►
Description
How Red Hat Enterprise Linux (RHEL) users and admins can benefit their organizations and improve their careers by learning how to use containers, Kubernetes, and Red Hat OpenShift.
Learn more at https://red.ht/leveluphour
A
Good
morning,
good
afternoon,
good
evening,
wherever
you're
hailing
from
welcome
to
another
edition
of
the
level
up
hour
here
on
openshift
tv,
I
am
chris
short
executive
producer
of
openshift
tv.
I
am
joined
by
the
one
and
only
the
illustrious
langdon
white
langdon
hello
good
morning.
How
are
you,
sir,
pretty
good.
B
A
Because
of
some
events
this
week,
but
yeah,
oh
right,
yeah,
so
yeah
the
the
coffee
is
flowing
here
on
channel
and
we've
got
the
lights
up
for
everybody
and
we're
here
and
we're
happy
to
be
amongst
you.
And
you
know
what
do
you
got
on
tap
for
us
today?
Lange
or
I
guess
what
do
we
have
on
tap
today?
Right.
B
B
And
let's
see
if
I
can
share
them,
you
know
one:
one:
technology
fail
at
a
time,
yeah
all
right,
all
right,
yay
slides
they
are
the
best,
although
I
heard
we
have
maybe
some
new
colors
and
stuff
rather
than
the
red
I
kind
of
like
the
the
busy
stuff
on
the
side,
but
so
we'll
see
we'll
see.
But
so,
oh
sorry,
let
me
let
me
back
up.
B
B
You
know,
for
example,
this
episode
where
we're
gonna
talk
about
service
mesh,
but
the
idea
is,
you
know:
let's
try
to
cover
all
things
that
you
know
kind
of
about
why
you
might
want
to
get
engaged
with
containers
and
then
what
we've
discovered
as
as
humans
right
is
that
as
we
work
more
and
more
with
containers,
we
discover
that
we
need
to
orchestrate
those
containers,
and
so
then
we
talk
about
as
our
namesake
channel.
We
talk
about
openshift
as
a
way
to
orchestrate
those
containers.
B
So
specifically,
today
we're
going
to
talk
about
kind
of
it's
not
really
orchestration
per
se,
but
ways
to
do
extra,
like
kind
of
give
your
application
extra
abilities
without
code
change,
and
arguably
you
know,
depending
on
how
how
much
risk
you
like
to
take
without
qa.
But
you
know
you
got
to
be
crazy,
so
all
right,
so
I'm
lagging
to
white.
You
can
find
me
at
langdon
with
a
one
on
twitter
and
chris
is
chris
short
and
he's
not
as
short
as
implied
by
his
tag.
A
A
B
Something
we
usually
hang
out
on
discord.
The
you
know,
thanks
to
the
lovely
restream,
you
can
go
to
the
discord
and
chat
with
us
for
the
show.
You
can
also
go
and
chat
with
us
there
for
other
stuff,
but
you
can
also
chat
in
twitch
or
the
youtube
live,
or
I'm
not
even
sure
what
role
we're
on
these
days.
Facebook
and
they're.
B
No,
no!
No!
On
twitter,
oh
completely,
blinking,
my
my
children
will
be
disappointed
all
right.
So
there's
our
discord
go
join
us
if
you
like,
and
then
kind
of
a
little
bit
about
the
show.
One
of
the
things
I
feel
like
I
was
looking
at
some
thing.
Whatever
recently,
the
level
up
hour
is
actually
a
show,
that's
part
of
a
kind
of
a
general
program
that
red
hat
is
offering.
B
B
B
Yeah
and
then
under
and
then
some
and
then
sometimes
there's
also
kind
of
access
to
licenses
for
openshift.
That
kind
of
stuff
so
go
check
it
out
because
that's
you
know,
there's
a
lot
more
to
the
program
in
a
sense
than
us
talking
here
and
you
know.
Of
course
we
have
the
sweet,
sweet
internet
points
which
nobody
else
does
but
we're
trying
to
change
that.
A
B
B
A
B
B
Yeah,
so
today's
episode
is
episode.
21
service
mesh,
hopefully
we're
gonna,
get
that
integrated
with
next
cloud.
What
I
realized
as
part
of
the
kind
of
prep
for
the
show,
is
that
it's
weird
that
you
know
the
reason
I
chose
nexcloud
is
because
it's
not
a
hello
world
scenario
right,
but
when
you
talk
about
service
mesh,
it
kind
of
is
because
it
doesn't
do
anything
like,
as
in
a
bunch
of
the
features
around
service
mesh.
