►
From YouTube: CNCF TAG App Delivery 2021-12-01
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
B
C
A
Okay,
we
have
a
nice
agenda
here
today,
conveyor
some
pattern,
collection
updates
and
some
potato
has
updates
and
walkthroughs.
B
A
B
A
A
I
guess
the
most
funny
commit
messages
that
are
out
there
all
right,
so
we
are
three
minutes
in.
Let's
give
people
one
more
minute
and
then
we,
for
the
sake
of
time,
join
start
talking
about
this
submission
work.
A
All
right,
yeah,
I
think,
let's
get
started
and
james
do
you
want
to
get
started
on
the
on
conveyor,
yeah
yeah.
Absolutely
thanks.
D
So
I
know
I
think
two
weeks
ago,
when
we
met
or
three
weeks
ago,
remember
how
often
this
is,
but
when
we
met,
I
know
I
had
we,
the
conveyor
community
had
applied
for
sandbox
status
and
then
I
know
we
got
back
a
message.
There
was
a
message
posted
saying:
was
this
too
specific
to
openshift,
and
so
I
drew
then
I
joined,
and
then
we
had
spoken
about
last
time,
but
we
didn't
have
everybody
here.
I
think
that
we
have
here
now.
D
That's
the
chairs
and
the
tech
leads.
So
I
I
linked
in
here
to
the.
If
you
click
on
the
link
that
I
put
in
the
agenda,
that
that
will
actually
take
you
to
kind
of
you
know,
I
try
to
respond
to
some
of
the
questions,
to
show
that
you
know
it's
not
specific
to
openshift
in
fact,
like
our
demonstrations
and
things
like
that
nature,
we're
done
on
upstream
kubernetes
and
using
those
things
so
and
anything
we
do
find.
Where
there's
challenges
or
issues
we're.
D
You
know
we
basically
want
to
make
this
not
specific
to
a
distribution,
and
then
I
tried
to
answer
some
of
the
other
projects
as
well
as
far
as
like
how
could
it
ever
move
from
sandbox
to
graduated,
and
so
we
don't
really
see
it.
Modernization
and
migration
is
like
a
one-time
thing,
but
as
an
ongoing
process.
That
will
probably
continue
to
happen.
D
So
we
expect
this.
We
expect
the
project
actually
long-lived
inside
of
people's.
You
know
clusters
and
then,
and
then
I
and
then
I
include
some
github
issues
from
from
outside
that
you
can
see
outsiders.
You
know
from
various
organizations
opening
this,
like,
I
think
digitalocean
was
one
of
them,
there's
a
couple
of
others.
Basically,
so
that's
kind
of
you
know
I.
D
A
Not
yet
to
be
fair,
I
think
we
can.
I
thank
thanks
for
posting
it
in
here.
I
think
we
can
follow
up
on
this
one.
Is
there
any
example
how
I
can
take
like
an
existing
service
using
conveyor
and
deploy
it
to
kubernetes?
I'm
just
wondering
anything.
We
should
give
back
to
the
toc
as
an
example
or
recording
of
how
this
actually
works.
In
practice
I
mean
given
what
the
project
does.
It
might
be
a
bit
hard.
Maybe,
but
maybe
there
is
something.
D
Yeah,
actually
right
in
the
middle
of
that
of
the
second
question
that
I
answered
on
there,
it
says:
there's
a
link.
It
says
this
end
to
end
demonstration
and
I'll
paste
it
in
the
chat.
Here
too,
you
might
don't
open
it
now,
because
it'll
start
playing
on
youtube,
but
but
right
in
the
middle
of
this
end-to-end
demonstration
up
a
little
bit
right
in
the
center
of
your
screen.
D
That's
like
a
45-minute
demo
that
we
do
where
we
kind
of
take
an
end
to
your
application.
We
break
it
down.
We
move
from
like
some
spring
boot
apps
over
from
cloud
foundry
into
kubernetes,
show
how
we
could
do
that.
You
know
show
how
we
could
re-host
the
vm
into
cube.
We
show
how
we,
you
know,
assess
and
analyze.
C
I
just
if
I
can
answer
ask
a
question
whilst
we're
looking
at
this,
I
think
one
of
the
things
that
is
most
prevalent,
in
my
mind
is
who
are
going
to
be
the
main
drivers.
C
Especially
like
move
to
cube
like
who
are
the
contributors
be?
Who
do
you
see
that,
as
will
it
be
folks
who
are
end
users
who
have
accomplished
and
want
to
add
features,
or
will
it
be
companies
who
are
trying
to
influence
migration
to
their
platform?
Do
you
want
to
maybe
shed
some
light
on
what
your
thoughts
are,
because
I
think
that
will
be
a
really
relevant
point
for
the
toc.
D
Yeah
yeah,
so
I
I
think
there
are.
There
are
end
users
that
will
contribute.
We
actually
have
one
financial
service
customer
who
who
is
using
not
moved
to
cube
but
they're,
actually
using
a
portion
of
the
tackle
project
inside
a
conveyor
to
migrate,
a
bunch
of
tomcat
applications
and
basically
build
like
use
it
to
build
rules
to
ver
validate
things
are
done
correctly
and
match
the
12-factor
apps.
So
we
do
have,
I
think
end
users
is
one,
but
I
think
another
one
is
actually
most
likely
global
system
integrators.
D
I
think
global
system
integrators
are
all
have
their
various
kind
of
migration,
factories
and
methodologies,
but
the
reality
is
they're.
Not
you
know
they
don't
there's
no
real
differentiation
and
it
takes
them
a
lot
of
time
to
keep
all
those
tools
and
that
tool
chain
up
to
date.
So
the
idea
would
be
that
they
can
leverage
these
as
a
more
common
tool
chain
as
well.
So
I
think
there's
there's
a
global
system
integrator
space,
and
then
we
would
love
to
have
more
other
vendors
and
platform
vendors
involved
in
it
too.
D
We've
had
the
we've
had
the
team
from
vmware
show
us
some
of
their
tooling,
but
as
far
as
I
understand
is
proprietary,
proprietary
and
licensed.
So
at
this
point
I
don't
think
they.
You
know
they
were
interested
in
kind
of.
A
Shame
with
them
can
you
please
mute
yourself?
Well,
not
speaking,
because
we
hear
you
coughing
here
to
not
interrupt
james
thanks.
A
The
other
thing,
thanks
for
sharing
the
the
details,
I
think
giving
us
some
time
to
follow
up
on
lego,
king
of
30
examples,
but
right
now
we
have
no
gsis
contributing.
So
this
is
something
where
you
would
want
to
go.
