►
Description
Hilliary and Jonny talk about the CI in CI/CD and tee up the next series of topics around how CI is done not just in the cloud.
A
Hello
get
options.
We
figured
out
why
I
have
technical
difficulties.
So
that's
why
we're
a
little
bit
late
today
and
it's
because
I
updated
Firefox
and
now
our
streaming
software
doesn't
want
to
work
with
Firefox.
So
I
had
to
go
back
to
Chrome
and
let's
see
how
my
CPU
performs
because
I'm
pretty
much
due
for
a
tech
refresh
at
this
point.
Well,.
A
Fam
in
terms
of
announcements,
I,
don't
know,
I
have
no
idea
what's
happening
in
the
good
ops
world.
Right
now,
to
be
honest
with
you,
I
made
I.
Think
people
pretty
well
know.
I
made
a
bit
of
a
change
in
my
wow.
My
dog
is
going
to
try
and
eat
my
cat
right
now
and
it's
about
to
happen,
and
you
all
will
get
to
hear
it
get
out.
Why
are
you
even
here?
A
B
Oh,
my
God,
why
are
you
back?
Isn't
it
funny,
like
the
you
know
like
when
you're
single,
you
never
think
you
hear
all
these
other
parents
and
people
are
like
you
know
whatever
with
their
kids
right?
Oh,
you
know,
Johnny
ate
the
Crayons
again,
you
know
and
you're
like
man.
That's
crazy!
I'll,
never
hear
that
and
then
you
have
kids
and
you're
like
dude
stop.
Why
are
you
eating
that?
What
are
you
doing?
Yeah.
A
A
B
A
Oh
my
God
you're,
so
heavy
right,
so
announcements,
I,
don't
have
any
I.
I
know
that
the
get
Ops
con
or
not
I
could
have
kind
of
see
the
argocon
co-located
event
for
kubecon
I
know
that
they
put
out
the
agenda
last
couple
days
and
that's
what
I
got
so
tell
us
if
we're
talking
tecton
today,
no
we're
not
talking
tecton!
Today
we
are
going
to
talk
about
continuous
integration.
A
So
what
we're
going
to
be
doing
for
the
next
several
episode
and
get
UPS
guide?
Is
we're
going
to
be
talking
about
how
lots
of
types
of
software
are
like
built,
shipped
and
and
delivered,
and
so
are
we
talking
about
what
continuous
integration
looks
like
and
what
delivery
looks
like
outside
of,
like
Cloud
native
environments,
a
little
bit
so
we'll
talk
about
it
inside
as
well?
A
I've
got
some
people
working
on
getting
us
a
I'm
working
on
getting
us
some
of
the
folks
from
the
kubernetes
community
to
talk
about
proud
and
wow.
This
is
levels
of
excitement
from
this
dog
he's
acting
like
he's
way
younger
than
he
is
so.
We've
got
that
coming
up
I'm
working
on
kind
of
scheduling,
the
dates
for
everybody
sit
down,
I
am
going
to
have
to
get
rid
of
him.
He
is
a
whole
problem
right
now.
I,
don't
know
who
let
him
in.
A
A
Then
you
know
like
I
rotate
it
out
so
like
which
ones
and
which
one's
outfit
I
don't
know
who
let
I
think
that
I
think
the
I
I
I'm
a
very
spoiled
person.
Who
has
someone
to
help
me
clean
my
house
once
a
week
and
I
think
she
let
him
in
because
he's
old
and
he
likes
to
be
in,
but
my
door
to
my
my
office
wasn't
closed.
A
A
B
A
B
Cool
yeah,
so
I
think
it
was
last
week
or
maybe
earlier
this
week,
Dev
spaces.
We
announced
it
on
the
asking
open
shift.
Admin
live
stream
yesterday,
but
devspace
is
updated
to
3.8
and
I
just
saw
an
announcement
about
the
migration
toolkit
for
virtualization
had
just
upgraded
to
I
think
it
was
V
2.5.
So
there's
some
good
stuff
coming
out.
If
you
didn't
catch
the
stream.
B
You
know
to
update
the
openshift
cluster,
to
put
the
shoes
are
out
there
and
set
them
up
out
and
get
a
git
lab
or
GitHub,
and
you
know
make
sure
that
they
have
access
to
the
repos
and
stuff
like
that
that
they
need.
So,
if
you
get
a
chance
to
check
it
out,
devspaces
is
a
really
cool
project.
I
didn't
realize
how
how
far
along
it
is
so
yeah
give
it
a
shot.
Give
it.
B
A
B
A
A
A
B
B
A
So
what
are
we
doing?
We're
talking
about
continuous
integration.
I
actually
have
some
slides
because
I
talk
about
continuous
integration,
all
the
time
and
we're
talking
a
little
bit
about
test
driven
development
because
for
the
next
several
episodes
it's
going
to
be
about
how
how
software
like
is,
is
delivered
across
the
realm
right
across
the
Galaxy.
Why
did
I
not
even
go
straight
there?
Obviously
I'm
under
caffeinated
today.
A
B
A
A
So
that's
great
news
very
excited
about
that
back
down
to
three
dogs.
Next
week
right,
continuous
integration,
I
swear
folks,
I'm
awake.
Do
you
remember
that
my
my
Twitter
handle
is
caffeinated
Integrations?
It
actually
only
works
as
the
caffeine
exists,
and
it
doesn't
not
until
we
shouldn't
quantities
this
morning
afternoon.
What
even
time
is
it
so
I'm
gonna
go?
Oh
I
forgot
to
just
screen
share
with
the
new
browser.
Oh,
let's
see
how
this
works.
It's.
A
I
am
not
a
long
distance
runner
on
the
four
dogs
needing
a
lot
of
exercise.