Are
traffic
shaping
canary,
releasing
that
kind
of
stuff?
B
We'll
talk
about
that
a
bit
more
in
a
minute.
But
next
cloud
is
actually
not
a
brilliant
example.
Sorry,
so,
okay,
but
show
notes
from
last
time.
If
you
are
a
long
time,
show
noter
or
you
know
someone
who
has
been
watching
the
show
for
a
while.
You
may
notice
that
I
rejiggered
how
the
notes
or
the
basically
the
repo
format.
B
A
B
B
A
B
A
B
B
B
A
B
That
would
be.
That
would
be
funny
we
could.
We
could
definitely
talk
to
ryan
jay
about
that.
I
bet
he
has
the
perfect
music
for
it,
yeah
all
right.
Let's,
let's
share
this
whole
desktop,
but
to
make
it
less
confusing.
We
will
not
put
your
picture
up
there.
Thank
you
yeah.
It
works,
although
it
still
threw
it
up
there.
So
now
you
get
the
little
box
on
the
right.
Let's
get
rid
of
that.
There
we
go
there.
B
A
B
A
B
Concept
a
long
long
time,
but
definitely
in
like
the
soa
days,
we
saw
kind
of
the
mesh
idea
kind
of
mixed
in
with
the
grid
idea,
but
the
idea
with
the
mesh
is
that
there
are
aspects
of
your
application
that
you
want
to
empower
without
necessarily
injecting
that
into
every
piece
of
code.
So
the
simplest
example.
I
think
that
everyone
kind
of
regularly
comes
to
is
this
one
here
that
you
can
see
on
the
screen.
B
Which
is
the
mtlsfying
your
application,
so
in
many
applications
right?
The
the
communication
between
the
various
points
of
your
application
are
not
secured
by
default
right.
Most
programmers
don't
do
that
because
it's
a
headache
and
it
requires
at
least
for
them,
overhead
of
both
performance
and
then
maintenance
of
that
code.
So
one
of
the
things
that
is
common
to
the
concept
of
a
service
mesh
is:
how
can
we
encrypt
all
those
communication
channels
without
it
relying
on
code
changes?
B
B
You
want
to
move
a
bunch
of
stuff
into
the
cloud
where
how
do
you
communicate
between
like
say,
different
clouds
or
different
vms
within
the
cloud,
if
you're
not
using
kind
of
a
dedicated
private
vpn-
or
let's
say
you
know
a
lot
of
the
time,
I
would
say
you
put
your
application
quote-unquote
in
the
cloud,
but
you
keep
your
data
inside
your
data
center.
Okay.
Well,
you
want
to
be
able
to
secure
those
communication
channels,
but
if
you
do
it
at
the
service
mesh
level,
you
have
a
couple
advantages.
B
B
B
Okay,
all
right
so
so
that's
kind
of
the
classic
example
of
like
a
simple
service
mesh.
However,
in
some
ways
it's
the
most
boring,
at
least
in
my
opinion,
because
you
know
yes,
it's
important
to
kind
of
encrypt,
all
those
channels
or
whatever,
but
it
doesn't
really
give
you
the
flexibility
of
like
a
service
application
that
you,
you
know.
You
see
a
term
like
service
mesh
and
you're
like
oh,
my
goodness,
let's
now
I
get
to
have
all
these.
A
B
The
mines
with
them
and
if
the
canary
died,
they
knew
that
it
was
time
to
get
out
because
the
air
was
bad,
basically
because
of
the
very
fast
metabolism
of
a
canary
they
die
more
easily
than
humans.
Do
so,
while
it's
a
terrible
story
and
I
feel
bad
for
the
canaries-
that's
where
the
term
comes
from,
but
so
the
idea
here
is
that,
instead
of
mass
moving
from
v1
to
v2
of
microservice,
a
you
can
instead
say
here
is
sorry.
B
Here
is
the
new
version,
but
I
only
want
to
give
it
to
say:
10
of
the
user
population
make
sure
everything
is
working
and
everything
doesn't
light
on.
Fire
then
start
to
roll
it
out
to
the
rest
of
the
application
users,
and
so
that's
where
the
canary
comes
in
and
where
it's
it's
really
kind
of
cool.
B
You
know
kind
of
development
style,
but
even
if
you're,
not
even
if
you're
a
more
traditional,
you
know
application
environment.
You
know
this
is
a
really
great
way
to
have
a
group
of
power
users
be
able
to
try
out
your
your
new
feature
set.
