►
From YouTube: WG Platforms - Project Meeting
Description
TAG web site: https://tag-app-delivery.cncf.io/
TAG Slack channel: https://cloud-native.slack.com/archives/CL3SL0CP5
TAG git repo: https://github.com/cncf/tag-app-delivery
TAG meeting notes: https://docs.google.com/document/d/1OykvqvhSG4AxEdmDMXilrupsX2n1qCSJUWwTc3I7AOs/edit
A
You
can
put
your
name
at
the
Top
If
you
want
of
the
notes
you're
attending
we
like
to
start
out
to.
If
anyone
is
new
I've
got
a
couple
new
faces,
I
want
to
introduce
yourself
and
kind
of
what
your
interests
are
just
so
we
can
kind
of
know
what
to
look
for
you
for
so,
if
anyone
wants
to
introduce
themselves
hey.
B
C
D
D
Yeah
I
was
on
here
because
after
being
remote
for
30,
like
10
years,
I
still
don't
know
one
too
many
of
myself.
So
hello,
everybody.
My
name
is
Fred
Frederick
Harper
I'm
from
Montreal
Canada
I
work
with
John
and
Jared
that
you
may
know
at
you.
First
I'm,
a
developer,
Advocate
I've
been
doing
that
for
a
long
time
and
you
know
looking
to
be
involved
in
the
way
or
another,
so
probably
something
around
developer
relation
skills.
It's
places
that
I
can
help.
D
So
anything
people
related
content,
organizing
stuff
connecting
with
people.
This
is
where
I
think
I
can
give
a
hand,
so
just
happy
to
be
here,
I
kind
of
like
here,
to
look
and
see
how
it
goes
and
see
where
I
could
help
and
head
a
possible
way.
So
have
you
have
any
idea
anything
specific,
please!
Let
me
know
I'm
in
the
cncf
and
fabinity
slack
and
spend
too
much
time
on
Twitter
so
anyway,
happy
to
be
here.
E
She's
got
a
a
maturity
model
that
she's
been
working
through
and
so
we're
just
kind
of
trying
to
see
where
we
might
be
able
to
help
and
contribute
in
any
way
and
I.
Also,
you
know
just
keep
tabs
on
what's
going
on
in
the
platform
engineering
ecosystem,
so
we'll
probably
be
regular
attendees,
not
sure
how
much
value
we
had,
but
we're
happy
to
to
chime
in
whenever
it
seems
like.
We
have
some
valuable
to
say
so,
I'm
glad
to
be
here.
Thank
you
guys.
Thank.
F
Oh
just
to
quickly
introduce
myself
as
well:
I'm
Simon
Foster,
based
in
in
London
and
I,
will
be
a
Solutions
architect,
I
guess
by
by
day
I'm
a
co-chair
of
the
cartographist
working
group
that
developed
the
cncf
maturity
model,
so
I've
I've
come
along
I
guess
that
was
my
main
introduction,
but
likewise
a
lot
of
my
my
work
is
involved
around
at
Deliverance,
so
I
thought
it
came
well
as
well
and
check
out.
Thank
you.
That's
all
thanks.
C
Okay,
I
can
go
next,
hello.
Everyone
thank
you
Abby
for
inviting
me
so
I
work
out
very
hard.
I
am
a
principal
architect
focused
on
applications
developer
on
my
previous
role.
I
was
helping
customer
high
profile,
customers,
adapting
containers
and
openshifts
I
learned
a
lot
from
there
and
I
am
right
now
in
my
new
role,
helping
products,
marketing
teams
to
understand
how
yeah,
how
how
developers
see
the
platform
and
also
I'm
all
about
adoption.
So
that's
I
think
I
would
like
to
contribute
and
hopefully
I
will.
C
A
G
Was
yeah
I
just
said
brief:
High
yeah
I'm,
sorry,
my
senior
devops
engineer
for
a
company
called
plan
view,
but
it's
a
lot.
More
of
my
role
has
turned
into
platform
engineering,
so
yeah
I'm,
just
kind
of
you
might
be
why
I
work
with
Mauricio
salentino
to
do
a
bit
of
work.
So
you
recommend
an
extra
fun
of
these.
A
B
A
Good
to
have
you
here
yeah,
we,
we
appreciate
it.
Okay,
so
I
was
thinking
today.
You
know
kind
of
the
to
start
talk
about
some
of
the
stuff
related
to
the
platform's
work
itself,
spend
about
half
the
time,
probably
on
that
maybe
more
and
then
I
wanted
to
bring
up
the
conversation
about
which
came
up.
You
know
a
lot
of
times
as
we
were
working
on
the
paper
which
is
going
into
the
capability
domains
a
little
bit
more.
A
H
Yeah
that'd
be
cool,
yes,
a
lot
of
faces
from
the
kickoff
call
last
week.
Weeks
are
really
long
right
now,
I,
don't
know
where
I
am
I
think
it
was
last
week
the
kind
of
takeaway
from
that
call.
It
is
recorded
it's
in
the
same
spot
as
all
the
other
calls,
and
it's
in
the
channel
is
that
there's
a
lot
of
enthusiasm,
but
people
haven't
picked
it
up.
Just
yet
is
what
I
would
say
is
the
short.
