►
From YouTube: TAG General Meeting - 2022-09-07
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
C
A
B
B
D
Any
topic
today
because
I
didn't
see
any
thing
on
the
agenda.
D
D
Oh
really,
oh,
that
that's
the
my
seven
yeah
I
see
oh
yeah,
so
I
saw
the
Knight
21
like.
D
Yep,
like
I,
think
I
think
maybe
you
can
help
just
like
host
a
meeting
because
like
it's
like
will
you
do
that?
Josh.
B
Yeah
sure,
actually,
just
as
you
said,
that
some
guys
started
doing
work
right
next
to
me
here,
but
yeah
I
know
that
we
have
Bill
Maxwell
here
for
to
talk
about
Acorn,
even
though
we
don't
have
a
ton
of
folks
here
palaces.
What's
your
first
name,
I
I
keep
forgetting.
C
It's
Mike,
okay,
there's
too
many
mics
where,
where
I
work
so
I
just
go
by
my
last
name.
Oh.
B
Okay,
cool
so
yeah
and
Bill
like,
even
though
it's
just
us.
This
will
be
posted
up
on
YouTube
afterwards.
I'll
take
some
notes
here
and
we'll
we'll
create
an
issue
too,
if
there's
significant
stuff
so
that
other
people
can
review
it.
Yeah.
A
B
Yeah
so
let's
go
ahead
and
kick
that
off
and
then
we
can
tackle
a
couple
other
things
on
the
agenda
after.
E
So,
just
at
a
high
level,
you
know
acorns
a
platform
for
packaging
applications
to
deploy
on
the
kubernetes
Clusters,
we're
really
focusing
more
on
the
application
layer
and
trying
to
minimize
how
much
kids
experience
and
knowledge
is
needed
to
at
least
get
started
and
get
going.
E
You
know
I
think
over
time,
as
as
people
evolved,
they'll
work
better,
but
you
know
out
of
the
box
secure
so
that
you
know,
operators
of
the
cluster
can
feel
confident
that
developers
that
are
deploying
these
applications.
You
know
they're
following
a
best
practice.
You
know
basically
putting
some
guard
rails
as
people
as
apps
go
out
to
production
there,
but
really
focusing
you
know
on
that
developer
experience
where
you
we
saw
with
Docker
people
were
able
to
get
up
and
running
quickly
on
their
laptops,
but
then
they
transitioned
to
kubernetes
required.
E
You
know
a
lot
more.
You
know
a
whole
new
set
of
contexts
to
bring
into
the
picture
to
deploy
that
application
being
able
to
do
some
validation
and
some
testing.
Ultimately,
we
want
to
get
to
the
point
where
this
is
packaged,
and
then
you
can
validate
that
it's
packaged
appropriately
passing
unit
tests.
Everything,
you
know,
integration
type,
testing
and
then
really
want
to
have
the
common
artifact,
so
everything
that
gets
packaged
into
an
acorn
image.
E
You
know
it's,
it's
both
the
the
Manifest
and
metadata
on
how
you
deploy
that
plus
all
of
the
container
images.
So,
if
you
think
about
scenarios
where
you're
trying
to
deploy
in
an
air
gap,
you
know
current
Solutions
sort
of
leap,
you
yeah,
you
have
the
like
a
Helm
chart,
for
instance,
and
then
you
have
to
pull
in
all
of
the
different
images
from
all
the
Registries
and
aggregate
that
yourself
yeah.
E
B
Yeah
like
what
what
artifacts
are
are
in
there,
is
it
just
kubernetes.
E
So
we
have
our
our
Acorn
file.
Actually,
this
is
good,
so
we
have
an
acorn
file
which
describes
the
app
we
use
terms
like
sort
of
containers.
Oh
similar
to
the
compose
you
know,
sort
of
the
docker
ux
of
you
know
you
have
a
container
and
here's
the
application.