You
know
without
submitting
you
know
all
the
or
subjecting
I
guess
you
know
kind
of
your
general
non-power
users
to
that
feature
set
until
it's
ready.
B
So
there's
a
lot
of
really
good
reasons
to
to
play
with
it,
and
you
know
I
highly
recommend
if
you
can,
if
you
can
pull
that
off,
if
you
have
the
service
like
services
to
be
able
to
roll
things
out,
that
way,
doing
canary
releases
is
really
nice
and
then
another
one
that's
related
and
sometimes
gets
mixed
in
is
what's
referred
to
as
blue,
green
or
red
black.
I
actually
looked
this
up.
A
Red
and
black
is
it:
it
is
an
old
encryption
kind
of
thing
right.
You
had
red
traffic,
which
was
the
you
know,
unsecured
traffic,
and
then
you
had
black
traffic,
which
was
the
secured
traffic.
So
an
encryption
device
would
take
the
red
traffic
and
turn
it
into
black
traffic
right.
So
that's
where
that
concept
kind
of
comes
from
so
encrypted
being
or
unencrypted
being
read
encrypted
being
black
meaning.
It
is
safe
to
transmit
that
now.
B
A
B
Was
one
of
them-
I
don't
remember
but
yeah,
so
I
went
and
tried
to
look
it
up,
couldn't
figure
out
a
good
answer,
but,
generally
speaking-
and
this
is
an
and
kind
of
reflective
of
the
fact
that
I
couldn't
find
a
good
answer
to
this-
is
that
this
can
kind
of
mean
what
your
organization
wants
it
to
mean.
But,
generally
speaking,
it's
the
idea
of.
Let's
give
this
version
of
the
application
to
this
set
of
users
and
that
version
to
another
set
of
users.
A
B
B
So
what
they
did
was
they
actually
rolled
out
a
version
of
the
application
that
allowed
you
to
select
check
as
your
payment
type
and
then,
if
you
selected
that
and
then
hit
submit
or
whatever
it
would
actually
come
back
and
say:
oh
I'm
sorry,
but
the
check
support
is
currently
unavailable.
It's
just
down
right
now
and
the
fact
that
zero
people
clicked
on
that
and
ran
into
that
error
is
how
they
determined
that
they
are
not
going
to
implement
check
by
payment.
So
this
is
kind
of
a
loose
example.
B
I
really
like
the
example,
so
that's
why
I
bring
it
up,
but
it's
not
quite
the
same
thing,
but
it's
kind
of
the
idea.
You
know,
I
think
a
classic
with
this
is
often
used
with
ux
is
like
you
know
where,
where
and
what
color
button
on
the
screen
gets
people
to
to
click
it
more
often,
and
that
way
you
can
have
a
different.
You
know.
Basically,
you
can
try
out
different
things.
B
You
know
I'm
a
big
fan
of
this,
and
so
what
you're
doing
here
with
these
tool
chains
with
a
service
mesh,
is
basically
providing
aspects
to
your
applications
and
one
of
the
new
features
in
the
2.0
version
of
service
mesh.
Is
this
ability
to
inject
this
here,
which
is
basically
that
you
can
decorate
or
provide
an
aspect
to
almost
anything
inside
your
application
or
actually
any
any
service
within
your
application?
B
So
it
has
to
be
one
of
those
kind
of
container
units,
but
not
only
can
you
kind
of
apply,
you
know
the
mtls
or
the
canary
or
the
whatever
feature
to
that
service.
You
can
actually
do
an
arbitrary
feature
by
using
webassembly,
and
so,
if
you're
not
familiar
with
web
assembly,
it's
often
referred
to
as
wasm
yep.
B
A
I
mean
the
the
idea
behind
webassembly
is
that
you
can
take
programs
that
you've
written
in
other
languages
and
run
them
natively
in
the
web,
browser
right,
and
that
means
c,
plus
plus
or
you
know,
rust,
whatever
kind
of
language
you
want
to
use.
You
can
then
take
that
program
and
say
all
right.
Let's
run
this
in
the
browser
now
at
first
to
a
security
person
that
might
sound
like
horrifying,
but
when
you
think
about
it,
right,
like
javascript,
is
just
actively
running
on
clients
everywhere,
all
the
time
right.
A
B
B
There
is
a
ton
of
stuff
and
yeah
so
suffice
to
say
it
lets
you
kind
of
natively
compile
like
native
to
the
web,
whatever
that
means
right,
compile
your
existing
languages.
So
what
that?