H
So
it
doesn't
mean
that
someone
might
not
come
along
with
a
really
good
argument
as
to
why
to
include
Technologies,
but
that's
going
to
have
to
be
a
separate
conversation.
As
of
now
that's
sort
of
the
rule
of
thumb.
The
other
major
decision
was
that,
while
GitHub
and
the
the
issues
format
is
something
that
we
want
to
use
in
when
we're
a
more
kind
of
polished
version,
we
thought
that
the
barrier
to
entry
and
the
access
to
good
conversation
was
best
done
in
Google
Docs.
H
So
we
are
going
to
maintain
that
which
is
the
Google
Doc
is
in
the
issue.
That's
linked
in
the
paper
there
I
guess
the
the
sense
that
call
I
have
added
a
lightweight
contributor
guide,
based
very
much
off
of
what
Simon
did
with
the
the
cloud
native
one.
So
thank
you
Simon
for
that
burst
of
inspiration,
but
it
is
more
lightweight
to
be
fair,
but
it
should
hopefully
give
people
some
insight
into
what
the
document
does
and
and
how
to
engage
and
all
those
if
it
doesn't
please
shout
that's.
H
My
intention
is
to
make
this
very
welcoming.
So
if
there's
anything
that
still
seems
confusing
when
you
read
it
or
if
it
makes
it
feel
more
difficult
to
contribute,
which
is
the
antithesis
of
what
I'm
hoping
for.
So
please
tell
me,
and
then
the
last
thing
I
would
say,
is
it's
been
very
quiet
so
far
on
engagement
actually
on
the
paper,
so
I
have
done
a
few
things,
so
maybe
try
and
encourage
people
to
jump
in
so
I've
added
a
few
to
do
items
throughout
the
document.
A
H
H
Yes
exactly
and
he
was
super
Keen.
His
one
big
takeaway
was
please
make
sure
that
this
doesn't
add
more
division
within
the
language
and
the
formats
and
the
whatever
Associated
his
other
big
takeaway
for
me
was
he
suggested
getting
it
onto
the
white
of
the
working
group
or
tag
website
as
soon
as
possible,
because
it
gives
an
anchor
to
people
of
like
right
now.
It
would
be
like
a
work
in
progress.
Come
see,
our
Google
doc
kind.
H
But
then
we
would
have
that
sort
of
thing.
I
am
optimistic
about
that,
but
I'm
also
conscious
that
that
gives
it
a
lot
of
weight
before
we
maybe
have
even
decide.
If
this
is
something
we
want
to
publish
so
I
don't
want
to
strong
arm
the
group
into
owning
a
maturity
model
by
putting
it
on
a
website,
not
that
we
couldn't
take
it
right
back
off
the
website,
but
I
do
think.
That's
a
conversation
of
making
sure
this
is
a
a
document.
H
A
H
There
exactly
and
I
think
they'll
need
to
work
just
to
be
clear.
Don't
I
have
not
made
any
big
edits
since
you
made
that
point
and
I
think
it
was
a
really
good
point
and
that's
part
of
what
I
think
we
still
have
work
to
be
done.
This
is
not.
This
is
not
targeting
like
a
next
month
publishing
or
anything.
This
is
still
in
the
kind
of
ideation
and
editing
stage
and
contributors
welcome
for
sure,
so,
yeah
I.
A
H
My
optimism
I
have
in
the
contributor
guide
if
we
want
to
release
October
30th,
which
would
be
a
week
before
kubecon,
which
would
be
similar
to
kubecon
Amsterdam
with
the
white
paper.
I
worked
my
way
backwards
to
say
that
we
sort
of
need
to
be
getting
into
GitHub
so
where
we're
looking
for
that
final
review,
rather
than
the
really
wide
open
conversations
on
in
September,
so
September
10th
I
think
that's
tons
of
time.
It's
almost
too
much
time.
H
I,
don't
want
people
to
think
like
oh
I'll,
just
throw
it
away
till
September
we'll
come
back
to
it,
but
yeah
I
think
that
that
leaves
plenty
of
time
for
us
to
have
really
in-depth
good
conversations.
So
I
want
people
to
to
feel
comfortable.
Throwing
out
big
problems
like
is
this.
The
right
format
is:
are
these
the
right
aspects
like
all
those
kinds
of
questions,
because
we
do
have
the
time
to
chew
on
that
as
a
group
yeah?
This
is
the
right
time
to
do
that
so
yeah,
okay,.
A
Just
pause
here
for
a
minute
in
case
anyone
else
wants
to
chime
in
okay,
moving
down
the
moving
down
the
list
here,
so
the
survey
Michael
luskin
has
been
driving
that
we
do
have
some
some
replies
and
that's
nice.
One
interesting
thing
is
that
all
the
replies
so
far
have
said:
they're
platform
Builders,
not
platform,
users,
the
I
think
we
want
to
publish
a
blog
post
just
about
what
the
purpose
is
of
it.
Oh
and.
A
Because
we
that
was
kind
of
we
put
it
out
there,
we
kind
of
advertised
I
know
it's
been
in
Jorge.
Thank
you
for
putting
it
in
row
in
the
Roadie
newsletter.
I've
seen
it
in
a
couple
other
places,
so
that's
great,
but
mostly
like
we
kept
it
not
too
loud,
because
you
know
just
give
people
a
chance
to
review
it
and
give
us
some
feedback.
