►
From YouTube: CNCF TAG App Delivery 2021-05-26
Description
CNCF TAG App Delivery 2021-05-26
C
C
A
Yeah,
I
noticed
I
my
my
company's
changed
since
the
last
house
attendance.
C
Yeah,
let's
maybe
give
people
at
least
a
couple
more
minutes.
Otherwise
we
can
spend
more
time
today
really
discussing
the
enablement
working
group
today
some
might
watch
offline,
whereas
I
might
have
not
realized
that
we
actually
put
it
one
week
off
was
mostly
because
of
me,
because
I
was
not
there
last
week
and
yeah.
The
official
tlc
vote
on
new
chairs
and
tech
leads
is
still
open
which,
by
the
way,
was
my
fault.
C
I
mean
it's
kind
of
like
an
awkward
thing,
but
I
was
wondering
why
is
nobody
voting
and
then
I
talked
to
amy
and
he
said
well,
if
you're
not
right
wrote,
nobody
is
voting,
says
no
okay,
but
there
is
now
an
issue
and
this
it
should
be
moving
forward
pretty
quickly
and
if
there
was
some
confusion,
so
the
internal
vote
was
done
with
the
current
chairs
and
you
see
liaison-
and
this
is
just
the
official
tlc
approval-
I
think
about
this-
maybe
the
wrong
turn.
There
just
need
the
official
okay
from
from
uc.
C
Yeah,
I
think,
even
while
we
get
started
here,
let's
just
while
we're
waiting
for
probably
a
couple
more
people
to
join
here,
just
jump
into
the
app
enablement
working
group
proposal.
This
is
actually
perfectly
in
time
because
next
tuesday
there's
the
next
to
c
meeting
where
the
working
group
should
also
present
an
update.
I
think
this
is
perfectly
in
time.
Maybe
we
can
also
spend
some
time
on
the
operator
white
paper.
If
we
still
need
to
you
can
pick
either
one
to
go
first,
maybe
operator,
one
white
paper
would
be
really
quick.
B
E
E
B
E
B
C
Yes,
because
I
muted
myself,
while
I
was
typing
it,
you
know
one
week
off
of
meetings
and
suddenly
you
have
to
learn
back
all
of
those
skills
of
muting
and
unmuting
you
what
I
wanted
to
say
I'll
put
it
into
the
toc
update
next
week.
So
if
you
can
send
me
the
final
link
of
the
entire
document,
I'll
just
put
this
into
the
toc
update,
then
they
have
it
available
as
well.
D
Who
can't
should
we
write
out
something
on
the
23
million
list?
Or
would
you
do
this
or
whoever
should
do
this
because
I
saw
the
security
take
a
take
also
created
a
white
paper
and
they
announced
also
the
public
comment
phase
on
the
qsc
mailing
list.
C
I
will
do
it
after
tuesday,
maybe
so
wait
for
the
feedback
from
from
the
tocs
I
put
it
into.
I
put
it
into
the
the
tag
update
and
I'll
say
we
wanted
to
start
the
public
command
phase
and
see
whether
there's
any
feedback
on
how
this
should
be
handled
from
them
and
based
on
their
feedback.
We
can
then
do
it
next
week,
wednesday.
C
D
But
the
security
tag
created
a
kind
of
a
procedure
for
publishing
white
papers,
and
I
think
we
should
follow
this
as
far
as
possible
because
I
think
they
know
you
know,
based
of
us
out
how
to
write
white
papers.
D
A
C
C
C
D
D
C
It's
which
is
fine,
which
is
what
working
groups
kind
of
exist
for
so
they
have
a
dedicated
charter
to
produce
something
which,
in
this
case,
was
this
white
paper
and
it's
done,
and
now
they
can
more
or
less
move
on
to
do
other
things.
I
think
that's
totally
fine.
I
would
just
make
it
official
and
then
ensure
that,
like
the
calendar
entries.
D
C
Yeah,
I
think
it's
fine
yeah,
but
then
we
would
just
remove
it
from
the
calendars
that
there
are
no
official
meetings
anymore.
So
that's
fine,
it's
more
as
an
administration
type
of
item,
but
we
should
not
forget
about
it,
because
otherwise
it
would
accumulate
it
and
also
to
justify.
Why
are
we
creating
yet
another
working
group?
We
could
say
well
we're
resolving
one
moving
on
to
a
different
topic.
A
So
I
think
that
the
format
here
is,
we
can
walk
through
it
with
you
and
you
can
pick
out
any
questions
and
we
can
revisit
anything
that
needs
to
be
clarified.
A
I
think
the
walkthrough
of
this
document
will
take
five
to
ten
minutes
and
jennifer,
and
you
know
thomas,
please
feel
free
to
add
to
anything
that
I
say,
I'm
gonna
intend
to
give
essentially
an
overview,
but
we
can
dig
into
some
detail
and
express
some
additional
thoughts
whilst
we're
on
the
sections,
so
the
application
enablement
working
group
culminated
as
the
result
of
many
interested
parties
who
wanted
to
find
a
common
vernacular
for
describing
the
space
between
our
minimal
infrastructure
deployment
and
the
applications
on
which
reside
on
them,
and
this
isn't
new
news
to
anybody
within
this
group,
but
I
thought
I'd
clarify
it.
A
The
problem
statement
essentially
expresses
here
that
there
is
a
very
nebulous
and
ill-defined
way
of
expressing
how
to
deploy
an
application
across
providers
and
when
I
say
providers
that
could
mean
physical
infrastructure
providers.