What
does
that
mean
to
you?
Okay?
Well,
that
means
that
you
can
take
you.
You
know
whatever
language,
you're
comfortable
in
and
write
the
feature
that
you
need
to
decorate
that
particular
aspect
of
your
component.
A
A
B
Well
and
rust,
rust
is
a
little
biased.
Just
because
I
mean
they
develop
the
language
to
build
firefox
right,
like
they
weren't
developing
the
language,
to
build
a
language
yeah.
So
so
I
think
they're
they're
a
little
bit
biased,
but
you
know
I,
the
the
number
of
people
who
I
hear
just
swearing
about
how
like
awesome.
It
is.
A
B
A
No,
it's
a
funny
story.
There's
a
friend
of
mine
eric
st
martin.
He
works
at
microsoft,
he's
huge
in
the
go
community,
but
he's
been
doing
a
lot
with
webassembly
and
and
rust,
and
he
has
streams
as
well
where
he
does
some
like.
Just
live
coding
of
hey
we're
going
to
make
this
thing
run
in
the
browser,
and
it's
like
some
rust
program
that
he's
coding
and
he's
been
doing
a
lot
of
web
rtc
stuff.
Actually,
so
it's
it's
pretty
cool
to
watch
him.
B
Yeah
I
mean
like
twitch
is
turning
into
a
really
interesting,
like
platform
for
quote-unquote
learning
of
various
tech
stuff,
like
the
number
of
like
live
coding.
Twitch
streams
that
I
kind
of
come
across
and
are
really
quite
interesting,
you
know
is,
is
impressive.
Maybe
we
should
maybe
we
should
have
a
show
talking
about
other
shows
at
some
point
in
the
future,
or
maybe
we
should
start
thinking
about
doing
some
more
live
coding
shows.
Maybe
if
we
get
enough
requests,
we
could
do
more,
live
coding,
I'll.
B
And
you
know
that's
a
like
if
you
think
about
it,
for
a
second
right.
What
I
was
talking
about
before
is,
like
you
know,
you're
kind
of
injecting
into
every
service.
Your
functionality
is
part
of
the
service
mesh.
Well,
that
seems
like
a
logical
place
to
put
hey.
Why
don't
we
do
some
tracing
and
you
know
kind
of
analytics
there
and
you
know,
then
we
can
get.
You
know
better
understanding
of
our
application.
Sorry,
I
got
distracted
by
the
by
the
chat
there
for
a
second.
B
A
B
Right
right,
yeah,
you
sound
fine
to
me
so
everything's,
better
with
a
demo,
so
I
was
gonna.
Try
to
show
a
demo.
Let
me
see
if
I
can
find
the
right
window
because,
I'm
you
know
technology
defeated
today,
nice.
Where
is
here,
it
is
no
maybe
come
on.
I
I
know
I
can
find
it
somewhere
up,
not
that
this
one,
this
one
yeah
this
one.
Let's
try
this
one.
B
And
let
me
just
make
sure
it's
actually
working
so
okay,
so
we
have
a
platform
for
doing
demos
that
we
tend
to
call
internally
rh
pds.
So
if
you
hear
me
say
that
it
stands
for
the
brilliant
red
hat
product
demo
demo
system
in
that
demo
system,
so
this
is
kind
of
like
what
we
use
to
talk
to
people
about
it.
We've
been
pushing
a
lot
more
of
this
to
kind
of
general
public.
If
you
go
to
learn.openshift.com,
you
kind
of
see
our
progress
thus
far.
B
And
actually
one
of
the
things
I
was
trying
to
do,
but
so,
if
you
go
to
learn.openshift.com,
you
can
actually
see
there's
a
bunch
of
stuff
here.
So
there's
like
serverless
there's
an
istio
one
and
just
kind
of
by
way
of
background
when
I
say
open
shift
service
mesh,
as
with
all
things
red
hat,
there
is
kind
of
a
product
name
for
a
set
of
tools
that
are,
you
know
an
upstream
open
source.
B
You
know
world
the
ones
that
make
up
openshift
service
mesh
are
called
istio,
kiali
and
jaeger,
and
so,
if
you
see
the
istio
here,
this
is
actually
using
the
upstream
talking
about
that
various
piece
but
you'll
see
more
stuff
kind
of
rolling
out
here
regularly
we're
trying
to
get
things
out
here.
As
often
as
we
can.
We
also
talked
actually
about
the
developer
sandbox
recently.