A
F
A
B
A
So
moving
on
to
the
next
one
prototypes
and
samples,
does
anyone
have
any
update
on
that?
I
know
we're
all
busy
and
you
know
we
all
want
to
build
these,
but
the
I
don't
know
has
any.
Have
we
made
any
progress,
Mauricio's
chimed
in
on
the
thread
that
he
might
have
done
some
stuff
I'm.
Sorry,
so
you
came
off
mute.
You
wanna,
yeah,.
G
H
Which
thread
is
evolving
on
that
so
I'm
aware
of
the
GitHub
issue,
which
I
don't
think
it's
had
a
lot
of
updates
recently.
Is
there
a
thread
some
a
different
issue
that
I'm
not
following
more
where
that
is
or
a
different
repo
or
a
different
thread?
In
slack,
maybe
I'm,
just
I've
only
been
watching
out
of
curiosity.
I
haven't
been
involved
at
all,
but
it
is
curious.
Curiosity.
A
Did
we
I
know?
One
of
the
things
that
we
had
discussed
is
that
it
was
very.
It
was
kind
of
broad
just
to
create
a
platform.
We
wanted
a
use
case
like
I,
don't
know
web
app
development
or
AIML
or
I.
Don't
know
what
did
you
did
you
kind
of
hammer
on
that
a
little
bit
like?
What's
the
target
user
yeah.
G
So
it
seems
kind
of
broad
right,
so
he
seems
to
have
started
that
link
I
sent
started
with
just
kind
of
pulling
a
few
tools
so
he's
using
cross-plane,
b-cluster,
k-native
and
Dapper
to
create
some
type
of
like
it
does
seem
like
just
kind
of
like
a
simple
web
app.
That's
able
to
do
just
basic
calls
across
Services
and
then
yeah
we're
just
kind
of
recording
it
within
the
service.
Catalog.
F
G
H
One
of
the
things
I
really
liked
about
the
potato,
which
I
realized
is
narrow
in
its
scope
and
potato
being
a
way
to
package
up
apps,
which
is
a
small
part
of
a
developer
platform.
It's
something
that
was
being
sort
of
is
still
I,
think
being
maintained
by
the
working
group
in
ways,
but
the
the
reason
I
really
like
that
is
because
it
was
a
problem
statement
that
could
be
solved
in
many
different
ways,
from
open
source
software,
around
kind
of
the
community
and
and
in
piecing
together,
open
source
software.
H
So
there's
a
few
I
think
there
are
a
couple
Solutions,
whereas
more
than
one
so
like
in
this
case,
where
V
cluster
plus
cross
plane,
like
that's
a
really
nice
way,
to
be
able
to
Showcase
like
two
tools,
working
together
to
solve
a
problem.
I
wonder
if
to
your
point
Josh
about
trying
to
make
this
more
specific
and
Valentina.
H
Your
your
experience
with
trying
to
like
root
out
like
problems
with
customers
and
really
being
able
to
understand
the
use
case
and
stuff
I've
been
using
a
framework,
the
jobs
to
be
done,
framework
which
I
can't
say
I'm
an
expert
in
so
people
may
have
a
better
framework
or
a
something
else,
that's
easier
to
grasp
or
or
whatever,
but
something
like
this,
which
we
could
say
similar
potato
the
job
to
be
done,
is
to
package
up
your
app
like.
What's
this
platform's
job
to
be
done,
is
it
to
release
through
multiple,
multiple
environments?
H
H
Here's
an
example
like,
and
that
way
we
kind
of
build
that
Community
engagement,
which
might
reduce
also
the
cost
of
Maintenance,
which
I
know,
is
a
big
concern
in
Mauricio's,
which
is
like
we
make
this
thing,
and
it
goes
out
of
date
the
next
day,
because
it's
you
know,
sandbox
level,
projects
that
are
updating,
frequently
right.
So
just
a
thing
to
throw
out
as
an
idea.
C
No,
no,
it
was
crazy
because
what
you
said
I
was
trying
to
like
understand
the
like
the.
What
is
the
use
case
and
it
seems
It's
like
we
are
trying
to
show
the
value
of
the
platform
through
different
prototypes.
Is
that
okay?
So
so
maybe
we
can
Define
like
okay?
What
are
the
use
cases
we
want
to
cover,
depending
on
the
personas?
Is
it
for
developers
so
or
devops,
or
maybe
we
look
at
cicd
or
containers
or
on
boarding
and
and
then
to
create
different,
like
Journeys
for
them?
C
G
Yeah
yeah,
I,
totally
agree,
I.
Think
I
think
he
intentionally
made
this
very
vague,
because
it's
a
little
bit
easier
to
narrow
in
than
it
is
to
sort
of
like
widen
out,
but
yeah
I
see
that
I
think
it's
a
great
potential
like
at
least
next
step
of
like
okay.
We
can
sort
of
take
this
and
figure
out
what
are
good
use
cases
and
kind
of
fine-tune
different
prototypes
depending
on
the
use
case,
foreign.
A
I
Is
it
possible
to
have
like
a
more
just
project,
self-initiated
like
description
whatever
they
have?