You
know
you
have
multiple
apps
and
multiple
here's,
the
ports
that
need
to
be
exposed.
E
Here's
here's
the
volumes,
we're
going
to
need
here's
the
secrets
that
are
going
to
be
used,
so
we
take
all
of
that
and
then
so,
if
you
need
to,
we
could
build
the
application.
The
the
the
docker
image
there
on
the
cluster
or
you
know
we
pull
them
from
remote
Registries
and
you
know
that
they
all
get
bundled
into
one
single
oci
artifact.
E
E
Yeah
so
in
the
cluster
I
think
yeah
yeah,
so
there's
a
there's,
an
operator
or
controller
an
API
server.
And
then,
when
you
go
Acorn
build,
we
actually
deploy
a
build
kit
container
with
a
registry
inside
it.
So
that
way
you
can
build
and
deploy
into
that
cluster
without
having
to
yeah
push
it
to
a
remote
registry.
B
Okay,
so
I
guess
that
oh
I
guess
that's
what
you
say.
This
is
the
acorn
operator
they
have
to
install
into
the
cluster,
which
has
which
sets
up
some
of
these.
The
capabilities
like
build
yep.
E
So
we
put
a
API
server
and
a
there's,
a
controller
element
that
just
basically
watches
custom
resource
and
then
does
the
right
thing.
So
the
the
nice
thing
about
this
is
you
the
end
user's
perspective.
You
know
they're
they're,
describing
the
application
that
sort
of
containers
and
and
volumes
and
secret
layer
and
behind
the
scenes
we
take
all
that
information
and
we
create
the
you
know
the
Ingress
and
the
services
and
the
deployments
and
everything
else
that
the
user
would
have
to
do
and
map
the
PVC.
E
You
know
just
a
nice
ux
on
top
of
it
without
requiring
that
upfront,
knowledge
of
okay,
here's
the
ammo
for
an
Ingress,
here's,
the
ammo
for
the
service,
here's
the
ammo
for
the
deployment
yep,
so
we're
leveraging
everything
within
the
kubernetes
cluster
or
kubernetes,
but
trying
to
present
it
to
the
end
user
to
kind
of
make
it
easier
for
them
to
onboard
and
get
up
and
running
quickly.
E
Yeah
no
everything's
done
in
the
Acorn
file
and
it's
it's
explicit.
You
can
change
some
Behavior
at
run
like
like
you
used
to
be
able
to
do
with
Doc
curse,
you
Docker,
run
minus
p
and
you
know
publish
a
port
or
something
like
that,
so
you
can
do
that
with
Acorn
as
well.
E
There's
there's
three
layers
that
we
sort
of
describe
the
network
there's
internal
to
the
acorn.
So
if
you're
deploying
a
database
with
the
same
thing
that
database
doesn't
necessarily
need
to
be
exposed
outside
of
you
know
that
that
context
within
the
acorn
then
there's
expose
so
that's
sort
of
like
a
cluster-wide
and
then
there's
publish,
which
is
either
service,
load,
balancer
or
Ingress,
depending
on
HTTP
or
not
foreign,
but
yeah
I.
E
So
this
is
just
a
little
python
app
that
I've
been
you
know
putting
together.
You
know
it's
just
flask,
nothing
too,
fancy
it's
just
gonna
Echo
and
have
a
page
to
do.
You
know
you
say
Hi
and
then
I'll
go
back
and
then
here's
the
docker
file
for
that
which
is
again
real,
simple,
set
a
couple
environment
variables,
move
the
source
into
it.
E
But
then
so
that
you
know
you
can
do
all
of
the
normal
python
things
on
your
local
laptop,
but
it's
always
nice
to
yep
mimic
what
would
run
in
production.
So
if
we
start
you
know,
this
is
where
we
start
stepping
in
so
here
is
a
very,
very
simple
Acorn
file
for
this,
and
what
we're
describing
is
the
the
containers
that'll
run.
