►
Description
Join Andrew Sullivan, Chris Short, and the occasional special guest for an hour designed specifically to help the OpenShift admins out there. Come with your questions, leave with solutions.
A
Good
morning,
good
afternoon,
good
evening,
wherever
you're
healing
from
welcome
to
the
open
shift
administrator's
office
hours,
I
am
chris
short
executive
producer
and
host
of
openshift
tv,
I'm
joined
by
an
only
all-knowing
vert,
wise,
andrew
sullivan.
I
call
you
all
knowing
vert
wise
but
you're
all
knowing
in
many
things.
This
is
why
you're
really
great.
B
Yeah
so,
first
of
all
welcome
to
the
administrator
office
hours
openshift
administrator
office
hour,
so
for
anybody
new
or
for
a
refresher.
This
is
intended
to
be
an
ask
me
anything
style
of
session.
B
So
come
with
your
questions,
whether
they're
about
how
to
install
openshift
how
to
use
openshift,
you
know
we
can
certainly
entertain
developer-centric
questions,
but
outside
of,
I
think,
both
of
our
knowledge
bases.
So
that
would
be
we'll
have
to
get
back
to
you
on
that.
B
Exactly
so
today,
I
wanted
to
talk
about
operators,
and
you
may
be
thinking
to
yourself
but
andrew.
You
just
said
that
we're
not
going
to
talk
about
developer
things,
so
operators
are
core
and
central
to
open
shifts
and
they
were
really
the
driving
force
behind
the
version,
change
of
openshift
3
to
open
shift
4.
B
and
it's
critical
to
understand
operators
within
openshift4.
So
that's
kind
of
the
central
topic
that
you
know
you
and
I
will
be
conversing
about
welcome
any
questions
around
that.
But
for
the
audience
you
don't
have
to
ask
us
questions
just
about
operators.
Ask
us
questions
about
whatever
you
want
and
we'll,
certainly
whatever
detract
and
and
move
away
from
that.
B
I
was
having
my
one-on-one
with
aubry
earlier
this
week
and
I
was
commenting
on
how
I
feel
scatter
brains,
because
I
have
so
many
windows
and
usually
each
window
is
a
different
thought
process
and
each
window
has
so
many
tabs
that
have.
C
C
A
B
So
you
know,
I
use
the
brave
browser,
so
I
they
recently
re-enabled
sync,
so
I've
been
using
sync
across
a
whole
bunch
of
different
devices
in
order
to
keep
my
bookmarks
in
order,
etc.
So
it's
it's
helpful.
Sometimes
it's
not
helpful.
You
know
like
if
I,
if
I'm
looking
for
something
non-work
related
or
something
that
is
a
distraction
right,
if
you
will
right
hey,
I
want
to
look
at
the
news
and
see
what's
going
on
in
the
world,
because
I
am
feeling
too
happy
or
something.
A
Right,
yes,
yes,
no,
I'm
not
too
happy,
but
the
did.
A
I
know
I
did
no
yeah.
The
problem
with
me
is
that
I
end
up
using
multiple
computers
and
multiple
web
browsers
every
day,
no
matter
what
right,
like
all
you
end
up,
opening
firefox-
and
I
have
something
going
on
over
here:
I'll-
have
multiple
windows
of
chrome,
I'll,
have
chrome
open
over
there
and
firefox,
and
then
over
here
on
this
box.
It's
safari-
and
it's
just
like.
B
Yeah,
the
the
nirvana.
A
B
So
I'm
going
to
do
something:
that's
I
haven't
done
on
this
channel
before
and
that's
that's
going
to
be
used,
slides
and
not
just
use
any
slides,
I'm
going
to
put
them
into
presentation
mode,
even.
B
So
again,
this
is
the
the
the
focus
of
the
topic
today.
This
is,
I
think,
about
20
minutes
worth
of
slides,
give
or
take.
I
am
I
I
don't
want
to
use
slides.
This
is
just
the
easiest
way
to
tell
the
story
and
then
we
can
dig
around
and
we
can
explore
some
things,
and
I
may
even
do
that
in
the
middle
of
talking
to
the
slides,
often
so
yeah
hold
on,
but
anyways
for
anybody
watching
again
don't
be
afraid
to
stop.
Ask
questions.
B
I
don't
mind,
taking
a
break
to
answer
questions
not
related
to
what
is
coming
out
of
my
mouth.
If
you
will.
B
Sometimes
I
don't
even
know
when,
anyway,
all
right,
so
let's
talk
about
operators.
So
the
first
thing
for
anybody
who
has
sat
in
a
sales
meeting
right,
who
has
had
a
meeting
with
their
account
team,
talked
about
openshift4.
You've,
probably
seen
a
slide
like
this
one,
where
we
do
a
lot
of
hand
waving
right.
We
talk
breathlessly
about
how
great
operators
are
and
all
of
the
problems
that
they
solve
and
they're
the
best
things
since
sliced
bread.
B
You
may
have
even
seen
sl
so
I'd
like
this
one,
similar
type
of
concept,
giving
you
the
ability
to.
I
like
the
sentence
down
here
at
the
bottom,
that's
one
that
we
use
pretty
frequently
yeah,
codifying
operational
knowledge
and
and
workflows
and
then
putting
that
directly
into
the
cluster
itself.
B
B
If
you
take
a
step
down
and
you
talk
to
you
know
a
more
technical
person,
oftentimes
it'll,
be
you
know
things
like?
Oh
well,
you
know
we
use
operators
for
all
of
the
openshift
components,
and
you
know
the
operators
operate.
The
cluster
and
you
don't
have
to
worry
about
like
great,
but
what
does
that
actually
mean
so
right?
Let's
start
with.
B
B
I
used
to
do
josh
atwell
and
I
used
to
do
a
presentation
that
would
talk
about
the
the
hand
grenade
of
deployment
right
where
the
the
applications
team
would
package
up
their
code
into
this
little
hand,
grenade
shaped
object
and
then
lob
it
over
the
wall
to
the
ops
team
and
say
good
luck
may.
B
B
B
Right
create
a
help
desk
ticket.
I
need
17
virtual
machines
that
have
these
characteristics
and
these
operating
systems
and
I
have
to
go
through
and
deploy
the
application
right
connect
in.
Maybe
I
use
ansible
or
something
like
that,
deploy
the
rpm
put
in
a
bunch
of
configuration
data
all
of
those
files
on
the
file
system.
Configuring
you
see
the
list
here:
firewall
ports.
A
B
Exactly
right
and
that's
the
next
step
right,
hey,
I
just
updated
the
application
where
the
apps
team
just
updated
the
application.
Now
what
yeah?
How
do
I
go
through
and
do
this?
Well,
we
ops
guys
we
solved
this
a
while
ago,
right,
automation,
yeah.
I
mentioned
ansible
right
how
many
of
us
actually
go
through
and
manually
deploy
applications
more
than
the
first
or
second
time
right.
There's
a
huge
number
of
tools
out
there
to
take
care
of
this
for
us
and
make
it
as
easy
and
simple
as
possible,
yeah,
so
welcome
to
2012.
B
We
have
now
applica
automated
the
deployment
and
the
life
cycle
of
an
application.
So
let's
talk
about
this
in
the
context
of
kubernetes,
so
with
kubernetes.
Now
I
take
that
application
and
I'm
well
putting
it
into
a
container
right
and
that
could
be
as
simple
as
and
again.
You've
probably
heard
this
story
for
the
last
five
to
seven
years,
depending
on
how
long
you've
been
in
contact
with
with
containers.
B
Right
now
I
take
my
rpm
and
I
install
it
to
a
container.
I
take
all
of
that
config
and
I
put
it
in
the
container.
I
you
know,
maybe
I
expose
it
through
environment
variables
or
whatever
it
is
right
and
then
the
kubernetes
part
of
that
is
just
translating
that
into
kubernetes
objects.
You
know,
hey,
I
deploy
my
my
container
using
a
deployment
so
that
way
I
can
scale
it
up
and
down.
I
pass
configuration
into
the
application
using
a
config
map
or
a
c
or
a
secret.
B
I
expose
it
using
services
and
routes
so
on
and
so
forth,
so
just
like
before.
We
want
to
automate
this
because
well,
we
have
the
same
problems.
We
have
the
same
life
cycle
that
we
have
to
deal
with.
How
do
I
you
know
when
the
next
version
is
released?
How
do
I
update
that
so
often
times
this
is,
you
know,
scale
down
to
zero
change
the
deployment
whatever
that
happens
to
be
and
then
bring
everything
back
up
right
and
we've
got
a
lot
of
automation
to
do
that
as
well.
B
So
if
we
have
automation
to
manage
our
application
lifecycle
inside
of
kubernetes,
there's
nothing
that
says
that
we
can't
containerize
that
application
or
excuse
me
that
automation
right
right.
I
can
put
an
ansible
playbook
into
a
container
and
I
can
deploy
that
container
right
back
into
kubernetes.
If
I
so
choose
you
misspelled
operators.
B
Go
ahead,
so
I
can
take
my
automation
and
I
can
deploy
it
into
kubernetes
right
alongside
the
application
code
itself.
So
how
do
I
access?
How
do
I
trigger
that
automation
as
the
administrator
as
the
user,
and
this
is
where
we
begin
to
get
into
some
new
territory
so
from
a
kubernetes
perspective,
there's
two
things
that
we
want
to
be
aware
of:
one
is
a
custom
resource
definition
and
this
simply
tells
kubernetes
hey
here's
this
thing
right.
B
B
Kubernetes
informs
the
controller,
hey
something
happened,
controller,
do
your
automation
and
maybe
that
automation
is
okay.
Well,
I
need
a
new
instance
of
mysql.
Let's
deploy,
let's
create
a
new
deployment.
Let's
request
some
new
pvcs
so
on
and
so
forth
or
hey
the
version
was
updated.
I
need
to
do
these
actions.
B
So
this
is
just
the
the
the
text
version
or
the
sorry
the
the
slide
version
of
what
I
just
said
here.
So
the
important
part
here
is
over
here
on
the
kind
of
right
hand.
Side
of
the
slide
of
I
have
that
operator.
I
have
that
automation,
that's
running
inside
of
my
cluster
that
interacts
with
and
executes
things
or
translates
them
into
those
kubernetes
native
resources.
B
B
So
if
we
have
a
dba-
and
you
know
yes,
they
are
sometimes
difficult
to
work
with,
but
that's
because
they
are
biased.