I
have
a
good
size
yard,
and
the
thing
is
when
you
have
a
pack
of
dogs
the
reason
for
keeping
a
pack
of
dogs
is
they
actually
exercise
each
other,
so
they
spend
several
hours
a
day
playing
and
wrestling
and
chasing
and
and
doing
all
of
the
dog
things
and
then
I
get
to
take
them
on
nice.
Calm,
leisurely
walks
because
they're
not
completely
over
energized.
B
B
A
So
yeah,
but
he's
he's
too
old
for
the
subway.
Now
that
would
bug
him
too
much,
but
when
he
was
younger
he
he
didn't
mind
it
but
yeah.
So
that's
that's.
How
I
do
it
is
keeping
a
pack
training
it
like
a
pack.
I
could
talk
for
the
about
the
theory
of
dog
training
for
forever.
That
could
be
end
up
being
this
entire
episode.
If
you
let
me
so
we're
not
gonna.
Do
that
ping
me
some
other
time
and
I
will
tell
you
all
about
dog
training.
A
I,
actually
just
sent
one
of
my
colleagues
videos
on
how
to
train
your
dog
for
calm,
because
one
of
my
dogs,
I
can
well
I.
Theoretically,
I
can
stack
treats
on
his
paw,
but
he
actually
really
hates
it.
When
I
do
that
so
he'll,
let
me
get
like
two
or
three
two
or
three
treats
high
and
then
he'll
yank
his
paw
back,
but
he
won't
eat
the
things
until
I
tell
him
to
which
is
just
him
being
like
exactly
that.
Much
of
a
jerk
yeah
anyway
I
digress.
A
I
could
talk
about
my
dogs
all
day.
I
have
dogs,
I,
have
cats,
I've
got
kids,
it's
it's
Mayhem
over
here.
It's
a
lot
to
talk
about.
Don't
get
me
started
right,
so
I
think
I've
said
this
on
the
stream
before,
but
for
folks,
who
don't
know
I
actually
spent
like
the
majority
of
my
career
in
quality,
engineering
and
quality
engineering
and
reliability.
Engineering
have
a
lot
in
common,
but
you
know
one
of
the
things
that
I
spend
a
lot
of
time
on
and
I
just
I
was
just
debugging.
A
Something
on
one
today
spent
a
lot
of
time
on
continuous
integration
systems
right.
This
is
a
place
where
I've
spent
a
huge
amount
of
my
career
and
so
continuous
integration
systems.
You
know
we
talk
about
tecton
I,
like
tecton,
because
when
you
get
checked
on
going
in
openshift
and
you've
got
the
nice
real
like
this
is
what
step
I'm
on,
and
this
is
what
I'm
doing
and
like
that
whole
visual
set
I
really
like
it,
because
it
just
makes
things
so
much
easier
to
see
through
gitlab
has
like
really
similar
functionality.
A
A
It's
like
still
that
really
big
important.
First
half
of
the
thing,
though
right
you
don't
really
need
like.
If
you
don't
have
CI,
what
are
you
doing
with
CD
right,
you're,
not
you're,
not
doing
one
without
the
other,
you
might
do
CI
without
CD.
Actually
you
really
yeah
like,
but
you
I,
don't
it's
kind
of
there's
a
reason.
It
comes.
First,
even
though
it's
not
alphabetical
so
CI
is
it's.
A
It's
such
an
interesting
thing
and
I
was
talking
to
one
of
my
friends
here
at
red
hat
and
she's
over
in
Rel,
and
we
actually
started
talking
about.
I
was
like
you
know.
How
is
I?
Guess
you
don't
do
get
Ops
with
you
know
with
something
like
like
a
Linux
like
kernel
right,
but
I
was
like
how
is
Rel
like
what's
what
what
is
like
the
involvement
with
get
out
with
get
and
testing
and
like?
A
Is
there
a
concept
of
continuous
delivery
with
with
those
types
of
things,
or
at
least
like
continuous
integration
and
the
answer
that
was
yes
and
she
started
showing
me
some
of
the
tools
and
the
way
they
test
things
and
so
forth,
and
some
of
those
things
are
I.
Think
internal
only
and
I
was
like
oh
dang,
I,
don't
think
I
can
talk
about
any
of
this
stuff.
A
I
was
like,
let's
do
a
little
Johnny.
We
were
talking
about
soccer.
I
was
like,
let's
do
like
a
little
bit
of
a
series
about
this,
and
let's
actually
look
into
how
this
it.
You
know
what
this
looks
like
in
spaces
outside
of
the
cloud,
because
we
we
we're
here
to
be
nerdy
about
interesting
things
and
to
learn
interesting
things
and
I,
just
kind
of
want
to
take
take
a
few
weeks
and
Branch
out
away
from
a
little
bit
of
what
we've
been
doing
this
whole
time
and
get
into
to
the
CI.
A
So
as
we
kind
of
like
I've
given
talks
about
CI,
a
lot
like
this
is
slides
from
a
talk.
I
gave
and
I
copied
these
slides
from
a
different
talk.
I
gave
a
few
like
a
year
prior
and
like
this
is
this.
Is
this?
Is
a
make
file
right,
and
so
a
lot
of
you
know
when
we
talked
about
like
I,
think
it
was
getting
get
up
that
episode
with
oh,
my
God
I
hate,
saying
his
name
because
I'm
really
bad
at
it.
A
B
A
A
So
with
FTP
FTP,
you
know
FTP
the
the
the
the
new
code
to
release
it,
but
you
know
this
is
like
just
a
little
I
I
made
this
make
file
for
for
a
web
app
and
I've
just
updated
it
to
be
current,
because
you
could
actually
run
these
things
and
it
would
work
and,
of
course,
mad
props
to
anybody
who
gets
the
the
reference
in
the
in
in
the
in
the
Echo
so
kind
of
like.
But
this
is
basically
what
happened
right.
