►
Description
Our 10th meeting dedicated to solving issues that arise with building, deploying, maintaining, supporting, and using open source CMSs on Kubernetes covers a short presentation from DDEV’s Nic Hima — exploring a possible paradigm that a CMS community can follow to define a custom resource definition (CRD) for a Kubernetes CMS operator.
Catch up with the group on GitHub: http://bit.ly/338dXC5
Why are we so hooked on K8s? Our story: https://bit.ly/3oOFTFd
A
Typically,
we
start
this
off
with
quick
introductions
and
since
it's
a
relatively
small
group,
those
should
go
pretty
quickly.
So
I'll
kick
it
off:
I'm,
Kevin,
bridges,
I'm,
the
CTO
and
founder
of
a
company
called
d/deaf.
We
basically
make
development
tools
and
complete
and
in
workflow
management
systems
we
do
a
lot
of
hosting
on
kubernetes
and
are
explicitly
focused
in
the
concept
of
advancing
developer
communities.
So
helping
people
understand
how
to
make
the
most
out
of
this
kind
of
technology
and
just
going
around
through
the
people
that
I
see
online.
A
C
D
Yeah
hi
I'm,
Nixon,
Maria
I
work
with
Nick,
Scherr,
previous/next
and
I
guess.
The
majority
of
our
day
is
spent
working
on
our
skipper
hosting
platform
yeah.
So
we
just
launched
that
some
months
ago
and
yeah
I
guess
we're
trying
to
bring
like
a
really
good
developer
experience
to
trim
it
Eddie.
So
yeah
like
a
lot
about
what
goes
into
making
our
CLI
tool
really
abstract
away
a
lot
of
the
pain
of
using
kubernetes
like
if
you
were
to
use
the
tool
like
how
long
or
something
but
there's
a
lot
of
implementation.
A
A
A
E
So
I
was
asked
to
try
to
put
together
lightning
talk
based
on
some
conversations
that
I've
had
I
would
pass
a
few
months
about
trying
to
find
a
path
forward
for
defining
a
communities
native
resource
definition
for
cms's,
so
I
mean
just
go
ahead
and
share
slide
deck,
real,
quick,
all
right,
yeah,
it's
a
little
rushed
I'll,
say.
But
again
my
name
is
nicholas
keen
and
today
I'll
try
to
impart
on
you
a
vision
of
defining
a
crazed,
native
CMS.
E
Before
we
get
into
that,
we
were
asked
me
I
recently,
graduated
UTSA
weights,
a
bachelor's
in
computer
science
and
somewhat
of
a
self-proclaimed
community's
journeyman
and
PHP
freshman,
so
I
get
some
stuff
wrong,
I'm,
sorry
and
also
an
active
communities
operator
developer
at
PETA
and
was
fortunate
enough
to
first
be
exposed
to
the
world
of
CMS
by
working.
Indeed,
of
and
working
closely
with,
you
know
trying
to
build
a
platform
around
CMS's.
E
However,
what
we
seem
to
lack
in
the
CMS
space
is
a
centralized
and
well-defined
easy-to-use
way
to
specify
all
of
these
distributed
resources
in
an
easy
way,
so
tools
to
provide
a
potential
solution
for
this
would
be
custom,
resource
definitions
and
operators
and
controllers
to
be
able
to
actually
implement
the
actions
required
to
get
your
cluster
or
site
in
a
state
that
you
can
host.
You
can
actually
host
traffic
a
little
disclaimer.
E
E
E
So,
as
I
mentioned
before,
create
C,
RDS
and
operators
do
provide
a
potential
solution
to
the
problem
and
do
also
provide
the
potential
to
meet
those
three
goals,
and
this
is
by
utilizing
several
the
existing
communities
paradigms
and
projects
that
are
out
there.
So
the
first
one
customer
resource
definitions.
E
You
know
since
since
that
I
guess
breakthrough
occurred
in
the
Korean
age,
landscape,
you're
no
longer
tied
to
simply
core
resources.
You
can
essentially
define
your
own
custom
resources
and,
on
top
of
that,
define
the
actions
that
are
performed
based
on
the
contents
of
that
resource.
So
that's
a
very
powerful
thing
if
you
think
that
you
know
as
long
as
you
have
a
way
to
define
something
and
it's
a
set
list
of
steps
to
you
know
implement
it,
then
really
it
just
comes
down
to
the
details.
E
The
second
one
is
the
provider
pattern
which,
if
you're
familiar
with
large,
open
source
projects
like
Kay,
ops
or
pops4
terraform,
the
provider
pattern
allows
for
the
ability
to
customize
logic
based
on
whatever
the
provider
of
said
resource
you're.