I
know
when
I
deploy
a
dba
or
db
instance
for
this
application.
I
need
to
do
these
things
to
make
it
work
as
good
as
possible.
I
know
with
my
infrastructure.
I
need
to
do
these
things,
because
the
storage
has
these
capabilities
and
compute
this
and
network
this
etc.
B
B
So
now
we
can
rinse
and
repeat
right.
We
can
create
an
entire
catalog
of
operators.
We
can
increase
the
capabilities
of
those
operators,
so
maybe
it's
going
from
just
basic,
install
and
upgrades
to
the
deep
insights
level
here
this
comes
from
the
operator
framework
website
by
the
way,
and
it's
just
it's
expanding.
What
that
automation
is
capable
of
doing.
Is
that
automation,
looking
at
the
the
metric
service
to
collect
metrics
and
be
able
to
make
decisions
on
your
behalf
for
preemptive
or
proactive
operations
etc?
B
B
Thank
you
jafar.
Yes,
so
the
problem
now
happens
or
has
now
been
changed
so
that
we
have
the
same
life
cycle
management
paradigm
shift
associated
with
the
operator
that
we
did
with
the
application.
Before
hey,
I
increased
the
capability
of
my
operator.
I
added
this
new
ability.
How
do
I
now
replace
the
existing
automation?
Remember
it's
just
a
pod,
so
maybe
it's
you
know
replacing
the
deployment
or
modifying
the
deployment
to
replace
the
pod
etc.
B
So
it's
really
important
to
you
know:
lifecycle
manage
those,
and
this
is
where
olm
operator
lifecycle
manager
comes
in.
That's
exactly
its
role.
I
define
using
a
cluster
service
version
or
csv
hey.
This
operator
needs
these
resources.
These.
It
defines
these
custom
resource
definitions.
It
needs
these
permissions
for
the
service
accounts
right,
these
role,
bindings,
etc.
B
As
well
as
has
these
other
dependencies,
and
then
we
rely
on
olm
to
go
through
and
make
that
a
reality.
So
olm
offers
that
catalog
of
here's
all
the
operators
that
are
available.
Please
deploy
this
operator
right
and
making
sure
that
everything
it
needs
is
there.
It
does
a
bunch
of
other
things.
I'm
not
going
to
go
to
details
here
because
I
could
spend
a
while
about
this.
This
slide
was
actually
presented
by
it's
either
daniel
or
rob
during
one
of
the
community
operator
framework
updates.
A
A
The
cncf-
I
can't
remember
if
it's
incubating
or
sandbox
at
this
point
but
yeah
it's
been
in
the
cncf
ecosystem
for
a
couple
months
now,
a
few
months
and
to.
B
So
let's
talk
about
this
in
the
context
of
openshift
and
in
particular
the
core
of
openshift,
so
when
we
think
about
openshift
and
if
you're
familiar
with
the
install
process,
really
the
first
thing
that
happens
after,
of
course,
installing
coreos
is
a
kubernetes,
cluster
gets
stood
up
and
then
we
deploy
operator,
lifecycle
manager
and
then
we
start
installing
a
whole
bunch
of
components
right.
The
console
the
registry,
the
routers
and
on
and
on
and
on
so
when
you
think
about
it.
Open
shift
is
operators
all
we've.
D
B
Is
yeah
all
we've
done
is
take
vanilla,
kubernetes
and
all
of
the
things
that
you
need
with
kubernetes
to
have
a
production
deployment
and
we've
done
the
automation
for
you
of
deploying
and
managing
and
life
cycling
and
configuring
all
of
those
different
assets,
all
of
those
different
components,
and
we
do
that
right.
Openshift
itself
is
controlled
through
the
cluster
version
operator.
A
Know
what
olm
is
this
operator
life
cycle
manager?
Think
of
it
as
like
a
cataloging
and
management
tool
for
all
the
operators
in
your
life
right.
B
B
A
Yeah
notice,
we're
not
we're
not
leaving
much
to
chance
here
right,
like
the
node
tuning
operator,
will
respond
to
events
constantly
right
like
if
there's
an
event
that
needs
to
reconcile.
It
will
do
so
same
with
networking
same
with
storage
same
with
the
whole
nine
yards
right,
like
ncd
being
installed
with
and
maintained
with
an
operator
really
allows
us
to
not
necessarily
worry
about
how
ncd
is
configured
when
we're
building
up
openshift
clusters.
Right,
like
we
know
that
it's
done
in
a
way
that
works.
B
B
They
then
used
ansible
and
used
ansible
to
go
to
each
one
of
those
nodes
and
deploy
and
configure
a
bunch
of
components.
Openshift
4
is
kind
of
the
same
right.
I
stand
up
a
bunch
of
core
os
nodes.
I
deploy
kubernetes,
I
deploy
openshift
on
top
of
it.
It's
just
those
last
two
things
all
happen
via
operator
right,
so
all
of
that
ansible
automation
that
we
used
to
use
is
now
operator
automation,
so
to
use
one
of
my
favorite
terms
rates
so
where'd,
my
cheese
go.
B
Yeah
yeah,
you
should
you
should
look
that
up
right,
so
it's
a
a
management
philosophy
or
a
an
evolution
thing
of
if
I
am,
and
it's
usually
applied
to
it.
Workers
of
my
my
life
is
in
a
steady
state
and
then
something
changed
right.
Somebody
moved
my
cheese,
it's
in
a
different
place
in
the
maze,
so
my
world
has
been
changed
and
you
can
either
you
know
stay
where
you
are
and
complain
about
my
cheese
having
moved
or
you
can
go
and
find
where
the
cheese
has
moved
to.
I
like
this
premise.
B
So
yes,
where
has
my
cheese,
moved
to
so
you'll?
Note
that
this
means
that
some
things
are
different
right
with
openshift
3.
It
was
all
ansible,
it
was
all
rel
I
could
ssh
into
the
node.
I
could
go
and
dig
through
playbooks
and
I
could
see
exactly
what's
going
on
right.
It
was
several
billion
or
so
configuration
files
scattered
across
a
bunch
of
different
machines
that
I
needed
to
configure
right
yeah.
So
I.
B
B
How
do
I
actually
administer
my
cluster,
and
this
is
where
things
are
both
better
and
slightly
more
confusing,
so
I
say
they're
slightly
more
confusing
because
well
they're,
not
different
projects.
Different
components
sometimes
do
things
slightly
differently,
and
you
can
see
that
so
here
I'll
I'll
pick
on
one
of
these.
So
let's
do
oc
get
cluster
operator.
B
Yeah
yeah
yeah
oc
get
co,
so
you'll
see
that
some
of
these
are
some
of
these
have
two
different
name
spaces.
So,
let's
pick
on,
let's
see,
let's
pick
on
machine
api,
so
oc.
B
B
Some
of
these
and
what
I'm
trying
to
show
here
right
here,
coupe
scheduler,
so
some
of
these
have
an
operator
namespace
and
then
an
application
namespace.
B
B
B
So
that's
one
thing
to
be
aware
of:
is
we
sometimes
need
to
understand
how
things
are
laid
out?
So
let's
look
at
a
couple
of
ways
that
we
can
cheat
so
first
and
foremost,
of
course,
is
simply
oc
get
cluster
operator
right.
I
can
see
each
one
of
these
and
I
can
get
details
about
each
one,
so
I
can
say
oc
details,
co,
slash!
B
B
So
there's
we
can
look
at
the
details,
either
output
is
yaml
or
by
doing
a
describe
of
a
particular
cluster
operator
and
it'll
tell
us
high
level
status.
So
we
can
see
here
the
most
recent
status
message.
The
registry
is
available,
reason
is
ready.
Status
is
true
down
here,
right,
we're
progressing
right,
it
says
reason
ready
and
then
status
is
false,
meaning
it's
not
currently
ready
so
so
forth.
If
there's
errors,
oftentimes
they'll
get
surfaced
up
in
through
here,
but
remember.
A
B
B
B
Project-
and
I
can
see
the
pods
that
are
running
inside
of
here-
if
any
of
them
had
errors
right,
I
could
do
a
standards
oc
logs
or
oc
describe
of
the
pod
to
get
what's
going
on
now,
where
it
gets
a
little
confusing
is
because
so
far
I've
done
nothing
that
isn't
standard.
Kubernetes
right
is
most
of
the
time
an
operator
or
excuse
me,
the
operand
is
configured
by
the
custom
resource
instance.
B
B
B
This
is
another
area
where
it's
sometimes
unevenly
applied
right,
sometimes
we'll
have
operators
that
everything
in
the
configuration
is
applied
through
the
custom
resource,
and
sometimes
it's
done
through
config
maps,
and
sometimes
it's
done
through
secrets,
and
sometimes
it's
done
through
right,
there's
a
lot
of
different
ways
that
that
can
be
done.
How
do
I
find
out?
How
do
I
know?
Well,
there's
a
little
secret
that
you
can
do
if
I
do
oc
adm
release
info
dash
dash
commits
whoa.
B
A
B
See
that
it
comes
from
this
particular
github
repo
and
many
many
many
times,
you
will
find
that
the
engineers
have
a
lot
of
troubleshooting
and
other
information
inside
of
here
right
here.
Here's
how
to
go
through
and
configure
all
of
the
different
aspects.
So,
even
though
the
docs
have
some
of
this
information
for
configuration,
they
don't
always
have
all
of
the
information
right.
B
B
B
B
So
this
is
one
way
that
we
can
from
the
cli
right
go
through
and
figure
out
what
precisely
I
can
and
what
are
valid
values
for
my
particular
configuration.
So
an
oc
get
my
config
for
image
registry.
You
can
see,
there's
the
cluster
one,
so
cluster
dash,
oh
yaml,
so
inside
of
here
we'll
scroll
up
just
a
little
bit.
B
We
can
see
underneath
my
spec.
I
now
have
a
bunch
of
information
in
here
so
down.
Here
is
where
I
configured
my
particular
instance.
So
we
can
see
right
now.
My
registry
isn't
deployed.
If
it
were
deployed,
it
would
be
using
this
storage
information,
it's
using
a
local
local
instance,
so
all
of
that
is
through
the
cli.
B
So
what
about
the
gui?
So
the
gui?
We
can
see
a
lot
of
very
similar
information.
If
we
come
down
here
underneath
come
on
scroll
down,
you
can
scroll
underneath
administration,
and
then
we
have
a
couple
of
different
areas.
So
one
we
have
under
cluster
settings
cluster
operators
right.