For
example,
I
was
just
confused
about
you
know
cosplaying
and
kcp
TCP
and
then
I
came
to
know
cross
planes
for
basically
infrastructure
provisioning,
whereas
kcp
is
a
workflow
provisioning,
right
management
so
and
there
and
even
TCP
have
additional
extensions,
because
there
is
a
kcp
edge
which
now
become
Cube,
Stella,
there's
a
new
name
for
that.
I
Do
you
think
it
would
be
good
to
have
like
a
kind
of
kind
of
link
to
that
section
of
the
the
platform
choices
under
each
category,
just
whoever
think
they
belong,
that
they
can
add
them
to
their,
not
not,
maybe
not
Adam.
At
least
they
have
a
link
to
their
project
and
give
their
description
of
the
their
technical
capabilities.
I
So
one
thing
I
find
because
I
I'm
a
database
guy,
so
one
one
person,
CMU
Andy
Pavlo
in
CMU
University
he's
have
a
lot
of
talks
from
different
database
providers
in
his
seminars.
What
I
find
is
in
order
to
not
have
a
key
maker,
it's
good
to
actually
have
understanding,
what's
what's
unique
of
the
architecture
and
design
and,
of
course,
the
implementation.
I
This
way
you
know,
users
can
make
their
own
choice
based
on
the
actual
technical
details
on
whether
that's
what
they
need
instead
of
a
marketing,
because,
most
of
time
the
documentation
is
either
too
marketing.
Ish,
there's
too
high
level
or
two
details
for
the
actual
developers.
There's
nothing
in
between
that
can
help
the
users
I
mean
the
developers
to
make
a
choice
of
what
they
was
needed
for
that
you
know
particular
project.
H
I
think
that
was
a
much
better
way
of
saying
what
I
said,
which
was.
Basically,
if
we
start
with
what
is
the
problem
that
we're
trying
to
solve
on
the
platform
with
the
platform
it
all
of
a
sudden
becomes.
We
can
have
example
things
that
solve
that
problem
across
different
Technologies,
which
therefore
allows
users
to
evolve
and
check
their
understanding
of
the
different
Technologies
and
and
why
this
one
technology
can
be
replaced
by
these
two
working
well
together,
so
and
whatever
things
along
those
lines.
Yes,
that
was
yeah.
That
was
very
much
but.
H
A
Yeah
I'm,
getting
it
better
too
now
I
I
pasted
in
the
Potato
Head
delivery,
repo,
which
I
think
is
kind
of
what
you're
saying
Abby
like
there.
We
it
Potato
Head,
the
actual
containers
are
built
somewhere
else,
but
this
just
demonstrates
five
different
ways
or
something
to
deploy
it
so
like
we
could
take
off
like
if
we
have
a
prototype
platform
for
building
a
simple
web
app
and
someone
does
it
hearing
cross
play,
but
we
could
have
a
repo
that
shows
here's
four
other
ways
you
could
have
accomplished
the
same
thing.
H
Yeah
I
think
there
are
two
layers
to
this
like
I:
don't
want
to
lose
the
app
that
I
I'm
hearing
from
Victor
around,
like
hosting
potentially
deeper
marketing
and
or
more
from
the
projects
that
we
think
are
well
within
the
Tag,
app
delivery
and
working
group
platform
domain
space
right
so
like
there's
an
Ask
here,
that's
like
I!
Don't
I
want
to
understand
more
about
these
specific
projects.
H
I
hear
coming
up
in
those
spaces,
which
I
think
is
one
thing
that
we
can
deliver
as
a
group,
you
know
if,
if
we
see
wide
value
there
I
think
there's
a
separate
thing
that
we're
talking
about
as
well,
which
is
like
this
prototyping
or
I.
Think
reference
architectures
is
what
we
talk
about
of,
like
that's
solving
problems
and
I.
H
Think
in
that
case,
I
guess
the
ask
right
now
is
whether
or
not
we
we
can
move
up
a
level
from
the
tech
do
we
should
we
be
talking
about
you
solve
your
platform
by
using
V
cluster,
or
should
we
be
saying
you
solved
your
plaque
like
you
have
a
problem
where
you
want
to
have
a
cost-effective
way
to
have
isolation?
You
can
do
that
with
the
cluster.
You
can
do
that
with
this.
You
can
do
that,
whatever
right
like
and
and
but
this
is
the
problem
like
we,
the
FAQs,
essentially
like
we
get.
F
H
G
Sorry
to
roll
us
back
a
little
bit
but
I
think
I'm.
It
helps
to
know
like
what
and
I've
said
this
I
think
the
last
one.
It
helps
kind
of
know
what
the
end
is
going
to
look
like
as
we're
sort
of
like
trying
to
head
there.
So
it
seems
like
there's
a
few
different
uses
for
this
prototype.
I
just
haven't
heard
like
I
guess
the
overview.
G
It
seems
like
it's
going
to
be
used
to
like
kind
of
showcase
difference,
like
capabilities
like
what,
like
a
platform,
should
look
like
or
could
look
like,
depending
on
different
use
cases
and
to
also
like
help
engage
with
some
of
these
projects
and
sort
of
like
help
people
to
kind
of
fill
in
holes
where
there
might
be
like
it's
just
me
guessing
based
off
of
what's
being
said,
but
is
there
anybody
that
can
get
like
a
succinct?
This
is
what
we
want
to
do
with
this
prototype.