E
E
You
can
has
args
if
you
need
to
build
targets,
but
exposes
that
then
here
I'm
saying
the
port
is
going
to
publish
Port
5000,
there's
a
built-in
Dev
mode,
and
so
I'll
show
that
first
so
like
when
you're
working
for
that
that
tight
loop
on
on
a
laptop
or
you're
making
code
changes,
you
want
to
see
that
reflected
quickly.
E
We
have
this
Dev
mode
which
we'll
watch
the
local
file
system.
Sync,
the
contents
up
to
the
container
and
you'll
be
able
to
run
update
that.
So
in
this
case
you
know,
there's
Auto,
reloading
and
so
you'll
be
able
to
see
the
updates
in
real
time.
E
This
is
that's
a
good
point,
so
yeah
sorry,
I
forgot
about
this
is
written
in
AML,
Acorn,
markup
language.
It's
a
you
know
it's
similar
to
yaml
and
Json,
but
what's
nice
about
it?
Is
it
has
these?
It
has
control
structures.
So
you
can
do
ifs.
You
can
do
merging
of
object
objects.
You
can
do
for
Loops.
E
You
know
I,
think
you
know
the
other
approach
would
be
to
do
yaml,
plus
a
go
template
right
and
so
that
from
our
experience
that
it
works,
but
it's
it's
a
bit
challenging
to
author.
Sometimes
it
can
get
complicated
and
then
so
we
like
this
style.
It
was
heavily
inspired
by
Q,
but
it's
its
own
thing.
E
A
E
It
does
have
control
structure,
so
you
can
do
ifs
and
you
can
do
for
Loops
merges
and
we've
got
some
built-in
functions.
That'll
help
you
out,
but
it's
all
sort
of
one
language
and
you
have
one
sort
of
syntax
to
flip.
Follow
I'm,
also
declaring
a
second
container
I'm,
not
going
to
use
it
right
away,
but
this
is
going
to
be
the
reddest
one.
E
E
No,
if
you
you
can
do
that
in
the
container
definition.
There's
a
sidecars
field,
which
then
you
can
add
more
container
specs,
and
then
you
can
also
do
knit
containers
that
way.
E
It
actually
behind
the
scenes,
it'll
be
its
own
deployment,
which
can
be
one
or
many
yeah
you
can.
You
can
set
a
scale
as
well
so
but
yeah,
each
one
will
be
discrete.
E
So
then
I
can
do
an
acorn
run
as
I,
and
so
this
just
gives
me
what
this
is
doing
now
is
I'm
running
it.
It's
going
to
build
the
container
and
then
it's
going
to
deploy
it
as
it
comes
up.
Acorn
is
automatically
assigning
a
DNS
name
to
it.
So
I
can
just
click
on
that
and
my
app
is
up
and
running,
and
so
you
know
this
is
just
real,
simple,
okay,
hi
out
in
the
you
know.
Let's
say,
though
you
wanted
to
show
a
history.
E
E
Apparently
it
picks
up
the
swap
file
and
said
that,
but
now
I
can
say:
I
have
a
history,
it's
keeping
tabs
on
this
so
and
all
that's
doing
now.
E
Is
it's
actually
talking
to
the
redditist
container
in
the
background
and
storing
that
and
then
pulling
it
back
up
and
sorting
the
keys,
but
you
can
see
like
it's
really
a
nice
tight,
Loop,
I'm,
editing,
I'm,
making
code
changes
and
you
know,
let's
say
I'm
happy
with
this
now
I'm
happy
with
this
now
so
then
I
can
say:
okay,
let's
stop
that
and
then
I
can
say:
okay,
let's
Acorn
build
let's
log
into
the.
E
Hopefully,
that's
right
so
now
I
can
say:
Acorn
build
T
Cloud,
not
you
know.
Do
the
full
name.
A
E
E
So
now
you
can
see
if
I
do
acorn
images.
I've
got
my
new
one
there
and
if
I
do
Dash
C,
you
can
see
now
I've
got
because
this.
