►
From YouTube: Kubernetes for Product Managers Part 2
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
Hi,
my
name
is
viktor
knight,
I
am
a
product
manager
here
at
gitlab,
and
this
video
is
to
prepare,
let's
say,
non-engineering
people,
product
people,
designers
and
whoever
else
interested
to
think
about
kubernetes
and
to
be
able
to
consider
it
for
various
products
they
want
to
build.
It
will
not
go
into
detail
about
networking
and
role-based
access
controls
all
instead,
we'll
focus
on
the
core
principles
and
approaches
built
into
kubernetes
and
the
enabling
powers.
A
Now
this
is
part
two
of
this
series
and
I'm
going
to
skip
over
the
first
half.
So
if
you
haven't
seen,
then
watch
that
and
let's
start
with
part
two,
so
we
already
spoke
about
decorating,
nature
of
kubernetes
and,
actually
so
to
say,
the
user
interface
of
kubernetes
is
just
a
rest
api
that
consumes
yaml
resource
definitions.
These
yama
definitions
follow
what
we
call
the
kubernetes
resource
model.
It's
very
defined
and
extensible,
with
some
clear
principles
to
follow.
A
I
already
spoke
about
the
electricity
nature,
but
I'd
like
to
repeat
again
that
these
resources
typically
capture
the
desired
state
at
the
same
time
within
kubernetes,
they
will
be
enriched
by
the
observed
state
2.
What
you
can
see
here
in
this
pack
and
status
field
dimension
api
is
easy
to
navigate,
as
shown
below.
A
Besides
the
commander's
resource
model,
the
next
topic
we
should
understand.
Oh,
I
will
go
back
because
I
forgot
something.
One
very
important
thing
here
is
that
the
api
is
actually
extensible.
I
will
speak
about
this
a
bit
more
later,
but
you
might
imagine
how
to
extend
it
there.
You
see
this
prefix
group
version
namespace
and
so
on
path.
What
if
you
can
change
the
prefix?
What
if
you
can
define
a
new
group,
a
new
version
and
all
these
things
enable
you
to
have
an
extensible
api
within
competitors
itself.
A
So
how
does
this
desired?
State
actually
works
behind.
All
of
that
there
is,
there
is
a
reconciliation
loop.
It
compares
the
current
state
with
the
desired
state
and
there
are
controllers
that
act
to
reach
that
desired
state.
This
has
some
immense
benefits
like
automatic
configuration,
different
radiation
and
auto
healing.
Whenever
needed,
you
can
read
more
about
controllers
and
the
provided
link.
A
It's
such
a
thing,
at
least
to
me
about
actions
and
reactions,
one
when
we
want
to
design
system
and
in
the
end,
this
way
we
would
define
an
invented
system.
So
that
means
that
whenever
an
event
happens,
the
system
might
react
to
those
events
like
an
increased
load
or
a
change
in
the
configuration.
A
Kubernetes
took
a
different
approach.
It's
levels
based
system
and
this
by
design
provides
increased
resiliency,
for
example,
in
this
drawing
when
the
first
spike
happens,
the
system
might
be
too
slow
to
react
and
actually
the
system
might
be
already
back
to
its
desired
state.
By
the
time
we
can
react
to
the
event.
A
So
until
now
we
spoke
about
the
beginning,
main
building
blocks
and
principles
behind
kubernetes,
and
one
thing
I
mentioned
is
that
cuba,
api
is
extensible
the
how
we
extend
the
qb
api
is
called
custom,
resource
definitions
or
short
for
crds,
and
this
really
enables
us
to
define
custom
objects
within
the
cluster.
Moreover,
we
can
connect
custom
resource
definitions
with
custom
controllers
that
will
act
based
on
those
objects
like
the
environment
kind,
for
example.
We
I
showed,
on
the
very
beginning
of
the
first
series
that
would
be
a
custom
resource
definition.
A
Moreover,
on
top
of
the
custom
resource,
definitions
and
controllers,
we
can
even
build
something.
That's
called
operators
and
operators
are
meant
to
manage
the
full
like
cycle
life
cycle
of
community
sources,
for
example,
the
heat
lab
operators
operator
was
released
a
few
months
ago
and
it
allows
the
installation
of
a
gitlab
instance.