That
can
mean
cloud
providers
and
this
ranges
from
taxonomy
of
word.
A
This
could
be
also
the
methodology
of
provisioning,
and
so
in
this
problem
statement,
we've
given
some
examples
of
some
simple
everyday
tasks
that
are
made
difficult
by
lack
of
ubiquity
across
these
systems,
so
developers
who
want
to
use
something
like
efs
as
a
backend
for
the
application,
and
also
use
that
same
efs
methodology
in
their
dev
environment,
a
challenge
with
having
two
separate
systems,
two
different
pipelines
and
ways
of
going
about
that.
There
is
not
a
simplistic
way
of
describing
this
or
there's,
not
a
simple.
A
I
would
say
best
practice
for
how
you
should
be
going
about
this
equally,
when
they
think
about
adapting
those
packages
to
use
the
storage
class
within
the
development
environment.
There's
a
differentiation
between
what
is
in
a
higher
tier
and,
as
you
can
see
in
these
points,
as
this
gets
shipped
to
the
next
stage,
there'll
be
no
storage
class
available
in
this
higher
environment
because
the
storage
cost
differs.
A
This
also
is
reminiscent
of
things
like
setting
a
storage
class
to
a
local
path
and
for
those
not
familiar,
I'm
using
the
example
within
kubernetes,
for
when
you
want
to
mount
a
data
volume
and
you
might
have
one
type
of
volume
locally
and
one
type
remotely,
and
this
this
is
a
typical
scenario
that
you'll
see
another
one
that
we
thought
was
quite
good,
was
ingress
and
actually
making
an
application
available
to
the
wider
world.
A
In
addition,
we
think
about
the
cicd
pipeline
and
how
there's
a
lot
of
institutional
knowledge
for
not
only
the
application
composition,
whether
that's
secret
management
injection
of
configuration,
then
also
deployment
onto
the
cloud
providers
or
or
self-hosted
providers,
underneath
that
and
as
we'll
go
on
to
describe.
There
are
some
attempts
from
various
vendors
and
from
various
working
groups
and
and
projects
to
solve
some
of
these
problems.
But
there
is
not
a
a
condensement
of
knowledge
and
examples
about
how
to
do
this.
A
In
addition,
there
is
an
assumption
that
infrastructure
is
always
available
and
we
believe
that
there
needs
to
be
a
way
that
we
can
describe
an
immutable
state
of
this
is
what
I
need,
and
then
we
can
start
to
look
at.
How
can
we
resolve
these
provisioning
issues
and
when
I
say
we
need
to
to
start
thinking
about
resolving
it
as
we're
going
to
describe
at
the
end
of
this
document,
we're
not
looking
to
build
a
panacea
here.
This
working
group
is
very
much
focused
on
what
are
the
real
problems.
A
There
are
very
few
ways
that
systems
agree
to
to
come
together
to
create
a
essentially
application
first
or
developer.
First
experience.
We
are
seeing
some
efforts
to
be
able
to
take
a
container
or
to
take
a
package
to
do
something
like
auto
devops
and
deploy
it
out,
but
there
are
many
different
variations
on
this
and
mileage
may
vary,
and
so,
lastly,
in
this
stanza
here
we
talk
about.
A
There
are
no
really
ubiquitous
best
practices
and
I've
said
that
a
few
times
in
various
ways
throughout
my
my
illustration
here,
but
part
of
the
goal
as
we're
going
to
talk
about,
is
to
really
look
at
what
are
the
emergent
practices?
What
are
vendors
building
towards?
What
are
the
things
that
projects
like
you
know,
argo
flux
kept
in
cross
plane
terraform?
What
are
these
things
doing?
How
are
people
using
them
in
the
financial
industry?
A
How
are
people
using
the
technology
industry,
and
so
this
really
segues
into
this
idea
of
what
are
the
combinations
of
tooling
available
out
there,
and
how
do
they
aim
to
resolve
this,
and
so
this
very
simple
then,
which
is
you,
know,
wonderfully
naive
at
the
same
time?
It's
very
powerful
because
it
does
illustrate
this
gap.
You
know
whether
it's
pollumi
with
flux,
whether
it's
argo
or
terraform,
being
able
to
deploy
a
workload
is
not
the
dominion
of
one
of
these
alone.
You
cannot
run
argo
cd
without
having
the
infrastructure
on
which
to
deploy
it.
A
Equally,
there
is
no
point
having
cross
plane
if
you
have
no
application
workloads
to
run
on
it,
and
so
there
are
some
intersecting
concerns
here,
and
we
think
this
is
exciting,
because
it's
an
opportunity
for
many
of
the
participants
within
the
the
projects
mentioned
here,
to
bring
to
light
how
they
can
see
compatibility
between
these
projects
and
again,
I
think
about
things
like
oam
in
the
back
of
my
mind
as
some
of
the
prior
art
today,
in
terms
of
how
can
we
start
to
bring
an
interesting
and
more
holistic
ecosystem
talking
about
prior
arts
and
things
that
have
already
been
set
up?
A
If
we
look
at
the
examples
of
known
solutions-
and
I
thank
robert
as
well-
I
don't
believe
he's
on
this
call,
but
he
helped
us
to
fill
in
some
of
these.
We
we
had
things
like
ketch
and
and
cuba
and
shipper
into
these
application
deployment
methodologies
on
top
of
these
very,
very
well
trodden
piles
of
terraform
and
the
now
very
exciting
cross-plane
proposition.