This
is
effectively
the
gui
equivalent
of
doing
oc,
get
cluster
operator
or
cgit
co.
B
So
I
can
come
here
to
image
registry
and
the
helpful
part
is.
It
tells
me
about
the
related
objects,
so
we
can
see
here.
Oh
I
have
this.
Let's
see
config
right,
so
I
can
look
at
my
cluster
config
and
I
can
look
at
the
yml
and
skip
the
exact
same
thing.
So
if
I
scroll
down
here
underneath
my
spec,
you
can
see
here.
Management
states
right
is
right
there.
So
there's
my
storage
configuration
nice.
B
So
two
different
ways
of
going
about
the
exact
same
thing:
one
is
through
the
gui.
One
happens
to
be
through
the
cli.
You
know
which
one
you
prefer,
which
one
you
like
better
it's
entirely
up
to
you.
I
tend
to
switch
back
and
forth
much
like
yeah,
sometimes
say
you
know,
cube
cuddle
or
cube
control
or
coupe
ctl
or
I'm
an
equal
opportunity.
B
You
know
no,
nobody
knows
which
one
they
like
so
all
right.
Let
me
rephrase
that
everybody
likes
something
different,
so
I'm
trying
to
I'm
trying
to
include
everyone
or
offend
everyone,
depending
on
your
perspective,.
B
B
Right
so
we
have
this
group,
so
I
just
search
for
registry
and
see
the
config,
so
this
one
is
slightly
less
useful
for
the
cluster
operators
right
so
with
the
cluster
operators
right.
I
have
that
you
know
the
that
really
helpful
related
objects,
so
I
can
see
all
of
the
things
related
to
it.
Whereas
if
I
come
down
here
to
the
custom
resource
definitions,
I
have
to
do
a
little
bit
of
reverse
engineering
myself.
Some.
C
B
So
I'm
gonna
come
in
here
and
I'm
going
to
go
to
this
guy
and
I'm
just
gonna.
I'm
gonna
break
this
okay
great.
So
let's
assume
that
I
have
I
know
so.
So,
let's
assume
that
I
am
attempting
to
configure
my
registry
for
storage
and
I
bungled
it
right.
I
I
just
didn't
configure
it
correctly.
So
how
do
I
determine
that?
B
So,
let's
scroll
up,
so
we
can
actually
see
the
heading
here
so
one
because
I
changed
the
configuration
this
column
for
progressing
is
currently
set
to
true.
That
means
it's
in
the
progress
of
changing
right.
It
is
not
currently
degraded.
Usually
that
is
I
don't.
Each
one
of
the
operators
has
a
different
set
of
conditions
for
degraded.
I
don't
know
precisely
what
those
are.
That
would
be
an
interesting
side
project
to
figure
out.
B
Yeah
and
the
the
available
is
also
can
be
confusing,
for
example,
if
it
is
unconfigured.
So,
for
example,
if
I
put
my
image
registry
into
an
unmanaged
state,
then,
according
to
its
configuration,
it
is
available
the
way
that
it
is
supposed
to
be,
even
though
the
registry
isn't
deployed
and
therefore
isn't
available.
B
Yeah,
so
essentially,
you
have
to
have
a
little
bit
of
context
awareness
around
what
this
is
reporting.
So
let's
look
at
our
see,
describe
co
image
registry,
so
we
can
see
a
couple
of
things
here.
So
one
notice
I
changed
my
my
registry
config
and
now
it's
saying
that,
there's
an
error
right
unable
to
apply
resources
unable
to
sync
storage
configuration
hey.
B
B
Your
princess
is
in
another
castle,
okay,
so
where
else
can
I
look
so
remember
that
an
operator
is
automation,
so
is
my
operator
having
difficulty
or
is
the
thing
it
is
operating
on
the
upper
end
having
difficulty
so
if
I'm
still
in
the
openshift
image
registry
project,
so
let's
say
that
it
was
the
operator.
I
can't
instantiate
something
for
some
reason.
B
And
I
can
look
at
the
logs
for
this
particular
guy
now,
in
this
instance,
it's
surfacing
up
really
good
information.
Not
all
of
the
operators
are
quite
as
well
behaved
as
this
one
of
surfacing
up.
You
know
hey
unable
to
sync
storage
configuration,
so
if
I
have,
for
example,
an
image
or
if
this
is
failing
to
deploy
or
if
this
is
failing
to
deploy
something
like
that
right,
I
can
look
in
the
operator
to
see
what
that
error
might
potentially
be
right.
B
C
A
B
So,
let's
flip
over
to
the
gui
right
same
sort
of
thing,
how
do
I
figure
out
what's
going
on
with
my
project
with
my
thing,
so
the
first
thing
that
we
can
do
is
we
can
look
at
our
operator
here
and
we
can
see
that
why?
Why
isn't
it
done
updating?
What's
what's
going
on
here,
right
right.
D
B
B
So
and
then,
if
I
scroll
down
here
to
the
conditions,
we
can
see
the
same
thing
right,
so
it
is
currently
in
an
error
with
a
progressing
status
of
unable
to
to
sync
storage
configuration
so
same
as
up
above
like
before.
If
I
go
to
my
related
objects
here,
I
can
check
out
all
the
different
things
that
are
in
here.
B
So
maybe
I
want
to
go
to
the
namespace
and
now
that
I'm
in
my
namespace
I
can
go
to
workloads
and
pods,
and
I
can
see
all
of
my
pods
that
are
running
inside
of
here
just
like
before
so
here's
my
registry
operator
pod.
I
can
click
here
and
go
to
logs
and
we
can
see
that
it
is
very
unhappy
with
me
at
the
moment
and
scrolling
a
lot
of
lines
by
so
again.
B
Another
way
to
look
at
the
logs
that
are
going
on
inside
of
there
same
thing
is
true
of
the
other
pods,
I'm
not
going
to
reiterate
or
redo
that
for
that
particular
pod,
because
it's
the
exact
same
process
so,
okay,
now
let
me
change
my
my
configuration.
Let
me
fix
my
configuration,
so
that'll
fix
you,
okay
I'll
show
you
related
objects.
Go
back
to
my
config,
go
to
my
yaml
scroll
down
here
and
put
our
correct
port
back.
B
And
we
go
to
the
registry,
we
can
see,
it's
already
updated,
wow
didn't
take
it
long
and
it's
already
up
and
available
and
ready
to
go.
I
could
review
all
the
logs
and
see
all
the
changes
that
it
took
effect
but
effectively.
It's
just
remembering
that,
even
though
it's
an
operator
at
the
end
of
the
day,
everything's
still
a
pod,
so
you're
kind
of
checking
two
different
places.
One
is
the
operator
itself.
The
other
one
is
the
operand
is
the
application
that
the
operator
is
trying
to
deploy.
A
B
I
don't
know,
I
don't
know
if
I
don't
know
if
christian's
paying
attention
I'm
guessing
because
he
hasn't
been
responding,
he's
not
so
christian
is
usually
my
go-to
for
the
security
side
of
things
and
determining
the
security
context,
constraints
which
is
sec
that
are
important
and
relevant.
It
is
not
one
of
my
areas
of
strengths.
A
Yeah,
I'm
digging
up
some
stuff-
let's
see,
oh
there's,
actually
a
blog
post
from
this
year
about
it,
how
to
manage
them
and
kind
of
explains
them
and
so
forth.
So
on.
So
let
me
toss
that
in
here
that
could
be
helpful
check
that
out
vadran
and
maybe
that'll
help
like
give
this
to
your
developers
in
the
sense
of
like
you
need
to.
B
C
B
And
folks
and
see
what
we
can
find.
So,
if
you
will
please
send
me
a
message
so
either
twitter,
practical,
andrew
or
email
andrew.sullivan,
no
siri.
I
don't
want
you
to
send
a
message.
B
C
B
Now
anyway,
so
yeah
send
me
a
message
just
to
connect
and
that
way
I
will
reply
back
with
any
resources
that
we
find
to
help
you
with
that
inside
of
your
organization,
so
happy
to
do
that,
and
then
maybe
we
can
use
that
as
a
topic
for
two
weeks.
C
A
B
All
right
yeah!
Well,
I'm
I
don't
know
if
north
carolina
will
do
that
or
not.
I
feel
like
we
probably
should,
because
things
are
a
little
crazy.
B
A
B
All
right,
so
that's
that's
my
crash
course
on
operators.
That's
awesome.
A
I
would
like
to
remind
folks
you
can
write
operators
an
ansible
right,
so
if
your
application
has
needs
that
are
like
external
to
a
cluster
right
like
I
need
to
flip
a
load
balancer
or
change
some
port
or
something
you
know,
dns
whatever
you
can
do
that
with
ansible
too,
and
you
can
use
a
helm
chart
as
well
as
writing
it
and
pure
go,
and
there
are
some
other
tools
for
writing.
There's
comp
cop
kopf
for
writing.
Operators
in
python,
there's
flan
for
writing
operators
in
bash.
A
B
B
I'll
pick
on
one,
so
this
is
openshift
fertilization,
so
this
particular
component,
the
ssp
operator,
is
an
ansible
operator.
A
B
A
B
Yeah,
so
let
let's
let
let's
look
at
or
or
take
a
step
back,
so
if
we
go
to
github.com
operator
framework.
B
So
if
you
look
at
the
operator
framework,
which
is
the
cncf
right
sandbox,
is
it
sandbox
or
is
it
yeah?
So
it's
the
cncf
project
yeah,
so
the
operator
framework
is
made
up
of
multiple
things.
So
first
is
olm
rate
itself
so
for
managing
controlling
you
know,
interacting
with
operators
not
necessary
to
deploy
and
use
an
operator.
It
is
for
managing
operators
themselves
incubating
projects.
Sorry,
thank
you.
B
So
the
second
one
is
the
operator
sdk.
So
the
operator
sdk
is
an
easy
way
of
getting
started.
Writing
your
own
operator,
it's
essentially
scaffolding
and
a
bunch
of
things
to
help
with
that.
It
is
not
the
only
way
to
do
it,
and
I
think
that
that's
really
important
to
point
out,
because
if
you
look
at
the
operator
sdk,
it
really
focuses
on
go
on
ansible
and
on
helm.
B
B
B
B
But
if
I
go
up
here
to
documentation
and
olm
documentation
operator
lifecycle
manager,
this
is
the
component.
That's
responsible
for
right!
Olm
is
responsible
for,
like
this
operator
hub
display.