A
It's
very
important
to
understand
the
difference
here.
What
an
operator
provides
because
one
could
install
gitlab
into
kubernetes
cluster,
even
without
an
operator.
This
was
very
possible
for
a
long
long
time,
but
the
operator
allows
the
installation
the
upgrade
the
management
and
the
teardown
of
gitlab
instances.
A
So
it
really
manages
the
whole
life
cycle
of
gitlab
within
the
cluster,
and
you
can
have
your
own
operators
for
anything,
because
these
are
just
custom
resource
definitions.
These
are
just
controllers.
Everything
uses
the
basic
patterns
of
quebec
itself
and
this
extensible
api
as
a
result
with
crds
and
controllers,
and
operators.
Kubernetes
is
actually
a
programmable
decorative
api.
This
is
the
real
power
of
kubernetes
and
something
that
we
should
start
to
leverage
more
within
gitlab.
A
I
hope
that
by
now
you
understand
how
companies
might
use
kubernetes
to
build
their
own
internet
platform.
They
can
extend
the
cube
api
with
primitives,
like
the
environment,
kind
that
we
that
will
do
all
the
heavy
lifting
needed
to
set
up,
manage,
monitor
and
tear
down
environments,
especially
mentioned
on
this
slide
cross
pain,
as
it
shows
the
power
of
kubernetes
very
well
cross
pain
is
relatively
end
project.
The
first
public
release
was
at
the
end
of
2018
and
it
allows
decorative
infrastructure
management
using
kubernetes.
A
You
install
cross
paint
on
that
cluster
and
you
can
use
cross
plane
to
spin
up
new
clusters
or
you
can
use
cross
plane
to
spin
up
a
cloud
sql
instance,
or
anything
like
that
in
this
regard,
it's
very
similar
to
trafform,
but
with
the
help
of
the
kubernetes
api
and
the
reconciliation
loop,
it
provides
even
more
benefits
than
api
than
tranform
does,
because
software
needs
a
state
storage
where
it
stores
its
own
state
like
cross,
pay
and
receive
this
from
kubernetes
itself
and
with
the
reconciliation
loop.
It
provides
automatic
drive
detection
out
of
the
box.
A
Moreover,
because
it's
part
of
your
cluster
itself
with
this
single
a
rest
api
driven
by
kubernetes
resource
model,
purely
yaml
based
hundred
percent
declarative,
it
allows
others
to
easily
build
upon
the
crossplan
ecosystem
and
extend
it
further,
for
example,
to
say
that
whenever
I
start
a
new
review
app
give
me
a
new
database
for
every
deployment
and
that
database
is
an
infrastructure
support
piece.
It
might
come
from
any
cloud
provider
as
well
now
take
a
moment
to
think
about
how
gitlab
could
operate
in
a
fully
decorative
way.
A
A
We
are
evaluating
something
here,
and
this
makes
it
much
easier
for
the
developers
to
reason
about
the
system.
Of
course,
there's
still
someone
who
provides
these
controllers,
these
operators
or
crds,
and
that's
the
typically
the
platform
engineer
but
or
gitlab
itself,
but
that's
a
different
question.
Similarly,
we
could
describe
an
environment
and
manage
its
whole
life
cycle
integrated
into
gitlab
and
the
cluster.
A
Finally,
I
would
like
to
share
some
more
links.
If
you
like
to
learn
by
doing
and
would
like
to
give
kubernetes
a
try.
First,
there
is
venture
desktop.
It's
a
docker,
desktop
alternative.
It
chooses
a
simple
kubernetes
cluster
that
you
can
start
using
immediately
cube
cattle
or
cube
controller.
It
is
the
main
command
line
tool
to
interact
with
the
kubernetes
cluster.
A
Kindliness
is
a
terminal
based
minimal
quantities,
dashboarding
tool.
Of
course
we
have
the
agent
to
connect
the
cluster
with
gitlab
and
finally,
I
started
to
learn
kubernetes
at
the
this
last
link.
Here
they
start
with
very,
very
simple
setups,
and
I
think
that
the
tutorials
can
be
followed
even
even
without
programming
skills,
actually.
A
So
I
hope
you
found
this
interesting
if
you
have
any
questions,
please
reach
out
to
me
and
let's
start
thinking
about
kubernetes
a
bit
more.