A
I,
don't
know
what
to
tell
you
about
some
of
the
startups
I've
worked
for,
but
I've
got
some
interesting
stories
and
I.
Think
some
of
that
comes
from
having
worked
for
very,
very
small
companies,
doing
things
in
very
interesting
ways,
but
that's
another
I
think
I
spent
a
whole
different
hour
on
that,
so
the
right
so
like
the
way
I
kind
of
have
always
phrased
this.
Is
we
basically
just
kind
of
we
had?
A
We
would
do
something
like
you
know
these
make
files
or
whatever,
and
basically
we
just
you
know-
I
have
to
do
some
bash
things,
so
we
start
automating
our
bash.
We
started
we
get
some
bash
scripts
and
so
forth,
and-
and
at
this
point
a
lot
of
CI
is
really
just
like
bash
orchestration,
That's
What,
I
Call,
it's
bash
orchestration.
A
We
still
have
to
do
it
all
with
bash
we've
just
created
like
orchestration
systems
for
our
bash,
so
that
it
does
the
things
that
we
want,
and
so
that's
kind
of
like
how
how
it
started
and
like
this
is
kind
of
like
how
it's
going
so
I
actually
borrowed
the
format
for
these
slides
from
one
of
red,
Hat's
other
slide
decks
and
I
I
feel
so
terrible.
I
think
I
got
it
from
Christian,
but
I
think
he
may
have
got
it
from
somebody
else.
A
So
someone
somewhere
deserves
a
lot
of
credit
for
the
the
slide
formatting
here,
and
that
someone
is
not
me,
but
I
couldn't
tell
you
who
and
so
I
actually
changed
what
their
had
was,
what
they
had
a
bit,
though,
because
they
kind
of
just
had
it
as
one,
and
so
you
have.
This
whole
thing
is
the
concept
of
continuous
delivery.
A
Right,
like
you've,
got
all
the
steps
and
you
can
look
at
continuous
delivery
as
kind
of
an
overarching
concept,
not
just
the
delivery
aspect
or
the
get
Ops
aspect,
and
that
kind
of
just
is
not
really.
You
know
it
just
depends
on
who
you
talk
to
I
suppose,
but
the
continuous
integration
part
is
right.
It's
like
build
tests,
you
know,
do
your
security
checks
if
you
have
to
do
any
lockdown
procedures.
A
Anything
like
that,
like
all
any
of
those
little
steps
that
come
into
you
know
creating
the
thing
you
want
to
release,
creating
the
the
software
package
that
you
want
to
promote
and
then
a
deployment
to
Stage,
another
set
of
tests
and
a
deployment
to
production,
and
so
a
lot
of
times.
People
continue
like
cicd,
like
you
deployed
to
stage
and
like
that's
it,
that's
your
CI
CD
and
then
production
might
happen
on
a
less
continuous
Cadence
going
all
the
way
to
production,
sometimes
like
I,
think
Christian
called
it.
A
Something
like
it's
like
a
CI,
CD,
plus
plus
and
there's
kind
of
you
know,
depending
on
like
the
who,
the
what,
where
the,
why
the?
How
of
your
organization,
like
call
yourself
done
at
whatever
makes
sense
for
your
team,
and
that's
fine,
so,
like
I,
said,
I
spent
a
lot
of
time
in
the
like
Ci
space
directly
right
I
obviously
do
touch
the
CD
space
as
well.
B
A
To
a
degree,
this
a
CI
is
a
very
living
breathing
thing,
because
you're
probably
making
changes
in
your
CI
system
actually
at
every
release
right
for
every
release,
you
probably
increased
your
tests
right,
there's,
probably
something
new
in
there.
You
might
have
fixed
something
you
might
have
improved.
Something
like
that
that
test
block
like
continues
to
grow.
You
can,
you
might
refactor
tests,
you
might
change
your
testing.
A
Most
people
use
FTP,
didn't
have
a
unit
background
I
think
that's
accurate
for
the
people
who
set
up
the
way
we
develop
or
redeploy
to
production
for
that
particular
company.
Our
hope9
I'm
just
gonna,
leave
it
like
I.
Don't
I,
don't
think
that
there
was
a
Unix
background
involved
from
at
least
one
of
the
one
of
the
persons
major
digression
so
right
so
you're
you're
changing
things
in
your
CI
system
a
lot
and
we
actually
were
having
like
discussions
on
like
what
goes
into
a
CI
system.
A
What
makes
for
a
good
end-to-end
test
right,
there's
so
many
types
of
testing
as
well.
You've
got
your
your
unit
test,
your
functional
test,
your
regression
tests,
your
integration
tests,
penetration
tests,
load,
stress,
Etc,
like
that's,
actually
like
a
really
big
thing,
and
then
Security
checks,
maybe
before
the
testing
and
maybe
after
the
testing
you
might
be
during
the
testing
a
lot
of
modern
CI
systems.
You
can
do
a
bunch
of
things
concurrently,
which
is
super
cool
yeah.
A
A
Now
you
have,
you
know,
build
pipelines
and
the
pipelines
have
tasks
the
tasks
of
subtasks
and
you
can
kind
of
Branch
out
some
stuff
and
in
Cloud
native
environment
right.
You
spin
out
multiple
copies
and
do
different
things
to
each
copy
right
and
you
can
get
a
lot
done
in
less
time,
but
it
always
what
you
are
doing
at
every
single.
A
So
we
didn't
have
to
go
into
any
systems
or
do
anything.
We
just
had
to
go.
Look
at
git
and
see
what
we're
on.
When
did
it
run?
What
passed,
what
failed,
and
so
you
could
actually
it
allows
you
to
narrow
down
like
okay,
here's
the
first
time
that
test
failed.
So
this
is
probably
the
commit
that
broke
this
River
and
that
created
this
regression
right.