This
is
it's
the
one
that's
collecting
all
of
those
catalogs,
it's
displaying
them
back
or
it's
used
to
display
them
back
here
right.
B
It's
the
part,
that's
deploying
those
operators,
so
it
can
be
helpful
to
have
access
to
all
of
the
information
relating
to
olm
when
you're,
trying
to
troubleshoot
or
when
you're
trying
to
learn
more
about
how
openshift
works
down
at
that
core
level,
so
tons
and
tons
and
tons
of
information
inside
of
here
about
how
it
works
about
things.
To
look
at
you
see
this
troubleshooting
right.
I
have
this
catalog
source.
You
know
how
do
I
troubleshoot
catalog
sources,
particularly
for
customers
who
are
doing
restricted,
network
or
offline
deployments
right
in
our
documentation.
B
We
say
you
know
oc
adm,
mirror
blah
blah
blah
blah
and
then
to
get
the
deployment
images
across,
and
then
we
have
a
separate
set
of
tasks
for
mirroring
all
of
the
operators.
So
if
you
want
to
dig
into
details
about
what's
actually
happening
there,
if
you
have
issues
it's
another
great
place
to
look
inside
of
there.
A
Cool,
I
would
encourage
you
to
check
out
operatorhub.io,
there's
a
lot
of
operators
there
and
they
are
open
source
as
well.
So,
if
you're
like
hey,
I
want
to
deploy
a
db.
Well
guess
what
procona
has
an
operator
and
has
an
operator
and
all
these
people
have
operators
to
deploy
various
databases.
A
B
B
So
this
is
the
upstream
community
operators.
Essentially
anybody
can
create
an
operator
and
submit
it
to
be
included
here
more
precisely.
If
you
go
to
github.com
operator
framework
and
you
browse
to
this
community
operators,
repo.
That
is
where
all
of
these
are
at
specifically
in
the
upstream
community
operators
subdirectory.
B
B
B
B
So
when
we
look
in
operator
hub
and
openshift,
we
have
community
operators
and
we
have
a
bunch
of
others.
So
community
operators
with
this
fancy
community
tag
are
the
ones
that
are
found.
So
I'm
back
in
my
operator
framework
community
operators,
github
repo
they're,
the
ones
found
in
this
community
operators
sub
directory.
B
So
simple
things
like
it's
not
going
to
run
using
a
root
user
right
so
back
to
scc
and
that's
basically,
all
of
the
validation
that
has
been
done.
So
that's
why
we
have
uneven
or
different
sets
that
are
listed
there.
So
the
third
type
of
operator
that
you'll
see
is
the
certified
operators
nice.
These.
These
are
not
found
in
the
operator
framework
github,
repo
they're
on
registry.redhat.com,
and
these
are
ones
that
have
gone
through
a
full
certification
suite.
C
B
You're
really
curious,
you
can
actually
see
all
of
that
if
we
look
here
in
certified
operator,
build
guide
and
I'll
post
this
link
in
here.
So
if
you're
curious
about
what
that
looks
like
and
how
to
go
through,
and
if
you
just
want
to
know
what
the
certification
process
looks
like
and
what
our
partners
go
through
to
do
that,
that's
a
great
resource
to
see
all
of
that
information.
B
B
A
I
feel
like
we've
covered
operators
very
well
again,
if
you
have
questions,
feel
free
to
reach
out
practical
andrew
on
twitter,
chris
short
on
twitter
c
short
at
red
hat.
That's
my
email!
A
I
forget
your
email
because
it's
just
a
sullivan
or
andrew
sullivan
andrew.sullivan
there
you
go
yeah,
feel
free
to
reach
out.
If
you
have
questions
about
operators,
we
can
get
you
in
touch
with
the
experts.
If
we
can't
answer
it
ourselves
and
yeah
anything
else,.
B
I
do
see
a
question
from
murat
operators
and
databases.
Any
production
hints.
A
Ooh
don't
write.
It
yourself,
use
someone.
B
Else's
so
in
all
seriousness,
it
depends
on
who
the
vendor
is
and
whether
or
not
you
trust
them.
So,
if
you're
using
a
third
party
right,
so
here
I'll
pick
on
come
on,
so
I
will
let's
get
rid
of
this
guy
and
then
database
crunches.
B
You
see
here,
their
capability
level
goes
all
the
way
up
to
autopilot,
but
autopilot
might
mean
something
different
to
you
than
it
does
to
them
so
again,
working
with
that
partner
to
understand
precisely
how
far
it
goes
and
how
to
integrate
mongodb
is
going
to
be
the
same
way.
Cockroach
db
is
going
to
be
the
same
way.
Couchdb
is
going
to
be
the
same
way
so
on
and
so
forth.
B
You
can
also
augment
those
capabilities
right.
You
can
effectively
create
automation
that
also
looks
at,
and
does
things
based
off
of,
what's
going
on
inside
of
one
of
those
databases,
if
you
so
choose
so,
like
always,
just
have
a
full
understanding
of
what's
going
on
right
feel
comfortable
with
what
is
and
isn't
happening
inside
of
the
cluster
in
the
application
you
know
kind
of
standard
stuff.
A
B
B
Yep
theirs
was
a
month
or
so
ago.
So
for
those
of
us
in
the
u.s
happy
thanksgiving,
I
hope
that
it
is
relaxing,
and
hopefully
you
can,
you
know,
see
your
family
even
if
it
is
remotely
for
the
rest
of
the
worlds.
We
will
see
you
in
two
weeks
and
continue
to
be
safe.
C
D
E
Hello,
michael,
thank
you
for
everything,
I'm
doing
great
thanks
for
this
platform
to
red
hat
and
as
well
as
ibm
get
to
be
in
you
know.
We
have
been
a
partner
of
red
hat
for
a
long
time
and
it's
it's
great
to
be
with
you
guys.
Okay,.
D
E
I
am
right
where
nearby
the
hopkinton,
where
the
boston
marathon
starts,
you
know
I
don't
know
how
many
of
the
folks
listening
to
this
know
this.
You
know.
Boston
marathon
starts
right
from
our
location
and
it
runs
up
to
boston.
So
it's
a
you
know
this
year
we
missed
it,
but
that's
the
location.
D
Well,
this
year
certainly
has
been
challenging
to
say
the
least,
but
but
you've
been
around
a
long
time.
You've
had
a
pretty
successful
career
right.
You
were
at
fog
panel,
actifio
eucalyptus,
a
lot,
a
lot
of
big
names,
pretty
pretty
successful
career
so
far,.
E
Humble
enough,
you
know
thank
you
for
for
other
companies
for
giving
me
the
opportunity
to
run
along
those
wonderful
time
frame
and
I've
met
tremendous
people
along
this
journey.
You
know
we
have
to
agree
to
the
team
to
make
other
people
successful,
so
you
know
similar
to
other
platforms
and
other
teams.
You
know
we
have
gotten
a
great
team
with
apprenix
as
well
so
far.
E
D
Okay,
well
tell
me
about
it.
I
mean
you're,
the
you're,
the
ceo
and
the
founder,
so
you,
you
started
this
company
and
I
think
you
were
ceo
and
of
a
lot
of
other
successful
companies.
So
hopefully
you
know
you'll
have
the
same
same
good
luck
here,
but
why
apranics?
What
what's
in
the
name
it
when
I
think
you
know
a
frannix,
it
doesn't
make
me
say:
oh
yeah,
I
know
what
that
is
like.
Why
did
you
pick
that
name.
E
Great
question,
so
you
know
a
panic
starts
with
something
to
do
with
applications
right.
We
always
had
the
thought.
Now
I
come
from
the
storage
background
and
cloud
infrastructure
background
like,
as
I
mentioned,
eucalyptus
was
private
cloud
as
a
service
compatible
with
amazon
web
services.
Back
in
the
day,
when
you
know
nobody
understood
cloud
as
such,
then
we
as
a
team,
started
with
a
multi-cloud
management
before
all
these
in
a
market
talking
about
multi-cloud.
E
So
we
had
this
multi-cloud
management
to
layer
on
top
of
the
typical
infrastructure,
you
know
be
it
on
aws,
obviously
on
eucalyptus
or
azure,
and
then
open
stack
all
right,
then,
when
we
look
at
all
of
that,
you
know
everyone
was
worried
about
okay.
So
I'm
going
to
put
this
virtual
machine
over
from
one
cloud
infrastructure
to
another
or
vmware
to
aws.
E
Those
aspects
started
back
in
like
2011
and
2012,
but
what
really
mattered
is
that
you
know
if
you
don't
have
the
applications
properly
running
on
these
platforms.
I
mean
these
are
like.
Initially,
these
all
hypervisors
and
you
know
start
putting
the
applications
over
and
workloads
will
start,
but
when
you
think
about
applications
as
such,
they
were
going
to
become
a
much
more
distributed
on
the
software
defined
infrastructures
right
kubernetes
wasn't
there
at
all.
E
I
mean
kubernetes
started
back
in
2014,
so
we
started
working
with
kubernetes
when
kubernetes
was
0.9
like
there
was
not
even
the
you
know,
the
the
networking
component,
the
flannel
was
just
getting
introduced,
so
we
realized
that
applications
are
going
to
be
the
important
component
to
integrate
with.
When
I
talk
about
applications,
I'm
not
talking
about
this
client
server.
You
know
you
have
this
virtual
machine
running
front
end
of
the
application
and
then
there
is
database
right.
Maybe
three
tier
application.
E
That's
not
the
case,
and
that
is
not
how
we
envisioned
the
world
will
be.
The
world
will
be
with
applications
with
a
lot
more
distributed
components.
Micro
services
were
not
even
born
at
the
time
we
started
talking
about
it,
that's
where
the
application
component,
so
application
to
infrastructure
management,
and
the
next
is
basically
an
automation
component.
Anything
to
do
with
automation
is
an
automation,
so
just
back
to
that
that
circling
back
michael,
it's
about
applications
and
then
automation
behind
the
scene,
everything
else
in
between.
So
that's
why
a
tranex!
It's
it's
a
name.
D
All
right,
okay,
that
makes
sense.
I
I
mean
things
around
here,
change
so
fast,
sorry,
you
know
you
were
saying
you
know
before
kubernetes
before
microservices,
when
you
were
saying
that
I
was
thinking,
I
was
thinking
openstack
back
when
geez.
What
was
it?
I
don't
know
2000
and
2014
when
openstack
was
the
rage-
and
you
know
the
openstack
summit
was
kind
of
the
the
kubecon
of
the
day.
