►
From YouTube: Keynote: Maturing Kubernetes Operators - Rob Szumski
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Keynote: Maturing Kubernetes Operators - Rob Szumski, Principal Product Manager for OpenShift, Red Hat
We are at the “third era of Kubernetes.” Users learned about containers and orchestration all around the day one of operations. Now we well understand the operations needs from day two and beyond. Next, it is about solving the problems around the entire lifecycle of a Kubernetes application. Using an Operator and the Operator Lifecycle Manager, learn how the build, testing, install, upgrades, and access controls can help you prep your Kubernetes infrastructure for resilience.
To learn more: https://sched.co/Gsxz
A
Everybody
robbed
some
ski
from
Red
Hat
and
today,
I
want
to
give
you
an
update
on
operators
already
heard
a
little
bit
about
this
today.
We're
gonna
continue
hearing
it
for
years
to
come.
I
was
part
of
the
team
at
core
OS
that
introduced
the
concept
of
operators
in
2016,
and
so
I'm
excited
to
be
here
today
to
tell
you
about
the
success
we've
had
and
how
its
powering
the
third
round
of
growth
for
kubernetes.
A
So
when
you
look
at
these
three
rounds
of
growth,
they're
kind
of
roughly
mapped
to
when
certain
features
came
out
in
kubernetes
and
the
first
round
of
this
was
stateless
applications.
I'm
sure
many
of
you
are
running
these
on
kubernetes
today
and
you're,
using
replica
sets
and
deployments
to
make
this
happen
scale
things
out
horizontally
now.
The
second
phase
is
where
we
are
today,
where
you've
got
stateful
applications.
This
is
roughly
mapped
to
when
staple
sets
were
introduced
into
kubernetes,
and
these
were
really
great
for
running
decently.
A
Simple
workloads
that
needed
stateful
storage,
underneath
the
hood
you've
got
the
container
storage
interface
and
all
of
the
the
work
there
that
takes
storage
and
mounts
it
from
pod
to
pod,
as
it
moves
from
node
to
node.
Now,
as
a
community,
we've
progressed
past
that
we're
now
running
these
really
complex
distributed
systems
on
kubernetes,
and
this
is
a
little
bit
of
a
challenge
because
there
aren't
the
primitives
in
kubernetes
at
the
application
layer,
and
so
you
need
things
like
data
rebalancing,
seamless,
upgrades,
smarter,
auto
scaling,
and
this
is
where
we
are
today.
A
I
want
to
talk
about
how
you
can
fulfill
this
on
kubernetes
and
the
answer
is
an
operator.
So
what
an
operator
is
is
a
piece
of
software
that
has
unique
operational
knowledge
embedded
into
it
by
the
experts.
So
if
you
think
about
everything
that
is
in
like
run
book
or
a
wiki
that
you
might
have
maybe
you've
automated
some
failover
with
like
a
bash
script,
all
that
is
unique
application
knowledge
that
it
takes
your
team
to
run
an
application.
A
What
you
can
do
is
model
all
of
that
in
software
and
put
it
into
an
operator
now
what
an
operator
does
under
the
hood
is
using
you,
these
kubernetes
primitives,
to
make
that
happen.
You've
got
all
the
stateless
workloads,
all
your
stateful
workloads
secret,
just
got
or
service
discovery
secret
management.
All
this
is
what
you
can
use
in
your
toolkit
to
build
an
operator
and
ultimately
run
your
distributed
system.
A
So
the
great
thing
about
an
operator
is
that
you
can
do
flexible
architectures.
You
can
run
any
distributed
system,
you
can
run
on
kubernetes
and
that's
anything
and
you
have
this
really
great
flexibility,
reinvents
all
kinds
of
tiered
architectures,
whatever
you
need
to
do
in
your
application
now,
most
importantly,
because
they're
using
kubernetes
under
the
hood
you're,
not
reinventing
these
core
concepts,
nobody
needs
to
reinvent
service
discovery.