H
I
think
we're
back
porting
a
little
bit
of
hard
work
done
by
Mauricio
and
if
I
don't
want
to
speak
for
him,
he'll
probably
watch
his
back
and
be
like
Abby.
You've
got
this
all
wrong,
but
I
think
the
thing
he
articulated
most
clearly
at
the
tag
event
in
Amsterdam
was
that
he
wants
tangible
things.
He
wants
working
code,
tangible
things
that
people
can
can
play
with
and
so
I
think
from
his
perspective,
that's
what
he's
making
Headway
on
and
I
think
he's
doing
a
great
job
of
it.
H
A
One
of
the
things
that
went
through
my
head,
as
we
were
talking
about
docs,
is
adrs
like
I'm,
not
saying
it's
adrs,
but
we
basically
need
to
document
like
Mauricio
I
also
am
kind
of
like
he
creates
cool
things,
and
but
we
can
go
and
explain
like
we
chose
to
use
crossbling
because
of
X.
We
chose
to
use
Vault
because
of
Y.
We
chose
this
and
then
just
have
docs,
maybe
in
that
repo
or
something
make
sense.
E
It
works
on
k3d
locally,
it
works
in
AWS
and
it
works
in
sibo
and
then
in
beta.
It
also
works
in
digital
ocean
and
vulture
clouds
and
like
it,
if
that's
ever,
a
need
of
this
working
group
to
like
have
consistently
creatable
ephemeral,
local
or
Cloud
kubernetes
environment
that
has
terraform
infrastructure
as
code,
that's
managed
by
Atlantis.
It
has
hashicorp
Vault
that
serves
as
both
the
secrets
manager
and
an
IDP
and
an
oidc
provider
that
extends
throughout
the
platform
tools.
Argo
CD
get
Ops
is
our
opinion.
E
E
F
F
A
That's
awesome:
okay,
let
me
go
ahead
and
move
to
the
next.
The
next
two
things
are
are
simple,
so
we'll
just
unless
there's
something
about
them:
the
translations
or
making
progress.
You
just
have
to
figure
out
what
Hugo
expects
nothing
really
to
update
there.
I'm
managing
the
graphic.
That's
yeah
I
need
to
figure
out.
I
need
to
reply.
That's
on
me.
A
J
Yeah,
it's
very
simple
basically
because
I
didn't
have
that
much
time,
but
I
did
reach
out
to
the
marketing
committee
from
the
cncf
and
then
Christy,
who
is
I.
Think
the
lead
of
some
department
for
marketing
at
the
cncf
said
that
she
would
set
up
some
time
with
me
to
explain
how
things
work,
but
we
haven't
booked
anything
yet,
but
I
think
she.
A
J
F
J
Think
the
way
of
gathering
content
is
not
going
to
be
a
problem.
So
what
I
want
to
figure
out?
What
is
it
that
we
need
what
kind
of
caters
we
need
to
prepare
it
in
advance?
Do
they
do
it?
Do
we
do
it
and
from
that
point,
it'll
be
relatively
easy
to
set
up
a
proposal
on
how
we
can
create
this
kind
of
content
pipeline
for
the
newsletter
every
month.
A
J
A
J
J
Makes
sense,
I
was
just
curious
because
I've
seen
that
they
do
have
some
way
of
of
letting
people
subscribe
to
specific
channels.
So
it
might
be
useful
for
Gathering
people
to
sign
up
into
our
newsletter
if
you
can
use
their
channels
because
they
they
have
this
kind
of
big
massive
page
with
hundreds
of
newsletters,
because
I
know
from
the
cncf
I.
Don't.
A
J
A
H
C
J
H
J
A
Cool
all
right,
I
brought
up
my
page
here,
because
I
think
it'll
be
easier
to
walk
through
with
at
least
something
to
read
so
looking
over.
You
know
when
we
we're
working
on
this
and,
to
be
honest,
my
one
of
my
motivations
of
working
on
the
platform
stuff
is
to
start
rationalizing
some
of
the
capability
domains
that
we
have.
A
I
want
to
share
a
couple
ones
that
would
be
I
think
particularly
appropriate
for
us
in
this
group
or
in
Tag
app
delivery
in
general
yeah.
They
want
to
hear
yeah
your
feedback
if
there's
others
so,
okay,
I'll
just
get
started
so
I
put
in
the
link
here,
there's
a
a
couple
issues
that
have
come
up
recently
around
this.
Just
first
of
all,
the
end
of
the
paper.
A
I
mean
I
every
if
ever
read
the
paper,
I
guess
at
the
very
end,
there's
a
sector
on
capability
domains
where
we
hypothesize
I,
guess
12
or
13
things
and
basically
we're
trying
to
help
platform
Engineers.
You
know
on
they
just
think
about
the
space
and
what
they
might
need
to
consider.
One
of
the
things
we're
asking
in
the
survey
you
might
have
noticed
if
you
went
through
is
like
which
ones
of
these
are
important.
Are
you
using
these?
A
Most
of
them
have
out
of
the
survey
respondents?
Half
of
them
have
chosen
I
think
almost
all
of
them.
Some
of
them
have
gotten
more
than
that
closer
to
70
or
80
of
the
respondents
yeah.
So
within
those,
some
of
them
are
tags
already
within
cncf
or
other
groups.