D
Yeah,
so
we
have
so
so
we
have
a
lot
of
people
in
the
community
that
are
like,
you
know,
have
joined
the
the
mailing
lists
and
all
those
sorts
of
things
that
are
from
with
pro
hcl
accenture,
so
on
and
so
forth.
We
don't
have.
As
far
as
I
know,
we
don't
have
any
pull
requests
from
any
of
them,
but
we
do
know
they're
you.
We
do
know
that
some
of
them
have
been
using
the
tools
just
based
on
their
questions
in
chat
and
other
things
as
well.
A
Yeah,
I
might
have
checked
this
myself,
but
I'm
asking
here
because
we
already
have
you
here:
you
have
an
adopters
and
defile
where
you're
listing
who's
using
it.
Maybe
and
some
of
them
are
mentioned
in
there.
D
Well,
I
you
could
see
actually
in
the
in
the
questions.
The
last
question,
where
I
sent
you
the
I
pasted
a
bunch
of
github.
You
know
issues
that
were
that
were
opened
by
folks
and
if
you
go
look
at
who
they
are,
you
can
see
you
know
like.
Like
I
mentioned
some
of
them.
I
think
one
of
them's
from
digitalocean
there's
a
couple
anyway,
you
could,
you
could
see
the
ones
that
are
open.
Those
are
all
open
issues
by
external
users.
C
I
think
what
is
challenging
is
that
there
are
it's
an
umbrella
of
projects.
You
know
under
a
single
moniker.
C
I
think
that's
the
thing
that
takes
some
getting
used
to,
because
you
know
you've
got
tackle,
forklift,
crane,
etc,
and
each
one
on
its
own
needs
to
have
a
community
base
and
and
pull
requests
in
the
current
format
of
the
toc.
C
It
might
be
worth
considering
bringing
them
into.
I
guess
the
the
conveyor
umbrella.
You
know
in
a
slightly
more
tight
sense
I
to
get
the
impression
that
they
distributed
components
at
the
moment
right.
C
Yep,
okay,
yeah,
I
think
it's.
I
think
that
the
apprehension
here
is
and
we're
just
trying
to
understand
it.
It's
more
of
a
human
thing.
It's
just
that
there
are
a
few
different
moving
parts
that
we
need
to
just
kind
of
figure
out
each
one
and
it's
a
bit
sorry,
it's
appeal
to
the
community
and
how
we
see
it
integrating
into
the
existing
ecosystem.
A
And
also,
what's
sometimes
challenging
with
like
these
umbrella
projects
and
we
had
it
with
other
projects
as
well
like
when,
when
arco
had
their
submission,
for
example,
they
have
a
lot
of
different
projects
in
inside
of
argo
as
well,
how
it's
growing
and
how
you
start
new
projects
in
there,
and
I
mean
for
sandbox.
It
doesn't
really
matter
that
much
honestly,
but
assume
you
move
later
on
to
incubation
and
graduated,
and
then
you
start
something
new
like
how
do
you
decide
whether
you
add
another
tool
in
there?
A
So
you
might
have
a
very
early
stage
to
already
sandbox,
but
then
it's
part
of
an
incubated
or
graduated
project
and
there's
no
right
answer
there.
But
that's
always
the
challenge
like
with
this
tool
and
collection
of
type
of
projects
that
sometimes
to
also
to
access
point,
make
them
a
bit
hard
to
grasp.
A
Like
also,
I
know
for
sandbox,
you
are
obviously
getting
a
lot
of
questions,
but
this
would
be
a
question
like
an
incubation
that
you
would
definitely
get
like
how
to
ensure
that
we're
not
just
adding
like
15
other
tools
at
a
very
early
stage.
How
would
they
use
understand?
A
D
Yeah,
I
know
that's
it's
it's
a
good
point.
We
do
have
like
a
governance
model.
You
know
on
over
on
github
that
you
know
we
haven't.
You
know,
just
being
that.
D
So
far
we
haven't
had
to
have
a
lot
of
those
conversations,
but
there
are
areas
like,
for
example,
with
move
to
cube
a
lot
of
the
underlying
technology,
actually
is
going
to
be
integrated
with
tackle
and
used
by
tackle
to
like
generate
manifests,
and
things
like
that
or
you
know,
and
and
do
some
of
that
work,
because
it
could
already
do
that.
So
I
I
could
foresee
actually,
where
either
a
there's
more
tools
that
you
know
I
would
love
to
see.
D
People
contribute
more
tooling
to
the
community
right
if
there's
things
that
are
valuable,
that
we
can
get
them
in
there
and
help
them.
That
would
be
one
thing,
but
then
I
also
see
where
some
of
these
might
actually
come
together
and
actually
merge
over
time
and
we
might
might
end
up
with
less.
D
I
know
todd's
doing
some
something
similar
with
his
team's
working
on
crane
inside
there
and
they're
doing
something
similar
as
well.
So
it
could
be
that,
in
other
words,
it
could
be
that
we
have
a
far
more
opinionated.
You
know
set
of
tools
right
now.
I
think
we've
just
brought
them
together
under
the
umbrella
and
they're
there,
but
over
time
I
think
the
hope
is
that
it
becomes
more
of
a
platform
and
an
opinionated.
C
Just
just
to
reflect
on
that,
I
think
that's
a
really
important
distinction
from
a
marketing
perspective.
I
think
that
that
could
really
help
the
project
in
terms
of
growth.
If
tackle
was
kind
of
your
your
foundational
point
and
you
had
modules
and
components
that
were
leveraged
off
of
that,
because
one
of
the
big
problems
understanding
this.
Is
that
not
understanding
it?
Sorry,
that's
the
wrong
word.
I
think
one
of
one
of
the
challenges
is
that
this
project
requires
state.
C
You
know
you
require
to
be
in
a
position
where
you
need
to
migrate,
whereas
there
aren't
a
lot
of
cncf
projects
that
I'm
aware
of,
if
any
that
I
can
think
of
that,
require
you
to
already
be
in
a
state
where
you
require
to
move
it
to
a
jvm
based
application
of
cloud
foundry
so,
and
that's
not
to
say
that
this
isn't
an
awesome
idea.
I'm
just
trying
to
say
that
this
is
a
particularly
challenging
concept
and
maybe
reducing
the
amount
of
moving
parts
or
to
at
least
simplify.
C
D
Yeah,
no,
that's
fair.
I
think
that
I
mean.
That
is
definitely,
I
think
the
aim,
but
that's
the
hope
of
when
we're
in
sandbox.
Those
are
the
things
I
think
we
want
to
figure
out,
but
we
don't
want
to
just
figure
them
out
on
our
own
either.