Nobody
needs
to
reinvent
secret
handling,
it's
all
done
by
the
community,
and
we've
done
it
really
really.
A
Well
now,
of
course,
running
these,
you
don't
want
to
be
tied
into
a
bespoke
set
of
tools,
so
you're,
just
using
coop
cuddle
at
the
end
of
the
day,
you're
looking
into
pods
you're
streaming
back
logs.
It
just
feels
like
regular
kubernetes,
but
you've
got
this
super
power
of
the
operator,
backing
it
and
then
once
again,
because
you're
using
kubernetes,
this
application
is
truly
hybrid.
You
can
take
this
operator
and
give
it
to
somebody
that's
running
on
one
kubernetes
provider
or
another
kubernetes
provider,
and
it's
gonna
work
the
exact
same
on.
A
So
you
really
get
this
truly
hybrid
application
that
can
run
anywhere
and
cubed
on
EU.
Earlier
this
year
we
introduced
the
operator
framework
to
help
all
of
you
make
these
operators
and
it's
a
set
of
tools
to
get.
You
started,
starting
with
our
operator
SDK,
and
this
is
a
set
of
cogeneration
and
best
practices
around
communicating
with
the
kubernetes
api.
How
to
structure
your
operator
cuz
it
the
end
of
the
day,
you're
building
this
reconciliation
loop.
It
has
all
this
logic
built
in
about
your
application.
A
A
So
we
have
our
operator
lifecycle
manager,
which
does
just
that
and
then
operator
metering
rounds
out
the
trio-
and
this
is
a
set
of
tools
to
collect
operational
metrics
about
your
operators
as
they're
running
your
applications
can
expose
these
metrics
up
and
we
can
collect
them
together.
So
you
can
best
run
your
clusters.
All
this
is
housed
in
a
vendor
neutral,
github
board
the
operator
framework,
we'd
love
for
you
to
get
involved,
check
out
the
code.
We're
gonna
work
on
that
in
the
open.
A
So
what
does
this
look
like
for
you
as
an
end
user?
The
best
part
about
kubernetes
is
this
self
service
and
you
have
access
to
these
really
rich
command-line
tools
and
api's.
So
here
we're
looking
at
a
production
namespace.
This
has
two
operators
installed
on
it
and
if
you
wanted
to
use
one
of
these
operators,
let's
say
the
MongoDB
operator,
you
submit
an
object
that
you
see
on
the
right.
This
is
a
MongoDB
replica
set.
A
A
Of
course,
you're
not
just
running
one
or
two
of
these
on
a
cluster,
but
you've
got
a
bunch
of
them,
so
our
operator
lifecycle
manager
helps
you
take
a
look
at
what
versions
you're
running
on
the
cluster,
how
they
need
to
be
upgraded.
You
can
see
we
got
a
number
of
namespaces
here
and
a
number
of
operators
installed
upon
them,
and
this
gives
you
a
good
idea
of
what's
going
on
how
you
can
secure
your
cluster
over
time.
A
So
it
was
a
quick
update
on
where
we
are
with
operators.
We're
really
excited
about
the
future,
so
I
want
to
encourage
you
to
get
started.
We've
got
a
getting
started
guide
on
github
for
using
the
entire
framework,
get
started
building
an
operator,
and
if
you
need
some
help,
we
have
an
operator
sig
that
meets
monthly,
and
this
is
a
great
place
for
you
to
gather
with
the
community
solve
problems
together.
A
It's
a
really
great
place
to
be,
and
then
of
course,
if
you've
got
time
later
today
we
have
a
session
four-thirty
about
operator
metering
specifically
and
unfortunately,
our
workshop
that
we're
running
tomorrow
morning
is
sold
out,
but
we're
gonna
be
doing
more
of
those
in
the
future,
so
stay
tuned.
We're
really
excited
to
see
what
y'all
built.
Thank
you.