So
obviously
observability
you
know
we
don't
need
to
tackle
security,
probably
not
much
either
networking,
but
these
these
seem
appropriate
for
us.
A
The
first
I'm
almost
wondered
thinking
about
where
to
start
but
yeah.
The
first
two
are
kind
of
about
our
interfaces
and
I
didn't
put
portal
first,
because
I'm
I
hear
about
backstage.
So
much
literally
heard
a
big
announcements
about
it.
This
morning,
but
I
put
apis,
that's
actually
probably
my
passion.
So
that's
why
I
put
that
first
and
then
what
I
mean
by
that
is.
A
A
You
know
you
could
call
them
I
I,
put
rendering
because
I
sometimes
think
of
Helm
charts
as
kind
of
this
way,
things
that
bring
together
lots
of
underlying
kubernetes
resources
and
then
create
another
API
on
top,
so
I
guess
a
couple
in
this
area
and
it's
TBD
there's
not
even
an
issue
open
on
this.
So
this
is
just
like
kind
of
exploring
this
with
you
all.
A
A
Yeah
I
guess
I'm
not
clear
if
I'm
asking
for
feedback
at
this
point,
but
yeah
if
people
want
to
chime
in
I
mean
if
we
do
want
to
pursue
these
I
guess
what
we
would
do
is
kind
of
what
we're
doing
in
these
ladder
too,
which
is
why
I
was
like
hesitating.
Maybe
I
should
describe
these
first,
because
these
have
momentum
already,
but
the
way
they've
started
is
a
document
that
kind
of
describes
existing
projects
existing
work,
ideas
that
we
might
tackle
that
we
might
pursue
to
make
those
domains
easier.
A
The
artifacts
one
has
gotten
even
further
they've
defined.
You
know,
goals
that
they
want
to
achieve.
They
want
to
make
it
easy
to
query
across
many
different
kinds
of
oci
packages
to
find
out
information
about.
What's
in
them,
and
things
like
that,
there's
another
use
case
I
have
in
mind
which
we
haven't
even
started
on
yet
and
then
that
can
either
you
know,
depending
on
where
we
are.
If
that's
like
got
enough
stakeholders
that
can
become
its
own
group
and
you
know
they
can
pursue
those
goals
on
their
own.
A
If
not,
you
know
I,
don't
we
could
pursue
something
different,
it's
kind
of
iterative
yeah,
so
I
guess
I'll
just
keep
reading
through
all
of
the
next
one
that
I
have
been
really
thinking
about.
A
lot
is
Portal
use
cases,
so
it's
another
interface
into
the
platform.
I
guess
hopefully
built
on
top
of
the
apis,
in
my
opinion,
but
you
know
this
is
a
huge
topic
for
the
industry
right
now
with
backstages
in
interest.
A
What
exactly
are
the
things
that
people
are
doing
there
so
I
put
down
a
few
hypotheses
here
that
people
use
backstage
to
keep
track
of
their
first
party
applications
and
products,
people
use
it
to
provision
and
template
new
projects
and
people
use
it
for
docs
and
info.
Each
of
these
feel,
like
you
know,
there's
plenty
of
plenty
of
opportunities
here,
like
there's
other
first
party
software
catalogs.
So
should
we
you
know
if
we
got
together
and
discussed
capabilities
of
portals?
We
might
you
know,
take
these.
Take
like
okay
software
catalog.
A
There's
this
one!
There's
this
one
there's
this
one:
how
are
they
similar?
What
could
we
do
to
make
them
all
great
like
have
an
application
type
in
kubernetes?
Let's
say
I,
don't
know.
Project
templates
like
backstage
uses
some
sort
of
Arcane
cookie
cutter,
sorry
for
saying
Arcane,
maybe
it's
not,
but
it
uses
its
own
thing.
It.
A
Oh
sorry,
right
so
I'm
behind,
but
that's
the
idea,
like
maybe
there's
a
templating
engine
which
should
be
developed
outside
of
Backstage
and
then
backstage
can
plug
it
in
or
maybe
there's
another
one.
That's
that's
great
that
some
other
company
wants
to
include,
or
maybe
just
a
GitHub
project
template
I
just
want
to
clone
it
and
go.
A
A
The
other
two
have
a
bit
more
momentum
in
increasing
order.
Actually
the
artifacts
is
the
furthest,
but
connections
keeps
coming
up
service
connections
and
service
bindings,
and
this
one
was
is
a
little
more
mature
than
the
first
two,
so
I
I
guess
it
was
me,
I've
been
pulling
folks
together
that
have
a
stake
here.
Some
of
the
weaveworks
folks
reached
out
about
Sops
Sops,
which
was
also
recently
accepted.
A
A
A
Are
there
patterns
we
could
apply
like
right
now,
Vault.
You
know
people
think
of
that,
but
there's
the
the
Key
Management
Systems
from
the
big
cloud
providers
and
others
can
we
Sim
I
guess
my
my
thinking
here
is
the
end
goal
is
to
make
it
easy
for
your
app
developer
that
any
workload
could
connect
to
any
of
the
many
different
service
types
we
deal
with
and
right
now
it
feels
like
every
everything
is
a
snowflake
in
this
area.
A
Each
every
backing
service
has
a
different
way
for
publishing
its
Secrets
every
or
you
know,
connection
Secrets.
