►
From YouTube: Ask an OpenShift Admin (E88) | GitOps for Admins
Description
GitOps is a popular topic, so popular that we have an entire livestream devoted to it! What if you're a cluster admin - how does GitOps apply to us? For this week's stream we're joined by Harriet Lawrence to understand the basics of GitOps, discuss how it is beneficial for administrators to do cluster configuration, and compare GitOps role/functionality for cluster config against Advanced Cluster Manager (ACM).
Join Andrew Sullivan and Jonathan Rickard Wednesdays at 11 a.m. ET for a full hour explicitly designed to help OpenShift admins. Each episode addresses your questions and covers the different tools that containers need to run in a cloud-native environment.
Want to learn more?
Try Red Hat OpenShift: https://red.ht/3uVxkya
Follow us:
Andrew Sullivan https://twitter.com/practicalandrew
Jonathan Rickard https://twitter.com/jrocktx1
OpenShift https://twitter.com/openshift
Subscribe to Red Hat's YouTube channel: https://www.youtube.com/redhat/?sub_c...
About Ask an OpenShift Admin:
Do you have questions? We want to answer them! Every Wednesday at 11am ET, hosts Andrew Sullivan and Jonathan Rickard sit down to answer your questions about Red Hat OpenShift.
About OpenShift:
Red Hat OpenShift is an open source container application platform based on the Kubernetes container orchestrator for enterprise application development and deployment.
OpenShift on redhat.com https://red.ht/3rYz5ZS
OpenShift on GitHub https://github.com/openshift/
C
B
A
C
A
B
I
feel
you,
you
know
you
and
I've
been.
You
know
for
like
what
a
month
now
I've
been
complaining
about
this
congestion
that
I've
got
so
I
yeah
I
feel
you
I
blame
the
kids
right,
you
know
they're
they
could
they
go
to
those
petri
dishes
known
as
schools
and
come
back
with
absolutely
absolute
monsters.
I
know!
Well,
it's
it's
good
to
be
back
on
the
bicycle
again
right.
B
We
had
an
unexpected
cancellation
two
weeks
ago,
So
for
anybody
who
was
looking
forward
to
our
topic
two
weeks
ago,
which
was
open
shift
for
kubernetes
administrators.
We
have
rescheduled
that
probably
to
next
week,
I
say
it
probably
because
Johnny
can't
tell
but
I'm
looking
at
him
in
my
interface,
because
he's
going
to
be
running
that
one
solo
so
because
I'll
be
I'll,
be
off
site
I'll,
be
at
the
red
hat
Tower
meeting
with
a
whole
bunch
of
IBM
research.
B
Folks,
so
gonna
be
interesting
day
could
be
awesome,
so
yeah,
we'll
I
I,
think
I've
said
like
four
or
five
times
now.
I
need
to
get
that
one
staged
and
I
actually
do
need
to
get
that
one
staged
I
keep
I,
keep
like
I
created
the
graphics
and
I
did
this
thing
and
that
I
haven't
actually
gotten
it
put
into
to
restream
here.
So
that
way,
it's
it's
ready,
but
so
yeah
next
week,
blah
blah
blah
openshift
for
kubernetes
admins.
We
are
our
topic,
man,
it's
like
it.
B
B
So
definitely
if
you
have
an
opportunity
to
go
back,
there
is
a
number
of
blog
posts
that
were
put
out
by
Red
Hats
that
were
put
out
by
IBM
that
were
put
out
by
all
kinds
of
people
on
all
kinds
of
things.
Talking
about
the
stuff
that
happened
there,
openshift
Commons
happened
the
day
before
it
was
a
day
Zero
or
a
day
negative.
One
event:
there's
lots
of
really
cool
presentations
there
as
well.
There
was
one
by
gosh.
B
Now
I
can't
remember
the
name
company
Informatica,
maybe
that
I
thought
was
interesting.
So
they
are
the
first
customer
of
the
hosted
Advanced
cluster
Security
Service,
so
they
are
they're
using
I
want
to
say
Amazon,
eks
and
ACS
together
in
order
to
provide
security
and
all
of
the
things
that
ACS
does
with
whatever
kubernetes
they
choose
in
this
case
eks.
So
I
thought
that
was
really
awesome
and
it
really
that
is
awesome.
B
Really
one
of
those
Testaments
to
I
mean
you
know:
yeah
ACS,
you
know
formerly
stack
rocks,
is
now
a
part
of
red
hat
and
a
part
of
our
portfolio,
and
it
has
an
openshift.
You
know:
Red
Hat,
openshift
Advanced
cluster
security,
but
it's
still
a
community
thing
right.
It's
still
still
able
and
still
possible
to
do
a
lot
of
other
things
with
it.
So
I
thought
that
was
really
awesome.
B
Yeah,
there's
a
there's
lots
of
stuff
going
on.
It
is
despite
the
fact
that
it
is
early
November
right,
normally
we're
kind
of
spinning
down,
and
you
know
everybody's
getting
ready
for
the
holidays.
One
of
the
folks
on
my
team,
the
last
two
years
he
has.
He
has
taken
PTO
from
Thanksgiving
through
the
New
Year.
B
Basically
it
doesn't
take
PTO
in
any
at
all
the
rest
of
the
year,
and
then
you
know
it's
basically
four
or
four
and
a
half
weeks
of
of
working
time
between
those
two
holidays,
man
I
wish
I
could
take
I
I
I.
Try
to
I
try
to
break
my
PTO
up
right.
You
know
you're
supposed
to
use
it
to.
You,
know,
take
breaks
and
recharge
and
all
that
other
stuff,
but
teach
their
own.
A
A
B
A
B
A
B
Yeah
I'm
the
same
way
like
I'm
sitting
here:
I
just
glanced
down
at
my
at
my
bar
here
and
I
have
I
have
three
unread
emails,
which
means
that
it
for
me.
If
it's
unread
it's
an
action
item,
it
means
I
need
to
do
something
with
it,
so
it
I
won't
say
it
gives
me
anxiety
until
it
gets
above
10
and
then
it
starts
getting
I
start
getting
Twitchy.
A
B
Yeah
and
then
you
know
all
of
the
chats
there's
so
many
chats,
Google
slack
chat,
you
know
iMessage
and
just
on
and
on
and
on
I've
I've
managed
to
avoid
IRC.
You
know
for
fortunately
for
better
for
worse
I
guess:
I
have
avoided
IRC
here
in
red
hat,
which
still
has
a
strong
Community.
A
lot
of
the
engineering
folks
will
use
RC,
which
I
think
is
really
interesting,
all
right
so
yeah,
so
We've
bantered
here
longer
than
we
normally
do
anyways.
B
So,
first
of
all,
thank
you
for
your
patience
really
appreciate,
really
appreciate
that
and
I
want
to
welcome
Harriet
Harriet
Lawrence
to
the
stream
today.
So
our
our
topic
for
today
is
one
that
is
near
and
dear
to
my
heart
and
I
hope
that
Christian
is
out
there
listening
somewhere
because
we'll
be
talking
about
get
Ops
for
admins,
and
this
is
something
that
you
know
for
anybody.
Who's
watched
the
stream,
we're
80
something
a
street.
B
You
know
88,
according
to
the
thing
in
my
window,
here,
88
episodes
in
and
I
think.
Since
the
very
beginning,
we've
talked
about
the
concept
of
of
get
ups
and
you
know
managing
applications
and
managing
configuration
in
a
controllable
way
and
all
of
that
other
stuff,
and
usually
when
we
talk
about
it
and
when
we
think
about
you
know
when
you
think
about
red
hat
live
streaming,
there
is
get
Ops
Guide
to
the
Galaxy
right.