A
B
And
you
can
use
that
to
go
back
and
look
and
like
okay,
this
is
probably
a
one-off
right
or
maybe
there
was
just
like.
Maybe
there's
a
switch
or
something
like
that
that
went
down
in
the
network
that
was
causing
this
blip
for
a
day
or
two
or
something
like
that.
I,
don't
know
I'm
just
making
something
up
but
like
over
over
a
period
of
time.
You
can
go
back
and
look
at
like
okay,
there's
a
reason
for
this
right
go
back,
correlate!
A
Exactly
exactly
you
know,
the
problem
with
continuous
integration
it
in
in
Old
monolithic
systems
is
sometimes
why,
while
the
pipeline
was
running,
a
new
commit
would
come
in.
That
would
have
it
actually
all
redeploy,
so
you
could
actually
see
CI
like
runs
fail
due
to
the
fact
that
it
was
redeploying
in
the
middle
of
the
other
system.
Trying
to
do
you
know
or
the
last
last
integration
trying
to
do
things.
A
Because
of
the
fact
that
you
can
be
like
okay.
Well,
it's
going
to
be
a
whole
new.
You
know
it's
just
going
to
be
a
whole
new
set
of
the
application
on
on
the
on
the
on
the
cluster
or
you
know
what
have
you
VMS
or
depending
on
depending
on
your
setup
right
but
yeah?
Sometimes
the
failures
would
be
just
in
monolithic
systems,
where
you
know
you
just
kind
of
everything
around
all
at
one
time,
or
it
only
deployed
to
one
one
like
environment
that
was
never
ephemeral
and
was
always
up.
A
You
could
definitely
have
the
environment
redeploying
in
the
middle
of
tests
running
and
because
of
the
way
that
the
systems
were
all
like
you.
Basically,
it's
just
a
series
of
callbacks
to
things
that
are
going
to
run
independently
of
of
without
really
true
or
contextual
knowledge,
and
that
can
still
happen
right.
It
can
still
happen,
but
yeah
it's
it's
all
these
things
are
kind
of
running
independently.
They
don't
really
have
like
knowledge
of
the
other
thing,
which
is
a
huge
it's
kind
of
like
a
gap
in
CI
technology,
so
yeah.
B
Yeah,
the
way
that
we
do
it
is
we
use
I,
think
pretty
similar
to
what
you're
talking
about
where
you
had
like
your
your
code
document
come
back
up
in
a
formatted
way
where
we
use
like
color
coding
so
like.
If,
if
the
deployment
of
an
application
failed,
then
it's
a
heart,
it's
red
right
like
it
failed.
B
It
was
the
app
itself,
but
if
it
was
something
on
like
infrastructure
side,
so
maybe
I
was
maybe
I
was
spinning
up
a
cluster
and
you
know
you
can
run
openshift
installed,
10
out
of
10
times
and
it'll
be
fine
one
day
and
then
nine
out
of
ten
times
is
fine,
but
there's
always
that
one
time
and
so
anything
on
the
infrastructure
side
that
gets
hung
up.
You
know
it'll
come
back
as
like
a
like,
an
amber
color
or
something
like
that,
and
then,
if
it
goes
all
the
way
through
the
screen
right.
B
So
we
know
it's
good.
So
that
way
you
kind
of
get
that
quick,
dashboard.
Look
at
like
okay,
hey
this
is
cool
and
then
yeah
anything
red
or
yellow
you
can.
You
can
dig
in
go
straight
to
the
Jenkins
job
and
take
a
look
at
that
or
you
know
it's
it's
it's
quite
nice
and
when
you
need
to
report
up
to
people
like
hey,
what
are
you
guys
doing
or
how's
it
looking?
You
can
be
like.
Oh,
let's
go
check
out
the
dashboard
yourself
in
like
there.
It
is.
A
Yeah
exactly
it's
nice,
like
I,
said:
if
you
look
at
things
like
tecton
right,
they
make
you,
you
can
visualize
all
that
now
like
it
does
it
for
you.
We've
done
everybody's
done
something
weird
to
get
that
kind
of
same
outcome,
but
it
ends
up
being
a
very
like
common.
Like
experience,
I
did
this
thing
to
make
it
obvious
when
there
was
a
problem
or
when.
B
A
A
This,
because
what
ends
up
happening
is,
you
might
have
some
of
the
testing
happen
really
regularly
and
then
for
the
other
types
of
testing.
If
load
is
just
testing
happen,
even
at
all,
they're,
maybe
done
on
like
an
interval,
and
sometimes
that
interval
is
not
automated.
I
will
say
I've
seen
like
both
ends
of
the
spectrum,
where
everything
is
done
like
as
as
much
as
possible
at
Red
Hat.
We
have
to
do
this
right.
All
of
these
things
like
that
they're
mentioned
here.
A
We
have
to
do
all
of
them
on
everything
all
the
time,
especially
the
security
and
the
vulnerability
and
all
that
other
stuff,
because,
like
we
need
to
be
able
to
lock
down
things
and
guarantee
what
we
are
providing
meets
very
high,
like
regulation
standards.
So
we
definitely
have
seen
like
our
spectrum
of
the
things
that
I've
definitely
seen
the
oh.
No,
this
is
totally
fine
spectrum
of
things.
A
I
I
find
that
with,
in
my
opinion,
with
like
most
modern
software
architectures,
because
everything's
microservice,
driven
that
what
your
API
does
is
really
like.
The
most
important
thing
to
test
and
I
ended
up
personally
favoring,
it's
like
even
with
things
like
one
of
those
things
like
uis
that
are
like
you.
A
You
make
a
change
to
your
UI
somewhere
and
you
can
break
all
of
the
tests
because
the
way
it
does
selectors
and
so
forth,
it
may
not
necessarily
work
so
what
I
have
found
that
I
I
prefer
is
I
actually
like
to
do
a
lot
of
the
integration
style,
testing
and
so
forth.