Trying
to
use
so
I'll
get
a
little
bit
further
into
that
in
the
next
slide
and
following
slides
and
then
the
last
one
is
the
current
application.
H
E
So
the
reason
why
I
chose
this
specific
example
is
because
it's
a
already
you
know
currently
in
use
example
that
provides
definitions,
we'll
a
centralized
definition
of
a
multitude
of
different
aspects
for
a
site,
chiefly
going
to
be
the
like,
where
you
would
get
the
source
of
the
code.
That's
going
to
be
running
on
your
site,
specifying
the
runtime
environment,
such
as
the
container
image.
You
would
like
to
run
the
code
in
which
is
these
two
right
here.
E
Also,
you
specify
the
required
resources.
So
not
only
are
you
specifying,
where
you're
getting
your
code
from,
but
also
where
you're,
storing
the
different
components
of
your
site,
whether
that
would
be
like
in
this
portion
right
here,
the
where
you're,
putting
the
media
files
for
your
site,
or
even
you
know,
defining
where
your
database
is
going
to
be
provisioned
and
hosting,
and
then
lastly,
is
going
to
be
the
runtime
environment
configurations
which
is
going
to
be
on
the
next
slide,
which
allows
you
to
again
inject.
E
You
know
whether
it's
environment,
variables
or
specific
configurations
into
the
site,
and
all
this
really
boils
down
to
providing
you
the
ability
to
run
your
site
either
locally
or
a
cloud
environment,
depending
on
what
you
define
a
site
resource
to
look
like
a
interesting
example
would
be
in
this
media
section
here
they
specify
GCS.
However,
going
back
to
the
provider
pattern,
you
can
also
specify
AWS.
E
If
you
would
like
to
or
AWS
is
whatever
their
block,
storage
implementation
is
or
the
same
thing
with
adders
or
if
you'd
like
to
you
know,
run
it
in
a
local
volume.
Then
you
can
always
define
whatever
local
wanna
run
it.
So
it
allows
for
a
very
high
level
of
flexibility
in
configuration
and
then
going
back
to
the
rental
environments.
Here
you
can
see
that
this
is
for
your
WordPress
configurations
and
then
also
your
runtime
variable
configuration
and
injection.
E
E
D
Was
great
Nick
I,
the
provider
pattern
is
actually
pre
blank
not
knowingly
familiar
to
me,
because
that's
kind
of
how
we're
achieving
portability
between
providers
like
providers
available.
So
you
know
we
have
essentially
like
you,
declare
a
database
and
then
you
like
in
doubt
in
our
drupal
operator
spec,
we
just
like
say
I
want
the
the
prog
database
and
so
that'll
in
reference
to
another
proven
Eddie's
object
called
database
which
the
controller
we
need
apply
it
onto
an
AWS
cluster.
D
That
controller
will
create
an
RDS
instance,
so
you
can
just
swap
out
that
controller
piece
on
different
providers
like
I'm
GCP.
You
can
create
you,
you
know
you,
google,
equivalent
and
yeah,
and
so
other
places
that
could
be
useful
in
your
example
would
be
for
like
edge.
You
know
like
what
I'm
CD
and
you
want
to
spin
up
what
certificate
see
you're
gonna,
deploy
to
that
yeah.
So
I
really
think
that's
it's
a
promising
approach.
E
The
there
is
a
slight
deviation
for
the
decoupling
of
the
environment
that
it's
running
in
so,
for
example,
like
that
operator,
whether
it's
running
on
AWS,
GCP
or
locally,
it
has
the
ability
to
provision
whatever
that
resources,
regardless
of
like
what
environment
it's
currently
running
in
as
long
as
you've
provided
the
correct
credentials
if
it
requires
care
credentials,
and
we
at
least
in
my
experience,
I've
kind
of
had
the
same
approach
you
had
where
you
have
a
separate
controller.
That
kind
of
manages.
D
E
D
Yeah
did
yeah
I
mean
yes,
certainly
as
it's
pretty
it's
ambitious
goal
that
I
have
like
a
CMS,
accreta
or
controller,
but
yeah
I
mean
it's
I,
guess
the
the
struggle
that
we've
had.
This
conversation
has
been
bouncing
around
about
how
to
define
even
a
Drupal,
a
Drupal
CID
and
it
again
like
the
opinions
and
the
you
know
what
you
know.
D
E
And
to
that
point
the
the
cool
thing
about
the
way
that
they
have
their
their
operators
set
up
is
that,
for
example,
if
you
wanted
to
do
a
drupal
version
of
set
operator,
you
know
you'd
still
have
to
define
out
where
you
get
your
code
from
where
you're
you
know
how
many
replicas
you're,
adding.