We
have
an
entire
stream
dedicated
to
this.
It's
usually
focused
on
the
application
side
and
I
really
liked.
B
You
know
Harriet
reached
out
a
few
weeks
ago
and
said
you
know
hey
let's:
let's
do
a
stream
and
I
said
yes,
please
I
want
to
I
want
to
reset
I
want
to
go
back
to
to
kind
of
Base
zero,
and
you
know
Andrew
I
said,
as
I've
said
before
we
started
here:
I'm
gonna
play
the
role
that
I
was
born
to
play
dumb
guy
and
like
how
you
know
what
what
is
get
Ops.
How
does
it
help
us?
What
are
some
things
that
we
can
do
as
administrators
to
take
advantage
of
that
tool?
B
C
B
C
Wonderful,
yes,
it's
brilliant
I
am
I
actually
interned
with
red
hat
about
10
years
ago,
when
I
was
still
living
in
Australia,
so
interned
in
the
Brisbane
office
as
a
technical
writer,
and
that
was
before
I
went
off
and
did
development
stuff
and
eventually
found
my
way
through
to
product
management
and
back
to
Red
Hat.
And
it's
wonderful.
That's.
B
A
B
Sometimes
it
can
be
a
bit
of
a
fire
hose
and
I
have
a
a
new
person
on
my
team.
Who's
been
with
red
hat
now
for
about
six
weeks
and
we
were
having
a
one-on-one
the
other
day
he's
like
I
get
all
of
these
meeting
invites
and
I,
don't
know
which
ones
I
should
or
shouldn't
go
to
he's
like
I'm,
just
accepting
all
of
them
and
I
I'll
figure
it
out.
I'm
like
yep.
That's
well
welcome.
C
B
B
It
was
a
huge
number.
You
know
which
is
a
and
most
of
it
it's
kubecon
if
you've
never
been
to
a
kubecon
before
it's
not
really
vendor-centric,
it's
Community
right,
it's
it's
all
the
Upstream
stuff.
So
to
me,
that's
really
a
testament
of
just
how
much
Upstream
you
know
community
and
project
involvement.
That
red
hat
has
that
we
have
all
of
these
people.
You
know
we
don't
go
and
give
a
here's
why
you
should
choose
red
hat
openshift
session
at
kubecon.
B
A
Yeah
I,
like
that,
it's
same
thing
at
the
open
source.
Summit
too,
you
know
I
mean
like
it's
just
it's
awesome
to
see
the
impact
that
red
hat
has
across
the
community,
but
also
how
much
you
know.
We
appreciate
the
community
too
right
and
I.
Think
that
that's
to
your
point
right
about
what
kubecon
is
we
get
to
see
what
the
community
is
doing.
We
get
to
appreciate
the
value
that
they're
providing
and
it's
just
it's
awesome.
It
really
is.
B
B
B
What
we
discerned
was
that
effectively
some
of
the
lists,
some
of
the
people
who
were
listed
as
maintainers
for
some
of
the
repos
was
out
of
date,
so
I
think
they
kicked
off
an
initiative
to
update
those
and
make
sure
that
all
of
that
gets
addressed.
So,
if
you're,
if
you're
one
of
those
Community
folks,
if
you
have
you
know,
submitted
something
if
you're
interested
in
contributing
you
know
and
the
openshift
GitHub,
hopefully
all
of
that's
fixed
now.
B
So
you
know,
if
you
don't
hear
anything,
if
you
don't
see
anything,
let
us
know
you
know,
feel
free
to
reach
out
and
focus
and
we'll
we're
happy
to
find
the
appropriate
people-
and
you
know
see
if
we
can
get
some
attention
on
those
things,
because
it
is,
you
know
the
core
of
the
essence
of
red
hat.
Is
you
know
open
source
and
community
and
all
that?
So
it's
a
it's
a
sad
day
when
that
isn't
reflected
publicly?
A
Yeah
so
they
yeah,
we
we're
we're
red
hat,
so
names
are
hard,
and
so
it
looks
like
Microsoft
Microsoft
excuse
me
has
been
rebranded
as
red
hat
device,
Edge
and
so
I'm
going
to
post
the
link
to
the
blog,
but
just
the
blog
is
just
a
the
official
announcement
that
red
hat
device
Edge
is
now
microchip.
So
this
is
going
to
be
a
thing
pretty
soon
and
it's
going
to
be
awesome.
B
Yeah,
this
is
one
that's
interesting
to
me
right
because
we
we
talked
about
microshift
a
couple
of
times.
We
had
the
microshift
team
on
we've
had
who
was
it
Diego
and
Henry
came
on
and
radio
is
Luis
I,
don't
remember,
you
know
they
came
and
showed
microshift
actually
being
used
and
all
that
other
stuff.
So
to
me,
the
big
thing
that
came
out
of
this
announcement
is
that
it's
a
real
product,
or
it's
very
soon,
going
to
be
a
real
product
right.
B
It's
it
went
from
this
side
project
science
experiment
that
engineering
did
into
now
it's
a
valid
and
real
part
of
our
Edge
strategy,
so
yeah
I'm,
learning
more
and
more
I'm
becoming
a.
What
did
we
call
it?
Johnny,
a
15-minute
expert
on
on
the
the
red
hat
device
or
the
red
hat
Edge
stuff
right.
Some.
B
Work
that
I've
got
going
on
so
yeah
so
definitely
check
out
the
blog
post.
Again,
it's
it's
microshift,
just
the
productized
name
of
it
and
all
that
and
again
look
looking
forward
to
great
things
happening
there,
including
I,
know
ACM
is
doing
some
work
to
improve
scalability
and
all
that
other
stuff
to
help
manage
a
bunch
of
larger
numbers
of
clusters.
I
think
the
last
one
I
saw
was
something
like
2000
single
node
openshift
instances
managed
through
a
single
ACM.
B
A
Yeah
excited
times,
I
keep
saying
it.
We
are
in
like
exciting
times
right
now,
with
everything
ACM,
and
then
we
had
Adele
on
a
couple
weeks
ago
with
the
hosted
control,
plane,
yeah
cheetah.
All
these
things
that
are
coming
out
open
shift
get
obviously
advancements
that
we're
making
with
that
and
some
of
the
things
that
are
coming
out
with
that
which
is
going
to
talk
about.
There's,
there's
all
this
exciting
stuff,
that's
happening
and
we're
right
in
the
middle
of
it,
which
is
pretty
amazing.
B
Yeah
yeah
I
finally
had
time
to
get
Hands-On
with
hypershift
or
hosted
control
planes.
It's
it
is
embarrassingly
easy,
that's
it!
It
really
is
it's
you
deploy
a
CL,
you
deploy
a
cluster,
you
deploy,
ACM,
you
add
a
credential
object
through
ACM
and
then
basically
it's
one
crd
right.
You
created
a
new
hypershift
or
a
deployment
object
and
it
deploys
everything
that
you
need
and
stands
up
a
whole
cluster
like
it's.
It
is
painfully
easy,
I
I
kept
expecting
it
to
be
harder
and
it
wasn't.
B
B
C
B
Share
my
screen
here,
real
quick,
so
what
this
does
is
it
gives
us
the
ability
to
do
a
Cube
CTX
right.
It
gives
me
this
option
to
select
which
context
I
want
to
use
for
my
clusters,
based
off
of
whatever
cubeconfigured
happens
to
be
pointed
out.
So
this
is
my
default
one,
and
most
of
these
clusters
are
not
real
clusters,
they're
all
old
old
clusters-
it's
just
for
examples,
but
it
basically
I
type
Cube
CTX.
It
gives
me
this
selectable
menu
of
you
know,
which
one
of
these
clusters
do.