B
And
acorn
images
is
that's
looking
at
your
local
repository
right,
it's
not
like
pulling
in
something
from.
E
Yeah,
it's
looking
at
the
local
one
so
that
I
have
that
there
and
then,
if
I,
do
Images
dash
C,
that's
the
container,
so
it'll
inspect
the
image
and
list
out
the
containers
and
as
long
as
it's
been
published,
I
can
then
do
something
like
acorn
images,
so
I
see
cute.
A
E
Tribute
installed,
that
would
work,
but
basically
you
could
dump
the
images
to
you
know
an
image
scanner
and
scan
them
periodically
if
you
wanted
to
or
if
needed,
to
just
for
security
reasons
yeah
but
Acorn
all
so
you
can
see
here,
I've
got
my
app
is
running.
It's
got
two
healthy
containers.
It's
under
this
falling
Darkness.
You
can
see
that
this
one's
running
and
this
one's
running.
Maybe
we
have
volumes
and
secrets.
E
You
know
it's
worth
noting
that
so
redness
is
a
staple
application,
so
I
could
say
in
cornball
or
volumes
and
then
say
you
know,
redis
data.
A
A
B
I,
don't
want
to
disrupt
you,
but
I
have
a
question
about
this.
I
can
ask
yeah
what
like?
What's
your?
What
would
happen
if,
like,
let's
say,
I
had
databases
and
caches,
and
things
like
that
as
services
in
my
cluster
already
and
I
just
wanted
to
leverage
those
with
say
a
crd
like
these
are
all
just
containerized
redis
and
that
fits
like
Docker
compose
but
like
what,
if
I
want
to
use
like
operator,
Based,
Services
and
stuff.
E
That's
a
great
question,
so
two
things
there's:
if
you
have
another
app,
that's
deployed
as
an
acorn,
so
you
had
like
a
red
as
Acorn.
You
can
just
link
them,
so
you
would
do
Acorn
run
national
already
installed
redis
colon
redis,
and
then
that
would
just
we
wouldn't
deploy
the
red
as
container.
E
We
would
just
say:
okay,
we'll
create
a
service
that
maps
to
the
redis
endpoint,
and
then
at
that
point
you
can
also
say
you
know:
here's
the
credential
or
you
can
push
your
printage
there's
a
couple
different
ways
to
do
it,
but
it's
also
a
way
to
manage
sort
of
the
way
the
secret
flows
through
there.
So
if
you
needed
a
credential
for
that
database,
that's
also
possible.
E
That's
cool
yeah,
we're
working,
we
haven't
done
it
yet,
but
the
other
thing
is
that
you
know
operators
or,
if
you're,
using
RDS
or
just
like
a
cloud-based
service.
What
we're
going
to
do
is
allow
you
to
create
it's
basically
like
a
DNS
entry
and
then
link
that
in
so
we'll
create
a
service
that
allows
you
to
link
it
in
and
then
it'll
do
the
right
thing.
E
E
E
And
password
so
what's
going
to
happen
in
this
case,
is
you
know,
since
it's
undefined
when
it
in
a
development
scenario,
let's
say
acorn,
look
at
that
and
then
generate
a
random,
username
and
password?
What
you
can
do
is
then
reference
that
here
in
like
and
then
you
can
say,
I'll
probably
Secret
password
is.
B
Yeah,
like
what
I
think
when
I
see
this
is
hey.
If
I
had
an
operator
which
always
wrote
down
a
secret
at
a
you
know
a
conventional
secret
name,
then
yeah
I
could
put
that
in
the
ends
here
and
and
I.
Guess
it
wouldn't
be
a
container.
It
sounds
like
you're
saying.
Services
are
like
a
reverse
proxy
that
you
inject
so.