So
it's
like.
We
think
that
the
sandbox
designation
can
actually
help
us
get
more
people
engaged.
You
know
it's
kind
of
a
chicken
and
egg
problem
in
that
in
some
regard,
right,
like
you
know,
there's
I
think
sometimes
there's
we.
D
C
Yeah
and
that's
why
we're
having
this
conversation?
Because
it's
really
useful
for
us,
I
think,
to
understand
like
what
your
aspirations
are
for
it,
because
as
soon
as
we
can
kind
of
figure
that
out
and
we
can
come
to
an
alignment
and
say
yeah,
actually
that
that
makes
a
whole
lot
of
sense.
So
this
is.
C
D
Yeah
I
mean
like
the
so
the
the
financial
services
customer.
I
mentioned
one
of
the
reasons
they
they
got
excited
about.
This
was
they
would
they
have
a
certain
global
system,
integrator
they're
working
with
today,
if
they
bring
that
global
system
integrator
in
to
modernize
their
apps,
they
also
get
this
giant
tool,
set,
that's
specific
to
that
global
system,
integrator,
right
and
and
half
of
it's
proprietary
and
licensed
than
whatever
else.
D
If
they,
if
they,
if
that
global
system
integrator
uses
the
conveyor,
tooling
and
gets,
and
does
this
now
suddenly,
all
of
that
is
open
source,
but
not
only
that
it's
the
skill
sets
are
presumably
more
broadly
transferable
to
another
global
system,
integrator
or
another
person,
and
so
suddenly,
like
you
know,
we
anyway,
I
think
that's
really.
The
value
of
this
is
how
we
can
kind
of
make
this
an
open.
D
D
Are
there
any
other
questions,
or
I
don't
know,
I
don't
know
what
the
right
next
step
is.
If
you
guys
see
it.
A
D
A
It
is
sandbox,
I
think,
once
we
have
like
all
everything
in
place,
it
should
move
forward.
Then
I'd
assume
rather
quickly,
so
it
shouldn't
hold
you
back
too
much.
It's
it's
different
for
an
incubation
project.
It
has
a
full
diligence
and
everything
but
yeah
thanks
for
preparing
material.
That's
that's
definitely
helpful
and
I
think,
in
a
project,
that's
not
directly
fitting
into
what
other
projects
are
doing.
We
often
also
need
some
time
to
know.
A
E
D
D
A
Okay,
so
looking
at
the
time,
I
would
move
over
to
josh,
hey
actually
josh.
You
have
two
two
agenda
items
today
that
makes
you
really
famous
today.
B
The
first
one
is
what
I
what
I
was
thinking
about,
sharing
with
you
all
today
and
then
I
realized
I
was
looking
at
last
week
last
time's
notes
I
was
like
oh,
I
said
I
would
walk
through
potato
head,
so
I
before
this
meeting
briefly,
you
know
made
sure
that
my
environment
works
and
I
can
demo
that
too
we'll
get
that
in
a
second.
So.
D
B
Thing
on
patterns
so
a
little
background
in
the
coordinate,
cooperative
delivery
working
or
whatever
it's
called
coordinated
delivery.
One
thing
we
really
want
to
do
is
we
want
to
kind
of
start
from
the
ground
up
and
and
build
a
collection
of
patterns
that
not
only
help
others
but
help
us
understand
the
you
know
the
principles
that
are
going
on
in
the
landscape
and
as
I
was,
I
proposed
it
to
our
group
a
couple
weeks
ago
and
then,
as
I
was
doing
it,
we
kind
of
realized.
B
This
is
actually
a
little
bit
broader,
and
so
that's
why
I
wanted
to
bring
it
up
with
a
whole
tag,
this
concept
and
you
you're
already
do
I
mean
we're
already
dancing
around.
It
was
potato
head,
it's
just
kind
of
formalizing.
Well,
I
don't
know
kind
of
it
is
formalizing
that
and
adding
a
little
bit.
The
idea
is
basically
hey
we're
gonna.
Let's,
let's
have
an
initiative
to
publish
and
maintain
a
library
of
recommended
practices
recommended
patterns
with
complete
docs.
B
You
know
in
a
standard
template
and
runnable
code
and
I'll
even
add
testable
code,
meaning
it
should
be
able
to
be
tested
in
ci.
So
that
if
people
want
to
change
it,
you
know
they
can
with
relative
confidence
yeah,
so
that
that
issue
is
basically
hey.
What
do
you
think
of
this,
and
should
we,
you
know,
officially
try
to
build
such
a
library
of
patterns
and
you
know
start
start
to
kind
of
formalize
what
that
process
would
look
like.
B
So
that's
that
one
thing
I'll
mention
on
on
that
also
is
that
the
observability
group,
open
telemetry
group,
is
doing
the
same
they're.
Looking
into
the
same
thing,
I
actually
was
going
through
all
the
cncf
channels
yesterday,
and
I
found
that
like
last
week,
they
had
a
bunch
of
meetings
about
this.
They
referenced
potato
head
and
google's
microservices
demo
and
a
couple
other
ones,
and
they
want
to
build
basically
a
playground
and
they're
already
also
kind
of
looking
into
this
in
a
library
of
patterns.
B
So
I
basically
I
wrote
in
their
slack,
you
can
see
and
like
hey,
if
you
guys
don't
use
potato
ad,
then
that
would
be
that'd
be
great
or
or
whatever
it
is,
but
they're.
The
relevance
they
hear
is
that
they're
also
looking
to
this
library
of
past
concept,.
C
Would
be
really
interesting
to
get
I
mean
I'll
talk
to
bartek
actually,
but
it'd
be
really
interesting
to
get
collaboration
with
tag
observability,
because
they
will
have
an
opinion
on
a
lot
of
this
stuff
and,
for
example,
introducing
open
tracing
that
will.
That
will
be
very
opinionated
on
whether
or
not
you
want
to
have
a
sidecar
running
or
not.
You
know
there
needs
to
be
some
some
conversation
and
there's
not
a
one-size-fits-all
approach.
C
So
I
actually
think
that
I
think
almost
the
number
one
goal
should
be
to
collectively
aggregate
these
different
sort
of
research
projects
into
library
patterns
and
try
and
say:
hey,
we'll,
we'll
app
delivery
will
orchestrate
and
you
know
we
will
we'll
help
you
to
build
out
your
goals.
But
let's
do
it
at
least
in
a
single
place
in
a
cncf
project.
That's
already
established,
let's
not
reinvent
the
wheel,
so
I
always
think
that
should
be
the
highest
priority
at
this
point
in
time
is
get
those
collaborators
in
early
on.