B
You
actually
want
to
point
at,
and
if
this
is
a
real
cluster
nope,
it
is
not.
So
it
would
immediately
switch
context
to
that.
So
I
can
do
Cube,
CTX
again
and
I
can
choose
like
this
one.
This
is
also
not
a
real
cluster
or
no
longer
a
real
cluster.
I
should
say,
but
my
point
here
is
that
it
gives
us
this
really
neat
ability
to
right.
I'm
not
going
to
wait
on
you
to
to
manage
our
clusters.
B
B
Admin,
because
that's
one
of
the
other
contacts
that
I
have
now
I
can
do
the
same.
Oc
get
node.
You
can
see
just
like
that.
I
switched
over
to
a
different
cluster
right.
I've
got
different
set
of
nodes
that
are
running
inside
of
here
different
set
of
things
that
are
going
on
so
I
thought
that
was
really
cool.
So
how
did
I
create?
How
did
I
actually
ingest
all
of
these
contacts?
C
B
B
Now,
if
I
do
an
echo
config
I've
got
all
of
my
all
four
of
those
there.
If
I
do
a
cube,
CTL,
config
View,
you
can
see
it
prints
off
all
of
these
different
things
that
are
in
each
one
of
my
cubeconfigs
right,
but
you'll
notice.
For
example,
there's
only
one
user
down
here:
admin!
That's
because
in
each
one
of
these
Cube
config
files
it
defines
a
cluster.
You
can
see
here's
the
cluster
name,
but
they
all
have
the
same
username.
B
So
if
I
wanted
to
do
this
correctly,
I
would
need
to
go
in
and
I
would
need
to
modify
each
one
of
these
to
have
a
unique
username.
Instead
of
all
of
them
using
the
name,
admin
and
then
I
would
should
have
four
admin
users,
but
anyway.
So
if
I
were
to
do
that,
then
what
I
do
is
that
same
Cube,
cuddle,
config
view
flatten
and
then
I
output
that
to
s
dot
cubeconfig
now
I
have
a
single
Cube
config
file
that
has
all
of
those
in
there
and
now.
B
I
can
simply
point
my
fig
variable
to
that
code
config
and
do
Cube
CTX,
and
you
see
it
pops
up
again
with
those
same,
if
my
set
of
options
that
are
available
there,
so
really
simple
way,
I've
been
looking
for
and
I
haven't
found.
Yet
in
my
cursory
searching
of
you
know
how
can
I
integrate
that
into
my
bash?
Prompt
I
would
love
to
find
a
way
to
to
have
so
that
way
when
I
set
a
context,
it
tells
me
which
cluster
I'm
connected
to
as
a
part
of
my
command
prompt.
B
So
if
anybody
knows
of
anything
that
would
be
great
but
yeah
other
than
that
I,
don't
think
I
have
any
any
of
those
clusters
open
here
other
than
that,
if
I'm
looking
at
the
console,
I
always
put
in
a
banner
so
that
way,
I
know
which
cluster
I'm
looking
at
but
from
the
command
line.
This
is
this
is
the
way
that
I
tend
to
use
it
and
I'll
dig
up
the
link
for
cube
CTX
and
I'll
I'll
post
that
into
the
chat
here.
A
Go
ahead,
I
was
gonna,
say
I,
don't
think
I've
heard
of
pure
line
or
Kube
CTX,
because
I
was
always
wondering
how,
especially
on
the
kubernetes
side,
how
you
did
that
right,
like
with
I,
mean
I,
guess
even
on
the
open
ship
side.
I
guess
normally
I'm
managing
a
single
cluster
at
a
time.
So
it's
not
that
big
of
a
deal
but
yeah,
that's
kind
of
cool,
that's
cool
that
they
have
tools
like
that.
B
Yeah,
yeah
and
I
know
there's
a
whole
bunch
of
them.
This
was
just
happens
to
be
the
one
that
I
stumbled
acronym
upon.
First,
as
a
recommendation
and
I
find
it
to
be
very
convenient
but
yeah
whatever.
Whatever
floats
your
boat,
Christian
Pure
Line,
all
right
I
will
stop
my
screen
share
here.
So
that
way
we
can
get
to
the
topic.
Du
jour
so
get
Ops,
so
get
Ops
is
and
Harriet.
Please
correct
me
at
any
point
in
time.
Get
UPS
is
a
feature,
slash
product,
I,
don't
think
as
a
product.
B
B
C
A
C
C
Just
get
as
a
single
source
of
Truth
and
there
are
four
kind
of
key
principles:
I
think
they
get
called
core
get
Ops
principles,
and
these
were
defined
by
the
get
Ops
working
group
I
think
last
year,
and
they
kind
of
give
you
an
idea
of
what
the
kind
of
constraints
are
for
saying
like
I.
Do
things
in
the
get
UPS
way
it
it?
None
of
them
like
say
like?
Oh,
it's,
it's
agrocity
or
it's
flux
or
it's
a
particular
product.
It's
just
like
the
general
overview.
C
So
the
first
one
is
that,
like
your
system
is
described,
declaratively,
so
you've
written
down
somewhere
and
usually
in
yaml
what
your
system
should
look
like
and
that
can
be
an
application
that
has
access
controls
or
more
relevant
to
the
viewers
of
the
stream.
It's
cluster
configuration
or
infrastructure.
C
So
that's
your
first
one.
Secondly,
we
want
this
state,
this
desired
state
to
be
versioned
and
immutable,
and
so
you
want
your
yaml
stored
somewhere
that
is
Central
and
accessible
and
is
a
single
source
of
Truth
for
the
rest
of
your
system.
You
usually
get
that
from
using
something
like
git,
but
any
system
that
complies
with
versioned
and
immutable
can
be
your
source
of
Truth.
C
C
So
you've
done
some
kind
of
approval
process
and
when
your
changes
merge
in
Git,
You
can
roll
them
out
automatically.
So
not
only
does
the
get
Ops
workflow
pull
the
stuff
out
of
Git.
It
will
pull
it
at
regular
intervals
to
check
for
changes
which
lead
into
the
last
one,
the
last
principle,
which
is
the
having
a
controller
that
monitors
your
repo
and
polls
for
changes.
B
B
You
know
some
sort
of
revision
control,
Source
control
system
to
to
manage
that
configuration
so
that
we
can
see
over
time
what's
changed
and
with
Git
Ops.
It's
now
a
core
requirement,
like
you,
have
no
option
other
than
to
you're
going
to
use
git
you're
going
to
use
this
so
that
you
can
see
precisely
what's
happened
when
who
approved
it
and
all
that
other
stuff,
which
I
think
is
not
not
a
bad
thing.
B
B
B
You
know
last
week
or
two
weeks
ago,
at
ansible
Fest,
they
talked
about
event-driven,
you
know
ansible,
where
now
you
can
kick
off
ansible
Playbook,
sensible
tasks
based
off
of
something
triggering
that
to
happen
a
web
hook,
for
example.
So
you
know
there's
ways
to
do
that,
but
here
it's
native
and
it's
a
core
component,
a
core
aspect
of
the
implementation.
C
Yeah
totally
I,
don't
like
I,
don't
think
any
of
these
things
are
new
and
I,
don't
think
even
get.
Ops
itself
is
new
I
think
people
have
been
using
these
workflows
for
ages.
They
just
haven't
been
talking
about
it
as
get
Ops
and
having
that
framework
to
talk
about
it
and
try
it
out
has
made
it
more
accessible
to
large
parts
of
our
industry.
A
Exactly
and
putting
the
putting
the
rules
around
it
right,
like
the
the
four,
what
whatever
they
are,
the
four
declarations
around
like
what
what
good
Ops
is.