E
Yeah
I
think
so,
there's
just
standard
kubernetes
services
and
so
we'll
map
them
and
do
the
right
thing
and
then
I
think
we
do
some
magic
for
certain
types
of
services,
but
I'd
have
to
unpacks
that
a
bit
more
but
actually
yeah.
So
if
the
so,
if
your
operator
created
a
secret
and
you
knew
the
structure-
that's
eager
to
have
these
keys
in
the
data
field,
you
would
Define
it
here
either
as
a
opaque
or
whatever
you
want.
E
But
then,
when
you
go
to
run
it,
you
do
Acorn,
run
minus
s
existing
secret,
and
then
you
bind
it
over
the
DB
creds,
and
so
what
that
does
is
instead
of
us,
generating
it.
We'll
use
the
existing
secret
and
then
map
that
in
and
as
long
as
the
the
the
keys
in
the
data
field,
map
up
will
just
automatically
put
the
right
stuff
where
it's
supposed
to
be.
E
Yeah
we're
we're
working
on
getting
it
so
yeah.
We
originally
started
with
q
and
we've
sort
of
moved
down
this
path
of
our
our
own,
so
we're
working
on
I
think
the
next
release
will
have
we're
targeting
to
get
the
the
AML
bits
and
the
actual
parser
and
everything
for
AML.
So
we
we
leveraged
queue
heavily
to
get
it.
The
initial
batch.
E
B
E
So
we've
created
a
build
kit
container
inside
the
cluster,
that's
been
deployed,
and
so
then
we
basically
just
use
build
kit
to
to
do
all
of
the
build.
We
pack
up
everything
from
the
local
in
the
context
that's
defined
by
the
user,
and
then
we
stick
that
into
the
container
I
think
now,
if
I
do
split
plane
vertically.
So
you
can
see
here.
E
So
you
can
see
now
I'm
creating
this
volume
that
would
get
mounted
it's
bound
to
the
redis.
It's
bound
and
it's
called
redis
data
and
it's
part
of
this
application.
I
generated
the
DB
credit
secret,
and
so
now
I
could
do
acorn.
E
So
if
I
needed
them,
I
could
get
the
generated
ones,
but
you
know
there's
just
sort
of
randomly
generated
and
if
I
go
into
the
redis
container
and
I
printed
the
environment,
you
would
see
that
it's
there,
so
you
can
Define.
It
once
share
it
between
the
two
different
containers
and
it
would
be.
You
know
up
and
running
there.
B
E
E
But
anyway,
you
can
get
the
output
which
will
give
you
the
custom
resource,
which
you
could
then
put
into
like
a
flux
and
sort
of
integrate
into
your
existing
cicd
flows.
Today,
we're
working
on
something
that
would
make
it
a
bit
more
native
where
you
can
commit
from
a
developer
standpoint.
E
They
can
commit
just
the
acorn
file
and
then
they
never
have
to
do
the
translation
to
the
kubernetes
Manifest,
even
though
that's
what's
happening
under
the
hood,
but
you
know
just
kind
of
keeping
that
that
flow,
nice
and
consistent
for
the
end
user
and
the
developer
in
particular.
E
You
know
we're
going
to
continue
to
expand.
You
mentioned
operators,
you
know
so
we're
we're
looking
at
the
way
like
the
best
way
to
sort
of
make
that
expose
some
of
those
capabilities.
So
you
could
deploy
an
operator
driven
database
or
something
like
that,
a
stateful
service,
and
then
how
do
we
plug
that
in
in
a
nice
way?
E
Yeah?
So
so
we're
looking
at
expanding
some
of
these
things?
You
know
it's
going
to
take
us
a
bit
of
time
as
most
projects
do.
B
E
E
E
Yeah
yeah,
we
have
the
concept
of
you
can
pass
in
arguments,
and
then
we
have
the
concept
of
profiles.
Even
so,
you
can
set
the
different
defaults
for
those
environments,
so
for
those
arguments.