A
It's
it's
extremely
interesting
in
how
the
end
user
is
is
actually
digesting
this.
The
technology
stacks
are
starting
to
outpace
people
quite
quickly,
and
so
what
we're
seeing
and
what
we're
hoping
to
understand
is
how
are
people
adapting
to
this?
We
have
infinite
choice
now,
but
being
able
to
describe
a
workload
and
deploy
it
and
actually
having
enough
infrastructure
to
do
just
the
right
thing
is
still
an
interesting
and
quite
complex
task.
A
So
with
everything
I
just
said
there,
how
does
this
actually
align
with
the
the
goal
of
of
the
tag?
Well,
in
many
sense,
the
the
definition
and
the
actual
explanation
here,
I
think
is-
is
quite
well
put
so
I'll.
Just
I'll
just
read
this
line
so
as
application
delivery,
often
coupled
with
underlying
infrastructure,
it
can
often
impact
package
formats
and
the
application
delivery
workflow,
and
this
we've
described
prior
to
this.
A
This
topic
around
this
idea
of
definition,
description,
parameter
configuration
all
falls
into
the
room
of
app
delivery
because
it's
part
of
the
idea
of
bundling
and
deployment
and
the
sort
of
delivery,
workflow
and
strategy
within
the
six
charter
topics.
So
there's
a
very
close
alignment
and
that's
that's
why
we
feel
that
natural
synergy
would
be
something
that
we
could
explore
within
this
group
and
really
help
to
leverage
some
of
the
interest
areas
of
the
wider
tag
and
to
really
shine
a
light
and
exemplify
some
of
the
industry
best
practices.
A
And
so
with
that
said,
you
know
what
are
the
actual
tangible
outcomes
of
this
working
group
capturing
the
current
landscape
practices.
So
that's
talking
to
end
users,
focus
groups,
conversations
interviews,
focus
on
balancing
the
involvement
of
personas
in
the
application
management
enablement
arena.
So
this
specifically,
is
a
very
interesting
point,
because
we
want
to
understand
that
it's
not
just
sres
who
are
building
infrastructure.
It's
not
just
developers
who
are
deploying
workloads.
A
There
are
some
hybrid
personas
of
folks
who
are
not
only
building
but
then
deploying
the
scale
of
your
company
will
have
a
magnification
effect
on
whether
or
not
you're
going
to
be
doing
one
or
the
other.
Often
it
might
be
a
detractor
and
you'll
just
be
doing
deployment
and
vice
versa,
either
way
that
has
a
significant
influence
on
the
future
of
the
of
the
ecosystem.
A
The
third
point
provide
providing
interoperability
examples
between
iac,
so
infrastructure
is
code
and
continuous
delivery
tools.
In
a
code
repository
this,
we
think
about
using
projects
such
as
potato
head
to
really
exemplify
how
are
people
using
the
the
tooling
within
the
landscape?
How
do
you
knit
these
things
together?
How
do
I
take
a
chart
or
a
manifest,
and
how
do
I
build
a
githubs
platform
that
can
consume
that,
on
top
of
some
infrastructure,
to
enable
me
to
present
my
business
and
finally
provide
successful
patterns
in
a
white
paper
based
on
practical
work?
A
Not
a
king
making
process
of
saying
this
is
how
it's
done,
but
more
a
white
paper
in
a
similar
vein
to
the
operator
group,
where
we
analyze
recurrent
patterns
of
behavior
and
best
practices,
and
so,
lastly,
more
of
an
admin
part
of
this
document,
the
working
mode
and
expected
outcomes
we'll
meet
to
discuss
plans.
Concepts
will
develop
demo
infrastructure,
as
mentioned,
and
create
essentially
an
mvp
for
various
patterns
that
will
invite
project
maintainers
and
contributors
to
to
get
involved
with,
and
this
will
be
inevitably
and
ultimately
for
the
end
user
benefit.
A
So
they
can
refer
to
this
project
saying
this
is
how
we
we
can
see
such
and
such
a
git
ops
process
on
top
of
this
cloud
provider
or
this
platform
working
and
as
mentioned
prior
to
this,
we
have
end
user
interviews,
landscape
radar,
contribution
of
ioc
and
ci
cd,
tooling
workflows
and
white
paper.
All
as
part
of
our
ambition,
and
so
that
really
covers
the
high
level
outline
of
this
document-
and
I
think
at
this
point
I'll
just
invite
thomas
or
jennifer.
If
you
want
to
add
anything
to
what
I've
already
said.
D
So
I
think
you,
you
said
everything
and
yes
for
me:
it's
a
really
really
nice
possibility
to
to
bring
infrastructure
and
application
developers
more
together
to
find
out
how
how
ic
and
app
delivery
tools
can
interact
and
so
on.
And
yes,
I
think
this.
I
think
this
would
be
a
funny
working
move.
A
Yeah,
so
I
think
with
that
said,
unless
you
think
any
other
points
we
could,
we
could
open
it
up
if
there
are
any
questions.
B
Yeah,
I
have
a
question.
One
thing
that
we
talked
about
this
is
a
while
ago
about
application.
Enablement
was
that
the
result
wouldn't
be
a
white
paper,
but
now
we're
saying
that
the
goal
is
a
white
paper,
so
yeah.
How
did
that?
Like
decision
happen?.
A
A
So
so
the
question
I
think
matthew
was
asking.
There
is
prior
to
prior
conversations.
There
hadn't
been
the
indication
of
there
being
a
white
paper
being
written.