B
Great
so
to
the
degree
that
any
of
you
can
influence
them
too.
I
mean
I
wrote
in
there.
I
know
scott
rigby
from
the
get
ops
group
has
met
with
them.
He
didn't.
I
don't
think
the
get
ups
group
really
has
this
thing,
but
he
but
but
he
has
a
connection
to
them.
Also
so
they're
yeah
they're,
already
kind
of
aware.
B
Yeah,
I
think
that
they
talked
with
chris
anishik
and
he
said,
learn
from
what
the
open
get
ops
project
did,
which
is
good.
The
governance
is
good.
The
what
I
can
see,
I
don't
think
opengl
actually
has
any
code
yet
so
that
might
that
might
be
where
we
partner,
like
the
actual
content,
like
they've,
built
this
kind
of
framework,
we
can
shift
potato
head
or
microservices
them
whatever.
It
is,
what
was
microsoft,
pet
shop
or
something
music
store?
C
So
it's
just
you've
got
to
walk
a
line
there
very
tentatively,
that
the
reason
that
you're
building
these
template
libraries
is
to
capture
the
current
practice
practices
not
to
set
them.
C
Is
another
contentious
point:
it's
not
that
it's
best
practice
it's
that
it
is
a
practice
so
like
using
customized,
helm,
etc,
etc.
These
different
tools
and
platforms.
We
can
just
create
a
collection
of
examples,
but
we
shouldn't
put
a
recommendation
behind
them
in
terms
of
what
is
the
best
practice.
B
Okay,
I
think
that
might
I
think
that
might
pop
up
like-
maybe
not
so
much
in
like
a
list
of
tools
like
argo
flux
and
captain,
but
like
when
we're
talking
about
integrating
a
bunch
together
like
with
cooperative
delivery.
Like
I,
I
actually
put
up
a
proposal
because
the
best
way
I
so
this
167
is
followed
by
168,
which
I
didn't
bring
up
here.
B
But
it's
an
actual
proposal
for
something
that
I
would
do
for
cooperative
delivery
and
the
proposal
is
about
deploying
generically
deploying
databases
using
crossplane
and
then
generically
getting
a
binding
using
service,
binding
operator
or
vault
or
whatever,
containing
a
secret
store
csi.
Some
form
of
that.
So
like
we're
kind
of
trying
to
develop
some
ideas,
not
necessarily
to
say
this
is
the
only
way
but
like
here's
some
ways.
Let's
start
the
conversation
about
this.
A
Yeah,
I
I
think
the
alex
is
pointing
wording
is
very
important
here
because,
as
soon
as
we
say,
this
is
a
best
practice
of
how
to,
for
example,
deploy
your
database
on
kubernetes
with
cross
plane
and
people
not
using
cloud
cross.
Planes
or
other
approaches
would
come
after
us
say,
and
they
will
always
always
bring
up
the
non
king
making
rule.
A
A
A
A
Area,
it
should
be
must
be
testable.
We
have
must
have
a
way
to
actively
update
that
it
gets
updated
properly
by
people
that
they
are
willing
to.
Also
then
take
the
time
to
engage
with
with
folks.
So
there
must
be
from
the
people
who
are
participating
must
be
willing
to
have
some
sort
of
time
investment
into
this,
because
nothing
worse
than
working
on
a
library
of
tutorials.
I
think
everybo,
every
one
of
us
knows,
and
then
you
figure
that
the
tutorial
is
outdated.
A
A
The
tools
that
were
used
in
there
are
outdated.
So
I
think
somebody
like
who's
actively
owning
it
and
for
us
having
a
means
saying:
okay,
this
is
still
valid
today.
I
think
it's
super
important,
but
but
maybe
that
anger
of
tutorial
might
be
better
just
proposing
it
here
to
for
all
of.
C
You,
let's
add
something
to
that:
I'm
just
I've
got
the
goals
and
non-goals
up.
I
think
you're
pretty
close
to
the
mark.
I
think,
if
my
advice
would
be,
is
go
ahead
and
do
it
like
create
a
bunch
of
different
patterns
for
deployment.
You
know
you
can
even
have
a
github's
pattern
in
there
like
we
do.
We've
got
some
argo
content.
C
Do
a
bunch
of
that
stuff.
That's
fine,
but
I
would
repurpose
add
a
preface
on
the
rubric
of
the
potato
head
repository
saying
this
is
a
collection,
a
creation
of
relevant
industry
patterns
that
that
are
being
used.
That
is
what
I
would
say
and
quite
specifically
that
and
then
the
real
work
isn't
that
the
real
work
is
the
day
two
exercise
of
looking
at
the
patterns
that
we
see
surfacing
time
and
time
again.
Okay
with
argo
and
flux
and
non-gitop
systems
such
as
you
know,
more
of
the
imperative
command
imperative
kind
of
pipelines.
C
They
all
have
to
accomplish
secret
management.
Somehow,
what
are
the
most
popular
ways-
and
that
actually
is
where
it
comes
to
this
idea
of
focused
questions,
and
so,
if
we
look
at
all
these
things,
we're
going
to
start
to
get
to
the
real
meat
of
what
this
working
group
about,
and
it's
looking
for
the
patterns
in
the
industry
and
then
producing
the
white
paper.
C
So
I
think
go
ahead
with
what
you're
doing,
but
I
think
it's
really
important
to
recognize
the
fact
that
it'll
be
part
of
the
journey
and
the
rest
of
is
is
going
to
be
then
reflecting
on
what
is
the
current
state
of
it.
So
I
would
expect
us
to
be
starting
a
white
paper
on
this
in
the
next
six
months.
Looking
at
what
we're
seeing.
B
C
Absolutely
and
I
think
just
to
meet
you
halfway
in
the
middle,
they
should
also
be
like
exemplars
of
of
what
works,
not
necessarily
what's
best
so
things
that
are
relevant.
They
they
are
functional,
you
can
just
download
it
and
it
works,
and
there
is
a
level
of
responsibility
of
the
curators
and
the
and
the
maintainers
to
keep
them
functional.
B
Definitely
definitely
I
yeah,
hopefully
I
I'll
just
add.
Hopefully
the
framework
is
high
level,
but
hopefully,
as
like,
I
actually,
you
know
what
let
me
jump
into
the
potato
head
demo
and
then
you'll
see
kind
of
what
I'm
thinking
so
that
that's
the
next
one
I
just
I'm
gonna
do
the
top
two
here.