So,
if
you've
got
like
Dev
test
prod,
here's
the
fields
that
need
to
be
filled
out
and
those
would
be
same
defaults,
we're
sort
of
kicking
around
the
idea
of
having
you
know
a
deployment
file
specific
to
the
deployment
unless
the
acorn,
where
you
would
basically
say
you
know
for
to
deploy
this
application.
E
Able
to
do
you
know
as
we
were
talking
you
got,
you
know
a
database
Acorn
or
another
Acorn
or,
however,
it
works,
but
basically
giving
you
a
way
to
describe
essentially,
similarly
what
you're
driving
with
like
a
values.yaml
per
environment.
You
know
you
do
so.
You
know
I,
think
it's
common
pattern
or
you
either
do
like
a
customize
or
different
values.
Files
we've
seen
both,
but
you
know
the
other
thing
is
I.
Think
we've
we've
integrated
a
few
tools.
E
You
know
to
accommodate
a
couple
different
workflows,
so
you
know
there's
a
few
in
the
space
of
you
know
being
able
to
do
the
development
life
cycle.
On
top
of
you
know,
on
the
laptop
with
the
kubernetes
cluster
like
a
tilt
or
something
like
that,
you
we've
got
that,
but
then,
ultimately
we
end
up
packaging,
the
whole
artifact
and
shipping
that
and
moving
it
through
production.
We're
not
just
focused
on
the
one
use
case
where,
where
it's
develop,
Etc.
B
Okay,
I
put
in
a
couple
other
ones,
I
mean
I,
know
the
idea
of
bundling
up
all
this
stuff
into
a
oci
image
that
the
idea
have
you
heard
of
cnap
I
just
wanted
to
bring
these
up
just
in.
E
Case
yeah
yeah,
yes,
I've
heard
of
it
I'm
personally
less
familiar
with
that
one,
but
I
think
it's
similar
in
concept.
You
know
I.
Think
again,
though,
the
the
piece
here
is
is
not
needing
the
kubernetes
knowledge
to
be
able
to
create
those
fun
pulls.
You
know
be
able
to
describe
what
that
application.
That
package
looks
like.
E
A
B
The
other
big
one
that
I've
seen
come
through
is
is
catch
which,
which
is
even
more
similar
I
feel
like
to
yours,
which
is
shipa,
is
the
commercial
entity
behind
it
s-h-I-p-a
and
they
they
don't
have
their
own
file.
You
I'm
not
supposed
to
make
Kings,
but
you're
you
have
a
little
bit.
It
looks
like
a
little
more
thought
in
your
in
the
acorn
file
and
stuff,
but
they're
it's
a
more
like
Heroku
kind
of
manifest.
B
Like
it
just
says:
what
is
the
hero
stuff
used
to
like
web,
calling
this
and
worker
cola?
This
is
like
two
files
or
something
two
two
lines,
and
then
it
tries
to
infer
lots
of
stuff
and
do
build
pack
builds
and
deployments.
E
Yeah
they
yeah
they're,
they're,
you
they're,
cue,
driven
I
believe
is,
is
sort
of
one
of
their
big
yep
yeah.
We
ran
into
them,
not
ran
into
them,
but
yeah.
We
we
well.
We
were
looking
at
Q
and
and
different
things.
We
definitely
came
across
them.
C
Cool
I
was
curious
as
more
coming
from
the
operator
side.
Do
you
all
plan
to
do
something
where
I'm
assuming
there's
custom
resources
behind
there?
That
says
container
equals,
deploy,
I'm
oversimplifying
probably,
but
is
there
somewhere
in
that
space,
where
you
think
there
would
be
a
place
for
an
operator
to
inform
and
say,
like
okay,
anytime
I,
see
a
container,
let's
say
I'm,
going
to
add
these
labels
or
something
like
that.