However,
now
in
the
initial
goals
we
have
stated
that
a
white
paper
should
be
one
of
the
products
and
he
was
just
asking
a
little
bit
around
that
decision
process
and
why
that's
there.
D
One
of
the
first
goals
of
the
or
the
first
abuse
of
this
one
of
this
working
group
was
that
we
don't
want
to
create
the
white
paper
at
first,
so
at
first
we
we
want
to
give
end
users
something
they
can
work
with
so
examples
of
configurations
of
how
to
how
to
implement
such
such
such
infrastructures
and
activity
with
things
and
afterwards
after
we
have,
we
have
felt
the
pain
ourselves
and
after
we
have
found
out
what
could
be
done
better
or
could
be
a
good
pattern
for
implementing
this.
D
Then
I
think
we
are
able
to
write
the
white
paper
and
could
provide
also
our
our
experience
to
the
to
the
end
users
and
yes,
make
their
lives
easier.
So
this
was
if
this,
if
this.
A
Just
to
add
to
that
very
briefly,
one
of
the
one
of
the
really
exciting
prospects
of
the
white
paper
for
tag
observability
was
that
we
saw
a
lot
of
vendor
contribution
as
well
as
end
user
contribution
and
it
actually
helped
to
fuel
in
a
sort
of
cyclical
sense,
the
interview
process
and
the
participation
within
the
working
group-
and
I,
I
think,
just
to
add
what
thomas
was
saying
about
this
kind
of
being.
A
Almost
like
a
last
step
before
we
wind
the
group
down,
it's
not
just
about
our
experiences
as
individuals,
but
hopefully
looking
at
getting
some
of
these
projects
and
maintainers
of
these
products
to
work
together.
To
say
you
know:
okay,
cluster,
api
with
crossplane,
etc.
You
know,
and
let's
use,
that
within
git
option
and
looking
at
how
we
start
to
see
that
become
something
we
can
actually
codify
into
a
practical
white
paper
that
exemplifies
some
of
these
techniques.
A
B
Great
yeah,
so
that
makes
sense
I
was
just
wondering,
like
you
know,
it
seems
like
a
project
like
this.
It
would
be
good
to
kind
of
formulate
some
sort
of
standard
or
or
something
like
that,
for
how
you
know
app
enablement
could
be
done
at
similar.
Like
do
you
think
you
sees
like
in
a
similar
way
to
smi,
tries
to
standardize
service
measures.
Do
you
think
this
could
be
something
you
wouldn't
like
investigate.
A
I
think
if
I
can
speak
for
everyone,
who's
been
involved
in
that
I
think
that's,
probably
one
of
our
non-goals,
actually
we're
not
looking
to
create
a
14th
standard
kind
of
thing.
You
know,
I
think
it's
going
to
be
an
area,
and
maybe
we
will
think
about
that
conversation
again,
but
initially
as
an
interesting
set
of
goals.
It's
definitely
not
something.
We've
stayed
away
from
for
a
variety
of
reasons.
I
don't
know
anyone
wants
to
add
to
that.
C
Yeah,
I
I
would
even
say
right
now:
we
wouldn't
even
in
detail,
know
what
the
standard
should
be
for
exactly
like.
If
you
would
start
with
the
standard
work,
you
would
not
necessarily
be
entirely
sure
what
you
want
to
standardize
and
and
also
to
your
point
like
the
way
I
see
it,
you
move
into
a
more
more
opinionated
way
if
you
look
at
higher
frameworks,
but
it's
cubellar,
it's
it's
it's
catch
and
others,
and
I'm
not
using
opinion
data
in
a
negative
way.
It's
just
a
specific
way
of
doing
things.
That's
that's!
C
I
think
that
the
the
question
yourself
to
find
okay,
where,
where
do
the
individual
approaches
should
where
should
they
start
and
where
should
some
complicated
ldp?
Maybe
it's
just
at
the
core
commodities
primitive.
So
we
say
every
tool
at
the
end
of
the
day
needs
to
output.
C
What
is
available
in
kubernetes
by
default,
which
it
technically
needs.
Otherwise,
it
can
run
it
with
some
exceptions.
Obviously,
because
if
you
look
at
crossplane,
obviously
it
uses
a
crd,
but
it
configures
a
system.
That's
in
highly
outside
of
kubernetes.
C
So
I
like
the
approach
I
like
capturing.
The
current
state,
I
think,
is
the
the
current
problems
to
be
solved
at
the
current
state
of
where
pro
projects
are
helpful
is
something
I
like.
I
know
we
started
working
or
we
had
the
intent
to
work
on
a
landscape
before
I
always
see
landscape.
As
a
very
interesting
thing.
C
Landscapes
are
a
way
to
get
a
lot
of
people
out
of
this
meeting.
So
if
you
would
work
on
the
application
delivery
landscape,
we
would
be
maybe
50
people
in
this
meeting
for
two
or
three
weeks,
because
every
a
lot
of
deaf
real
people
and
a
lot
of
marketing
people
will
suddenly
be
tasked
to
participate
and
ensure
that
their
logo
is
there.
C
And
if
I
look
at
pretty
much
every
landscape
out
there,
it
isn't
usually
helping
me
a
lot
like
there
are
like
five
tools
in
there
like
five
logos
in
there
say
four
cd.
But
what
does
this
mean
they're
like
a
couple
of
logos
in
their
form,
let's
say,
templating
and
then
and
and
and
and
other
steps
in
there?
That's
why
I
think,
having
what
tools
and
what
tools
can
do
overview
that
I
would
not
necessarily
call
landscape.