You
can
do
mock
objects
into
the
unit
testing
and
do
a
lot
there
and
then
the
thing
that
I
think
matters
the
most
and
is
the
thing
I
see
done.
A
The
least
is
negative
test
casing
negative
test
casing
is
I
know
how
I
expect
this
to
work
now.
I
need
to
figure
out
how
it
fails
and
that's
where
you
start
testing
how
your
software
responds
to
bad
situations.
So,
if
you
imagine,
for
example,
that
you
have
an
integration
with
an
Amazon
API-
and
this
is
something
I've
actually
seen
and
they
change
the
way
their
API
like
formats,
the
response
they
remove
a
field,
they
add
a
field
or
something
like
that
right.
A
How
do
you
want
your
software
to
behave
when
that
happens,
and
are
you
testing
for
how
it
behaves?
If
something
like
that
happens,
are
you
so,
and
that
can
all
be
done?
You
can
do
it
in
the
unit
tests,
and
so
that's
that's
where
I
like
to
put
I
actually
will
put
as
much
as
like
humanly
possible
into
the
unit
testing
side
of
things
and
just
basically
like
what
happens
when
this
gets
something.
That's
too
long.
What
happens
when
this
gets
something?
That's
too
short.
What
happens
when
this
gets?
Something?
A
That's
out
of
order.
Does
you
know?
Is
this
able
to
is
a
system
able
to
handle
changes,
because
sometimes
you,
even
if
it's
you,
don't
have
a
dependency
on
external
API,
sometimes
developers
within
you
know
some
of
the
same
groups
may
make
a
change
that
seemingly
shouldn't
break
anything,
but
actually
can
because
of
the
way
something
they
didn't
directly
touch
during
their
feature.
Improvement
actually
like
works
well.
B
If
you
remember
not,
too
long
ago,
openshift
install
I
think
it
was
4.12
like
one
of
the
early
versions
of
4.12.
It
might
have
been
413
I
can't
remember,
but
the
AWS
API
had
blocked
ACLS
on
S3
buckets
like
a
certain
type
of
ACL
on
S3,
and
it
was
just
something
that
they
made
a
change
and
like
they
announced
it
and
I
I.
Don't
know
what
happened
it's
just
it.
B
So
it's
got
a
lot
of
the
old
artifacts
from
like
when
they
first
stood
up
AWS,
so
like
the
way
that
you
provision
a
bucket,
they
made
it
more
restrictive
so
like
if
you're
going
to
be
in
USC's
one,
you
have
to
use
USC's
one
you
have
to
like
use
certain
parameters
and
stuff
like
that,
where
everything
else
just
kind
of
works,
so
it
it's
not
what
you're
saying
isn't
a
like
a
severe
Edge
case.
That's
a
legit
reality
that
could
happen.
Pretty
quickly
happens.
A
All
of
the
time,
all
of
the
time,
and
and
it's
things
like
you
know,
a
lot
of
times
with
companies,
what
will
happen
is
somebody
creates
the
like
AWS
account,
for
example,
and
but
whatever
email
address
is
used,
maybe
that
person
isn't
exist
anymore
or
goes
to
a
bucket
that
nobody
really
checks
like
the
way
of
actually
reaching
account.
Owners
is
really
really
difficult,
and
so
sometimes
you
know
even
with
best
efforts
of
like
trying
to
tell
people
hey.
This
is
changing
we're
deprecating,
this
API
et
cetera.
A
They
just
don't
get
the
notifications,
because
it's
not
actually
set
up
to
reach
a
real
human
happens
all
the
time,
because
what
with
account
ownership,
you
know
it's
really
actually
like
I.
Don't
think
anybody
has
actually
truly
solved
this
well
in
a
concept
of
here's,
okay,
somebody
goes
and
actually
owns
this
AWS
account,
or
this
is
your
account
or
this
red
hat
account
or
whatever
right
and
like
so
they
get
notified,
but
they're,
not
there
anymore,
or
it's
not
a
real
email
address
or
like
it's.
It's
not
great.
A
B
A
It's
like
this
is
like
an
Institutional
problem.
I
think
everybody
struggles
with
this
I
know
that
we've
seen
it
here.
It
has
happened
to
us
at
Red
Hat,
not
knowing
something
some
places.
My
husband
worked
like
it
happened
to
them.
They
didn't
know
something
they
couldn't
do
a
very
major
function
that
they
needed
to
be
able
to
do
to
satisfy
their
customers,
and
it
was
a
whole
thing,
and
so
this
it
just
happens
it
just
does.
So.
A
That's
why
you
need
to
know
what
will
happen
like
what
will
happen
to
your
software
in
those
situations.
How
are
you
going
to
fail,
because,
if
you're
failing
out
with
like
a
500,
that's
pretty
bad,
so
you
need
to
be
writing
graceful
failures
into
your
software.
You
need
to
be
checking
for
how
software
fails
during
the
the
testing
phase
and
that's
what
negative
testing
means
is
like
I
I
tested
that
everything
works.
A
The
way
I
want
it
to
work
now,
I'm
going
to
look
for
how
it
fails
now,
I'm
going
to
see
what
happens
when
I
do
bad
things
to
my
software,
that's
negative
testing.
It's
tested
the
unhappy
paths
or
the
unexpected
things,
some
that
could
happen,
maybe
that
a
customer
does-
or
you
know,
like
I,
said
if
you've
got
an
integration
or
a
dependency,
that
the
dependency
does
and
with
microservices
and
the
way
that
we
use
we're
cobbling
together
things
from
like
a
bunch
of
different
sources.
A
B
A
I,
don't
if
you've
never
heard
the
term
negative
testing
before
I
feel
very
sad,
yeah
I'm,
just
gonna
leave
that
I
feel
very
sad.