E
E
It's
so
we
can
work
on
basic
infrastructure
as
a
service
and
then
kubernetes
on
top
of
that,
then
the
application
right.
So
we
look
at
the
entire
spectrum.
You
know
going
back
to
openstack
days.
You
know
I
talked
about
eucalyptus,
you
could
this
pretty
much
created
in
a
sense
open
stack,
because
you
know
eucalyptus
had
the
licensing,
different
licensing
aspects
right,
that's
how
openstack
was
born
so
with
our
multi-cloud
management.
At
the
time
we
supported
openstack,
but
the
issue
has
always
been.
Where
do
you
get?
E
This
hardware
stack
right
in
order
to
run
openstack,
you
need
the
hardware,
and
that
is
when
you
know,
when
you
look
at
the
public
cloud
infrastructure,
you
get
the
entire
stack
right.
It's
a
hyper
converged
infinitely
scalable,
hyper,
converged,
infrastructure
and
openstack
created
the
market
in
terms
of
putting
that
software
definition
for
your
server
infrastructure,
network
infrastructure
and
the
storage
infrastructure.
That's
what
openstack
brought
in
so
our
solution.
E
You
know
pretty
much
works
on
all
the
public
cloud
infrastructures
and
you
know
given
openshift
works
on
all
of
these
three
as
a
services
we
work
with
every
one
of
them
and
vmware
is
another
component
right.
So
if
openshift
works
on
vmware,
we
work
on
vmware
plus
openshift,
and
we
look
everything
from
that
application
down
the
infrastructure.
We
take
care
of
that.
D
Okay,
well:
how?
How
are
things?
How
are
things
going?
We're
certainly
I
mean
normally
it'd,
be
like
okay,
you're
fours
four
years
into
it.
What's
your
exit
strategy,
you
know,
are
you
gonna
ipo
you,
you
know
how
are
you
doing,
but
I'm
talking
about
now,
the
way
you
know,
I'm
sitting
here
in
my
house
wearing
clothing
that
that
normally
I
wouldn't
be
wearing
you're.
Presumably
you
know
in
your
in
your
house
in
holliston.
D
How
has
the
last
eight
months
had
an
impact
on
how
you
folks
are
seeing
the
need
for
what
you're
offering
is
it
you
know?
Are
you?
Are
you
seeing
a
drastic
decrease
in
sales?
Has
it
changed
the
the
way
sales
are
happening?
Is
it
changed
the
way,
and
you
know
the
types
of
solutions
that
customers
need,
given
that
we're
probably
going
to
be
in
this
in
this,
you
know
work
mode
for
quite
some
time.
E
To
be
frank,
michael,
it
has
been
a
at
the
positive
side
for
us.
You
know
for
a
lot
of
folks.
It
has
been
a
tremendous
kind
of
pressure
financially
as
well
as
otherwise.
I
think
this
pandemic
has
created
or
opened
the
eyes
of
a
lot
of
enterprises.
E
You
see
this
is
it's
it's
all
becoming
cloud
first.
So
when
we
talk
about
cloud
application
resilience,
it
started
resonating
more
this
year
compared
to
last
year,
so
we
pretty
much
doubled.
Our
number
of
customers
doubled
our
revenue
this
year
and
we
have
closed
multi-million
dollar
deals
this
year
and
gartner
has
recognized
as
a
cool
vendor
in
this
space
and
we
certified
our
product
on
the
marketplace.
E
The
red
hat
marketplace-
and
you
know
it
has
been
a
a
good
year
for
us
overall,
and
we
are
lucky
to
be
in
this
position,
so
it
has
been
a
positive
one
for
us
and
hopefully
for
others
in
this
space.
Hopefully
they
also
reflect
the
same
feeling,
but
you
know
I.
I
cannot
talk
about
other
folks
that
are,
you
know,
dealing
with
the
data
center
infrastructure.
D
Sure
yeah,
thanks
for
reminding
me
I
forgot
to
put
in
a
gratuitous
plug
for
for
for
my
team
here,
we've
been
working
with
you
folks,
so
my
team,
at
red
hat.
We
work
with
software
partners
to
help
them
test
and
you
know
certify
their
their
software
as
a
red
hat
certified
container.
D
You
know
when,
when
you're
talking
about
containers
and
customer
production
use
of
containerized
applications,
you
want
to
make
sure
that
the
components
inside
that
container
are
fully
supportable.
So
you
know
people
can
go
out
and
build
a
container
using.
You
know,
upstream
community,
linux
or
two
fred's
and
a
shed
linux,
but
if
that
container
breaks
they
get
to
keep
both
pieces.
So
you
know
we
we
work.
My
team
specifically
works
with
software.
D
Vendors
to
you
know,
take
and
ensure
that
the
components
inside
that
container
are
you
know:
red
hat
enterprise
linux,
you
know
components
plus
their
software
tested
together,
certified
and
then
so.
Whenever
there's
a
cve
that
comes
out,
you
know
the
the
customers
who
are
using
those
containers
know
that
you
know
that
the
defect
repair
will
will
happen
and
the
changes
will
make
it
back
into
the
upstream.
You
know
community,
so
they
won't
be
left
holding
the
bag,
and
then
I
think
you
guys
also
have
an
operator
right.
E
D
E
Yeah,
so
it's
basically
in
in
terms
of
introducing
our
technology
right.
It
is
the
application
resilience
that
matters.
So
when
you
think
about
an
application,
namespace
or
multiple
namespaces
that
make
up
an
application
environment,
there
are
so
many
moving
parts
to
these
applications
right
anytime,
when
you
deploy
a
new
container
or,
and
the
parts
automatically
vacate
from
a
node
to
another
node,
your
application
connectivity
keeps
changing.
E
Now
when
you
think
about
holistically
protecting
that
environment,
I'm
talking
about
you
know,
I
am
obviously
comparing
with
the
regular
backup
and
recovery
type
of
scenarios
and
contrasting
that
with
application
resilience
right.
So
when
you
talk
about
application
resilient,
that
is
a
higher
order
compared
to
the
infrastructure
level,
talk
which
is
about
hey.
How
do
I
back
up
that
node?
How
do
I
backup
this
particular
vm
that
is
running
a
set
of
parts
that
is
running
a
set
of
containers?
E
So
when
you
think
about
that,
you
know
it's
a
different
approach,
but
the
problems
are
same,
except
that
we
look
at
the
entire
environment
and
their
name
space,
and
also
we
look
at
from
the
entire
open
shift
perspective.
So
that's
the
difference
and
from
that
point
on
you
know
you
talk
about
any
sres
that
are
managing
application,
specific
infrastructure.
E
You
know
combining
multiple
micro
services
running
on
openshift
platforms
and
not
only
that,
but
also
looking
at
when
the
openshift
platform
itself,
because
it
relies
on
underlying
infrastructure
right.
If
that
underlying
infrastructure
runs
into
problem,
then
entire
openshift
cluster
runs
into
problem
right.
So
we
approach
that
way
as
opposed
to
regular
backup
and
recovery.
D
And
then
you
said
that
the
operator
is
available
for
openshift
4
and
is
in
the
red
hat
marketplace.
The
red
hat
marketplace
is
a:
is
a
b2b
web
store?
If
you
will
that
we
launched,
I
think
back.
It
was
in
may
of
this
year,
so
you
folks
have
a
commercial
offering
of
your
of
your
software.
It's
it's
as
an
operator,
it's
in
the
catalog,
so
people
can
go
there
and
click
to
buy.
E
Absolutely
so
we
worked
with
red
hat
marketplace
team
from
the
beginning,
I
think
when
they
approached
we
had
the
catalog
operator
already
certified
at
the
catalog
world,
so
we
could
go
to
open
shift
platform
can
search
for
the
catalog
available,
but
from
a
commercial
aspect
perspective
we
waited
until
the
red
hat
marketplace
became
ga.
So
as
soon
as
we
were
one
of
the
core
vendors,
when,
when
we
looked
at
you
know
the
openshift
marketplace
at
the
time,
like
only
four
five
vendors.
E
Now
it's
like
hundreds
and
hundreds
right,
so
the
we
worked
with
the
team
to
launch
it
certified,
go
through
all
the
legal
aspects
of
it.
So
all
of
that
background
information,
it's
all
taken
care
of
by
red
hat.
So
what
customers
can
directly
go?
Is
that
go
through
the
marketplace,
search
for
application
resilience
or
even
you
can
search
for
backup
and
recovery,
locate,
apprenix
and
then
go
ahead
and
buy
it
directly.
E
All
the
commercial
aspects
will
be
handled
by
that
hat,
so
you
know
from
an
sr
perspective
for
them,
it's
straightforward
right
directly
from
the
open
shift
marketplace.
Commercial
aspects
are
already
handled
all
the
pricing.
All
of
that
is
very
straightforward.
It's
per
node,
we
don't
complicate
any
of
that,
and
billing
comes
through
as
part
of
the
marketplace.
So
you
don't
have
to
worry
about.
A
panic
is
getting
built.
None
of
that
right.
It's
very,
very
straightforward.
D
D
D
E
D
E
E
C
E
Of
down
time,
for
that
matter,
now,
what's
the
best
way
to
get
back
into
business
after
some
kind
of
a
downtime
happens,
you
know
you
obviously
look
for
backup
tools
or
recovery
tools
and
worst
case
scenario.
You
look
for
you
know,
dr
as
such,
so
from
our
point
of
view,
when
you
look
for
bringing
back
your
environment,
meaning
providing
that
resilience
from
any
downtime.
E
What
we
do
is
that
you
can
recover
any
container
or
any
resource
from
your
cluster
at
any
point
in
time
in
any
region
of
the
cloud
right.
So
when
you
have
that
kind
of
a
powerful
capability
in
a
service,
so
it's
provided
as
a
service,
so
you
don't
have
to
look
at
how
to
run
this
service
from
an
infrastructure
perspective
or
managing
openshift
cluster
and
application
center.
You
don't
have
to
worry
about
any
of
that.
All
you
have
to
worry
about
is.
E
E
D
That
that
that
I
remember
trying
to
get
get
on
the
off
ramp
in
the
mornings
and
it
was
just
backed
up
out
onto
the
highway
but
so
anyways.
I
I
worked
with
some
really
well.
My
job
at
the
time
was
working
with
software
companies
to
help
them
get
their
apps,
supported
to
open
vms,
digital
unix
and
alpha
nt.
If
you
remember
that.
A
D
I
remember
that
there
were
like
some:
there
were
some
really
big
mainstream.