So
there's
two
major
updates
that
potato
had
oh
before
I
jump
to
that
is
that
does
anybody
have
anything
else
to
say
on
the
patterns,
collection,
stuff.
C
It
would
be
nice
to
get
kind
of
a
check-in
of
what
we've
got.
You
know
like
number
like
we've
got,
15
different
types
of
patterns
and
they
are
on
a
scale
of
you,
know
everything
from
configuration
management
to
actual
deployment,
because
you
know
there's
quite
a
lot
of
stuff
in
there
right.
C
So
maybe
maybe
in
the
next
couple
of
tag
meetings
you
know
in
a
couple
weeks,
maybe
after
the
holidays,
we
get
sort
of
an
assessment
of
where
we
are.
D
B
I
just
I
did
you
know
I
did
this
for
microsoft,
with
their
sdk
for
go
and
we
built
this
library
of
runnable
tests
and
things
like
so
it
just
it's
like
a
natural
kind
of
thing.
I
just
I
don't
know,
I'm
surprised,
we
don't
have
it
and
I
well,
we
do
have
it.
I'm
surprised,
it's
not
more
of
a
thing
with
cncf
to
be
honest,
like
let's,
let's
get
vulnerable
examples.
B
Okay,
so
let
me
walk
you
through
two
things
here,
so
I
have
two
two
themes.
I
wanna
share
potato
head:
the
adoption
of
multi-service
in
cube,
cuddle,
helm
and
customize
the
addition
of
tests
runnable
tests
for
those
three
I'm
not
gonna
go.
I
did
also
do
flux
and
catch
so
far,
but
I
won't
do
those
we
could
do
those
another
time
and
then,
if
we
have
time
and
you're
interested,
I
can
show
you
the
refactor
and
that
now
has
end-to-end
tests.
B
It
actually
spins
up
a
cluster
and
things
like
that
too.
So
I
guess
at
its
simplest,
the
idea
here
cube
cuddle.
I
actually
am
using
a
template
now
for
these
readmes.
I
didn't
check
it
in
yet,
but
I'm
thinking
I
would
eventually
they
all
kind
of
say,
deliver
with
whatever.
B
Then
they
say:
here's
how
to
deliver.
Here's,
how
to
test
here's,
how
to
test
the
api
endpoint,
so
they
kind
of
walk
you
through
it
and
then
there's
a
test
file
which
you
have
the
option
of
running.
If
you
have
an
up
cluster
already
so
like
here's
potato
head,
I
can
delivery
cube
cuddle
test.
So
I
have
a
running
cluster
here
and
this
will
just
go
ahead
and
spin
that
up.
B
So
yeah
for
cube
cuddle,
it's
very
straightforward.
I
mean
it's
yeah,
it's
just
manifest
in
order
to
make
the
test
work,
you
have
to
substitute
in
another
image
name,
but
it
comes
up
and
then
you
can
call
this
endpoint.
If
you
wanted
it,
you
see
all
the
containers
are
running,
I'm
not
going
to
call
the
api
there
the
same
thing
I'm
going
to
go
ahead
and
just
delete
that
whole
name,
space
oops
and
it
works
kind
of
the
same.
B
I
should
we
should
look
at
the
chart
because
that's
where
things
start
to
change
with
the
with
the
multi-service
I'll,
just
quickly
show
you
that
so
I
mean
we
could
have
done
this
differently.
B
I
basically
broke
it
into
six
deployments
and
six
services,
and
then
I
have
some
common
properties
that
are
in
the
values,
the
ammo
and
then
again
test
sh,
we'll
just
deploy
that
out,
and
I
could
run
it
for
you
here
and
it
would
delivery
chart
test
oops
and
it
will,
you
know,
do
the
same
thing
and
go
through
and
install
it
and
then
finally,
customize
is
the
is
the
other.
B
I
consider
these
kind
of
the
bases
for
everything
else,
so
that's
that's
why
I'm
walking
through
them
now
here,
I
kind
of
stripped
it
down
to
the
most
basic,
so
we
have
a
base
which
you
could
deploy
directly
and
we
have
an
overlay
which
represents
kind
of
a
production
deployment
and
then
the
test
again,
just
basically
substitutes
in
you
know,
testable
container
images
and
the
same
thing
here.
You
could
run
delivery
customize
test.
B
Yeah,
so
all
those
those
kind
of
come
together,
oh
yeah,
and
then
I
was
gonna-
show
you
around
this.
B
The
idea
of
these
tests
is
to
ultimately
put
them
into
whatever
kind
of
ci
cd
system
we
have
so
like
right
now
and-
and
this
you
know,
there's
work
to
maybe
refine
this,
but
there's
test
cube
cuddle
and
it
calls
test
sh,
there's
test,
helm,
test,
sh,
here's
customize,
here's
catch
and
here's
flux,
so
kind
of
you
know
this
is
I'm
just
kind
of
putting
this
pattern
in
as
I
go
like
a
readme,
that's
following
a
certain
template,
a
test
which
spins
up
and
in
the
case
of
customize,
I
even
had
to
delete
everything
and
then
putting
that
test
into
the
github
workflow,
so
that
when
people
make
changes,
you
know
recommended
changes
in
the
future.
C
C
So
are
these
actually
performing
any
tests
or
they're
just
called
tests?
Shell.
B
Yeah
sure,
like,
let
me
show
you
just
the
basic
cube
cuddle
one
you'll
see,
let's
see
deployment
blah
blah
blah.
B
And
then
it'll
wait
see
this
is
already
met
because
I
actually
had
it
up,
but
it
was
waiting.
It
was
a
cute
cuddle,
wait
this
one's
only
checking
for
potato
main
but
like
it
was
waiting
for
yeah.
So
in
other
words,
the
other
images
could
have
failed
here
and
you
wouldn't
have
known,
but
you
could
submit
a
pr
for
that.
But
here.
E
B
And
actually,
if
you
walk,
if
you
go
through,
so
it
doesn't
call
well,
if
you
go
through
the
instructions,
I
could
actually
walk
you
through
curling
here
this
api
endpoint,
how
you
could
do
that
it
doesn't
do
that
here,
but
I
will
show
you
if
you're
interested,
I
can
keep
going
the
potato
head.
Refactor
has
been
the
next
project
I'm
working
on,
and
here
it
does
include
a
complete
end-to-end
test,
so
it
tests
the
api
endpoints.
I.
C
Mean
I
I
I
would
love
to
I'm
just
conscious
of
anything
else
on
the
agenda,
but
I'd
love
to
go
through
this.
I
I
really
like
you
put
it
into
a
single
project,
because
I
see
that
we've
got
three
separate
folders
for
the
application
components
at
the
moment,
so
you
actually
have
you
put
them
into
a
single
package.