That
is
it.
That
is
a
term
that
people
should
know.
Almost
every
single
like
serious,
like
Day
Zero
sav-zero,
like
cve,
would
have
been
caught
with
negative
testing.
Yeah
like
if
you
think
about
heart,
bleed.
What
are
they
basically
just
threw
a
negative
test
at
it
to
see
what
would
happen?
Oh,
it
turns
out.
It
gives
me
a
bunch
of
stuff
all
right,
cool
right
on.
B
A
Same
thing,
same
thing,
like
negative
testing,
is
one
of
the
most
important
aspects
to
creating
security
like
today,
you
can
get
involved
in
the
security
of
your
company
by
adding
negative
test
cases
to
your
unit
testing
like
just
go,
do
it
now.
That
is
how
you
today
can
make
a
positive
security
impact,
because
that's
a
major
digression,
though
anyway.
B
Real
quick
before
we
move
on
because
Angela's
making
some
good
points
here.
So
Angela
said
that
chaos
engineering
is
really
a
great
way
to
help.
You
discover
some
of
those
negative
tests,
because
you
know
you'll
start
seeing
things
that
you
can't
imagine,
and
then
you
know
testing
for
the
negative
case
in
the
bad
case.
B
A
A
The
what
do
you
call
it?
That
was
the
slide
I
meant
to
go
to,
but
whatever
we're
actually
because
it
gets
us
into
exactly
test
driven
development
right
and
so
the
process
of
defining
your
test
before
you
write
your
software,
so
I
think
a
lot
of
us
right
when
we're
going
through
the
process
of
writing.
Writing
some
software.
A
A
A
It's
really
just
about
getting
something
together
that
that
works,
and
one
of
the
way
like
test
driven
development,
like
figuring
out
like
how
you
would
define
whether
or
not
a
step
works
for
me,
helps
with
the
iteration
so
that,
instead
of
getting
as
I
mean
I
think
we've
all
literally
done
this
like
we
write
something
and
it's
like
250
lines
of
code
and
then
after
that,
I
was
like.
Okay,
this
all
works
and
then,
when
we
go
back
and
we
refine
and
it
gets
down
to
like
90
right
yeah
all
the
time
all.
A
A
What
are
my
steps
to
get
there
and
how
would
I
validate
each
step
Works
and
so
the
idea
of
kind
of
like
designing
your
unit
tests
ahead
of
defy
of
actually
writing
your
code?
It
just
kind
of
helps.
You
make
sure
that
you're
kind
of
getting
all
the
things
together
up
front,
and
then
you
actually
start
like
really
like
executing
those
unit
tests
and
then
start
getting
into
like
the
builds
and
everything
else
like
that
right.
So.
A
Test
driven
development
like
for
me
test
driven
development,
continuous
integration
kind
of
were
like
simultaneous
movements
like
from
the
perspective
of
where
I
was
when
I
started
my
career
and
what
I
was
working
on
at
the
time.
Those
two
things
really
like
came
up
together.
They
were
really
coupled
and
then
like
continuous
delivery,
sort
of
came
after
and
then
you
know
because
CI
and
CD
have
been
like
so
tightly
coupled
but
I.
A
Think
I
think
you
can
really
like
say
like
test
driven
development
is
like
a
really
big
like
Cornerstone
to
to
why
all
these
things
we're
doing.
It
is
a
lot
about
like
wanting
to
have
faith
in
what
we
were
putting
in
front
of
in
front
of
customers,
so
I
think
that's
it
for
my
slides.
I
have
like
really.
A
Bad
architecture
diagrams
in
here,
which
I
literally
just
did
because
one
this
presentation
was
done
for
some
other
group
where
I
talked
about
other
things
and
two
because
I'm
really
bad
I
am
not
good
at
these
folks.
I
I,
just
somebody
who's
great
at
diagrams,
needs
to
like
teach
me
how
to
be
good
at
diagrams,
because
I'm
not
I,
don't
know
how
to
make
things
pretty
yeah.
But
the
point
is
that
kind
of
finishes.
B
Yeah
I
I,
with
with
diagrams
always
try
like
my
old
manager,
actually
taught
me
this
like
it
was.
It
was
telling
a
story
right,
so
he
gets
up
on
the
Whiteboard
and
he's
like
hey
look.
Everything
starts
here,
and
so
he
would
start
at
the
bottom
block
and
he'd
write
in
what
it
was
and
I
have
some
little
pointers
out
to
the
side
and
then
he's
just
going
to
build
up
on
top
of
it.
B
A
A
B
B
Start
like
I'll,
try
and
explain
it
back
and,
and
so
like
I'll
go
through
it.
If
I
get
stuck,
then
I'll
go
back
and
review
what
that
thing
is,
but
that
helps
me
commit
it
to
memory,
but
then
it
also
like
somewhere
deep
in
the
recesses
of
my
brain
right.
It
kind
of
puts
it
puts
that
map
together.
You
know
so
it's
like
okay,
I
kind
of
see
the
flow
and
then
I
can,
when
I'm,
going
back
I'm
thinking
about
it.
I
can
think
in
that
particular
spot.
You
know
so.
A
So
anyway,
we
can
skip
this
Slide.
The
point
is
it's
a
bad
slide,
but
it
was
helpful
for
that
presentation
and
I.
Don't
care
about
any
of
these
other
slides
I
already
share
this
in
popular
opinion
and
I.
Think
the
slides
duplicate
I
know
why
they
duplicated.
Actually
it
was
a
whole
that
was
the
whole
thing
anyway.
B
Well,
and
with
with
the
whole,
you
know
like
focus
on
supply,
chain
and
stuff
like
that
right,
like
everything
that
you're
talking
about
right
now,
it's
I
think
it
was
everybody
knew
about
it
right.
