work
right
now.
Is
you
get
the
service
deploy,
but
then
how
do
you
get
credentials
and
it
was
so
nice
in
pcf
and
in
cloud
foundry
and.
B
But
there
is
also
some
kind
of
service
cataloging
humanities.
I
think.
A
Yeah,
but
but
everybody's
kind
of
moved
away
from,
I
feel
like
open
service
broker
and.
B
B
B
A
C
Yeah
just
having
this
is
what
people,
what
what
kind
of
looked
like,
where
we
were
going
and
then
all
of
a
sudden
something
else
came
and
then
why
is?
Why?
Is
that
now
the
preferred
way
of
doing
that
and
stuff
like
that?
That's
a
good
exercise
just
to
understand
what
kind
of
drives
for
people
how
people
want
to
work
with
these
things.
A
Thomas,
can
you
make
the
notes
publicly
editable,
like
the
others
drop
yeah?
I.
B
B
Furthermore,
we
should
try
to
get
some
people
to
the
to
the
next
meeting
and
to
get
some
get
some
exam
get.
Let's,
let's
try
to
find
one
who
wants
to
present
something
and
I'm.
B
And
also,
I
think,
hong
kong
said
he
wants
to
do
something
with
kubrilla
and
open
the
open
application
model.
Yeah.
That's
also
like.
C
And
and
and
and
I
you
know-
I've
been
looking
that
they
have,
they
got
a
terraform
operator,
that's
kind
of
doing
so
it
was,
I
was
kind
of
interested
in
because
it
actually
does.
What
I
wanted
to
do,
which
is
point
to
where
terraform
is,
is,
and
then
stuff
happens.
Not
you
know
so
so
you
know
that
was
kind
of
cool,
but
I
kind
of
looked
through
there
and
yeah
it's
it's
mostly
china-based.
You
know
alibaba
people
that
are
doing
things
there.
C
A
B
C
That
might
just
be
you
know,
because
of
you
know
stuff
to
do
there,
there's
a
lot
of
things
to
do
for
a
lot
of
people
these
days
and
and
it
can't
be
part
of
every
single
project
and
if
whoever
you
know
was
the
the
one
that
wanted
to
to
help
with
that
kind
of
fell
out
or
or
you
know,
got
too
much
stuff
to
do,
then,
all
of
a
sudden
the
entire
company's
gone
from
there.
This
is
the
same
with
my
company.
C
A
So
interesting
because
I
brought
up
crossplane
as
basically
at
my
company,
which
I
might
not
be
staying
at
that
much
longer,
but
it
they
were
having
trouble
scaling,
we're
having
trouble,
scaling,
infrastructure
services,
and
so
I
brought
up
crossplay,
but
that
was
the
fear
also
was
like.
Well
that's
interesting
who's
going
to
maintain
that,
like
there's
a
lot
of
like
I
guess
that
goes
back.
To
kind
of
this
theme
of
like
infrastructure,
development
is
changing,
and
that
makes
people
nervous
it's
hard
to.
How
do
we?
How
do
we
make
them
feel
better.
C
Yeah
and
a
lot
of
these
things
are
defined
in
such
a
way
that
you
should
be
able
to
just
change
how
you
do
this
like
with
not
a
big
issue
like
you
could,
if
you
have
things
in
terror
from
now,
and
you
want
to
do
something
else,
you
can
do
that
you
can
take.
You
can
do
the
move,
you're
not
stuck
there,
just
because
that's
what
you
started
with,
and
but
people
fair
change,
so.
A
A
Like
a
thousand
percent
agree,
and
that,
like
that's
like
one
of
the
reasons
why
I'm
like
just
look
at
the
interface
and
then
you
could
do
whatever
implementation
you
want
underneath
like
I've
seen
companies,
they
don't
even
realize,
like
hey,
you
changed
the
url,
you
just
messed
up
all
my
docker
images.
You
just
stopped.
C
C
C
It's
probably
going
to
be
more
about
that
later
on,
but
obviously
we're
still
in
kind
of,
even
though
devops
have
been
a
thing
for
a
long
while
there's
few
of
us
ops
that
are
devs,
so
to
speak,
so
you
know,
there's
still
there's
still
that,
and
you
know
that's
why
terraform
is
popular
because
it's
not
actually
programming,
it's
very
easy
to
read
even
for
those
who
don't
understand,
but
obviously,
as
soon
as
it
gets
complicated
with
with
loops
and
and
doing
logic
things
they
get,
you
know
they
fall
off,
but
that's
why
I
have
a
job,
but
you
know
for
me:
I
I
couldn't
care
less.
C
If
it
was
written
in,
go
or
terraform
or
c
sharp
or
whatever
you
know,
I
can
still
read
it
and
I
can
still
edit
that,
but
there
are
few
you
know,
people
that
work
with
devops
that
actually
are
devops
they're,
more
ops,
but
they
kind
of
have
to
do
yaml
stuff
or
something
whatever
that's
called
yamalops.
C
So
you
know
getting
some
of
these
examples
out
there
and
kind
of
show
that
it's
not
really
that
hard.