I
think
that
that
helps
kind
of
like
box
it
in,
and
you
know,
kind
of,
like
all
right,
Thou
shalt
do
these
things
to
make
sure
that
you're
following
the
Sprint,
you
know
it's!
A
Otherwise,
it
it'll
sprawl
right,
because
you
give
any
you
give
any
nerd
and
it's
right,
we're
gonna,
take
it
and
we're
going
to
take
that
and
more
and
we're
gonna
try
and
add
our
own
little.
You
know
Special
Sauce
to
it
so
I
think
that,
having
that
having
the
framework
with
it
and
then
just
having
the
tool
that
can
go
out
and
reconcile
things
kind
of
like
you
know,
everyone
used
to
hit
on
puppet.
You
know
and
Andrew
just
brought
the
stuff
right.
A
Everybody
used
to
hate
on
puppet,
but
at
the
end
of
the
day,
right
puppet
was
pretty
rad
because
it
would
go
out
and
if
you
had,
if
you
had
made
a
modification
to
something
that
was
being
monitored,
it
would
change
it
back.
You
know
and
I
I
think
that
that's
that's
powerful
right,
especially
when
you
talk
about
configuration,
drift
and
and
keeping
your
clusters
in
sync
when
you
start
managing
multiple
clusters
or
hundreds
of
clusters
like
you
need
something
like
this.
B
Yeah,
so
I
guess
get
Ops
solves
the
problem
if
you
will,
or
or
in
a
kubernetes
native
way,
IT
addresses
the
problem
of
declarative
configuration
and
I
want
to
be
clear
that
a
lot
of
folks
also
refer
to
kubernetes
itself
as
a
declarative
configuration
system
right,
the
the
yaml
objects
that
you
apply,
you're
declaring
to
the
control
plane.
You
want
it
to
look
that
way,
but
get
up
supplies
that
to
an
application.
B
So
you
define
this
thing
as
a
set
of
object,
objects
that
need
to
be
managed
that
need
to
have
that
revision,
control
that
configuration
management
against
them
as
opposed
to
individuals.
So
I
guess
my
question
where
I'm,
rambling,
eventually
rambling
here
too
is
so.
What
is
that
unit
of
of
control?
Look
like
you
know.
If
I
want
to
as
an
administrator,
you
know
be
able
to
apply
specific
machine
configs
or,
if
I
want
to
make
sure
that
machine
sets.
B
You
know
this
machine
set
exists
with
this
particular
scale,
so
I
think
these
are
important
things
you
know
going
back
to
the
whole
point
of
what
was
it?
The
third
one
you
know
sorry
I'm,
looking
at
the
thing
right,
approve
changes
right
can
be
applied
automatically.
You
know
hey,
we
need
to
scale
this
cluster
up
from.
You
know
two
nodes
to
ten
nodes.
Well
that
has
cost
associated
with
it.
B
C
Yeah,
so
you
mentioned
application
sets.
That
is
a
kind
of
newer
Concept
in
Argo
CD,
it's
like
so
the
smallest
one
is
an
application.
You
make
an
application
CR
and
an
application
set
is
a
way
of
saying
I
keep
using
a
generator
you're
saying
Hey
I
want
this
one
application,
but
I
want
like
six
replicas
or
I,
want
it
in
these
different
environments
or
getting
it
you're
like
templating
out
your
applications
and
pushing
them
out
while
you
need
them.
A
It's
something
to
key
in
on
there
yeah.
You
mentioned
the
generators
right
so
with
an
application,
you
generally
Define
or
you
declare
your
repo
the
path.
You
know
the
cluster,
all
those
things
right
and
it's
kind
of
like
in
that
single
application
resource
with
an
application
set.
You
define
this
generator
and
there's
there's
a
bunch
of
them.
I
think
there's
like
they're,
six
or
seven
of
them.
A
The
one
that
I
have
experience
with
is
the
git
generator
and
you
essentially
say
hey
in
my
repo
go
look
in
that
directory
and
then
it'll
just
pull
in
all
of
those
any
directory
that
you
have
in
there.
So
if
you
have
Helm,
charts
or
whatever
that
are
sitting
in
that
in
that
directory,
it'll
pull
all
of
them
in
and
then
it'll
it'll
create
the
individual
application
resources
for
them
and
then
roll
them
out.
A
So
there's
a
lot
of
power
as
far
as
like
being
able
to
like
dynamically
generate
some
of
these
declarative
things
right
where
I
can
just
say:
hey,
go,
go,
look
there
and
then
deploy
these
things
and
manage
them.
For
me,
after
that,
like
that's,
that's
that's
one
way
and
then
the
other
one
I
think
that's
pretty
popular.
Is
his
list
list
generator
and
that's
where
you
it's
literally
like
you're,
creating
an
application
right?
A
You
define,
you
know
the
different
components
of
that
application
and
then
you
just
let
Argo
do
its
thing
and
roll
them
out
and
it'll
create
the
individual
application
objects.
So
it
it's
pretty
cool
like
there's,
there's
a
lot
of
cool
stuff
that
you
can
do
and
having
the
visual
aid
along
with
it.
You
know
the
green,
yellow
red
type
thing
I,
think
that
that's
that's
big
for
teams
that
don't
have
a
lot
of
experience
with.
You
know
kubernetes
and
application
deployments
in
infrastructure
deployments
and
stuff
like
that,
so
it
it's
pretty
red.
C
Yeah
absolutely
and
I
really
love
the
like
Dr
aspect
of
keeping
everything
in
a
single
source
of
Truth
as
well
like
if
you're
defining,
not
just
like
your
application
deployments.
If
you've
got
all
your
code
or
your
like
infrastructure
and
everything
there,
your
cluster
goes
down,
no
worries,
I'll
go
bring
you
up
a
new
one.
Everything
goes
down,
you've
got
it
all
defend
you
just
like
kick
it
off
and
off.
It
goes
and
recreates
everything
for
you.
B
You
know
hugely
hugely
beneficial,
so
Johnny
I'm,
gonna,
I'm,
gonna
toss
you
under
the
bus
I'm
going
to
put
you
on
on
the
spot
and
ask
ask
if
you
have
a
an
environment
available,
so
we
can
look
at
a
few
things
and
then
Harriet
I,
don't
know
if
you
know
the
answer
to
orphans
and
apologies
for
butchering
anybody's
name
question
here:
why
why
we?
Why
don't
we
have
the
get
Ops
operator
available
in
the
developer,
sandbox.
A
A
C
B
A
Yeah
is
that
better
yep,
okay,
cool
yeah,
so
this
is
openshift
kid
Ops
and
and
so
what
I've
done
is
I'm.
On
my
team,
the
validated
patterns
we
deploy,
multiple
Red
Hat
products
together
integrated
together
using
the
get
Ops
framework.
So
all
the
things
that
Harriet
was
talking
about
the
four
principles
of
get
Ops
the
pillars,
that's
what
they're
the
four
pillars
of
good
Ops.
A
A
It
manages
all
the
cluster
level
resources,
so
I
mean
literally
like
anything
that
you
would
need
like
a
a
cluster
admin
type
thing
for
that's
what
that
that
Argo
does
and
then
we
have
a
a
hub
organ
CD
that
deploys
the
actual
application
resources.
So
what
I'm
going
to
do
is
I'm
going
to
go
into
the
cluster
level,
because
when
we
Define
and
declare
our
applications,
we
have
the
cluster
Argo
CD,
actually
defining
the
application.
A
A
A
So
we
have
this
application
here,
called
the
multi-cloud
get
Ops
application,
and
what
this
says
is
this
is
essentially
our
meta
app
and
then
it's
going
to
deploy
some
of
these
other
applications.