There.
B
Oh
yeah,
this
refactor.
If
I
walk
into
the
refactor
yeah,
I
put
them,
I
I've
kind
of
unified
everything
into
one
still
kept
them
under
a
subpac
subdirectory.
I
talked
that
over
with
thomas.
B
And
then
I
well-
and
I
should
show
you-
I
pretty
significantly
refactor
the
internals-
to
make
maine
a
little
more
just
generic
and
then
put
more
stuff
into
these
things
into
into
packages
and
to
share
packages
so
like,
and
this
is
actually
going
to
be
pretty
helpful.
I'll
give
you
an
example
here.
The
service
discoverer,
which
exists
here
now
see
if
I
can
find
it
is
this.
It
yeah
looks
something
like
this.
B
I've
actually
turned
it
into
an
interface,
and
you
know
there's
right
now:
a
static
service
discoverer,
which
is
basically
what
we
had
before,
but
this
provides
a
great
opportunity
for
cooperative
delivery
to
demonstrate
injection
of
service
discovery.
You
know
dynamic
service
discovery,
the
service
handler
I
factored
out.
So
this
is
how
the
main
thing
calls
off
to
all
the
parts.
B
So
it's
got
the
handlers
extracted,
here's
the
part
handler
which
would
and
I've.
Oh,
that's
also
a
generous
there's
a
lot
of
stuff
here,
so
I
genericized
the
part,
the
part
command
so
basically
for
any
given
part,
it
uses
the
same
entry.
The
only
thing
is,
it
looks
to
see
if
I
can
find
it
to
know
what
part
it's
supposed
to
serve.
I
think
it.
I
have
to
look
again
desire,
part
image,
name
yeah.
B
It
looks
to
the
to
the
url
that
was
sent
to
it
and,
in
fact,
what
this
guy
serves.
B
Version
part
number
there:
it
is
so
I've
actually
moved
some
of
those
things
into
version,
so
part
number
actually
now
is
dynamically
controlled.
This
is
a
change.
This
is,
this
is
probably
the
biggest
api
change.
Is
that
I
added
this
environment
variable
which
then
says
hey
serve
one
serve
two
series
or
four,
so
it's
dynamic,
as
opposed
to
recompiling
for
different
versions.
B
Yeah
thanks
and
then
this
one
actually.
So
if
I
switch
over
to
my
other
shell-
and
I
do
say
actually
I
put
it
into
make,
so
I
do
a
make
test
services.
This
actually
will
go
through
and
build.
C
C
B
C
Exactly
that
and
also
you
know
just
make
as
powerful
right,
so
you
can
do
a
lot
of
stuff
with
the
flags
and
arguments
and
make
file
that
would
be
a
bit
clunky
to
do
in
the
shell
scripts.
B
Yeah
some
other
interesting
stuff
that
opens
up
like
I'd
like
to
add
a
handler
that
does
crud
operations
over
a
database
for
my
cooperative
delivery
stuff,
because
I
want
to
use
that
as
an
example
so
like
now
with
this
refactoring,
it
won't
be
hard
to
add
another
handler,
that's
kind
of
one
of
the
goals.
B
It
won't
be
hard
to
reason
about
hey.
I
want
to
inject
open
telemetry
into
a
hand.
Oh
okay,
well
I'll,
just
go
over
to
lost
it,
but
where
is
this
I'll
just
go
over
to
this
part
handler
and
add
in,
like
you
know,
open,
telemetry
or
whatever
tracer.startspan,
or
something
or
add
to
spam
or
whatever?
It
is.
B
Yeah
so
here
in
the
background,
I've
got
this.
It
did
work
last
week,
oh,
and
I
should
show
you
actually.
This
is
forget
about
showing
it
to
you
on
my
own
local,
so
it's
gonna
go
into
github.
B
That
these
chest,
these
checks
are
running
and
passing,
and
I
actually
have
there's
actually
two
now
so,
where
can
I
go
to
the
actions.
C
B
Yeah,
exactly
that's
that's
what
I'm
hoping
so
here.
You
can
actually
see
this
random
background,
like
you
asked
before
alex
a
weight
readiness
here.
I
did
await
the
readiness
of
all
of
them
and
that's
what's
happening
on
the
left.
Here
too,
you
can
see
then
we'd
guess
you
know.
I
just
did
a
cube,
cuddle
get
deployments
test
and
then
I
test
the
apis
and
if
they
fail
it
lets
me
know,
and
then
I
just
get
cube
cuddle
logs
for
each
just
for
curiosity's
sake
and
it
all
passes.
B
C
B
B
B
Here
number
120,
so
I
kind
of
said
so.
I
got
a
couple
more
to
add
thomas
had
added
in
a
version
incrementation
facility,
and
I
need
to
add
that
and
and
trivia.
B
Okay,
I
think
that's
all
that
I
was
going
to
say
I
put
in
the
new
tutorials
for
qualcomm
customize.
I
showed
you
that
the
refactored
core,
with
the
end
to
end
test,
showed
you
that
if
we
want,
I
mean
I've
done
catch
and
flux
on.
My
short
list
is
argo
and
captain,
but
I
haven't
gotten
them
yet,
so
I
can
show
you
them.
You
know
some
future
time
you're
getting
the
idea,
though-
and
I
mentioned
this
already
above-
that
it
would
be
nice
to
get
observability
group
also
partying
with
us
here.
B
C
I
could
build
an
example.
I
mean
I'm
cautious
about
suggesting
this,
but
I
could
build
an
example
for
nomad
that's
a
hashicorp
product,
so
I'm
not
entirely
sure
whether
I
want
to
because
I
suppose
we
have
some
event.
We
have
some
vendor
products.
A
Yeah,
I
think
they
are
fine,
so
we
always
we
had
this
also
when
we
were
looking
into
catch
and
others.
Okay,
I
think
it's
fine
to
have
things
in
there
that
people
can
actively
try
so
actually
worry
motive.
We
can
automatically
test
them,
also
like
what
what
just
just
shown,
which
kind
of
means
they
need
to
have
at
least
an
open
capabilities
for
us
to
test
it.
I
mean
if
it
would
be
like
super
proprietary,
we
couldn't
test
it,
then
I
would
not
put
it
in
there.
C
B
B
What
about
all
the
various
ways
you
have
to
to
build
clusters,
which
that
that
your
question
about
nomad
kind
of
reminded
me
of
that
it's
maybe
that's
a
a
little
different,
not
that
it's
not
so
valuable,
but
I
just
wanted
to
call
out
that
this
christian
hernandez
has
been
starting,
something
not
sure
if
that's
that
might
dance
around
that
being
able
to
deploy
kubernetes
in
a
few
different
ways.