E
We're
we're
thinking
about
a
way
to
do
sort
of
a
plug-in
system
so
that
we
could
basically
do
what
you're
describing
so
that
you
know
from
both
a
policy
side
of
things
and
also
like
it
might
be
a
way
to
add
monitoring
or
something
like
that
without
having
the
end
user
have
to
be
very
much
aware
of
like
oh,
you
need
this
Monitor
and
this
annotation
in
this
label.
It's
like
what
is
that
so
yeah,
so
so
we
are
looking
at
ways
to
do
that.
E
A
B
All
right
so
moving
down
the
agenda
here
feel
free
to
stay,
built,
we're
happy
to
help
them
happy
to
have
you.
We
we're
going
to
talk
about
what
I
put
on
I
put
this
on
the
agenda,
but
I
don't
know
if
we
have
the
right
people
to
talk
about
it,
kubecon
Detroit
plans,
hongchao
I,
don't
know
if
you
know
anything
about
it.
D
I
think
Thomas
and-
and
they
have
a
top
for
that,
like
personally
I
yeah,
because
due
to
the
time
zone,
I
I,
understand
I,
normally
just
attended
the
China
One
yeah.
D
Yeah
definitely
like,
like
last
time,
I
talked
with
Thomas.
We
couldn't
speak
at
the
Chinese,
oh
yeah
event
about.
D
B
Okay,
so
I
guess
you
know,
I
will
I
gotta
follow
up
with
Thomas,
because
we're
gonna
do
something
we
got
to
do
it
now
we
gotta
get
going
okay
moving
to
the
next
multi-tenancy
white
paper.
I,
don't
have
any
updates
on
that.
Alex
is
kind
of
the
lead
on
that.
So
I'll
leave
that
and
then
the
Cooperative
delivery
article
I'll
just
bring
it
up.
I
posted
that
in
our
chat
just
to
kind
of
asynchronously
talk.
B
You
know
we
started
the
Cooperative
delivery
working
group
last
I,
guess
November,
so
it
hasn't
really
gotten
that
much
traction
so
we're
trying
to
you
know
I
want
to.
We
want
to
publish
something.
Maybe
that
would
get
traction
or
maybe
to
fold
it
back
into
this
main
mainstream
conversation.
You
know,
there's
not
much
point
in
me
and
one
or
two
other
people
are
here
anyway
having
a
separate
group
for
that.
B
So
that's
kind
of
the
the
impetus
for
writing
up
that
article
to
kind
of
share
where
we
are
maybe
bring
a
few
more
people
and
then
maybe
reabsorb
it
into
this
Main
discussion.
I
mean
I
feel
like
a
lot
of
the
stuff
today,
like
even
with
Bill,
was
sharing,
is
kind
of
about
how
do
you
I
I,
ask
about
coordinating
infrastructure
that
might
be
part
of
it,
but
yeah
so
I
guess
I
I'm
just
bringing
that
up.
So
we
can.
If
anyone
wants
to
talk
about
it
here,.
B
And
then
the
one
other
I
guess
I
put
down
some
thoughts
there
about
the
article
that
was
just
a
first
draft.
One
is
platforms,
it's
you
know
where
do
plat,
how
do
platforms
fit
in
Tag
app
delivery?
You
know
in
our
landscape
which
we're
trying
to
rationalize
a
little
bit
so
I
thought
this
was
an
opportunity
to
kind
of
ask
people
for
opinions
on
that.
Try
to
develop
an
opinion
on
that
on
where
we
think
this
fits
and
what
what
platform
engineering
means
and
yeah.
A
C
I
know
we're
we're
in
the
process
of
investigating
options,
for
how
do
we
we're
driving
people
towards
Helm
right
now?
How
do
we
try
and
it's
a
lot
of
disjointed
activities
that
a
developer
needs
to
do
in
order
to
get
all
the
different
things
that
they
need
to
deploy
their
app?
So
they,
you
know,
need
something
in
Azure,
a
storage
account
and
then
they
need
a
postgres
database
and
then
they
need
to
write
their
Helm
chart
to
deploy
the
app.