B
So
we
have
this
new
developer,
sandbox
as
well,
which
unfortunately
doesn't
do
service
mesh,
which
is
where
I
was
going
to
demo
this
originally
right,
which
I
was
kind
of
sad
by,
but
it
is
in
progress.
So
if
you
look
at
developer
sandbox,
you
can
go.
Get
your
own
open
shift
instance
that
you
can
then
kind
of
run
whatever
you
want
on,
however,
not
service
mesh,
so
yeah.
If,
if
you
couldn't
tell,
I
had
a
number
of
frustrating
moments
trying
to
put
this
this
show
together.
B
So
what
I
wanted
to
show
you
here,
though,
was
you
know.
A
picture
is
worth
kind
of
a
thousand
words
and
let's
see
if
we
can
make
things
work.
Well,
I'm
not
sure
it's
not
going
to
pop
into
a
new
window.
Here
we
go
so
what
we're
going
to
do
is
basically
so
we
have
service
mesh
set
up
in
in
this
demo
environment,
and
so,
if
you
want
to
see
how
it's
kind
of
set
up,
let
me
go
to
the
correct
project.
B
Not
there,
though-
and
you
can
see,
there's
this
this-
this
isn't
a
great
picture.
Let
me
show
you
this
one,
so
basically
I
have
all
these
different
services
this,
you
know,
they're,
not
brilliant,
there's
this
curl
service,
there's
this
customer
preference
whatever.
So
I
have
all
these
different
services
in
a
project
in
openshift
and
they're,
just
kind
of
doing
their
thing
right.
B
The
application
hooks
together,
but
you
know
like
as
as
pretty
as
this
is
like
it,
doesn't
actually
really
show
you
very
much
about
how
the
application
is
kind
of
connected
together.
So
what
I
wanted
to
show
you
is
kind
of
how
you
can
decorate
this
with
service
mesh
and
then
you
can
see
kind
of
how
the
application
connects
together.
However,
the
best
way
to
do
that
is
kind
of
like
one
step
at
a
time.
So
if
you
look
at,
let's
see
it's
oh
shoot.
It's
sorry
thinking.
B
Yes,
yes,
and
I
have
successfully.
However,
I
was
having
massive
problems
with
it
in
the
kind
of
couple
days
leading
up
to
the
show,
I'm
not
sure
if
it
was
me
or
it
was
crc
or
what,
but
I
was
just
having
a
lot
of
trouble
getting
my
crc
to
run
it
was.
It
was
actually
before
I
even
installed
service
mesh
so
basically
to
get
service
mesh
2.0,
which
gives
you
kind
of
all
the
stuff
I
just
talked
about.
You
need
openshift,
4.6
crc
is
running
openshift
4.6.
B
So
some
of
the
feature
set
is
not
here
yet,
but
the
things
that
I
think
are
cool,
like
the
things
like
the
broad
strokes
of
them
are
here
the
thing
that
isn't
like
some
of
the
stuff
that
isn't
here
is
like
the
webassembly
stuff,
which
I
wasn't
really
gonna
demo
per
se,
because
I
think
that's
just
very
specific
and
so
not
very
demo-able.
In
a
sense
you
know,
maybe
we
could
talk
about
it
in
a
future,
show
why
it
might
be
useful,
so
there
should
be
yeah.
It's
admin.
A
A
A
B
Yeah
one
of
the
things
I
was
struggling
with
in
my
banging
my
head
against
the
wall
was
I
needed.
So
if
you're
using
kind
of
a
container,
you
know
like
from
the
red
hat
catalog
or
whatever-
and
it
has
you
know,
it
uses
certain
environment
variables
to
set
it
up.
So,
like
we
discussed
this
in
the
past,
show
it's
like
the
mysql
container
or
actually
it's
mariadb
container
expects
these
four
environment
variables
right.
One
is
my
sql
host
one
is
my
sql
password,
etc,
so
password
user
database.
A
B
B
B
B
No
wait:
how
do
I
want
to
do
this?
Sorry,
I
I
confused
myself.
So
if
you
yeah
see,
we
have
a
terminal
right
here
and
so
now
I
can
do
in
here
and
say:
grep
minus
I
host
and
you
see
I
have
a
mysql
host
so
because
the
container
is
expecting
that
environment
variable
name,
hey,
I
got
it
and
it
has
the
ip
that
I
want,
rather
than
some
hard-coded
ip.
So
okay,
so
kyali
is
what
I
wanted
to
show
you.
So
this
is
part
of
open
shift
service
mesh.