So
it's
deploying
the
rgo
CD
resource,
there's
some
deployments
and
stuff
like
that
in
here
and
then
pretty
soon
I'm
going
to
be
able
to
scroll
and
find
the
other
applications.
But
you
can
see
that
all
the
resources
that
get
spun
up
and
created,
obviously
sync
and
sync:
okay-
is
good.
A
Sometimes
you'll
see
things
like
degraded
and
sync,
okay,
which
just
means
that
the
sync
from
the
get
repo
is
good
to
go
so
that
that
reconciliation
between
what's
in
get
and
what's
on
the
cluster
is
is
good,
but
it
might
be
degraded
because
maybe
the
resource
didn't
deploy
correctly.
So
maybe
there
was
a
misconfiguration
somewhere.
So
that's
just
some
of
the
things
that
you'll
see
and
I'm
just
gonna.
A
So,
oh
here
we
go
so
here's
we
as
part
of
this
deployment
we
deploy
ACM.
So
you
can
see
we
have
our
ACM
application
and
then
we
have
an
application
here
that
a
guy
on
my
team
named
Marty.
He
wrote
and
then
we
integrate
with
external
secrets
and
hash
Court
Vault
for
our
secrets
management.
So
you
can
just
see
like
we've
got
our
top
level
applications
and
in
the
definition,
we're
just
saying,
hey,
look
I've
got
I,
need
this
app
to
clear
it.
A
Here's
all
my
here's
all
my
parameters
that
I
want
with
it.
It's
going
to
this
GitHub
URL,
here's
the
path
within
my
within
my
git
repo,
where
I
can
find
this
application
resource
or
where
I
can
find
the
contents
for
this
application,
and
so
then,
just
a
and
I'm
sorry
for
jumping
around
here,
but
like
just
imagine
that
this
is
all
one
Argo,
so
it
deploys
this
application
and
what
we
can
see
now
is
we
have
another
application
here
for
ACM
and
all
the
all
the
resources
that
are
associated
with
ACM.
A
So
you
can
see
like
you
get
this
really
nice,
visual
representation
of
your
application
and
what
it
looks
like
on
the
cluster.
And
if
any
one
of
these
was
you
know
out
of
sync
or
had
an
issue
or
anything
like
that,
then
you
would
see
the
you
know:
it'd
either
be
like
yellow
shield,
or
you
know,
a
broken
heart,
which
is
my
favorite,
it's
kind
of
funny
or
the
green
check,
because
everything's
good
to
go
here.
A
Yep
it
does,
and
so,
but
the
thing
about
this
is
that
I
did
not
declare
each
one
of
these
objects
right.
So,
as
you
know
like
with
an
operator
when
you
deploy
an
operator
that
operator
will
deploy
whatever
right,
config
map
secrets,
you
know
deployments
Etc,
so
the
things
that
I
will
generally
deploy
with
our
with
our
framework
is
I'll,
create
a
subscription
I'll
create
if
there's
an
operator
group
required
or
deployment
or
whatever
for
an
application.
B
So
I
I
have
a
question
and
I
don't
know
Harriet
or
Johnny,
or
you
know,
maybe
one
of
the
other
folks
I
know
from
the
get
Ops
Team.
That's
watching
there
I
think
there's
some
overlap
between
get
Ops
and
kind
of
the
way
that
we
talk
about
or
present
operators,
and
you
know
even
things
like
Helm
charts.
So
you
know
because
I
can
deploy
an
application
right.
B
I
use
a
home,
repo,
I
go
and
I
say
you
know
him
deploy
and
it
instantiates
my
application
and
it
manages
a
lot
of
those
different
components
and
stuff
like
that.
So
do
you
have
any
any
thoughts?
You
know,
suggestions
around
like
comparing
those
different
ways
of
deploying
applications.
Maybe
Pros
cons,
strengths,
weaknesses,
type
of
of
scenarios.
B
And
just
to
be
clear,
I
keep
saying
applications
but
I
use
that
as
a
blanket
term
to
mean
like
really
any
configuration
in
the
cluster
right.
I
know.
We
keep
saying
that
because
that's
the
literally
the
language
that's
used
inside
of
Argo
to
Define,
as
you
pointed
out
here,
yet
to
define
the
the
managed
object.
B
A
Yeah,
so
so
this
is
like
the
gnome
versus
KDE
right.
This
is
like
the
it's.
It's
very
opinionated
because
there's
people
like
you
know
Gerald
is
a
Christian
around
here
and
they're,
both
big
fans
of
customized,
where
our
team
we're
very
heavy
Helm
focused.
But
so
you
could.
You
can
use
Helm
I
like
Helm,
because
I'm
very
familiar
with
ansible
and
so
Helm
templating
is
very
it's
not
the
same,
but
it
uses
a
lot
of
the
same
kind
of
like
structure.
So
to
me,
it's
it's
more
familiar.
We're
customized!
A
In
my
opinion,
like
you
have
to
do,
you
have
your
bases,
and
then
you
have
your
overlays
and
stuff
like
that.
So
it
seems
like
in
some
cases,
there's
a
lot
of
duplication.
I,
think
that
there's
things
that
customize
make
some
very
easy,
where
Helm
is
a
little
bit
more
difficult
and
then
the
way
the
Argo
utilizes
Helm
makes
using
home
a
little
just.
A
You
can't
necessarily
do
out
of
the
box
with
Argo
and
Helm
templates
so
that
that's
where
I
think,
like
things
like
customize
kind
of
they
kind
of
solve
that
problem
a
little
bit,
but
the
other
way
is
kind
of
like
how
we
did
it
back
in
the
day.
Andrew,
you
remember
you
just
you
built
something.
It
worked,
you
do
a
OC
get
minus,
oh
yaml,
carry
that
to
a
yaml
file,
and
then
you
just
reuse
that
same
config.
A
If,
as
long
as
Argo
CD
or
openshift
get
Ops
knows
that
path
is
there
and
then
that
resource
is
sitting
in
that
path.
It
doesn't
have
to
be
a
Helm
template.
It
doesn't
have
to
be
a
customized
chart
or
anything
like
that.
It
could
just
be
a
yaml
file,
that's
sitting
in
there.
That's
already
completed
so
it's
it's
kind
of
like
picture
poison
me
personally,
I'm
all
about
reusability.
A
A
A
One
thing:
so
we
have
this
application
here
and
it's
called
config
demo
and
again
this
is
just
it's
a
simple
application.
It's
a
deployment,
and
so
what
we
do
is
we
declare
the
way
that
we
do
our
framework.
Is
we
essentially
declare
what
applications
we
want?
So
we
we
create
a
values
file
and
we
say:
hey
I
want
config
demo
deployed
in
this
name
space
Here's.
A
What
then
my
repo
where
it's
at
and
what
happens
is
Argo
CD
will
pick
this
up
and
it
says:
oh,
hey,
look
I'm
supposed
to
deploy
this
application,
that's
not
on
the
cluster!
Yet!
So
let
me
do
this
reconciliation
and
it'll
roll
it
out
to
the
cluster
right.
So
in
August
CD.
This
is
what
it
looks
like.
It
says:
hey,
here's!
Here's
your
map
of
your
or
here's,
your
topology
of
your
application,
but
if
we
jump
over
to
openshift
and
we
kind
of
take
a
look-
let
me
go.
B
There
we
go
I
I,
think
that
also
brings
up
another
an
extension
to
my
previous
question,
which
is
how
ACM
fits
in
because
I
think
there
are
some.
You
use
the
word
opinions
before
and
I.
Think
that's
the
right
word
of
you
know:
how
do
you
manage
applications?
How
do
you
distribute
you
know
these
things
across
multiple
clusters?