Backup
vendors
I
mean
backup
is
not
new.
Backup
has
been
around
longer
than
us.
So
why
do
we
need
a
pranix?
I
mean
why
you
know
it
backup
and
data.
You
know.
Disaster
recovery
has
been
a
core
piece
of
enterprise
computing
for
as
long
as
the
word
enterprise
computing
has
been
around.
So
what's
different.
Now
why
can't
people
just
use?
D
E
Absolutely
so,
when
you
think
about
backup
has
been
there
forever,
compute
has
been
there,
for
our
storage
has
been
there,
for
our
network
has
improved,
but
the
difference
is
that
applications
are
different.
So
when
you
think
about
applications
that
used
to
run
on
mainframe
backup
was
there
to
protect
the
infrastructure
for
those
applications,
client
server
came
in,
you
know,
pc-based
revolution
came
in,
the
internet
came
in
right,
web
two-way
companies
came
in
and
then
cloud
was
born
to
give
the
enterprises
any
infrastructure
at
any
point
in
time,
infinitely
scalable
type
of
one.
E
So
suddenly,
what
has
happened
is
that
the
application
infrastructures
have
changed.
The
applications
have
become
a
lot
more
distributed
and
lot
more
dynamic,
and
if
you
think
about
open
shift,
what
openshift
is
doing
underneath
is
that
it's
automatically
provisioning
the
infrastructure
for
your
applications
using
cube
specs.
So
you
you
send
in
hey.
I
need
more
storage
at
this
spec
level
right
and
what
it
is
doing
is
that
it
is
eliminating
bunch
of
you,
know,
support
tickets
and
you,
you
kind
of
clarify
the
support
ticket.
Oh,
what
is
the
size
of
the
vm
that
you
need?
E
E
All
of
that
has
changed
when
you
go
to
the
cloud,
so
you
cannot
say.
Oh,
I
am
going
to
take
my
backup
system
that
used
to
work
for
mainframe
and
physical
servers
and
the
vmware
environment
right
and
then
say:
oh
I'm
going
to
use
the
same.
That's
a
fundamental
thing
right!
When
you
go
to
the
cloud,
you
don't
talk
about
vms
anymore.
You
don't
talk
about
storage
anymore.
You
don't
talk
about
networking,
you
don't
talk
about
load
balancer!
You
simply
use
an
api
and
are
even
defined
in
your
spec
cube
spec.
E
What
you
want
to
use
the
infrastructure
automatically
gets
provisioned.
Now
you
think
about
it.
When
you
have
to
do
the
backup
in
that
infrastructure,
you
cannot
obviously
use
the
legacy,
approaches
your
surveys,
because
you
don't
want
to
even
deal
with
infrastructure
anymore
right.
You
want
to
deal
at
the
application
level
and
then
the
rest
in
terms
of
protection
and
recovery
should
be
automatically
taken
care
of.
Just
like
your
kubernetes
right,
you
don't
ask
kubernetes.
Oh,
can
you
please
figure
out
where
that
next
node
comes
from
right,
application
and
sres?
E
You
know
introducing
a
term
sorry
site,
reliability,
engineer
that
really
focus
on
micro
service
or
application
specific
reliability
right,
they
don't
say:
oh,
I
need
to
have
x
size
of
the
vm
provisioned.
For
my
part,
that's
not
the
case
anymore.
So
we
are
asking.
Why
should
backup
breeze
that
way?
And
when
you
bring
in
backup
dr
used
to
be
a
separate
product
altogether?
In
order
to
do
vr,
I
should
be
able
to
take
the
entire
application
infrastructure
and
bring
it
up
in
some
other.
E
That
is
not
the
case
anymore.
It's
like
my
application,
is
running
now
the
expectation
for
these
applications
like
when
I'm
running
on
amazon.com
amazon
web
services
right.
I
expect
the
reliability
to
be
like
amazon.com.
Okay,
am
I
not
running
on
aws,
okay,
so
I'm
running
on
google
cloud,
okay,
gmail
is
running
all
the
time.
Right.
Outlook.Com
is
running
all
the
time,
so
my
expectation
when
I
go
to
the
cloud
is
that
my
applications
should
continue
to
run.
Nobody
talks
about.
Oh
my
application.
Downtime
is
x
and
sre
struggle.
E
Look
at
the
moving
parts
within
kubernetes,
environment,
too
many,
and
then
now,
if
you
put
in,
can
it
can
you
put
in
more
sres
because
there
are,
there
is
more
moving
parts,
it's
impossible
to
do.
So
if
you
rely
on
the
legacy,
backup
and
recovery
you'll
be
let
down
at
some
point
you
can.
You
can
kind
of
work
through
it,
but
you
will
have
to
do
so.
Many
kind
of
you
know
gymnastics
if
you
will
to
figure
out
the
moving
part
with
our
planets.
E
You
don't
do
any
of
that,
because
your
system,
your
application
environment,
is
protected.
You
go
and
mess
around
your
firewalls
right
security
groups
between
the
nodes.
You
go
and
configure
your
gateways,
you
go
and
configure
load
balancer
or
your
config
maps
within
your
open
shift
is
changing
all
the
time.
E
So
how
do
you
keep
track
of
these
configurations,
along
with
your
data
right
and
that's
why
we
come
in?
Don't
have
to
worry
about
your
changing
configuration
in
your
production
environment.
Just
take
it
as
somebody
is
monitoring.
All
of
somebody
is
backing
up
like
over
a
period
of
time
like
a
time,
machine,
application,
environment
time
machine
is
captured
along
with
your
data,
along
with
your
containers.
So
at
any
point
in
them
you
can
recover
them.
E
It
is
software
as
a
service,
it
runs
on
one
of
the
cloud
platforms
with
all
the
you
know,
compliance
and
stock.
All
of
that-
and
we
provide
you-
know-
enterprise
level
multi-factor
authenticated
signed
through
ssos,
all
of
the
good
things
that
a
typical
enterprise
look
for,
and
you
know
you
don't
really
worry
whether
as
a
service
goes
down
or
not.
You
know
it's
a
funny
thing
that
a
partner
came
in
on
the
other
day
and
he
had
a
backup
software
that
he's
using
for
protecting
some
of
his
client.
E
The
first
thing
that
the
partner
did
is
this
backup
software
runs
on
the
cloud
right
runs
on
one
of
the
cloud
platforms.
He
used
a
pranix
to
protect
the
backup
software
right.
If,
if
the
backup
software
as
an
application,
runs
on
the
cloud
that
application
could
go
down
any
time
right,
so
he
used
apprenix
to
provide
the
resiliency
for
the
backup
software
itself.
Can
you
can
imagine
that
it's
kind
of
you
know?
That's.
D
E
D
Good,
creative
thinking,
so
is
there
how
if
you
have
hundreds
or
thousands
of
nodes
and
pods-
and
you
know,
applications
distributed
across
multiple
clouds
because
you
know
openshift
allows
people
allows
developers
to
build
it
once
deploy
everywhere.
So
you
know
being
sort
of
the
layer
of
abstraction
between
all
the
clouds.
D
E
E
Now
you
are
making
those
applications
much
more
distributed,
so
that
kubernetes
and
openshift
can
very
easily
scale
horizontally
right
and
making
it
easier
for
applications
to
run.
Now.
Your
front-end
components
are
great
in
terms
of
containerization,
because
you
embed
all
your
dependency.
Libraries,
nice
packaging
mechanism
called
containers,
and
you
run
them
on
open
shift
using
red
hat
linux
right
perfect.
E
Now,
the
back
end
of
that
application
inevitably
relies
on
data
infrastructure
right.
The
data
infrastructure
is
also
changing.
It
is
not
like
you
know,
a
client
server
applications
relying
on
oracle
database
anymore
right
and
every
microservice
has
its
own
data
infrastructure.
You
could
be
using
mongodb,
you
could
be
using
cassandra
for
something
you
could
be
using
a
very
simple
file
based
database
right
or
you
could
be
using
cloud
rds
relation
database
services.
E
This
databases
need
their
own
maintenance
schedules
and
you
know
upgrades
and
all
of
that
and
a
lot
of
these
enterprises
are
offloading
that
work
right
after
moving
to
the
cloud
to
the
cloud
platform
providers
right,
so
you
have
cloud
sql,
you
have
rds,
you
have
aurora,
you
have
serverless
databases,
aurora
serverless
right.
You
have
redshift,
you
have.
You
know
you
have
unlimited
possibilities
for
the
data
infrastructure
right
and
then
you
end
up
consuming
other
services.
E
E
Suddenly
you
have
something
called
a
hybrid
application
right,
an
application
dynamic
enough
and
allowing
the
devops
pipeline
to
change
and
deploy
the
changes
or
the
feature
requests
really
really
fast
at
the
container
level.
And
then
you
have
the
data
infrastructure
in
the
back
end
now
going
slower
a
different
pipeline,
different
devops
pipeline
right,
changing
differently.
Now,
if
you
think
about
that
environment
is
what
we
call
it
as
a
hybrid
application.
E
D
When
you
say
sre
I
mean
I,
I
love
acronyms
just
as
much
as
the
next
person,
but
for
everyone.
Why
don't
you
share
with
us
what
what
an
sre
is.
E
Oh
yeah,
thank
you
for
asking
that
so
sites
reliability,
engineering
as
a
concept
as
a
a
new
way
of
managing
the
application
infrastructures
right.
E
You
saw
operations,
the
system,
managers,
system,
operators
and
so
on.
Right
then,
cloud
operations
came
in
cloud
operations.
You
know
is
the
word
taking
care
of
the
vpc
virtual
private
cloud,
taking
care
of
the
networking
infrastructure
and
all
of
that
and
slowly
the
world
has
moved
to
application
specific
infrastructure
call
it
as
a
micro
services
infrastructure
and
a
particular
site.
Reliability
engineering
person
will
take
care
of
that
application,
specific
infrastructure
providing
reliability
for
that
application
infrastructure.
This
concept
overall
came
in
because
google
started
kind
of
pointing
out.
E
You
know
how
they
ran
their
infrastructure,
their
their
google
infrastructure
in
the
back
and
right
for
applications
and
industry
has
caught
on.
Particularly
you
know,
along
with
you
know,
kubernetes.
E
E
Absolutely
so
let
me
let
me
share
my
screen
here.
I'm
going
to
share,
I
share
everything.
E
D
E
D
E
Absolutely
michael,
so
you
know
I'm
going
to
just
introduce
the
concept.