B
I
think
one
of
the
things
that
is
kind
of
in
progress
is
being
worked
on
is
like
how
to
make
this
a
little
bit
more
of
a
unified
experience.
It's
somewhat
better
in
4.6,
it's
not
a
lot
better.
It's
still
like
in
my
mind
it's
still
kind
of
disconnected,
but
the
feature
is
the
feature
set.
Is
there
if
you
know
what
I
mean?
So
you
know
who
cares
all
right?
So
what
we
can
do
is
do
this
correctly
and
I
think,
let's
look
at
not
that
at
this.
B
Oh,
did
I
actually
copy
that
the
only
problem
with
these
terminals
is
ctrl
shift
c
and
ctrl
shift
v,
don't
work
all
the
time,
so
you
know
I
can
never
remember.
If
my
fingers
just
did
it
automatically
or
not
so
yeah
so
jupiter
hub
user
is
the
problem
child
and
needs
to
be.
I
believe,
oh
boy,
user
one.
So
let
me
just
fix
that.
A
B
B
B
B
B
And
then
find
traceus,
so
you
can
not
only
see
the
kind
of
like
traffic
flow
that
we
were
seeing
in
kylie,
but
what
jaeger
lets
you
do
is
actually
dig
into
any
given
call
and
actually
dig
into
that
individual
trace.
So
you
can
actually
see
exactly
what's
happening
exactly
what's
being
called
and
you.
B
A
B
Very
hard
problem,
usually
what
you
end
up
with
is
lots
and
lots
of
print
line.
Statements
throughout
your
code
writing
to
some
sort
of
log
file,
and
then
you,
basically,
you
know,
do
entail
minus
f
on,
like
40
different
files
across
all
your
different
services,
trying
to
figure
out
where
exactly
your
code
is
going.
This
is
a
this
is
not
fun,
and
so
the
ability
to
if
you
notice,
there's
no
like
there's
not
really
any
awareness
of
either
jaeger
or
kiali
in
the
application
itself.
B
Right,
which
is
like
kind
of
amazing,
and
that
was
kind
of
what
I
wanted
to
show
with
next
cloud.
So
if
we,
if
we,
you
know
we're
already
at
44.
but
actually
and
by
way
of
introduction
next
episode,
what
we're
going
to
do
is
we're
going
to
talk
to
a
team
at
red
hat
who
actually
works
in?
What
do
we
call
them
now?
Application
services,
but
basically
the
the
kind
of
java
stack
team,
but
they
have
developed
an
application.
B
That
is
what
they
call
the
red
hat
cool
store,
which
is
one
of
those
things
where
it's
a
toy
except
built
as
a
real
application,
so
that
it
is
a
set
of
microservices
that
are
a
mixture
of
java
and
node.js
that
you
can
deploy
as
a
set
of
microservices
and
they're
going
to
come
and
introduce
that
to
us
next
time.
And
then
what
I
want
to
do
is
switch
over
to
that
from
the
next
cloud
demo
so
that
we
can
kind
of
say,
hey.
B
A
A
B
What
we
may
do
is
come
back
and
revisit
this
stuff
when
we
can
kind
of
show
off
some
of
those
aspects
so
because
this
is
still
pretty
demo-esque.
You
know,
if
you
know
what
I
mean
and
but
what
I
want
to
you
know
like,
but
we
can
do
some
really
cool
things.
I
was
just
looking
for
the
one
that
shows
some
canary,
oh
and
actually
one
I
didn't
even
mention.
Is
this
weighted
routing
idea?
So
if
we
change
this
one
to
using
that's
is
that
the
same
one?
B
Oh
yeah,
it
is
what
we
can
actually
do.
Is
we
can
change
the
weights
of
the
different
pathways
so
like
if
you
notice
back
over
here,
where'd
my
window
go.
I
have
three
different
versions
of
this
recommendation:
engine
right
so
but
they're
all
getting.
You
know
an
equal
percentage
of
the
let's
actually
switch
to
request
percentage,
they're,
getting
each
three
33
and
a
half
percent
of
the
traffic
coming
into
it.
What
if
we
want
to
do
the
canary
canary
scenario,
we
can
make
recommendation
v3.
B
A
A
B
I
well
I
don't
really
see
an
overhead
of
openshift
service,
mesh
versus
jaeger,
kiyali
and
istio.
Sorry,
I
was
like,
what's
the
last
one
yeah
so
like
I
mean
you're,
basically
it's
the
same
overhead.
I
mean.