C
C
That's
why
I
think
we
should
combine,
maybe
goals
with
we
should
mix
goals
with
deliverables
as
okay.
What
do
we
want
to
achieve
for
us
internally,
like
for
moving
the
application
delivery
space
forcing
capital
forward?
What's
the
output
for
an
end
user
that
might
make
sense,
it
is
usually
not
always
done
that
way,
but
in
this
case
it
might
be
helpful
and
maybe
they're
the
same
anyways.
C
C
C
C
What
can
I
use
and
how
will
this
product
or
did
this
project?
Do
it
also
something
that
I
found
interesting
from
a
was
what
came
out
in
a
couple
of
other
cncf
tech
creators
that
people
usually
don't
look
like
for
that
one
tool
to
do
everything
it's
more
like
a
best
of
breed
and
mix
and
match
type
of
approach
like
what's
the
input
and
output
of
a
tool?
C
A
Absolutely
and
it's
interesting
as
well
like,
I
think
that
your
point
around
this
may
lead
to
a
standard.
I
think
that's
absolutely
true.
You
know
the
idea
of
mix
and
match
is
so
poignant
right
now,
because
that
certainly
is
how
people
are
selecting,
and
I
come
back
to
the
idea
of
observability.
A
You
know
and
the
idea
of
the
hotel
system
came
out
of
things
like
open,
metrics
and
open
census.
You
know
the
convergence
of
multiple
standards
into
something
that
was
a
little
bit
more
ubiquitous
and
a
little
bit
more
democratized.
It
could
be
the
case
that
we
find
that
there
are
some
quite
opinionated
projects
like
crossplane
that
have
xdr.
I
think
it's
called
xrd
I
forget,
but
you
know,
and
then
we
can
see
that
there
are
parallels
in
other
projects,
and
it
might
be
that
something
comes
out
of
that.
A
So
I
think
this
is
going
to
shine
a
light
on
something
that
is
fairly
prevalent.
Lots
of
people
are
doing
it,
lots
of
people
trying
to
create
stuff.
We
saw
lagoon
again
a
few
weeks
back
as
a
presentation,
and
I
think
that
we
will
start
to
start
to
shake
the
tree
a
bit
doing
this.
I
think
it'll
be
a
very,
very
useful
exercise.
C
They
go
two
paths
I
mean
not
not
to
bash
in
any
technologies,
but
they
were
like
a
lot
of
folks
who
were
focusing
massively
on
cloud
foundry
when
kubernetes
was
emerging,
so
that
might
not
feel
that
good
today,
anyways,
I'm
not
saying
it's
a
bad
technology.
I
think
it
had
some
a
lot
of
the
things
we
are
building
right
now
in
the
kubernetes
space
reminds
me
of
a
lot
of
things
that
were
already
there
in
cloud
foundry
and
some
projects
made
it
over
like
like
cloud
native,
build
packs.
E
So
you
are
saying
like
we,
I
mean
what
you
said
before
when
alex
was
putting
the
comments
that,
like
adding
to
our
like
output,
like
capabilities
and
like
requirements
and
pros
and
cons
of
the
of
each
and
also
use
cases
like
I
guess
the
white
paper
could
contain
some
or
if
it's
some
other
thing
that
we
decided
to
do
like
use
cases
because
of
the
interview.
E
So
we
will
have
some
of
these
like
data
and
then
be
able
to
then
out
surface
like
the
the
capabilities
and
what
are
what
workarounds
people
are
are
taking
now
or
doing
now.
So
yeah
I
if
that
was
what
you
meant
with
the
capabilities.
E
When
you
said
we
could
use
the
interviews
to
create
those
use
cases
and
yeah.
C
I
think,
just
by
exactly
just
when
people
tell
you
what
they're
using
a
certain
product
for
we
use
it
only
for
deployment.
No,
we
don't
use
it
for
modification.
C
You
can
use
certain
products
for
a
lot
of
things
and
only
for
a
small
subset,
and
I
think
that,
because
just
describing
the
space
or
just
describing
cloud
negative
application
delivering
everything
that
needs
to
happen
is
complex
by
itself
and
there
is
no
no
model
at
the
very
beginning
of
the
working
group
harry
built
up
this.
This
model
that
you're
referring
to
as
well
regarding,
like
definition,
packaging,
delivery
and
and
so
forth.
C
We
didn't
cover
infrastructure
definition
in
there
that
much.
C
A
It's
interesting
as
well
like,
I
think.
One
of
the
things
that
keeps
going
through
my
mind
right
now
is
that
this
is
essentially
driving
home.
A
a
barometer
of
how
interoperable
is
the
cloud
native
ecosystem
right
like
how
many
of
these
projects
have,
inter
interoperability
with
each
other
and
we're
just
sort
of
polarizing
them
between
oops
between
application,
delivery
and
infrastructure
concerns.
But
really
it's
a
far
more
introspective
question.
A
You
know
how
are
people
using
these
tools
is
essentially
going
to
answer
the
question
of
which
you
know
what
sort
of
direction
is
the
future
heading
us
heading
in,
and
you
know
a
lot
of
prevalence
around
this
idea
of
going
to
small
clusters
that
are
self-managed
with
git
ops
based
approach
is
really
giving
us
a
bit
of
an
emergent
trend,
point
of
view
through
this
working
group.
So
again
that's
one
of
the
outcomes
that
I
think
is
going
to
be
probably
more
anecdotal,
but
I
think
it's
super
vital,
like
you
said,
for
strategic
decisions.