I
know
we
talked
a
certain
aspect
of
cloud
application:
resilience
how
the
old
backup
and
recovery
systems
are
not
sufficient
and
then,
given
the
time
frame,
you
know
I
looked
at
the
possibility
of
demoing
live.
You
know.
I
typically
do
that
during
the
webinar.
For
this
particular
situation,
I
have
recorded
one
of
the
effective
application
container
application
webinars,
so
I
will
cut
through
that
at
the
end
of
this
presentation.
E
So
if
you
look
at
what
we're
talking
about,
I
need
to
take
an
open
shift
environment.
Now
people
go
to
openshift
or
overall
container
based
applications
and
they
are
going
to
grow
multifold
and
all
these
namespaces
and
various
application
environments.
E
The
potential
possibilities
are,
you
know,
fast,
application
delivery,
that's
no
brainer,
not
a
problem
and
better
infrastructure
utilization.
You
know
when
you
can
pack
so
many
containers
on
a
single
virtual
machine.
That's
perfect!
The
utilization
always
kind
of
changes.
You
know
the
virtualization
era
promise
the
utilization
of
hardware
better,
but
in
the
end
you
know
it
didn't
it
achieved
a
lot.
It
simplified
a
lot,
but
at
the
end
of
the
day,
from
an
application
perspective,
the
control
wasn't
completely
given
to
the
application
developers.
E
So
you
know
open
shift
and
the
underlying
kubernetes
environment
promises
it
along
with
containers,
and
they
are
all
running.
As
I
mentioned
on
the
cloud
infrastructure,
I'm
putting
two
of
the
infrastructure.
I
didn't
have
time.
You
know
space
for
adding
more,
but
if
you
think
about
it
that
infrastructure
believe
it
or
not,
runs
on
a
lot
of
the
services
that
are
actually
honestly
taken
for
granted.
Nowadays
right,
you
know,
people
take
it
for
granted
that
I
can
create
an
instance.
E
E
That
has
to
grow
and
then,
when
you
put
open
shift
in
the
open
shift
has
to
scale
that
machine
level
scaling
for
the
nodes
so
that
your
parts
can
be
vacated
at
any
point
in
time
and
move
over
as
a
bunch
to
another
node,
because
that
particular
node
is
running
out
of
infrastructure.
So
that
is
the
complexity
of
it.
E
D
E
No,
so
what
we
do
is
that
we
directly
go
to
the
cloud
account.
If
you
will
right,
we
ask
for
a
few
permissions
from
our
service
perspective.
It's
very
standard
in
terms
of
the
discovery
and
we
have
been
certified
by
the
cloud
providers.
Google,
as
well
as
aws
from
a
security
perspective.
It's
very,
very,
very
standard
permissions
level
service
account
level
permissions.
Once
you
give
us
those
permissions,
we
simply
go
and
discover
all
the
infrastructure.
I
will
show
you
in
a
moment
how
the
entire
thing
kind
of
gets
discovered.
E
Great
question:
absolutely
so,
when
you
think
about
that
kind
of
a
complex
application
environment
think
about
the
downtime
situation
right,
it's
on
an
average
again
gartner
puts
his
numbers
concert.
Gartner
has
always
been
a
very
conservative
firm
when
you
think
about
your
application
here
it
could
be
your
platinum
application
and
that's
taking
care
of
millions
and
millions
of
dollars
of
revenue
right
or
any
of
these
tiers.
E
E
You
know
you
will
have
to
think
about
what
all
the
bigger
problems
that
introduce
the
downtime
in
the
cloud
environment.
You
know
it
could
be
as
simple
as
a
bad
pipeline
devops
pipeline
coming
in
and
changing
it.
You
know
you
cannot
be
watched
watching
over
every
one
of
the
changes
that's
coming
into
your
environment
or
it
could
be
cloud
misconfiguration.
Oh
somebody
changed
the
firewall
and
when
was
that
it
was
like
an
hour
ago
or
15
minutes
ago,
or
yesterday
everything
seemed
to
work.
E
Last
week
I
went
on
vacation
thinking
that
everything
was
fine
and
dandy,
and
then
I
come
back
now.
The
application
is
down.
I
have
no
idea
what
really
happened
and
then,
when
you
operate
particularly
on
the
public
cloud,
I
noticed
that
openshift
released
as
a
service
on
aws.
I
saw
the
news
yesterday
so
think
about
that.
Most
of
the
complexity
of
managing
that
kubernetes
environment
openshift
environment
is
taken
over
by
red
hat
as
a
service.
That's
perfect,
but
it
is
operating
on
a
shared
public
cloud
infrastructure.
E
We
meaning
that
you
have
nice,
enabled
problem
right.
You
know
on
the
side,
it
could
be
some
kind
of
bitcoin
based.
You
know,
mining
going
on.
God
only
knows
right.
Nobody
knows
what's
happening
on
the
public
cloud
infrastructure,
so
you
know
it's
in
the
same
network,
same
rack
and
you
are
going
to
be
affected
by
the
noisy
neighbors.
So,
overall,
if
you
think
about
it
right,
your
downtime
could
cost
anywhere
between,
depending
on
the
tier
of
application
anywhere
between
you
know,
140,
to
300
to
500
k
per
hour.
E
So
how
would
your
application
come
up?
And
on
top
of
that,
they
introduce
this
lock-in
mechanism?
They
are
capturing
the
data
right
in
the
data
center.
That
is,
okay
is
there's
so
many
storage,
vendors,
you
pull
the
data
and
then
you
keep
it
in
a
backup
environment,
secondary
storage.
If
you
will,
but
when
you
go
to
the
cloud
you
are
running
on
the
same
storage,
engine
pbs
or
you
know,
persistent
discard,
blob
storage
and
why
suck
the
data
out
and
put
it
in
another
data
store
and
slow
down
the
entire
process?
E
When
your
application
is
done,
it
doesn't
make
any
sense.
So,
on
top
of
that,
you
don't
really
understand
the
dependency
mapping
between
the
components,
because,
like
you
mentioned,
why
can't
we
use
client
server
error
created,
backup
and
recovery
tools
in
the
cloud
the
native
world,
I
mean
when
the
infrastructures
change,
there
was
no
kubernetes.
E
There
was
no
open
shift
at
the
time
right
and
now,
there's
openshift
with
the
changing
infrastructure.
There's
no
point
in
using
similar
architecture
or
just
because
I
protected
virtual
machine.
Now
I
can
protect
containers
that
doesn't
mean
anything
because
containers
alone
they
are
not
living
in
their
own
isolated
world
containers
are
connected
all
the
time.
So
the
parts,
the
config
maps,
the
secret
that
you
have
to
maintain
and
for
I'm
not
even
talking
about
the
certificate
nightmare
that
you
have
to
manage
right.
E
E
You
need
to
use
infrastructure
as
code
who
writes
those
infrastructure
escort.
That's
too
complex,
that's
why
gartner
is
saying
you
know,
even
if
you
started
with
some
kind
of
a
backup
right,
you
know
it's
better
to
look
at
your
entire
application
environment.
Look
at
from
an
application,
resiliency
perspective
right
and
things
will
become
a
lot
more
clearer
and,
yes,
you
have
to
do
backup.
Yes,
you
have
to
do
disaster
recovery.
These
are
like.
E
Yes,
it's
a
table
sticks
right.
Nobody
at
some
point.
The
people
who
start
talking
about
backup
right
backup
is
a
chore.
I
have
to
do
it.
Let
somebody
else
do
it
some
service.
Take
care
of
that.
Let
me
worry
about
application
resilience
right.
So
that's
the
aspect
that
we
talk
about
and
when
you
think
about
operanics
forget
about
all
of
that.
You
know
infrastructure,
infrastructure
level,
talk
nomenclature,
you
know
you
don't
talk
about
vms.
You
don't
talk
about
nodes.
You
talk
about.
E
How
do
I
protect
that
application?
Environment
right
virtual
machines,
containers,
combination
of
those?
That's
what
because
hybrid
applications,
containers
plus
infrastructure,
open
shift,
running
your
container
world
and
the
rest
of
the
infrastructure.
So
we
look
at
when
we
discover
that
infrastructure.
We
discover
everything
within
that
application:
environment,
the
connectivity
between
the
components
of
that
environment
right
and
when
you
recover,
you
can
click
and
recover
this
particular
resource
or
the
entire
environment
at
any
point
in
time.
This
is
what
I
call
it
as
an
application
environment
time
machine.
E
C
E
Where
I
have
to
think
about
multiple
zones,
every
zone
is
a
data
as
a
huge
data
center
right
and
when
you
think
about
spreading
out
that
kubernetes,
even
the
open
shift
infrastructure
spreading
out
between
three
different
zones
when
the
cluster
goes
down,
your
entire
application
environment
is
down
right.
So
how
would
you
bring
up
that
entire
open
shift
environment
in
another
region
of
the
cloud?
E
What
are
the
connectivity
pieces
around
it?
You
will
have
to
take
care
of
that
right.
So
that's
what
application
resiliency
is
all
about
when
you
think
about
doing
that,
guess
what
you
know
we
released
in
such
a
way
that
it's
a
sas
meaning
one
and
two
are
one-time
activities.
You
subscribe,
go
to
the
marketplace,
openshift
marketplace
or
any
other
cloud,
provide
marketplaces
and
give
permission
to
discover
it's
read-only,
and
then
you
simply
add
your
policies.
I
want
to
protect
every
hour.
I
want
to
protect
every
15
minutes
I
want
to
protect.
E
You
know
every
week,
retain
it
for
seven
or
four
weeks
or
five
weeks
depending
on
a
month
and
then
protect
every
month.
Retain
it
for
12
months,
protect
every
year
return
for
five
years.
All
of
the
traditional
kind
of
a
policy
oriented
mechanisms
you
simply
key
in
policies
operating
takes
care
of
it,
creating
that
application
environment
time
machine.
E
Then
you
simply
go
and
pick
up
that
right
and
having
done
that
now,
you
can
graduate
to
a
level
where
you
talk
about
continuous
resilience
right
now
that
you
have
achieved
that
backup
and
recovery
based,
you
know,
take
care
of
our
practice.
Worry
about
my
dynamic
ever-changing
application
in
mind,
a
protected
place
once
you're
done
with
that
you
graduate
to
the
pipeline
oriented
approach,
so
you
already
have
cicd
established
with
your
tools,
go
ahead
and
use
any
cicd
pipeline
tools
and
structure.