Maybe
it's
slightly
more
for
the
integration
with
openshift,
but
the
the
value
of
being
able
to
do
things
like
encrypted
communication
between
all
of
your
channels,
with.
B
You
know
in
this
kind
of
way,
so
you
can
kind
of
say:
okay,
let's
make
sure
that
aws
stuff
is
actually
working.
So
let's
give
it
10,
you
know
and
then
kind
of
say:
okay
now,
let's
scale
up
and
make
sure
it
keeps
working
like
I
don't
know,
I
I
think
I
think,
having
something
like
a
service
mesh
assuming
now.
This
does
make
the
assumption
that
your
application
is
relatively
service
oriented.
B
You
know
if
you
have
a
monolith,
or
you
know,
or
it's
just
like
it
and
kind
of
why
I
was
arguing
against
doing
next
cloud
this
way
in
the
in
the
deployment
we
have
been
working
with,
it's
just
two
pieces
right:
it's
it's
a
web
server
and
a
database
that
that's
probably
not
worth
it
right
right.
You
know,
whereas,
as
soon
as
you
start
having
you,
because,
basically
what
happens
is
like
once
you
get
to
like
it's
some
weird
number
right,
but
it's
like
five
services.
All
of
a
sudden,
you
have
20.
B
like
it
just
kind
of
explodes
as
soon
as
your
application
starts
thinking,
service
oriented
you
know,
or
micro
service
oriented.
If
we
want
to
use
what
the
cool
kids
say
that
that
service
orientation
explodes
right
like
look
at,
you
know
the
the
famous
bezos
you
know,
internal
memo
of
the
two
pizza
teams
and
every
service
must
be
accessible
from
the
outside,
like
aws
would
be
like
it
just
it
explodes
right.
I
mean
that's,
that's
what
happened
there
is
that
you
know
it
was
essentially
an
organic
growth
to
say,
okay.
B
Well,
we
need
this
feature
on
the
retail
site,
so
let's
do
it
as
a
service
and
then
the
retail
site
can
use
it,
but
then
we
can
offer
it
to
other
customers
and
that's
kind
of
how
how
aws
grew.
I
excuse
me,
I'm
sure,
there's
some
high
level
planning
to
it,
and
maybe
more
so
now,
but
initially
it
was
just.
B
We
need
more
services
for
the
retail
site.
Let's,
let's
build
those
services
in
such
a
way
that
they
could
be
consumed
by
consul
customers
themselves
right.
As
a
result,
a
service
explosion
with
service
explosion
comes
need
to
manage
them,
and
I
think
a
service
mesh
does
a
really
nice
job
of
that.
I
don't
know
that
was
a
long-range
way
of
saying
hotshot.
What
do
you
think
do
you
agree.
A
Yeah
like
watching
some
videos
from
who
was
with
johnny
borkowitz,
I
freaked
out
probably
just
butchered
his
last
name
but
brookwoods
from
the
go
field.
Right
like
I
went
to
a
meet-up
once
years
ago,
and
he
was
telling
me
like
showing
us
how
to
implement
tracing
in
an
application
he
was
deploying
to
aws
and
it
was
just
like
wow.
You
have
to
add
so
many
lines
of
code.
Okay
like
this
is
why
service
mesh
exists.
B
Right
right
and
just
to
kind
of
hotspots
last
point
about,
like
you
know
the
package
of
things
you'd
want,
if
you're
doing
micro
services
well
yeah
I
mean
this
is
where
it's
important
to
say
not
like
not
just
micro
services
like
kind
of
any
kind
of
services.
But
if
you
go
back
and
watch
episode
like
it
was
like
12,
it
was
the
one
with
scott
mcbride.
It's
like
this
is
a
little
bit
of
the
same
problem
of
when
to
orchestrate
is
like,
if
you
do
it
too
early
you're
doing
it
all.
B
This
you've
got
all
this
overhead
with
very
little
benefit,
but
if
you
do
it
too
late,
it's
really
hard
to
to
apply
to
an
existing
application,
because
you've
already
worked
around
a
lot
of
these
problems.
So
you
know
it
is
still
different.
You
know
different
scenarios
or
different,
like
kind
of
trade-offs
that
you
want
to
make,
but
be
careful
of
your
developers
going
and
implementing
all
of
this
stuff
by
hand
in
their
individual
applications.
B
You
know
is
one
aspect.
The
other
thing
is,
I
would
say,
like
we
probably
have
documentation
regarding
straight
up
scalability,
I'm
just
not
familiar
with
it
off
the
top
of
my
head,
so
it
is
definitely
something
you
could
file
a
ticket
and-
and
we
could
kind
of
get
back
to
you
on
or
you
know,
we've
got
to
take
it
under
advice.