It
was
one
of
those
things
where
it's
like
yeah
I
know
what
cicd
is
I
know
what
you
know,
maybe
not
test
driven
development
for
everybody,
but
for
I.
Think
by
and
large
everybody
knew
like
the
concepts
of
what
was
going
on,
but
nobody
really
cared.
You
know
it's.
A
B
Not
outside
the
the
development
team
right
they're
like
okay,
look,
you
gotta,
build
server,
that's
really
great!
That's
awesome!
But
now
now
that
people
have
been
getting
owned
right
by
some
of
these
groups
and
some
of
the
software
that
they
put
in
they're
like
okay
I,
do
need
to
pay
attention
to
this.
I.
Do
need
to
understand
this
and
I.
Think
that
that's
why
it's
becoming
like
a
bigger
topic
for
people
that
aren't
necessarily
developers
and
why
they
are
trying
to
understand
it.
So
that
way
they
can.
B
You
know
like
where
is
this
happening
at
like
what
does
it
mean
when
this
happens
and
stuff
like
that?
And
so,
when
you
know
Johnny
and
Hillary
go
to,
and
we
say,
hey
look:
this
is
broke
because
of
this.
It's
not
just
this!
So
what
it's
like!
Okay,
we
got
to
get
this
done
because
I'm
about
to
push
this
thing
to
production
and
I,
don't
want
to
push
it
out
there
with
you
know
a
back
door.
That's
gonna!
Yeah!
Have
all
my
customers
get
owned
exactly.
A
And
so
here
the
thing
about
it
is
one
of
the
reasons
I
want
to
touch
on.
This
is
exactly
the
supply
chain
thing,
because,
as
get
Ops
is
getting
a
lot
of
attention
for
being
able
to
know
like
what
is
what
is
in
the
software
right.
Github
has
a
lot
of
attention
for
that,
but
you
can
get
a
lot
of
this
information.
This
artifact
collection
and
out
of
a
CI
system
where
CD
and
get
Ops
is
not
necessarily
going
to
be
an
option.
A
So
things
like
the
Linux
kernel
like
I
was
mentioning
is
one
thing
we
want
to
look
at
on
other
things
right.
So
the
CI
systems
and
kind
of
the
power
of
CI
systems
is
get
Ops,
eat
enough
and
I
care
about
it
enough
that
we're
going
to
spend
some
time
talking
about
it
over
the
next
few
weeks,
and
so
like
I,
said
I
just
wanted
to
spend
like
today,
kind
of
setting
the
stage
for
like
what
is
continuous
integration.
A
What
belongs
in
a
continuous
integration
set
up
and
I'm
going
to
start
getting
into
some
like
real
examples
of
this
outside
of
outside
of
like
git
Ops,
which
I'm
pretty
excited
about,
because
I
was
looking
through.
You
know
coming
up
with
you.
You
know
this
because
you're
both
here
and
on
ask
an
openshift
admin
coming
up
with
like
the
set
of
topics
and
like
what
are
we
going
to
talk
about?
A
What's
going
to
be
useful,
what's
going
to
be
engaging,
what's
going
to
make
and
what's
going
to
provide
value
right
because
we're
doing
this
to
we
don't
monetize
these
videos
right?
We
literally
just
do
this
to
provide
additional
value
to
the
community,
to
our
customers
to
other
red
Hatters,
and
we
do
this
in
our
spare
time
right.
B
A
B
B
And
I'm
sorry
I
didn't
mean
to
cut
you
off
there,
but
like
a
lot
of
times
when
I've
been
talking
to
customers
about
good
Ops-
or
you
know
other
red,
Hatters
and
stuff
like
that
right,
there's
there's
still
that
fundamental
part
of
like
Source
Control
Management
that
people
aren't
familiar
with
right.
How
do
I?
How
do
I
and
my
team
work
on
the
same
repository?
B
It's
if
you
don't
know
how
to
use
git
or
if
you
don't
know
how
to
use
the
source
control
manager
like
it's
going
to
be
very
difficult
to
simultaneously
work
on
the
same
projects
with
other
people,
because
you
were
literally
going
to
be
stomping
on
each
other,
so
you'll
you'll
be
in
the
situation
where
you
know,
since
Hillary
can
type
well
he'll
be
typing,
you
know
we'll
send
everything
to
her
she'll
merge
everything
in
and
then
push
it
up.
It's
it's
just.
A
You
know
there
are:
there
are
many
projects,
even
here
at
red
hat,
that
don't
use
git,
they
use
other
source
control
and
other
source
Control
Systems
have
other
feature
sets
there's
some
some
similarities
right.
There's
some
things
that
you
can
pretty
much
guarantee
is
going
to
exist
in
a
source,
control,
industry
control
system,
but
some
of
the
concepts
of
branching
strategies
and
so
forth
that
we
get
familiar
with
and
get
don't
exist
and
I
think
I.
Think
the
rail
space
is
like
that.
A
A
No,
it's
something
else
with
something
else,
but
yeah.
So
this
is
why
it's
kind
of
an
important
thing
and
we're
talking
a
little
bit
more
broadly
I-
want
to
learn
about
this
stuff
more
because
it's
like
a
nerdy
thing,
I'm
interested
in
and
I
get
to
decide
what
we
do
on
this
stream.
So
you
all
get
to
do
the
nerdy
things
I'm
interested
in,
because
I
did
ask
I
said
viewers.
A
B
A
good
one
and
I
think
like
it
again,
not
necessarily
being
gadopsy
like
100.
It's
it's
still
part
of
the
core
right
like
you
have
to
know
how
to
do
this
and
it's
important,
and
even
if
you
don't
know
at
least
you
get
to
look
at
some
tools
and
well
that's
kind
of
cool.
Maybe
I
can
make
that
work.
In
my
you
know
my
organization.