What
are
your
host
names
that
you
want
for
your
site?
Where
do
you
want
to
store
your
media?
E
Your
database,
all
that
kind
of
stuff
and
the
thing
that
distinguishes
like
the
drupal
operator
from
the
wordpress
operator,
is
going
to
be
kind
of
like
the
quote:
unquote
secret
sauce.
That
itself
distinguishes
WordPress
from
group,
so
there's
me,
I
would
see
that
there
would
be
a
lot
of
common
logic
between
the
two
operators.
As
far
as
you
know,
managing
and
provisioning
all
of
the
resources
required
for
the
sites
and
the
differentiation
and
logic
is
going
to
come
through
and
be
more
CMS,
specific
and
configurations
and
management.
A
A
You
know
with
some
basic
questions
to
kind
of
fill
in
a
little
bit
and
the
first
one
that
we
have
up
is:
has
anyone
have
any
recent
challenges
or
celebrations
that
they'd
like
to
share
what
kind
of
big
accomplishments
have
you
seen
lately
anything
that
we
can
help
give
feedback
on
it
sounds
like
skipper
was:
is
gaining
steam,
so
that
might
be
one
to
talk
about,
but
let's
go
ahead
and
kick
that
off.
How
about
you,
Nick
hema,
since
you
were
speaking
sure.
E
Not
CMS
specific
I
have
seen
in
the
past
month
or
so
two
or
three
more
open-source
operators
come
out
that
utilize.
The
application
CRT,
which
again
I,
think
that
a
CMS
operator
would
be
a
great
use
case
for
something
like
the
application
CRT
simply
because
there
are
so
many
different
resources
or
underlying
core
resources
that
a
CMS
site
needs
to
kind
of
couple
together
and
the
application
CRT
does
a
fantastic
job
of
doing
that.
Yeah
cool.
D
Lastly,
I
was
able
to
go
to
the
kubernetes
forum
in
Sydney,
so
that
was
like
a
sort
of
mini
coop
con
and
yeah.
That
was
quite
a
opening
because
you
know
in
our
little
a
little
drupal
kubernetes
ecosystem.
You
know
like
kind
of
aware
of
what
everyone's
doing,
but
when
you
look
at
the
broader
scope
of
use
cases
yeah,
it's
pretty
clear
that,
like
that
helm-
and
you
know
open
policy
agent
and
the
stuff
in
that
realm
is
kind
of
where
everything's
going.
What
big
organizations
are
deploying
out
so
yeah.
A
C
Yeah
so
I
for
me
this
week
recently
it's
been
around
I
kind
of
alluded
to
it
before
it
was
about
trying
to
make
applications
containments,
and
that
is
logging.
So
I
have
been
down
the
rabbit
hole
on
logging
and
trying
to
provide
a
way
for
like
in
the
past
in
Drupal.
You
just
enable
syslog
module
when
you're
done
so
and
that
like
without
going
too
deep,
is
all
about
into
this.
C
Like
lovely
socket,
that's
on
that
Jeanne's
provided
by
the
system
of
Damon
fooled
slash,
dev,
slash
blog,
but
you
don't
have
that
in
the
containerized
world
and
slash
dead
is
very
he's
very
well
used
like
it's
a
system
path,
so
you
can't
exactly
just
Maps
a
storage
over
and
provide
a
something
to
say.
I
have
been
going
very,
very
deep,
I,
actually
hatched
my
muscle
for
Alpine,
like
the
Lipsy,
so
I
rewrite
it
anyway.
Basically,
I'm
on
this
journey,
where
my
end
goal
is
I
want
to
be
able
to
make
it
cylindrical
applications.
C
Just
it
enables
this
love
module
and
that's
it
and
all
that
kind
of
stuff,
especially
when
you're
running
at
scale
lots
and
lots
and
lots
of
science.
That's
a
that's
a
pretty
big
burden
so
see
our
ultimate
goal
enables
this
mode
module
and
have
some
kind
of
psychic
are
sitting
there.
So
yeah
it's
a
bit
of
a
deep
one.
It's
a
little
bit
of
a
struggle,
but
but
I
once
I
have
something
I'll
definitely
be
sharing.
It
I
think
it's
one
of
those
patterns
that
that
we
could
all
use
for
our
applications.
C
A
F
A
H
It's
Kim
here:
okay,
jumpy,
yeah,
I,
think
one
of
the
challenges
is
definitely
trying
to
keep
up
with
the
release.
Okay
looks
like
it's.
This
seems
to
be
my
seventh
six
months
right,
so
there's
always
new
features,
and
you
know
evaluating
things
I
think
in
terms
of
like
the
stability
like
I,
think.