B
And
that
is
one
of
the
things
in
service
mesh
two
that's
handled
for
you
automatically.
That
is
not
in
whatever
it
was
1.11,
maybe
that
we're
working
with
right
now
or
that
I'm
showing
you
right
now,
so
so
that's
a
definite
reason
to
go
and
look
at
service
mesh
two.
You
know
if
you're
interested
in
that
minimizing
that
particular
pain
right
all.
B
Points
so
I
will
say
and
caution
to
the
regular
attendees
of
the
show.
I
did
rejigger
some
of
the
points
and
the
reason
I
rejiggered
some
of
the
points
is
because,
as
I
said
at
the
beginning
of
the
show,
we
do
have
maybe
hopefully
soon
some
non
just
intrinsic
value
to
the
points
coming
up.
Oh
shoot.
That
is
entirely
the
wrong
code.
I
just
realized
well.
B
I
could
I
could
so
please
if
I
type
chat,
that's
not
gonna
help
anybody.
So,
as
you
can
tell
I
was
doing,
there
was
a
lot
of
prep
work
for
this
episode,
and
so
the
kind
of
like
the
the
niceties
are
the
stuff
I
dropped
right.
So
like
the
last
two
times
thing
and
and
the
episode
update
so
here
is:
where
did
the
chat
window
go
here?
It
is
this.
Is
the
correct
code
or
the
the
deep
link?
A
B
Like
to
do
exactly
do
keep
in
mind
the
way
I
calculate
the
points
right
is
by
mapping
your
email
addresses.
So
please
try
to
make
sure
you
use
the
correct
email
address.
If
you
submit
it
incorrectly,
you
can
a
submit
it
again
correctly
and
it
will
be
fine
or
contact
me
on
discord,
and
I
will
correct
any
errors
that
you
may
have,
but
thanks
again
to
netherlands,
hackham
for
rocking
the
2800
points
and
narendev
staying
strong
with
2600,
I
don't
think
I've
seen
narendra
today.
I.
B
Okay,
all
right
so
and
then
noah
friction
with
2400
points,
joe
fuzz
with
1800
and
then
det,
cone
and
kudo,
who
is
starting
to
move
up
in
the
rankings,
but
we
will,
like.
I
said
I
rejiggered
it
a
bit,
I'm
thinking
about
adding
some
more
activities
as
well,
that
you
can
also
get
points
for
we
are
trying
to
get
some
other
people
to
also
be
giving
out
points
on
the
show
yeah.
B
A
B
Let's
see
okay,
what
else
did
I
want
to
show
here?
Not
that
not
that
either?
Oh,
yes,
it
is.
This
is
one
of
the
funny
things
about
the
terminal.
Is
that
sometimes
it
comes
in
like
straight
up
and
down,
but
if
you
readjust
the
the
window
it
fixes
it
which
I
find
amusing.
I
I
find
much
of
modern
web
development
insane.
So,
let's
see
I
don't
know
what
is
there
like,
I'm
kind
of
feeling
like?
B
A
B
Well,
with
with
like
a
real
application
right
like
right,
you
know
so
you
know
the
the
the
cool
store
has
been
designed
to
be.
First
of
all,
a
you
know,
kind
of
all
microservices
be
you
know
we
have
lots
of
ancillary
stuff
that
we
can
then
build
into
it
as
we
go
along
so
yeah.
So
maybe
we
should
call
it
a
day
here
unless
we
have
any
other
questions.
A
A
A
A
B
A
B
Won't
we
be
having
an
episode,
I'm
just
looking
at
calendar
yeah
we'll
have
an
episode
either
the
day
of
chinese
new
year
or
the
day
after
so,
let's
all
hope
for
or
sorry
the
day
before
or
the
day
of,
I'm
not
sure
which
I'd
have
to
look
it
up
of
chinese
new
year.
So
hopefully
we'll
be
able
to
say
happy
chinese
new
year
and
everything
will
be
much
much
better
than
it
has
been
since
this
new
year.
A
A
Open
shift
administrator's
office
hours,
it's
it's
an
office
hour,
so
please
feel
free
to
bring
your
questions.
It's
geared
towards
administrators,
but
we're
going
to
be
discussing
as
kind
of
our
starting
point,
disconnected
openshift
environments.
So
if
you're
part
of
the
people
that
need
to
air
gap,
your
things,
let
us
know.