B
B
Yeah
and
it's
so
funny
that
so
many
questions
are
answered
that
way
with
kubernetes
yeah,
so
somebody
asked
yesterday
in
one
of
the
chat
rooms
like
oh,
how
do
I
do
capacity
forecasting
with
openshift?
Well,
what's
your
application,
like
is
like?
What's
the
application
doing?
Well,
we
don't
know
what
the
application
is.
Well,
it's
gonna
be
hard.
B
B
Some
things
you
can
do
like
monitoring
can
do
so
much,
but
at
the
same
time
you
know
if
the
app
you
have
to
work
with
the
application
team.
You
have
to
understand
that
oh
yeah
we're
gonna
substitute.
You
know
database
a
for
database,
B
and
oh
yeah.
Database
B
has
a
you
know
a
much
heavier
resource
footprint.
Well,
if
you
don't
know
about
that
ahead
of
time,
then
you
know
it's.
B
A
Well,
you're
you're,
fine
you're,
fine,
the
one
thing
that
I
like
about
ACM
and
and
get
Ops
together
right
is
they're
both
very
like.
Oh
now,
I
can't
think
of
the
word.
I
just
had
like
complete
Flatline,
complimentary,
complimentary
yeah.
Thank
you
and
because
ACM
has
that
get
Ops
component
right.
It
has
that
aspect
where
it
can
roll
things
out
and
then
obviously
you
have
openshift
get
Ops
and
I.
Think
I
think
that
the
way
that
when
you
use
them
together,
you
know
so
what
we
do
on
our
team
is.
A
We
will
we'll
roll
out
an
open
shift,
kit,
Ops
and
then
we'll
roll
out
an
ACM
and
then
once
we
start
talking
about
like
multiple
clusters,
right,
we'll
roll
out
ACM
policies
and
then
that
policy
will
push
out
like
that.
Initial
openshift
get
Ops
instance
on
that
remote
cluster
or
that
Hub
spoke
cluster
and
then
essentially
what
will
happen?
Is
it
essentially
it
does
all
the
construction
work
right
like
it
goes
down.
A
It
digs
up
the
roads,
it
lays
down
the
asphalt
and
then
it
says:
hey
I'm
done
I'm
ready
to
go
and
then
what
happens
is
the
applications
start
getting
deployed?
And
it's
just
a
nice
complimentary
toolkit
that
you
can
have
and
it's
powerful
because
back
to
Hillary's
Point
earlier
about
the
single
source
of
Truth.
Now
it's
like
okay,
I
have
a
cluster
go
down?
Okay!
Well,
let
me
quickly,
let
me
just
get
off
to
spin
up
a
new
cluster
through
ACM
and
then
then
have
assembly
down
some
of
that.
A
You
know
some
of
that
construction
zone
stuff
and
then
have
Argo
come
in
behind
it
and
lay
down
the
rest
of
the
environment,
and
so
now
you
start
talking
about
these
rolling
environments
that
you
can
kind
of
spin
up
and
spin
down
pretty
quickly
and
you've
got
a
lot
of
flexibility
and
a
lot
of
power
in
your
hands.
Right
now
and
and
the
concern
of
like
oh
man,
I
just
had
this
node
go
down
in
the
middle
of
the
night
or
you
know
this
Zone
went
down
in
the
middle
of
the
night.
It's
like
okay!
C
A
Exactly
and
a
good
point
right
with
artists
or
with
ACM,
you
get
you
get
that
governance
right
so
that
observability
across
all
the
Clusters
and
then
you
get
that
governance
and
insight
into
like
okay
I've
got
these
applications
deployed
to
this
cluster
I've
got
it's
just
yeah,
it's
man,
it's
just
powerful
and
it's
it's
awesome
and
just
the
the
strides
I
mean
on
the
open
shift
get
outside
too
and
and
ACN
like
the
strides
that
they've
made
as
far
as
capability
and
getting
stuff
out
like
it's.
It's
incredible.
B
Well,
I
think
the
way
I
interpret.
It
is
the
same
way
that
we
talk
about.
You
know
the
differences
between
what
ACM
does
with
managing
security
policies
and
configuration
versus
what
ACS
does
versus
what
like
the
compliance
operator,
does
right
and
that
they
all
three
have
a
role
to
play.
They
just
do
it
at
different
levels
and
in
different
places.
So
you
know
ACM,
you
know,
deploys
Argo
or
openshift
git
Ops
and
then
applies.
B
A
The
scenes
I
so
I
I
stood
up
another
cluster
when
we
first
started
because
I
think
what
I
think
some
of
the
value
that
you
can
see
is,
like
you
know
just
talking
about
some
of
the
policies,
so
we
can
see.
I
have
two
violations
right,
so
I've
added
this
cluster
I've
created
some
policies,
ACM
policies
right,
and
so
it
it's
just
a
little
bit
slow
but
we're
getting
a
live
view,
as
things
are
starting
to
be
applied
to
this
remote
cluster
right.
A
A
We'll
also
start
seeing
that
config
demo
application
that
we're
just
going
to
looking
at
get
deployed
out
to
the
remote
cluster,
and
we
should
start
seeing
some
updates
yeah
here
we
go
so
pretty
soon,
we'll
start
seeing
the
console
updates
and
then
that's
it
and
all,
and
just
just
so
you
know,
like
all
I
did
to
get
this
ready
is
I
deployed.
The
cluster
and
I
ran,
make
install
and
all
it
just
deploys,
Argo,
CD
and
ACM
and
those
subscriptions
and
then
lets
to
application
resources
itself
from
Argo
deploy
all
the
other
stuff.
A
B
B
A
Then
we
also
have
that
ability
as
well.
So
that
way
you
can
kind
of
scrape
out
the
ACM
aspect
of
it
and
just
do
the
Argo
Vargo
type
deployments,
and
that's
that
I
think
that's
really
where,
like
an
application
set
with
your
generators
and
stuff
like
that,
that
would
really
become
like
a
powerful
tool
in
that
regard,
but
yeah
just
to
say,
like
when
I
say,
make
install
like
I
I,
didn't
change.
Anything
I
didn't
update
any
file.
I
literally
cloned
the
repo
down
I
deployed
the
cluster
I
exported.
B
So
we've
got
about
15
minutes
left
so
for
the
audience.
Please
any
questions
that
you
have
go
ahead
and
send
those
in
we'll
we'll
do
our
best,
we'll
answer
all
the
questions
that
we
can
and
if
we
don't
know
the
answer,
then
we'll
go
and
we'll
find
the
answer,
because
that's
what
we're
here
for
I
just
I'm
just
now,
realizing
as
I
say
that
that
I
didn't
say
that
at
the
beginning
of
the
stream,
like
I,
normally
do
so
yeah,
please
go
ahead!
B
Ask
any
questions
that
you
got
you're,
also
more
than
welcome
to
reach
out
to
Johnny
or
I.
So
if
you're
not
watching
us
live
if
you're
watching
a
recording,
you
can
reach
out
to
us
via
email
or
social
media
or
all
of
the
normal
things.
So
if
you
want
to
reach
out
to
me
it's
very
simply
a
practical
Andrew
on
all
of
the
various
platforms
and
then
andrew.sullivan
redhat.com
or
if
you
prefer,
to
harass
Johnny
Jrock
tx1
on
the
various
platforms
or
Johnny
Jonny
redhat.com.
B
B
Out
to
us
I
know
he
he
got
a
nice
story
and
I
actually
I
haven't
I,
haven't
announced
it
yet,
but
I
have
Sully
at
redhat.com,
very
nice,
yeah,
yeah
I,
don't
know
I
think
when
I
first
joined
you
know
four
plus
years
ago.