There
are
so
many
of
them
right.
E
Openshift
comes
with
a
packaged
approach
as
well,
so
you
could
use
a
built-in
open
shift,
a
ci
cd,
a
tool
set
and
what
you
could
build
with
operanics
is
continuous
resilience,
so
continuous
integration,
continuous
deployment
or
delivery.
Along
with
that
you
introduce
you
could
introduce
continuous
resilience
with
a
single
line
of
code
integrated
into
your
pipeline.
That
will
take
care
of
protecting
your
environments
at
any
point
in
time
you
could
roll
back.
Even
if
you
had
a
bad
deployment,
you
could
recover
it
or
completely
fail
over
to
another
region
right.
E
All
of
that
will
be
taken
care
of.
So
from
a
demo
perspective.
I
cooked
up
this
on
top
of
gcp,
because
we
have
traditionally
we
use
gcp.
You
could
do
it
exactly
that
on
aws
or
azure,
not
a
problem
for
demo
purposes.
Again,
it's
just
a
demo
purposes.
We
have
a
kubernetes
based
application,
running
microservices
here
and
you
have
the
data
infrastructure
running
cloud.
Sql,
like
I
mentioned
it's
a
hybrid
application.
E
Then
you
have
some
mongodbs
and
things
like
that
running
in
one
of
the
regions
of
the
cloud
with
the
cloud
load
balancer
connecting
the
applications
is
running,
the
data
is
copied
to
the
other
region,
so
thinkness
or
asynchronous
copy,
totally
fine.
Make
sure
that
you
know
it's
copied
now.
If
that
thing
comes
down
our
resources
come
down,
you
could
bring
it
back
again.
E
So
that's
the
capability
that
we
have
in
terms
of
creating
an
application
resilience
and
providing
that
application
resilience
when
you
operate
your
applications
on
openshift,
plus
the
data
infrastructure
in
the
backend
right
with
that
you
know
now,
if
you
look
beyond
the
legacy
approaches,
you
know
mainframes,
perfect
infrastructure.
For
that
time,
client
server
break
virtualization,
based
vmware
environment,
perfect.
They
are
satisfied
to
available.
When
you
go
to
the
cloud,
it's
different,
a
lot
more
dynamic
when
you
think
about
10,
vm
based
environment.
E
Now
we
are
talking
about
80
different
resource
right,
so
you
know,
when
you
go
to
the
open
shift
level,
then
you
have
lot
more
config.
If
you
look
at
the
config
maps,
it's
very
complicated
right
and
a
lot
of
moving
parts
and
the
many
many
pipelines
are
introducing
changes
to
that
infrastructure
right.
When
you
think
about
that,
you
don't
not
only
think
about
just
the
region,
one
you
can
now
have
one
region,
two
with
multiple
zones
or
you
can
even
have
additional
protection
in
another
service,
account
completely
right.
Think
about
that
level
of
resiliency.
E
That
is
unheard
of
in
the
data
center
days
and
unheard
of
even
in
the
cloud
days,
and
unless
you
know
how
to
make
use
of
all
these
wonderful
capabilities
right
and
that's
what
it
is
in
terms
of
application
resilience,
so
you
can
get
started
very
easily,
but
before
that,
I
let
me
also
give
you
that
entire
demo,
so
I
recorded
the
demo
if
you
think
about
it.
This
is
the
application
that
you
know
it's
a
very
well-known
distributed
application.
E
Thanks
to
we
worked
the
open
sourced
it,
you
can
go
and
buy
stocks
all
day
long.
You
know
tone
stocks
right
and
you
know
that
application
needs
to
run
continuously.
So
if
you
look
at
the
shopping
cart
here,
there
are
267
items
running
and
we
spin
off
an
automated
in
a
population
of
that
sox
shopping
cart.
So
you
know
you
know
we
automate
using
a
test
script.
E
It
keeps
filling
up
that
shopping
cart
item,
so
to
simulate
that
your
e-commerce
application
runs
continuously
and,
in
the
back
end,
that
application
looks
very,
very
dynamic.
If
you
look
at
the
complexity
of
the
application,
that's
it
runs
on
the
data
infrastructure
with
the
dual
node,
but
that
application
has
so
many
moving
components.
E
If
you
look
at
the
the
dynamism
of
that
application
is
pretty
complex
right,
it
keeps
changing.
If
you
look
at
the
processes,
that's
the
type
of
application
that
you
see
on
the
application
environment
perspective,
but
if
it
comes
down
to
protecting
that
environment,
as
I
mentioned
you
know,
sock
shop
is
that
application
name.
You
have
the
backend
infrastructure,
it
runs
on
the
instances
that
cloud
platform
provides
and
you
have
the
load
balancers.
E
You
have
the
cloud
sql,
which
takes
care
of
the
back
and
data
infrastructure,
and
on
top
of
that,
you
have
the
container
resilience.
So
if
you
look
at
operanics,
you
have
container
resilience
that
provides
the
resilience
for
your
open
shift
based
applications,
dynamic
applications
and
the
external
assembly,
consisting
of
your
data
infrastructure,
is
protected
using
the
cloud
reasons,
so
both
of
them
will
be
protected
using
a
pranax.
This
is
what
I
call
as
a
hybrid
application
and
in
that
hybrid
application
you
think
about
it.
E
You
know
at
any
point
in
time
in
the
timeline,
we
call
it
as
an
application.
Environment
timeline,
operandix
automatically
captures
that
timeline,
and
you
can
go
and
say
recover.
My
entire
cluster
recover
the
entire
cluster
in
on
the
east
coast.
This
application
runs
on
the
central
region
and
we
recover
on
the
east
coast
and
then
get
the
external
service
at
that
particular
point
in
time.
E
So
when
you
think
about
this
very
complex
and
so
many
moving
parts,
so
many
different
configurations
right
and
operations,
folks
do
not
have
time,
even
if
you
put
the
operators
together
on
the
open
shift,
you
will
have
to
deal
with
bunch
of
operators
right
with
a
pranax
single
operator
connected
to
our
service,
takes
care
of
all
the
complexities
for
you.
So
when
an
application
is
done
all
of
the
back
end,
if
you
think
about
it,
cumulative
engine
in
the
back
and
automatically
brings
back
then
from
a
recovery
perspective.
E
It
comes
back
in
about
you
know
about.
You
know
two
to
three
minutes
and
it
writes
all
the
infrastructure
score.
So
if
you
think
about
it
turn
off
infrastructure.
This
is
a
very
simple
application.
It
has
about
less
than
thousand
lines
of
infrastructure
score
that
is
automatically
written
and
brought
back
up
as
a
single
unit,
so
that
entire
thing
comes
back
completely
on
the
recovery
side.
So
all
your
applications
will
come
up.
E
D
Yep,
you
were
talking
about
the
importance
of
time
in
in
your
demo
right
there.
I
do
want
to
point
out
that
we
have
just
a
little
under
two
minutes
left.
E
Got
it
I,
I
was
rushing
through
this
demo
part
michael,
thank
you
for
pointing
me.
You
know
we
have
only
two
minutes,
so
you
know
the
point
that
we
want
to
convey
is
that
you
are
putting
together
as
a
customer
of
open
shift,
a
complex
and
dynamic
application
infrastructure
right,
consisting
of
many
containers
consisting
of
many
namespaces,
maybe
may
spread
out
across
many
different
clusters
right.
E
You
know
I
have
seen
hundreds
and
hundreds
of
clusters
in
an
enterprise
environment
right,
it's
not
just
one
cluster,
I
learn
anymore
and
when
you
run
that
kind
of
a
complex
application,
that
is
one
level
of
complexity.
You
know
you
don't
have
to
worry
about
the
skill
set
for
your
sras
or
operations,
folks,
second
of
all,
think
about
the
application
complexity
going
down
into
the
data
infrastructure
right.
E
Think
about
your
application,
basically
right
when
you
think
about
application
resilience,
think
about.
Can
I
recover
any
of
my
containers
at
any
point
in
time
any
of
my
cloud
resources
at
any
point
in
time
in
any
region,
and
how
fast
can
I
recover
really
defines
the
capability
for
your
teams
to
recover
from
the
downtime
and
then
downtime
really
costs
a
lot
of
money
depending
on
your
application
types.
So
you
know
that's
the
difference
that
apranix
brings
in
going
to
go
back
to
that
naming
michael.
E
That
operanics
means
applications
to
start
with
right
and
we
look
from
the
application
down
to
the
infrastructure
and
how
we
can
provide
application
resilience
is
all
we
talk
about,
breathe
and
die
on
right.
Just
that,
so
you
know
yes,
backup
is
necessary.
We
take
care
of
that.
Yes,
disaster
recovery
is
necessary.
We
take
care
of
that
right,
and
all
you
have
to
worry
about
is
applications.
E
Is
my
application
resilient
or
not?
So
let
me
stop
here
and
see
if
there
are
any
other
questions
from
you
michael.
This
has
been
a
wonderful
experience.
Talking
to
you.
D
I
don't
have
any
other
questions.
I
I
know
that
I
kept
interrupting
you
here
and
there,
but
I
was
it
was.
It
was
pretty
interesting,
pretty
interesting
talk
track.
So
I
do
we
are.
We
are
running
out
of
time
here,
so
I
just
want
to
put
up
the
our
call
to
action
slide
here.
D
Please
stop
by
the
red
hat
marketplace:
you
can
you
can
check
out
the
apranix
certified
operators
there
for
openshift4
in
the
marketplace?
Govind
rangasami
ceo
of
aprenix,
thanks
for
coming
today.
I'm
really
glad
that
you
were
able
to
be
part
of
our
show
here
on
the
openshift
container
briefings
and
I'm
mike
waite
from
red
hat
and
we're
wrapping
up
another
another
show
this
week.
So
thanks
for
coming,
everyone.
D
E
E
It
was
fantastic,
I
think
you
streamlined
it
much
beyond
my
expectations,
so
really
really
appreciate
it.
D
I
well
normally
what
we
do
is
we
get
a
half
an
hour
call
a
week
or
two
in
advance,
and
I
just
kind
of
ran
out
of
time-
and
I
was
talking
with
cyrus
this
morning
and
I
was
like:
do
we
have
his
front
and
back
slides?
Do
we
have
do?
We
have
the
talk
track
and
I
was
like
we
better
get
that
stuff
going
here.
So
anyways
yeah,
thank
you
and
we
will
get
a
the
recording
off
to
you.
We're
gonna.
Take
your.