C
That's
also
a
good
point.
Usually
it's,
however,
hard
to
get
the
end
user
interviews.
I'm
not
saying
it's
not
a
good
idea.
I
think
we
should
just
clearly
come
up
with
a
plan
on
how
to
reach
the
end
users,
because
usually
the
cncf
well,
the
hdf
has
a
good
relationship
with
the
nd
with
its
end
users,
but
it
is
often
very
hard
to
get
them
to
engage
even
more
because
they
like
ask
for
service
old
all
the
time.
C
C
C
We
might
find
a
bit
more
openness
here,
and
I
think
this
this
it
also
jumps
into
my
goes
into
my
next
comment.
I
would
also
actively
invite
the
projects
that
are
part
of
this
to
that
working
group
and
have
them
discuss
it.
Some
of
them
are
obviously
already
in
there.
Some
of
them
are
not.
A
E
Regarding
the
strategy
to
reach
out
to
people,
I
guess
so
far
we
are
trying
to
talk
to
the
people
that
we
kind
of
like
have
more
connections
with,
like
I,
for
example,
used
to
be
part
of
the
end
user
group.
When
I
used
to
work
for
condonast,
they
are
an
end
user.
E
So
I
have
ex-colleagues
that
are
willing
to
talk
to
me
and
we
can
do
interviews
so
and
there's
vendor
now
in
vmware,
also
my
team,
because
we
use
as
well
so
we
so
that's
that's
how
like
I
will
get
started
from
this
side
and
maybe
other
companies
that
are
not
yet
end
users,
but
I
know
that
are
intending
to
do
that,
so,
like
basically
reaching
out
to
people
that
we
know
that
working
companies
and
that
are
willing
to
share
some
things.
A
Yeah,
in
addition,
like
there's
the
the
get
ops
working
group,
isn't
there
that
has
a
lot
of
interested
parties
that
would
probably
be
quite
quite
happy
to
set
apart
what
their
particular
offering
is
and
who
potentially
some
of
their
their
customers
are,
with
their
permission,
of
course,
to
give
kind
of
almost
like
case
study
examples
of
of
how
they
are
using
their
their
tooling
and
product,
so
that
might
be
worthless,
exploring
as
well.
A
I
think
tracy,
who,
who
tracy
reagan,
who
you
know
she's
from
cdf,
was
on
one
of
these
calls
a
few
weeks
back
and
she
might
be
a
really
good
person
to
contact
she's,
a
good
relationship.
I
believe,
with
with
a
lot
of
these
folks.
C
C
I
mean
this
is
not
to
steal
your
thunder,
but
that's
what
you
usually
see
them
if
you
have
projects
jumping
in
there,
because
you,
as
you
have
already
seen
with
the
operator
working
group,
you
have
always
a
massive
drop-off,
that's
kind
of
what
I
learned
here
and
what
you
will
also
see
some
sometimes
in
the
working
groups.
C
E
Thank
you
so
yeah.
I
I
agree
with
that
and
yeah.
We
experienced
that
in
the
working
operator
working
group
for
sure,
like
we
know
the
pain.
What
I
I
suggest
with
that
like
I
intend
to
do
more
in
this
this
time.
Around
learning
from
that
is
to
we
should
have
what
we
need
like
know,
know
what
we
need
before
reaching
out.
E
You
know
the
information,
so
the
user
interviews
like
we
we
need,
we
need
to
kind
of
just
get
it
like
set
first,
so
we
can
when
reaching
out
and
people
are
willing
to
give
their
attention.
You
don't
like
need
them
to
be
like
consistently
participating
for
a
long
time
and
then,
like
maybe
lose
the
momentum
you
know
like.
So
we
should.
We
should
try
to
know
exactly
what
information
we
need.
E
E
C
Yes,
so
what
I
learned
over
like
working
on
standards
and
these
kind
of
things
for
a
very
long
time,
people
usually
contribute
if
they
have
a
very
personal
interest
in
the
things
that
they
are
doing,
not
saying
that
people
don't
do
things
altruistically,
but
the
closer
is
something
is
aligned
to
what
you're
working
on
anyways
they're
more
likely
to
account.
You
are
contributing
to
it
because
you're
you
always
doing
the
work
that
you
want
to
do.
Anyways.
C
C
They
will
join
this
discussion
because
the
problem
that
they
have
they
want
to
listen
to
other
people
who
already
worked
on
this.
If
they
built
something
interesting,
they
might
be
willing
to
present
what
they,
interestingly
built,
and
it's
way
more
practical.
C
That's
also,
why,
like
you,
put
the
white
paper
to
the
end,
like
also
with
the
app
delivery,
so
it's
the
operator
working
group,
a
white
paper
is
always
very
hard
because
it,
it
really
doesn't
add
the
people
that
doesn't
provide
a
lot
in
equivalent
and
value
to
the
people
who
write
it
then
don't
necessarily
get
out
of
it.
It's
a
lot
of
work
to
get
there,
but
once
you
have
the
high
level
idea,
you
got
everything
out
of
it.
There's
still
a
lot
of
work
left.
C
I
think,
if
you
can
think,
especially
in
the
beginning
of
tasks
that
are
low
to
medium
input
required
from
people
but
have
high
value
for
them,
or
actually,
I
think
the
best
we
can
pick-
that's
how
you
usually
keep
people
engaged
like
if
you
would
discuss
like
three
ways
on
the
comparison
of
the
three
ways
and
how
you
can
manage
infrastructure
in
a
kubernetes
environment,
you
might
invite
the
project
people
over.