Somebody
else
had
it,
but
now
it's
mine,
all
mine,
anyways.
Yes,
so
yeah,
please,
you
know,
send
in
whatever
questions
you've
got
we'll
we'll
do
our
best
to
answer
those.
B
That
being
said,
Johnny
I
wanted
to
put
you
on
the
spot
again
so
can
you
can
you
show
an
example
of
like
I
want
to
create
a
config
map?
I
want
to
I
want
to
apply
like
some
cluster
level
configuration
like
walk
me
through
that
process
of
going
from
config
map
yaml
to
Argo
components
and
and
how
we
actually
do
that
and
for
some
please,
please
ask
your
question.
A
C
B
A
Dude,
my
daughter
only
wanted
to
hit
the
houses
that
had
full-size
candy
bars.
I
should
go
up
and
she's
like
she
couldn't
like
scope
it
off
and
then
yeah
she's,
given
because
she
wanted
candy
anyways.
But
then
you
know
like
if
they
had
full-size
candy
bar
she's
in
a
Full
Tilt
getting
up
to
the
porch.
So
he
was.
It
was
pretty
funny
so
and
what's
that
I'm
sorry.
A
A
It
helps
fine,
so
when
we
Define
or
we
declare
our
application,
we
have
to
tell
it.
We
essentially
say
where,
in
the
repo
that
resource
lives
right,
so
essentially
this
common
hash,
Court,
Vault
or
even
config
demo.
Here
right,
it
lives
in
in
charts
all
config
demo
right.
So
if
I
go
into
charts,
all
config
demo
I
just
have
I'm
at
the
base
of
like
a
normal
Helm
template,
and
so,
if
I
go
into
templates
here,
I
can
see
that
I
have
all
of
my
resources.
A
A
Yeah,
let's
see
here
and
so
I'm
just
gonna
just
to
keep
it
simple
I'm
just
going
to
create,
like
you
know,
just
a
config
that
it's
just
a
config
knot,
that's
going
to
be
exactly
the
same
as
the
other
one,
but
it's
just
going
to
have
my
name
in
it
and
I'm
gonna.
Do
a
kit
switch
so
that
way,
I
don't
push
to
Main.
A
Okay,
well,
anyways,
that's
what
I
would
do
so
it's
it's
good!
It's
it's
context
specific!
So
if
I
do
like
a
get
remote
or
get
branched,
I
have
my
Branch
set
to
main
I.
Don't
want
to
push
to
the
main
branch
up
there,
because
I
don't
want
to
that's
our
public
repo
but
anyways.
What
I
would
do
is
I
would
push
this
in
and
then
it
would
just
get
deployed.
A
If
autohill
is
set
to
true,
then
it'll
say:
look,
there's
a
difference
and
that
files
config
and
it'll
it'll
puppet
it
right.
It'll
it'll
set
it
back
to
the
normal,
but
just
generally
speaking,
if
I
wanted
to
add
a
new
file
or
a
new
like
subscription
or
whatever
then
I
would
just
add.
The
file
here
get
add,
get
commit
push
it
into
my
repo
and
then
the
the
controller.
The
getups
controller
will
pick
up
the
difference
and
then
reconcile
and
push
that
file
out.
So
it's
there's
a
three
minute
like
default.
Reconcile.
C
B
A
C
Well,
you're
cracking
that
down.
We
have
a
really
cool
feature
coming
up
in
the
next
release
of
openshift
get
Ops,
which
is
support
for
Server
side
apply,
so
you'll
be
able
to
manage
way
more
stuff
in
openshift
from
githubs,
because
you'll
be
able
to
do
patching,
which
is
very
exciting
to
administrators.
C
B
C
A
A
So
Andrew
you
see
this
is
the
application
resource
and
you
see
like
because
I've
declared
this
path,
anything
that
I
add
underneath
that
path
is
going
to
get
ingested
so
once
once
I
I
create
the
new
file
or
the
new
subscription
or
new
whatever.
It
is
right.
New
deployment
as
soon
as
I
push
it
into
git
and
git
is
looking
at
this.
You
know
or
Argo.
A
The
controller
is
looking
at
this
directory
right
and
it
says:
oh,
hey,
look,
there's
a
new
deployment
out
there
and
then
it
will
automatically
start
that
reconciliation
to
get
it
out
on
the
onto
the
cluster,
but
I
don't
have
to
change
anything
within
the
application
itself.
So
the
the
spec
of
the
application
doesn't
change.
I,
don't
modify
my
my
git
base.
The
reason
why
this
is
an
issue
for
me
to
update
this
is
because
I
don't
want
to
update
main
off
of
our
our
public
repo
I
I
wasn't
expecting
to
make
any
change
as
much.
A
So
that's
that's
on
me.
I
should
use
my
own
remote,
but
but
yeah.
So
this
will
it's
pretty
nice,
because
this
gives
you
the
flexibility
to
add
things
in
and
then
you
can
test
things
as
you're
as
they're
starting
to
roll
out
and
pull
them
out
and
you
can
test
if
your
values,
you
know,
maybe
maybe
you're
declaring
some
some
new
values
and
stuff
like
that,
you're
testing,
some
logic
in
your
home
template.
A
It's
pretty
simple.
To
add
these
things
in
test
that
stuff
out
watch
it
fail
or
watch
it
pass.
You
know
that
type
of
thing,
and
then
you
know
it's
pretty
easy.
It
just
makes
it
makes
life
pretty
simple
and
that,
like
there's,
two
different
there's
like
Helm
charts,
there's
like
git
applications
and
then
there's
like
Helm
charts,
and
so
this
is
a
git
application
and
then
like
a
Helm
chart
would
be
like
a
a
chart.
A
That's
sitting
out
in
like
a
public
repository
somewhere
that
you
can
come
in
and
you
essentially
add
the
same.
You
can
override
it
the
same
way
where
you
add
some
of
these
values
that
are
specific
to
your
environment,
almost
like
you're
doing
like
a
set
on
the
command
line
for
a
home.
You
know
Helen
install
set
this
set
that
whatever.
B
A
Yes,
there's
not
capability,
that's
that's
within
the
helm,
sprig
stuff
right.
So
if
if
you
are-
and
let
me
just
kind
of
show
you
so
before
we
move
on
to
that
really
quick,
though
the
the
only
way
that
you
can
do
this
right
now
for
application
dependencies
is
set
up.
So
you
apply
sync
Hooks
and
sync
waves
right.
A
That's
right
now
the
only
way
that
you
can
actually
do
the
the
application,
dependency
and
application
sets
you'd
lose
that
capability
and
so
I'm,
quoting
like
Gerald
and
Christian
here.
So
I
could
be
quoting
it
wrong,
but
I'm
once
you
get
an
application
says
you
kind
of
lose
that
flexibility
and
so
right
now
there's
there's
ways
to
get
around
some
of
those
inner
application
dependencies
either
using
sync
waves
and
jobs
and
stuff
like
that.
A
A
Imperative
so
we
see
like
there's
there's
this
key
here
called
imperative
right
and
so
essentially
I'm
just
saying:
hey,
look
if
I
have
a
job
that
I
need
to
run
manually
right,
like
I
need,
I
need
to
actually
interact
or
interrogate
my
cluster
for
something.
Then
we
create
imperative
jobs,
and
this
is
within
our
framework,
but
to
go
back
to
your
question
about
using
that.
If
statement,
if
I
go
into
common.
A
And
I
look
at
the
job,
if
not
equal
values.
Okay,
so
ignore
that
one,
if
greater
than
length
equal
values,
cluster
group
comparative
jobs
right.