Some
of
these
things,
it
is
actually
kind
of
easy.
You
just
need
to
to
sit
down
and
experiment
a
little
bit
and
you'll
you'll
find
a
way
and
it
it
does
work.
So
I
think
it's
thing
is
that
there's
a
lot
of
value
in
this,
this
working
group
and
the
data
that's
gonna,
get
collected.
A
C
And
that's
sort
of
what
terraform
kind
of
is.
You
know
because
you
say:
name
equals
this
all
right
cool.
Where
do
you
want
it
to
go
location
equal
this
and
there
you
go
hopefully
yeah?
Hopefully,
hopefully
it
goes
there
that
will
be
but
yeah,
but
at
the
same
time
you
know
people
are
just
stuck
in
the
this.
C
It's
very
obvious
in
norway,
I'm
not
sure
how
it
is
other
places,
but
in
norway
we
with
you
know
the
average
I.t
person
works
in
you
know,
with
15
year
old
technology
in
mind
that
they
can't
get
out.
They
can't
keep
up
with
everything
again
great.
For
me.
You
know
I
got
a
job
forever,
but
but
helping
people
understand
these
things
is,
I
think,
is
vital,
for
you
know
at
least
here
in
in
my
country,
for
things
to
move
on
and
be
better.
A
C
Yeah
I
keep
on
bringing
up
the
example
where
I
did
something
for
a
customer.
It
was
greenfield
deployment
and
the
old
service
provider
didn't
actually
tell
us
which
ip
ranges
we
could
use
or
not
and
we
sent
ip
wrenches
and
they
said
yeah.
That
sounds
good.
I
think,
and
then
we
deployed
everything
and
it
turns
out
it
was
overlapping,
stuff
and
different
end.
So
it
took
30
minutes
and
I
brought
down
the
entire
asher
environment
and
up
again,
we
redeployed
everything
because
it
was
you
know
it
was
terraform.
C
I
don't
care.
What's
going
going
down,
it's
going
to
come
up
again,
just
kill
it
all
really
yeah,
but
you
know
that's
kind
of
that's
why
infrastructure
has
code.
That's
that's
why
you
don't
do
things
manually
if
you
did
that
manually,
you
know
good
luck,
you
know
getting
everything
off
those
those
virtual
networks-
and
you
know
cleaning
up
all
that
this.
C
B
Yes,
in
the
currently
I'm
teaching
a
class
at
university
and
infrastructure
is
cold
and
defeating
microservices
and
so
on,
and
it's
kind
of
funny.
If
you
tell
them
yes
and
if
you
want
to
spin
up
500
virtual
machines
in
asia
or
sorry
in
aws,
just
do
it,
but
in
the
in
about
10
or
15
years
ago
you
would
have
ordered
one
two,
three
four
five
servers
you
would
have
to
wait
for
one
or
two
months
and
just
like.
C
And
even
when
I
started,
even
though
you
know
vms
and
stuff
like
that,
all
right
cool
thanks,
josh.
C
I
was
just
I
was
just
saying
that
the
even
not
that
long
ago
I
I
haven't
been
in
it
for
ages,
but
like
seven
years
ago,
I
you
know
even
then,
having
things
locally
and
not
out
in
public
clouds
or
something
that
scales
that
amount
you
know
would
mean
that
all
right
we
got
to
a
certain
point,
but
at
a
certain
point
it
comes
like
the
machines.
We
don't
have
any
more.
We
don't
have
any
more
resources,
so
we
actually
need
to
order
machines
and
then
the
quick
change.
Well
yeah.
C
It's
gotta
be
a
couple
weeks
because
we
need
to
get
the
machines
in
so
everything
like
that
now,
just
the
way
you
can
scale.
So
it's
really
awesome.
B
So
it's
also
also
funny
from
from
a
kind
of
a
mindset
perspective
because
I
talked
to
students
and
we
created
some
kind
of
lab-
and
I
said
after
the
after
the
whole
lab
yes
and
now
dropping
virtual
machines,
and
they
said
why
should
we
drop
it
because
you
don't
need
it
anymore?
So.
C
Yeah
and
just
that
that
the
the
the
way
that
you
can
now
just
use
and
then
delete
when
you're
done,
you
know
both
to
save
cost
and
and
and
that
way
of
thinking
in
a
more
you
know,
stateless
manner,
so
to
speak.
C
You
know
really
really
helps
you
know
getting
progress,
if
you,
if
you,
if
you,
if
you
can
just
repeat
it
and
then
change
whatever
actually
created
it,
it
gets
so
much
more
done
than
you
know,
watching
out
and
and
building
that
server
piece
by
piece
over
a
couple
of
months.
Silly,
it's
silly
that
we
have
to
do
that
before.
Yes,
it's
good
yeah.
In
retrospect,
it's
kind
of
that's
just
weird,
but
yeah
all
right,
cool.
C
Yeah
yeah,
but
automation,
you
know
that's
great
great
great
drug
to
be
on
automation,
cool!
Do
you're,
gonna,
open
the
the
the
dark,
the
dark.