A
A
A
B
A
Still
using
it,
people
are
using
other
things,
they're
using
closed,
Source
ways
of
doing
Source
control.
There
are
those
are
those
that
that's
a
thing:
I,
don't
know
what
everybody's
doing
exactly
kerforce.
Actually,
you
know
what
that's
I
actually
was
literally
just
looking
at,
because
I
was
like
I
wonder
if
I
can
find
a
really
good.
A
Oh
I
did
use
CVS
once
there
we
go,
I
was
actually
just
looking.
I
was
looking
to
see
if
I
could
find
a
diagram
about
about
get
branching
strategies,
because
there
was
a
book
on
continuous
integration.
I
read
10
years
ago
that
had
some
cool
diagrams
and
honestly,
when
I
like
quickly
like
just
did
a
little
I,
don't
use
Google
I
use
a
kojo,
but
I
did
a
little
Lakota
search
for
for,
like
CI
branching
strategies.
This
this
perforce
document
showed
up.
A
So
this
is
not
me
endorsing
them.
It's
just,
ironically,
that
that
was
what
I
pulled
up
to
look
for
to
see
if
there
was
a
good
diagram
and
then
Angela
mentioned
perforce.
So
there
we
go
there,
we
go
there.
It
is
right.
So,
what's
next
right,
I
think
we
can
talk
about.
What's
next
further
questions,
I,
don't
know
if
they're
requested.
A
She
is
in
India,
though,
so
what
we'll
have
to
do
is
actually
we're
going
to
have
to
do
a
special
streaming
time
to
be
able
to
capture
her,
and
this
might
be
a
thing
that
we
have
to
do
to
get
some
really
neat.
Experts
in
these
fields
is
change
our
streaming
time
a
little
bit
so
keep
an
eye
on
what
I,
when
I
post
the
episodes
to
to
the
calendar
I
actually
will
change
them.
A
So
our
public
calendar
on
our
streaming
calendar
will
get
updated
based
on
like
what
what's
gonna
actually
happen
and
I
will
post
it
on
LinkedIn
and
Johnny
will
probably
post
it
on
Twitter
and
because
I
forget
Twitter
exists.
Sometimes
I
literally
only
use
it
for
this
and
I
still
forget
that
it's
there,
and
so
is
it
actually
even
still
called
Twitter
I.
Think
we
keep
asking
this
and
I
still
don't
know.
I.
A
I,
don't
anyway,
very
social
media
I
do
have
blue
skies.
I
will
try
to
be
using
it
more.
If
you
follow
me
on
Twitter
I
did
post
my
blue
skies
like
profile
Link
in
there
I
think
so
that
you
can
follow
it
and
then
I'll
try
to
actually
remember
to
use
this,
but
yeah
so
I'll
be
I
will
be
trying
to
do
my
best
to
communicate
the
changes
for
the
special
episode
times
and
I
had
no
idea.
You
were
in
Tokyo.
Maybe
you
told
me
this
once
before
or
hope.
A
A
So
yes,
we're
gonna
we're
gonna
we're
gonna,
potentially
dance
the
times
around
a
little
bit
so
that
we
can
get
some
pretty
cool
people
onto
the
stream
to
talk
about
some
critical
topics.
That's
really
what
I'm
getting
at
and
I
I
would
I
will
communicate
the
change
try
to
do
it
a
little
bit
faster
that
I've
been
kind
of
doing
it.
I've
kind
of
lazily
post
like
what
the
episode
is
like
the
day
before,
because
the
date
times
haven't
changed.
A
For
these
special
episodes,
I
will
be
communicating
as
as
rapidly
as
I
can,
when
it's
all
locked
in,
so
that
folks
can
adjust
the
schedule
or
a
lot
of
people
watch
this
on
a
delay
and
catch
the
recording
later
and
I
appreciate
them
too,
because
I
never
know
who
they
are
and
then
suddenly
somebody
will
send
me
a
message
and
be
like
oh
yeah,
I
watch
this
regularly
on
this
particular
episode
when
I
watched
it
I
saw
this
and
I
was
like
really
cool
and
I'm
like
I.
Had
no
idea,
you
were
a
viewer.
B
A
That's
wait,
so
we
don't
need
to
know
who
everybody
is
it's
totally
fine
anyway.
So
yeah
like
look,
keep
an
eye
out.
If
you
already
follow
me
great,
if
you
don't
follow
me,
I'm
sorry,
look
for
the
calendar
events
to
move
to
email.
Me
message
me
LinkedIn
message
me
whatever,
if
you're
curious
about
when
and
what
the
next
episode
is
going
to
be
and
when
it's
going
to
be-
and
you
haven't
seen
anything
feel
free
to
harass
me
and
I
will
tell
you
and.
B
A
B
B
Good
right,
I
mean
it's
and
that's
I
think
like
I
wish
we
could.
It
was
easier
for
us
to
bring
customers
on
or
or
even
other
partners
and
stuff
like
that.
I
mean
it's
not
bad
for
partners
but
to
bring
customers
on
because,
like
come
online,
come
tell
us
about
your
complaints.
Right
come,
maybe
not
like
the
super
crazy
complaints.
You
know
like
where
they're
dropping
bows
on
on
everybody,
but
hey
I,
struggled
with
this,
but
I
got
to
figure
it
out
and
here's
how
we
did
it.
A
I
got
I,
got
reached
I
reached
out
through,
like
three
levels
of
proxy
to
a
customer
to
ask
if
they
could
come
on
and
talk
about
something
pretty
cool
that
they're
doing
and
they're
basically
like
we
would
love
to
Legal
says
no
I
was
so
sad.
I
was
like
oh
I,
understand
but
yeah.
If
customers
want
to
come
on
and
talk
about
how
they're
using
things
and
what
they're
doing
I
mean
we
are
super
here,
for
it.