I
don't
know
I
haven't
gone
deep
into
it,
but
that's
what
he
said.
C
One
suggestion
is
that
this
is
this
is
quite
a
pivotal
project
for
the
tag
in
the
corporate
delivery
working
group.
Perhaps
we
could
have
a
road
map
meeting
at
some
point
or
catch
up
around
the
roadmap
of
potato
head.
B
C
Here
and
it's
actually
larger
than
the
tag,
it
crosses
the
tags
which
is
nice
just
because
it's
convenient
you
know,
but
it's
also
it's
established
and
there's
a
lot
of
active
participation
on
it.
So,
as
I
said,
it
would
be
incredible
if
we
get
folks
from
chaos,
testing
working
group
and
the
tag
observability
to
participate
on
this
as
well.
E
Hey,
I'm
pretty
good
had
to
come
in
late
today,
but
nice
to
see
you
all.
B
Yeah
well
one
of
the
now
I
don't
think
I
saw
you
before,
but
one
of
the
things
that
I've
mentioned
is
that
tag
observability
seems
to.
I
think
you
talk
to
them
about
a
patterns,
library
or
playground,
thing
that
they're
trying
to
spin
up-
and
I
know
and
I've
been-
and
so
I
mentioned
them
in
their
thread
like
I've
been
pursuing
that
for
app
delivery
and
I'm
kind
of
thinking
like
hey.
You
guys
have
this
whole
scaffolding
already
around
open,
git
ops.
Maybe
this
library
fits
in
into
what
you've
already
created
so.
E
Yep,
exactly
and
and
had
been
talking
with
matt
about
that
for
a
while
too
yeah.
In
short,
in
short,
I'm
happy
to
talk
with
with
them
and
just
kind
of
show.
Okay,
here
are
the
steps
that
we
went
through
specifically
for
opengl,
or
excuse
me
for
get
up
to
work
and
group
and
open
get
ups.
What
we
did
for
governance,
because
that
was
drawn
from
that
was
drawn
from
like
the
findings,
I
don't
know.
I
want
findings,
but
let's
say.
E
Discussions
with
other
with
with
people
in
other
cncf
projects,
several
of
which
I'm
part
of
and
some
that
I
I
haven't
been
to
try
to
understand
what
what's
worked
over
time
and
what
hasn't
as
well
just
so
we
could
kind
of
avoid
some
commonly
known
pitfalls.
E
Really
quick
one
one
small
note
about
that-
is
that
the
there's
a.
E
Oh
I'm
so
sorry,
I
think
it's
tag,
tag,
contributor
strategy
and,
and
then
there's
also
a
maintainer
circle.
That's
meeting
a
little
less
regularly
now,
but
when
there
are
people
that
want
to
connect
with
the
with
with
folks
in
the
maintainers
circle,
that's
kind
of
like
people
who
are
interested
in
this
across
the
different
projects
within
cncf.
E
Usually
sometimes
there
is
an
effort
to
make
time
for
that,
and
and
meetings
will
happen,
and
so
that's
one
of
the
places
I
I
was
able
to
also
get
reviews
on
the
governance
from
folks
there,
which
was
really
nice.
We
didn't
really
change
a
whole
lot
by
that
time,
but
there
were
a
few
things
that
were
mentioned
that
were
were
valuable.
B
C
I
think
one
of
the
things
that
maybe
it's
just
a
personal
aspiration,
but
I
would
really
like
to
try
and
use
potato
head
project
as
a
lightning
rod
for
some
of
the
other
tags
who
are
struggling
to
get
momentum
behind
some
of
the
initiatives
and
tangible
work.
They
want
to
do
it's
a
historic
problem
in
like
tag
observability,
you
know,
there's
a
lot
of
people,
you
know
too
many
chefs
or
too
many
cooks.
They
call
it
right,
but
nobody
actually
wants
to
do
the
work
right.
C
A
A
So
I
think
usually
they
are
reserved
to
companies
doing
them,
but
I
have
to
ask
amy
whether
there's
any
chance
we
potentially
can
do
something
out
of
tech
app
delivery.
Maybe
you
can
always
set
up
a
webinar
as
some
of
the
people
who
we
are
here,
but
we
did
quite
an
official
channel
and
josh
you
just
let
us
know
when
you
think
okay,
the
project
is
not
available
where
we
really
want
to
let
people
jump
on
it
and
maybe
getting
close
to
christmas
season
or
holiday
season.
B
C
A
B
A
Honestly,
I
would
reach
out
to
those
projects.
I
mean
it's
also
some
value
for
those
projects
that
they
can
get
it.
So
it's
like,
I
would
what
I
did
with
regarding
helm.
I
just
reached
out
to
the
home
community.
Like
can
you
take
this?
I
mean
if
you
really
want
to
be
there,
you
should
be
actually
taking
it
yeah.
They
might
have
a
better
recommendation
just
to
ask
them
hey
dude.
A
Can
you
want
to
update
the
example,
because
the
funny
thing
is
that
all
projects
have
demo
apps,
we
take
kind
of
like
the
burden
out
of
maintaining
them
to
some
extent,
but
we
want
to
have
the
examples
in
like
one
central
place,
so
I
think
usually
they
should
be
open
to
this.
A
I
think
the
best
way
to
maybe
get
this
done
is
create
an
issue
for
those
projects.
That's
what
I've
learned
in
the
past.
If
you
create,
if
you
write
on
slack,
they
might
or
might
not
reply,
but
there
are
people.
I
was
looking
at
issues
that
are
coming
in,
like
creating
them
on
both
sides-
cross-linking
them.
A
That
was
what
was
usually
working
best
for
me,
because
then
you
have
people
looking
at
it
and
that's
also
the
way
I
would
handle
it
going
forward
like
if
your
test
cases
break
you
get
a
modification
from
us
that
we
select
this
release
of
your
project
test
cases
break,
and
I
think
that's
how
it
also
took
like
the
day
two
operations,
part
of
alex
that
alex
mentioned
earlier.
A
E
Super
quick
question:
just
on
that
one
note.
I
know
that
was
an
example,
but
but
I'm
also
I'm
a
hell
maintainer
as
well
and
and
also
fairly
interested
in
a
space
where
the
the
limits
of
the
helm,
the
version
of
hell.
You
know
the
current
version
of
helm's
client
really
is
outside
of
the
scope
of
any
sort
of
anything
running
in
the
cluster.