Every
every
actual
workload
has
a
different
way
of
finding
them.
You
know
applications
not
Properties
or
something
and
nvars
and
etc,
etc.
So,
there's
some
days
in
there
we
don't
have
a
group
there.
So
it's
more
like
check
out
the
issue.
A
If
you're
passionate
about
it,
let's
keep
talking,
you
know
and
and
let's
get
it,
let's
get
a
group
together
and
we
can
Define
what
we
want
to
achieve
yeah
last,
but
but
not
least,
actually
this
one
came
up
and
I'm
just
writing
it
because
great.
If
there
are
people
that
are
passionate
about
it,
then
let's,
let's
facilitate
it
and
enable
it.
So
the
basically
this
came
out
of
oci
there's
a
group
there.
A
That's
trying
to
use
I
mean
oci
itself
is
deep
in
working
on
supporting
things
other
than
container
images
or
file
system
layers.
If
what
it
comes
down
to
in
a
config
file,
but
we
needed
a
group
that
was
actually
bringing
together
the
implementers
and
and
developing
examples
and
listing
schemas
and
just
Gathering
so
like
right
now
we're
we
had
our
initial.
We
had
an
initial
media
to
go
over
the
charter.
The
charter
is
almost
ready
to
go
pretty
clear
path.
A
The
goal
is
to
enable
searching
across
many
different
kinds
of
oci
packages
for
s-bombs
and
attestations
and
vulnerabilities
and
blah
blah.
That's
just
to
start
and
the
way
we're
gonna
do.
It
is
documenting
or
gathering
up
schemas
for
the
like,
Helm
charts
can
go
in
an
oci
I
just
had
a
chat
with
duasm
people,
maybe
I
shouldn't
call
them
the
wasm
people,
but
whatever
the
group,
people
that
are
working
on
wasm,
they
also
webassembly.
A
They
want
to
embed
binaries
into
those
packages,
maybe
not
all
file
system
layer,
the
you
know,
Opa
packages,
things
there's
a
lot
like
image
package
operator
framework
also
all
have
their
own
patterns
for
bundling
things
into
an
oci
package
and
then
unbundling
and
deploying
so
we're
just
Gathering
those
up
at
that.
At
this
point,
we're
almost
gonna
start
the
Ouija,
though
the
working
group
so
yeah,
so
that
one's
pretty
mature.
A
If
you
are
passionate
about
participating
there
come
join
hashtag,
Ouija
artifacts,
the
group,
the
last
one
I,
don't
I
I
am
not
as
sure
what
to
say
about
it
but
multi-tenancy.
That's
come
up.
Alex
Jones
brought
this
up
a
long
time
ago
and
there's
been
a
couple
threads
about
a
paper
on
it
or
so
I
didn't
write
exactly
what
we
would
want
to
achieve.
I
know
this
comes
up,
but
yeah
it's
another
one.
That
would
be
relevant
for
our
group.
A
I
I
guess
you
know
I
think
when
we
first
talked
about
it
was
like
multi-tenancy
Beyond,
like
the
cluster
Beyond
compute,
like
within
applications
and
backing
services,
so
yeah
I
didn't
want
to
skip
that
one.
So
that's
in
here
yeah,
so
I
guess
I'll
pause
here
and
I
kind
of
yeah.
What's
your
thoughts
on
this
is?
A
J
Yeah
I
think
this
sounds
great
and
all
of
it
resonates
with
in
what
I've
seen,
because,
usually
when
people
try
to
stand
up
backstage,
they
basically
have
to
touch
in
all
the
capabilities
of
the
platform.
So
I
definitely
seen
all
of
this
capabilities
or
or
issues
or
people
trying
to
tackle
all
of
these
when
sending
backstage.
So
it's
it
definitely
sounds
like
it's
the
right
path.
G
Yeah
I
also
think
it's
just
a
good
way
to
cultivate.
Innovation,
like
I,
think
you're
trying
to
hit
on
just
a
lot
of
different
avenues
that
there
could
potentially
be
gaps.
So
yeah
I
think
you're
absolutely
right
that
the
hard
part
is
finding
the
people
to
do
it
and
think
about
it
and
consider
how
to
solve
some
of
these
problems.
But
I
think
it's
a
good
way
to
approach
like
finding
the
problems
themselves.
E
Like
you
know,
when
you
have
a
when
you
have
a
software
development
shop
and
they're
building
apps,
you
know
it's
it's
it's
easy
to
kind
of
think
about
an
app
as
being
something
that
runs
in
kubernetes,
but
it
it
isn't
always
you
know.
Sometimes
you
need
a
an
RDS
database.
You
need
some
kind
of
data
storage
in
the
cloud
and
there
are
some
nuances
about
that
that
are
maybe
worth
considering
as
a
a
pillar
of
capability
domains.
E
Maybe
another
thing
that
jumps
out
is
you
know,
and
probably
closer
to
the
get
op
sphere
of
things,
but
Secrets
management
from
an
application
standpoint
from
an
application,
delivery
and
application.
You
know
ecosystem
standpoint
like
Secrets
management
is
regularly
one
that
feels
like
a
pillar
to
me
and
I
and
and
I.
Don't
know
what
of
those
may
just
kind
of
slide
under
portal
and
may
just
slide
under
you
know
other
other
pillars
that
you
guys
have
already
identified,
but
just
wanted
to
to
call
this.