C
C
We're
looking
at
like
terraform
controllers
that
could
go
and
spin
up
that
infrastructure
and
spit
out
that
secret
and
in
a
way
that
they
can
consume
that
all
via
you
know
their
Helm
chart
and
you
know
not,
and
they
have
their
terraform
sitting
somewhere,
calling
a
module
that
either
an
operator
has
created
or
whatnot,
but
trying
to
trying
to
simplify
that
aspect
of
like
hey
kubernetes,
isn't
going
to
contain
everything
today,
where
we
work
that
they
need
to
deploy
their
app.
C
So
we're
really
early
on
in
that
journey
and
we've
seen
examples
of
what
looks
like
other
folks
have
have
tackled
with
with
that.
But
yeah
I'll
be
interesting
to
see
if
we
can
get
that
all
functioning.
B
C
Yeah
and
probably
custom
resources
that
say:
okay,
I
need
to
create
a
storage
account.
So
then
the
terraform
controller
knows
that
it's
it's
got
to
do
that
for
them
and
return
the
secrets
and
on
the
helm
side
we
can
kind
of
templatize
that
for
them
a
little
bit
so
they
can
expect
like
okay,
their
their
storage
account.
Secrets
will
look
like
this,
so
reference
them.
This
way
and
you're
Helm
chart
yeah.
B
C
C
So
the
platform
feels
like
it
shouldn't
it
shouldn't
just
be
kubernetes,
although
it's
kubernetes
is
kind
of
that
entry
point
into
all
those
things.
The
platform
is
all
of
it.
I
think
yeah
that
those.
A
B
C
Yeah
we're
still
it's
still
a
very
hard
problem
to
solve,
and
we
we
have,
you
know
pretty
sure
people
that
are
are
tackling
different
aspects,
but
we're
finding
like
hey
it's
it.
It
continues
to
extend
into
a
lot
of
other
teams
in
a
lot
of
people's
areas.
Is
how
do
we?
How
do
we
unify
that?
That
view
so
that,
that's
to
me,
what
makes
it
feels
like
a
platform
is
like
okay,
I'm,
taking
something
and
unifying
it,
so
that
somebody
can
tackle
that
thing.
C
B
Helpful
to
come
to
Common
understandings
and
try
to
establish
some
common
terms
and
ideas.
That's
that's
the
idea.
Okay,
so
I
guess.
My
feedback
here
is
I'll
iterate
on
that
doc.
If
anyone
wants
to,
you
know,
write
something
in
I.
Think
it's
open
to
comment
to
anyone.
If
you
want
to
be
an
editor,
you
know
just
send
me
a
request
and
you
can
do
whatever
you
want
and
go
from
there.
C
And
I
I,
oh
I
I,
do
have
some
Potato
Head
stuff
that
I
I've
been
meaning
to
contribute
the
Ingress
resources
out
outdated.
Now
that
they've
deprecated
extensions
beta,
so
somebody
tried
to
use
the
Ingress.
It
won't,
at
least
for
I,
haven't
actually
looked
at
other
places
other
than
the
helm
chart,
but
but
yeah
we
would
have
to
bump
that
up.
So
I
do
have
like
a
branch
working
on
that
that
I
need
to
put
a
pull
request
in
for.
C
B
C
It's
it's
a
what
five
hour
drive
or
something
like
that,
but
I
I
have
obligations,
so
I
won't
be
able
to
make
it
to
that
kubecon.
But
hopefully
maybe
the
next
one.
D
No,
it's
too
far
away
from
me
now
yeah,
but
actually
I
attended
the
very
first
qcon
I.
When
I
was
in
coex,
it
was,
it
was
free,
like
I
just
enter
it
like.
No
one
stopped
me
for
the
first
one
yeah.
D
Yeah
yeah
because,
because,
like
my
company
course
was
the
sponsor
and
and
I
know,
a
lot
of
like
people
are
attending
at
that
time.
Yeah.