E
E
A
That
was
the
I
think
from
opti
operator
framework,
so
you
can
write
like
a
fully
fledged
operator,
but
the
very
easy
way
on
using
an
operator
with
the
operator
framework,
which
is
the
example.
That's
in
there,
it's
just
more
or
less
use
reusing
a
helm
chart.
It's
not
the
ideal
way
to
be
fair
to
use
an
operator,
easy
packaging
type
of
approach,
and
that's
due
to
the
history
of
the
project,
because
I
just
wanted
to
show
people.
You
can
only
use
an
operator
to
update
something
using
a
crd.
A
E
Yeah
that
makes
sense.
You
know
what
one
of
the
things
that
I'm
planning
to
do
you
know
not
to.
I
don't
think
it's
a
conflict
of
interest
in
any
way.
I
think
it's
it's
very
helpful,
because
I'm
also
a
flux,
maintainer,
the
the
flux
controller,
formerly
in
flux,
one.
E
There
was
a
flux,
helm
operator
that
that
really
managed
the
life
cycle
of
of
helm
releases
and
the
metadata
around
home
repos,
and
things
like
that
and
in
in
in
the
current
version
of
flux,
it's
more
distributed
so
that
there
are
multiple
controllers
rather
than
writing
on
the
operator
framework
it
it's
just
using.
Basically
what
the
operator
framework
is
using
too
at
this
point,
which
is
like
you
know,
coup
builder
on
top
of
controller
runtime.
E
You
know
so
at
this
point,
there's
really
not
a
whole
lot
of
difference,
except
for
some
of
the
life
cycle,
things
and
that's
not
so
so,
basically,
like
the
so
influx
2
the
control.
It's
really
controller,
not
an
operator
at
this
point
not
on
the
operator
framework,
and
I
don't
know
if
that
means
that
there
should
be
no
overlap
there.
E
You
know
or
or
if
you
think,
that
that's
it's
good
to
connect
those,
because
those
projects
are
so
quickly
aligned.
Now
you
know
code
builder
and
and
the
operator
framework
I
don't
know.
B
B
E
Okay,
well,
the
reason
I'm
okay.
Well,
that
makes
sense,
because
the
reason
I'm
mentioning
it
is
that
that,
at
this
point
the
there
were
a
couple
of
controllers
that
that
you
held
sdk.
I
know
rancher
matt
farina
from
ranchers
been
working
on
some
cool
things
there
and
the
the
flux
home
controller
is
makes
very
full
use
of
the
helm
sdk.
E
E
There's
documentation
out
there,
it's
just
it's
just
there
aren't
a
lot
of
examples
really
good
examples.
I
think
cluster
api
is
a
really
good
example
and
we've
been
refactoring
the
flux
controllers
to
better
use
of
of
of
conditions
and
even
wrote
a
package
to
help
other
people
use
that,
whether
they're
doing
git,
ops
or
not
so
that
might
I
don't
know
if
that
is
something
you
want
to
be
related
to
to
the
operator
stuff.
But
if
it
is,
I
could
link
to
that
and
there's
really
good
documentation
around
it.
E
Planning
on
on
on
adding
better
links
to
that
from
the
helm,
documentation
too
in
the
sdk
section.
B
Okay,
maybe
this.
E
E
Yes,
but
I
can't
I
can't
volunteer
to
do
it
right
right,
meow
I
can
I
I
I
can.
I
definitely
will
would
be
interested
in
looking
at
it
and
I
think
that
it's
in
the
long
run
yeah.
I
don't
know
what
the
scope
of
it.
I
don't
know
what
the
scope
of
it
is,
but
I'm
I'm
I'm
happy
to
look.
E
So
if
anyway,
I
agree
with
what
alois
said
about
the
issues
like
adding
an
issue,
there
is
probably
a
good
idea,
but
but
I
just
wanted
to
mention
that
for
the
helm
team
most
likely
unless
there's
something
specific
with
just
exemplifying
how
to
use
the
sdk
most
likely,
the
helm
team
would
just
be
like.
No.
I
don't
want
anything
to
do
with
this,
because
it's
it's
not
nothing
to
do
with
it,
but
but
just
out
of
scope
for
the
design
of
the
helm,
client.
A
No,
no!
No.
We
have
this
discussion.
I
think,
to
josh's
point
we're
currently
looking
for
somebody
from
an
app
delivery
perspective,
who's
taking
care
of
the
operator
example.
So
we
okay.
I
think
we
are
very
pragmatic
here,
and
that
was
also
the
the
approach
that
we
were
taking,
because
the
idea
is
we
help
people
learn
app
delivery
on
the
task
of
of
doing
it
and
while
documentation
is
super
important
and
usually
we
don't
have
enough
people
for
documentation.
A
I
think
right
now,
josh
is
like
the
the
lonely
coder
right
now
so,
together
with
help
from
some
others
who's
trying
to
do
this
and
he's
in
desperate
need
of
some
help
of
people
who
can
help
him
write
some
of
this
or
support
him
on
this
task,
so
that
we
for
the
writer
community
enable
them
to
see
how
they
can
use
it
and
again,
it's
a
very
simple
operator
that
just
deploys
a
couple
of
micro
services,
but
still
it's
work.
That
needs
to
be
done.
I
think
that's
what
we
want
him
to
support.
E
Okay,
then,
in
short,
sorry,
I
guess
I
could
have
summarized
that
a
little
quickly
and
I'll
be
super
fast,
because
then
we
need
to
leave.
But
I
also
know
a
few
people
on
on
the
team
that
I'm
working
on
at
waveworks
who
are
working
with
operators
every
every
day
and
especially
doing
some
of
the
work
for
for
flux
and
get
ups
for
the
operator
hub.
E
B
B
A
Okay,
I
think
then
we
are
already
over
time
and
done
for
today.
I
think
it
was
a
productive
meeting,
thanks
josh
for
sharing
and
really
appreciate
your
work
here.
You
would
definitely
deserve
a
potato
head.
T-Shirt
doesn't
need
to
find
somebody
who's
building
a
christmas
hat,
so
we
need
a
designer
now
on
the
team
as
well.
So
that's.
A
Okay
thanks
everybody
yeah
we're
right
on
the
mailing
list.
We
can
have
another
meeting
in
two
weeks
from
now,
and
it's
going
pretty
close
to
holiday
season
and
we'll
see
how
many
people
will
have
actually
time
to
join
but
being
conscious
about
it.
It's
getting
to
the
end
of
the
year.
Let's
see
how
many
people
are
have
time
to,
or
I
have
other
commitments
and
discuss
this
about
the
mailing
list.