So
this
is
saying
if
I
have
a
job,
that's
greater
if
I,
if
I
actually
have
something,
that's
defined
then
use
this
otherwise
ignore
it.
So
if
if
this
is,
if
this,
if
there's
nothing
there
to
find,
then
it
won't
create
anything
and
so
yeah
you
can
use
if
statements
you
can
use.
If
else
you
can
also,
you
know
where
you
can
do
like
I
can't.
A
C
We
certainly
recommend
it.
Our
openshift
get
Ops
operator
is
specifically
for
doing
this
with
openshift,
like
you
can
use
auger
CD
with
non-openshift
clubs,
that's
much
broader
than
just
openshift,
but
for
openshift
clusters
we
recommend
you
use
the
openshift
get
Ops
operator.
You
get
all
that
extra
openshift
support
the
kind
of
security
standards
that
we
set.
We
adhere
to
a
lot
of
good
stuff.
B
Yeah
and
I'll
answer
that
slightly
differently
of
if
you,
if
you
want
to
use
Argo
CD,
then
yeah
I
would
say
it's
a
best
practice
to
install,
but
Argo
CD
is
one
way
of
addressing
the
the
problem.
If
you
will
of
how
do
I
deploy
applications,
how
do
I
manage
applications?
How
do
I
to
your
point
at
the
very
beginning,
right,
declaratively
and
in
a
controlled
manner?
You
know
manage
those
you
know
across
not
just
one
but
multiple
clusters.
B
So
if
your
question
is,
you
know,
is
it
a
best
practice
to
use
Argo
CD
the
question
there
is
well
maybe-
and
it
depends
on
your
goals
and
your
requirements
and
which
one
of
the
multitude
of
tools
that's
out.
There
best
meets
your
needs
and,
as
Harriet
said,
you
know,
we're
certainly
biased
in
that.
If
you,
if
you're
using
openshift-
and
you
want
to
use
Argo
CD-
then
yes,
certainly
you
would
have
you
would
deploy
it
into
openshift
I,
don't
know.
C
A
C
C
You
have
to
manage
application
resources
that
were
inside
the
control
plane
cluster,
but
with
the
next
release
of
overshift,
get
UPS
you'll
be
able
to
do
those
outside
of
that
one
namespace
inside
your
one
cluster,
so
which
is
a
really
awesome
step
towards
a
better
multi-tenancy
story
for
us,
and
we're
really
excited
to
kind
of
make
some
progress
in
that
direction.
Yeah.
B
C
C
I'm
going
to
be
really
annoying
and
say
it
depends
again,
it
depends
on
your
access
requirements.
So
if,
if
you
need
to
divide
stuff
up
by
like
who
can
who
needs
to
use
these
applications?
Who
who
needs
access
to
them,
buy
like
team
or
something
like
that
and
you
need
them,
you
need
Agro
CD
at
that
level
to
do
your
R
back.
Then
you
need
to
like
divvy
it
up.
You
have
like
one
Argo
per
you
can
have
a
per
cluster.
You
can
have
an
aggro
per
namespace.
B
One
yeah
I
think
my
first
reaction
off
the
top
of
my
head
is
failure
domain
and
maybe
that's
because
I've
been
burnt
by
that
one
too
many
times
of
you
know
what
happens
if
I
have
one
one
Argo
instance:
that's
deploying
and
managing
you
know,
applications
across
you
know
two
or
five
or
ten
or
fifty
clusters.
You
know
what
happens
if
that
one
Argo
instance,
you
know,
goes
down.
How
is
that
going
to
affect
you
know,
operations
across,
whereas
with
multiple
Argo
instances,
you're
limiting
the
failure
domain
there.
B
Yeah,
well,
it's
it's
funny
because
you
know
we
talk
about.
You
know:
oh
yeah
use
you
know,
get
UPS
or
or
use
whatever
mechanism
to
recover
all
of
your
your
openshift,
your
your
kubernetes
ammo
files
and
all
that
and
we
hand
wave
over
things
like
persistent
data
and
all
of
that
yeah.
You
know
what
happens
if
I
need
to
redeploy
all
these
things
and
I
have
you
know
gigabytes
upon
gigabytes
of
container
images
that
I
now
have
to
wait
to
pull
back
down?
What
about
my
registry
yeah
yeah?
B
B
So
so
I
would
recommend
reaching
out
to
there
and
it's
like
4
000
of
our
closest
friends.
You
know,
there's
a
great
Community
there,
there's
lots
of
people
who
are
asking
questions
who
are
willing
to
help
with
kind
of
all
kinds
of
different
aspects.
Thank
you
for
posting
that
Johnny,
so
I
would
recommend
going
there
because
Johnny
and
I
are
not
at
least
I'm,
not
Savvy
on
pipelines.
Johnny
might
be.
A
A
People
like
when
I
talk
to
people
like
people
that
are
like
narrow
focused
are
great
because
they're,
like
they're
experts,
in
that
one
thing,
but
people
that
can
kind
of
see
the
whole
picture
together
are
to
me
more
valuable
because
that
devops
transition
is
going
to
end
up
going
to
you're
going
to
go
from
a
devops
admin
engineer
into
a
SRE
right
and
sres
need
to
be
a
big
picture.
You
know
need
to
have
that
big
picture
mindset.
So
that's
that's
my
advice.
A
C
C
We've
just
dropped
it
in
It's,
a
Wonderful
ebook
by
by
good
friend,
Christian
Hernandez,
who
I
believe
is
on
our
twitch
stream
at
the
moment,
lovely
overview
of
everything
about
get
Ops
and
all
different
facets,
and
that's
a
really
great
introduction.
So
you
should
definitely
check
it
out.
It's
free
and
it's
on
the
developers.redhat.com.
B
Yep
yep,
we
know
Christian
a
little
bit.
Let's
see,
chandan
is
Argo
CD
cni
development
for
kind
of
vxlan
and
IP
encapsulation
in
the
same
cluster,
so
Argo
CD
would
be
independent
of
that.
So
it
would
basically
control
the
configuration,
so
you
would
be
able
to
apply
config
or
something
like
that,
but
it
wouldn't
impact
whether
or
not
you
can
use
like
vxlan
with
encapsulation
or
anything
like
that.
It's
a
layer
on
top.
If
you
will
work
two
layers
on
top.
B
B
So
any
other
questions
that
anybody
has
please
don't
hesitate
to
reach
out
as
Johnny
and
I
have
said
many
times
here
on
the
stream,
the
interactivity
that
we
get
from
you
all
the
questions
that
we
receive
both
here:
Live
While
We're
on
the
stream,
as
well
as
through
email
and
slack
and
all
of
the
other
places
and
things.
That's.
Why
we're
here?
B
That's
what
we
love
doing
so,
please
don't
hesitate
to
reach
out
if
you
have
anything
so
I
say
that,
because
we
are
coming
up
on
the
end
of
our
stream,
I
know:
Harriet
has
it's
late
in
the
day
for
you,
because
you're
across
the
pond,
as
they
say
and
also
I,
think
you
have
something
else
coming
up
here
in
just
a
couple
minutes.
So,
okay,
we
want
to
make
sure
that
you
have
time
to
to
get
to
that
and
maybe
enjoy
a
bite
of
dinner
beforehand
or
supper.
B
B
I've
I've
said
many
times
that
well
I
I
know
how
to
hand
wave
and
talk
about
get
Ops
at
a
high
level
and
and
understand
the
value
prop,
and
all
of
that
the
implementation
side
is
is,
is
a
little
bit
foreign
to
me.
So
thank
you
for
coming
on
today
helping
to
explain
that
Johnny
appreciate
you.
Demoing
and
yeah
so
I
will
hand
over
to
you
Johnny
for
any
last
words,
Harry.