Those
two
out.
A
E
I
And
then
you
as
John's
first
meeting,
we're
actually
probably
nearer
to
the
community
or
to
the
overall
kubernetes
community,
so
I'm,
actually
learning
about
API,
Machinery
understanding,
that's
the
foundation
of
everything.
I
actually
find
it
useful
to
understand
a
little
bit
so
that
you
know
at
least,
for
example,
during
the
security
review.
Sometimes
it's
good
to
understand
how
individual
components
interact
with
each
other.
I
So
I
guess
that's
for
that
API
part
and
then
also
in
what
what
John
said
exactly
so
I'm
always
wondering
for
you
know,
kubernetes
is
definitely
taking
off
and,
however
right
now
people,
you
know
the
big
public,
Cloud
providers,
the
the
the
private
Cloud
providers,
VMware
or
openstack,
and
all
that
The
Sims,
like
virtual
machine,
is
still
the
main
technology
used
to
build
a
cloud
right
so
talking
about
capability
domains.
What
is
the
gap
between
from
from
here
to
in
the
future?
I
Is
public
cloud
provider
is
going
to
use
kubernetes
alone
to
build
future
cloud
and
what
is
the
gap
there?
I
know
it's
not
an
app
delivery
question.
It's
an
overall
kubernetes
question,
but
what
is
the
gap
and,
and
is
there
a
Gap?
What
what
do
we
solve
the
problem
right
now?
As
John
said,
we
probably
need
the
external
service,
whether
it's
you
know
rdas
or
whatever.
It
is
that's,
probably
where
the
Gap
is.
I
Not
necessarily,
for
example,
again
going
back
to
database
because
that's
where
I
talked
almost
right,
so
traditionally,
kubernetes
is
known
for
not
so
secure
as
secure
at
least
a
stateless
platform
to
build
application
middle
tier
not
for
database.
Now
there
is
initiatives
such
as
you
know
the
dok
database
on
kubernetes
community,
that's
putting
the
database
on
the
on
the
cluster,
but
there
are
I
think
there's
still
a
lot
of
work
to
be
done
to
make
it
really
Mission
critical.
I
Just
like
you
know,
because
some
older
Mainframe
and
Oracle
database
can
run
for
10
years
without
being
shutting
down
right.
Is
that
going
to
be
even
possible
to
be
running
on
kubernetes,
so
those
kind
of
things
and
also
like
like
having
like
a
Secret
store
it
is
it
guaranteed
to
be
because
secret
is
super
important
right,
because
if
you
lose
secret,
you'll
lose
everything
right.
So
virtual
machine
in
the
cloud
is
already
kind
of
mature
so
that
those
data
is
guaranteed
to
be
not
lost
right.
I
A
C
I,
don't
know
yeah
so
yeah
with
this
try
to
make
it
quick
with
the
secret.
It's
not
so
configurations
management.
So
not
everything
goes
into
the
cigarette,
but
I
agree
completely
like
how
what
I've
seen
is
developers
trying
to
understand
how
my
current
configurations
that
work
on
any
environment
that
is
not
kubernetes
will
work
now
on
kubernetes
and
that
could
be
secret
management
configurations.
C
All
of
that
around
it,
and
the
other
thing
that
I've
seen
yeah
how
we
can
abort
developers
or
other
users
into
the
platform
like
what
is
the
process
for
me
as
an
organization
to
go
through
this
process,
and
this
is
very
important
for
adoption,
because
if
we
can
have
users
end
users
connecting
into
the
platform
I
know,
we
have
idps
as
well
that
kind
of
abstract
that
it's
it's.
You
know
it's
very
important
for
adoption
of
the
platform
as
a
whole.
So
let's
see
what
I'm
saying
I
wanted
to
add:
yeah.
A
C
E
Yeah,
just
one
other
thought
to
throw
on
the
table
is
that
of
ephemeral,
environments
and
I
know
that's
a
a
a
a
bit
of
a
grenade
to
throw,
but
there
there's
like
there's
a
distinct
and
seemingly
ever-present
need
for
shops
that
are
adopting
kubernetes
to
just
want
that
that
ephemeral
environment
for
their
micro
services
that
they're
building
and
shipping
right
like
while
the
developer's
still
on
their
Branch.
They
want
to
see
it
they
want
to.
E
But
anyone
that's
tried
to
implement
it
like
you
know
that
there
are
layers
about
it.
If
it's
a
front-end
app,
it's
pretty
easy.
If
it's
a
front-end
app
with
a
back-end
app.
Does
the
ephemeral
environment
include
both
of
those
what
happens
if
there's
State
on
the
back
end
so
like
maybe
in
terms
of
just
like
identifying
app
delivery
pillars
like
that
that
that
may
very
well
be
another
one
to
consider.
A
A
Are
you're
passionate
about
and
you
feel,
like
you
know,
there's
a
number
of
projects
and
you
want
to
try
to
bring
them
together
like
feel
free
to
open
an
issue
in
you
know
either
a
Google,
doc
or
open
an
issue
in
GitHub.
Let
us
know
about
them,
and
you
know
we
can.
We
can
discuss
it.
There,
yeah,
like
with
the
with
the
artifacts
group.