C
That
would
be
a
by
the
way,
great
video
that
we
as
a
as
that,
as
the
tag
would
have
that
we
can
share
online.
That
would
draw
additional
interest
to
what
we're
doing
easy
to
produce
high
value
and
for
the
people
who
are
initially
contributing
very
low
effort,
so
that
that
that
might
be
something
to
think
about,
because
that's
easy
like.
If,
honestly,
if
you
would,
I
ask,
for
example,
to
cross-plane
people.
C
How
are
you
doing
it?
Then
you
talk
to
a
couple
of
other
people
who
do
infrastructure
based
on
on
kubernetes.
They
will
be
very
willing
to
share
what
they
do.
Some
high
level
concepts
you
put,
we
put
them
all
in
one
meeting.
You
have
five
presentations.
You
put
this
out
to
the
wider
audience
point
into
other
resources
that
we
have.
I
think
you
create
some
momentum
that
will
eventually
also
lead
to
to
some
critical
mass,
and
you
pick
like
one
topic
after
the
other,
it
still
needs
to
be
curated.
C
Obviously
somebody
needs
to
bring
them
on
board
and
motivate
them,
but
these
are
usually
the
easier
thing,
the
plus
side
for
you.
If
you
eventually
want
to
write
an
overview
of
the
of
the
of
the
landscape
and
eventually
white
paper,
that
a
lot
of
the
research
work
for
you
that
you
otherwise
needed
to
do
is
done
by
other
people,
so
there
might
be
like
a
win-win
situation
in
there.
C
I
know
the
observability
group
is
looking
into
doing
blog
posts,
some
type
of
video
series
and
these
these
kind
of
things,
and
that
might
be
something
like
listing
the
top
10
problems
that
are
on
top
of
your
mind,
that
we
want
to
discuss
and
how
people
manage
it.
I
mean
you
have
obviously
also
the
experience
what
people
usually
run
into
just
just
start
there
we
can
always,
as
we
go
along
change
change,
which
ones
we
want
to
work
on.
C
If
one's
becoming
more
important,
we
can
browse
it
up
before
it
becomes
less
important
to
prioritize
it
down.
A
Yeah,
I
think,
that's
great
advice,
especially
the
low
effort,
high
value
contributions,
part
I
I
think
as
much
as
we
can
to
incentivize
projects
to
participate.
It's
going
to
be
really
key,
and
actually
this
is
why
I
think,
perhaps
to
matthew's
question.
A
This
is
why
maybe
the
the
white
paper
is
actually
valuable
as
well
is
that
they
can,
from
a
marketing
perspective,
see
that
they've
been
you
know,
given
the
opportunity
to
look
at
their
interoperability
with
other
other
technologies
and
share
the
stories
of
how
their
users
are
sort
of
operating,
and
you
know
potentially
be
part
of
the
the
publicized
group
that
are
looking
at
working
towards
not
only
exemplifying
this,
but
also
sort
of
solving,
eventually,
for
perhaps
some
of
the
more
practical
challenges.
C
So
they
have
a
reason
why
they
exist
and
then
it's
more
or
less
about
curating
them
and
what
they
want
to.
So
we
obviously
have
to
go
in
a
couple
of
rounds
like
the
problem
with
application.
Delivery
on
kubernetes
is
too
complex.
It's
not
the
granularity
that
you
want
to
have
there.
I
mean
that's
still
a
bit
superficial,
but
as
you
go
deeper
and
be
more
specific,
I
think
you
can
really
get
there.
I
know
this
is
actually
making
more
work
than
it
takes
out
of
it.
C
C
C
C
A
This
is
fantastic,
so
I
think
we
will
need
to
meet
again
this
week
and
we
might
make
some
small
adjustments,
perhaps
to
you
know,
based
on
your
comments
here.
I
think
me
personally
the
thing
that
you
know
jennifer
and
yourself
said
that
actually
really
resonated
with
me
something
I
hadn't
thought
about
was:
how
do
we
actually
drive
active
participation
and
high
engagement,
and
you
know
make
use
of
that
while
we
have
it
early
on,
so
that's
something
that
I
think
we
need
to
think
about.
A
As
for
the
document,
if
you
think
in
general
it
looks
pretty
good
I'll
make
some
small
tweaks
to
it,
and
we
can.
We
can
then
get
it
ready
for
the
sort
of
presentation.
C
C
So
usually
we
can
publish
materials
this
we
just
need
to
let
the
cncf
know
not
that
it's
a
big
deal
per
se,
but
some
of
the
marketing
obviously
is
always
reserved
to
paying
members.
While
all
the
projects
we
engage
here
are
obviously
cncf
projects
per
se
mostly,
but
would
you
say,
notice
more
project?
That's
not
a
cncf
project.
That's
something
interesting
to
show
most
likely.
Not.
C
So,
if
somebody's
really
interested-
and
we
find
people
who
want
to
write
about
it
like
after
an
interview
which
would
be
positive,
like
which
stack,
are
you
running?
What
are
you
using
tools
for?
I
think
a
lot
of
people
want
to
learn.
They
want
to
hear
what
other
people
are
using,
what
works
for
them?
What
doesn't
work
for
them?
C
C
A
C
I'm
not
worried
about
the
name,
I'm
always
more
worried
about
about
the
work.
That's
done
somewhere
honestly
and
having
like
a
solid
execution
plan
for
something
having
the
best
name
in
the
world
and
not
knowing
what
to
do
is
usually
worse
than
the
than
the
other
way
around.