All
of
the
features
that
we
need
now
are
pretty
stable,
but
yeah,
just
just
it's
it's
a
fast-moving
project
and
knowing
you
know
when
to
sort
of
take
advantages
of
new
features
and
when
to
wait.
H
H
Is
there
and
I'm
looking
forward
to
you
know
seeing
what
Minneapolis
is
going
to
be
like
in
terms
of
ramping
up
discussions
about
this
I
mean
it
was
never
really
a
topic
of
discussion
hosting
generally
like
there's,
obviously
I
was
think
about,
is
if
your
pecans,
a
venture
in
the
past
yeah,
but
I
think
this
is
kind
of
a
great
forum
to
talk
about
open
source
hosting
platforms.
You
know
those
kinds
of
things
and
how
it
could
be.
A
Well,
in
Amsterdam,
we
had
a
really
fantastic
panel
that
included
Michael
Schmidt,
facilitated
florian,
Lou
Retton
from
wonder
crowd
was
there
I?
Was
there?
We
had
a
gentleman
named
Brad
Jones
from
Denver
who
works
with
an
independent
agency,
that's
also
running
kubernetes,
and
production
and
Thomas
was
facing
on
the
last
name
right
now,
but
from
David
Chiu
good.
Thank
you
bye.
He
was
there
as
well,
representing
amazing
I
believe,
and
it
was
a
standing-room-only
conversation
previously.
When
we've
talked
about
kubernetes
that
Drupal
cons,
it's
been
very
lackluster
as
far
as
attendance
or
interest.
A
The
very
first
time
that
we
had
any
type
of
a
conversation
related
to
goober
Nettie's
in
this
fashion
was
bought
that
we
did
and
Kubek
on
Seattle.
We
had
I
think
maybe
four
people
that
attended
that
bus
in
Amsterdam.
We
repeated
that
bath
concept
and
had
over
twenty
people
show
up
with
a
very
diverse
range
of
implementations
that
they're
trying
to
understand
and
trying
to
bring
in-house.
A
So
I'd
say
that
the
and
you
you
know
the
the
people
in
the
Drupal
community
are
quickly
noticing
that
kubernetes
is
something
to
pay
attention
to
and
they
are
talking
about
it
and
coming
up
with
some
very
good
use
cases.
Some
very
good
scenarios
that
we
all
have
challenges
with
the
statefulness
of
Drupal
is
a
challenge.
The
concept
of
managed
versus
non
managed
kubernetes
clusters
came
up
quite
a
bit
being
able
to
understand.
A
You
know
when
doesn't
make
sense
from
a
business
perspective,
to
engage
in
something
like
kubernetes
and
and
why
would
you
even
want
to
do
that
to
begin
with
very
good
conversations?
Ellie
posted
a
link
to
the
YouTube
video
of
that
panel.
If
you
get
a
chance
to
take
a
look
at
it,
I
think
it
was
pretty
informative.
I,
don't
know
exactly
what's
going
to
be
happening
with
the
next
couple
con,
but
we
are
planning
on
doing
the
same
type
of
a
boss
at
a
minimum.
A
I
A
Continuing
along
you
know,
are
the
questions.
I
think
we're
almost
coming
up
to
time,
but
a
kind
of
goals.
How
about
things
that
you'd
like
to
accomplish
over
the
next
year
with
kubernetes
and
I,
can
start
that
one
for
me,
you
know
it's
just
continuing
to
to
understand.
You
know
what
cute
native
really
is
and
staying
as
closely
aligned
to
it
as
possible
being
cognizant,
then
you
know
the
kubernetes
environment
right
now
that
the
ecosystem
is
inundated
with
tools
and
and-
and
you
know
all
these
startups
that
are
promising
all
these
great
things.
A
So
it's
very
easy
to
latch
on
to
the
latest
greatest
startup
or
producer
of
a
tool.
Thinking
that
that's
going
to
solve
a
lot
of
problems
in
reality,
well,
I
think
it's
a
better
approach
to
try
and
stay
as
cube
native
as
to
minimize
the
amount
of
technical
debt
that
we're
incurring
as
we
try
to
solve
some
of
these
more
complex
problems.
A
C
C
A
Excellent
well
well,
go
ahead
and
update
the
meaning
of
notes.
At
the
end
of
this
we'll
go
through
and
pull
these
transcripts
in.
There
was
a
link
here
in
YouTube
for
the
Drupal
con
panel.
That
was
mentioned,
as
well
as
a
talk
from
Drupal
South
by
by
Nick
sue,
so
we'll
get
those
added
to
the
document
so
that
people
can
take
a
look
at
those,
but
thank
you
very
much
for
your
time
enjoy
the
rest
of
the
year.