C
C
C
Yeah,
so
you
you
read
it,
you
kind
of
understand
it,
but
if
somebody
forced
you
to
put
it
in
like
one
single
sentence,
what
the
application
and
the
overwhelming
working
group
is
doing,
it
is
still
incredibly
hard.
I
think
we
all
are
here.
We
kind
of
have
this
agreement,
what
it
is
about,
but
like
that
usual
back
in
the
days,
and
hopefully
today's
soon
to
come,
when
you
sit
next
into
an
airplane
and
they
would
ask
you
what
it
is
or
we're
having
a
not
so
amazing
airplane
dinner.
A
Yeah,
that's
a
good
point.
I
mean
we
we
did.
We
did
like
the
idea
of
pattern
kept
coming
up.
Recurrently
pattern
experience.
You
know
these
kind
of
tax,
this
kind
of
taxonomy
of
words,
I
think
we're
getting
close
to
it.
C
And
and
again,
this
won't-
and
this
won't
really
restrict
us
from
moving
forward
here,
and
I
think
we
can
also
have
these
discussions
right
now
as
part
of
the
main
meeting
here
here
as
well
more,
there
are
no
project
reviews
right
now
schedule,
so
we
can
use
this
meeting
slot
and
it
might
also
draw
more
people
here-
hopefully,
hopefully,
there's
more
end
users
yeah.
C
To
some
of
those
presentations
where
people
show
how
they
do
certain
things,
I
think
a
side
by
side
of
cheaper
and
or
cheaper
and
the
cubewella
would
be
nice,
like
in
one
session
that
we
give
you
both
the
same
problem
show
how
you
do
it.
It's
not
competing.
A
Yeah,
I
mean,
I
think,
that's
quite
a
compelling
way
to
get
people
to
to
participate
actually
having
these
comparison
meetings,
where
we
can
really
dig
into
how
the
same
problem
is
solved
in
different
ways.
C
I
was
just
wondering
why
I
could
actually
use
cluster
api,
for
example
with
a
commercial
distribution
like
openshift,
or
I
could
even
use
cluster
api
to
configure
my
zero
cluster.
I
mean
cluster
api
is
great,
but
can
I
use
it
for
anything
except
for
little
kubernetes.
A
A
Okay,
perhaps
it
was
an
error,
but
yeah
essentially
like
just
like
you
were
saying
it's
all
the
glue
between
as
well,
which
is
the
concern
of
the
the
working
group.
Is
you
know
what
is
the
toil
involved
in
actually
describing
how
you
provision
on
that
particular
infrastructure
for
this
workload,
and
so
I
think
that's
something
that
we're
going
to
learn
as
we
go.
C
So
that's
why
I
think,
as
we
described,
maybe
you
have
a
couple
of
personas
in
there
but
say:
let's
describe
our
persona
as
somebody
who's,
maybe
either
a
developer
or
or
a
platform
operator
who
understands
how
kubernetes
works
and
how
you
can
maybe
configure
plane
kubernetes,
that's
their
skillset
like
they
have
a
cka
ckd,
that's
what
they
have.
That's
what
they
know.
E
C
C
C
A
And
I
feel
like
we're
almost
getting
into
the
arena
of
understanding
a
metric
for
this
or
understanding
some
sort
of
recurrent
way
of
measuring
this,
because
you
know
learning
those
technologies
is
one
thing
doing.
Those
on
a
vmware
based
solution
is
another
and
then
doing
it
on
a
cloud
provider
is
yet
another
that
kind
of
institutional
knowledge,
development
and
toil
for
the
particular
snapshot
or
your
vertical
is
something
that's
very,
very
difficult
to
express
it's
very
difficult
to
measure.
A
But
yet
it's
something
we
all
feel
right
like
you,
there
are
some
technology
stacks
that
just
click
and
they
just
click
together
in
a
way
that
works,
then
interoperability,
I
keep
coming
back
to
is
something
that
I
feel
we
can
start
to
think
about
quantifying,
maybe
not
with
a
with
a
scalar,
but
at
least
we
could
do
it
in
a
kind
of
a
you
know,
a
heuristic
where
we
can
say
you
know
how
was
your
experience.
C
Yeah-
and
I
think
that
makes
sense,
because
that's
what
I
sometimes
find
interesting
when
people
say
well,
you
can
use
an
ingress
to
easily
expose
a
service
in
your
kubernetes
cluster
and
the
answer
is
yes,
that's
theoretically
true,
but
if
you're
running
in
aws
in
gcp
or
in
azure,
your
experience
is
entirely
different.
To
really
see
the
thing
on
the
internet.
C
E
What
do
you
think
about,
for
example,
so
when
I
got
to
the
level
of
ck
sticker
cka,
I
still
didn't
know
much
about
like
crds
at
all
like
it
was
before
and
then
like
in
my
previous
job,
for
example,
I
used
aws
terraform,
so
I
I
had
to
create
some
resources
and
stuff
like
that,
like
prometheus
and
stuff,
but,
like
I
didn't,
have
to
think
about
custom
resource
definitions
and
stuff
and
some
stuff
like
crossplane
now
like
people
who
have
this
knowledge
more
is
like
easier
to
use.
C
C
But
you
brought
up
an
interesting
tool
point
which
tools
are
people
usually
familiar
with
when
they
come
to
these
problems.
Those
are
two
what
alex
said
and
then
in
this
diagram
they
come
with
terraform
knowledge.