►
Description
Come hang out with Kris Nova as she does a bit of hands on hacking of Kubernetes and related topics. Some of this will be Kris talking about the things she knows. Some of this will be Kris exploring something new with the audience. Come join the fun, ask questions, comment, and participate in the live chat!
A
Hi
hi:
what's
up
everybody,
this
is
Chris
Nova
coming
at.
You
live
from
the
t,
GI
case
studios
here
in
downtown
Seattle,
Washington,
happy
Friday,
everyone
how
how's
everybody
doing.
Let
me
pull
up
the
chat
here,
so
we
can
do
my
favorite
part
of
the
day,
which
is
looking
at
the
chat.
Oh,
the
chat
window
is
messed
up
in
my
screen
here.
Let
me
see
if
I
can't
fix
this
really
quick.
A
A
Okay,
let
me
adjust
my
OBS
for
everyone:
okay,
that
should
be
a
little
bit
better.
Tell
me
what
folks
think,
let's
see
and
let's
go
ahead
and
start
saying,
hi
to
chat
in
chat.
Why
people
tell
us
if
the
new
chat
window
looks?
Ok,
if
you
can
even
read
it
or
not,
it's
kind
of
hard
for
me
to
see
what
folks,
what
you're
seeing
on
your
end
over
there
on
the
other
side
of
the
Internet,
but
anyway,
it's
good
to
see
you.
A
So,
let's
see
first,
this
week
was
Olaf
hi
Olaf
how's
it
going
happy
Friday
from
Denmark,
bobs
monk
says
good
evening
from
Lagos
Nigeria
Ivan
says:
let's
see
good
evening,
Keith
Martin
says
good
evening.
Everyone
from
rotterdam,
hep
tio,
says
good
afternoon.
We
have
Duffy
on
the
line
today,
helping
out
so
shout
out
to
Duffy
for
helping
me
up.
George
is
taking
the
afternoon
off
and
we're
hanging
out
having
some
fun
and
Duffy
got
to
see.
A
I
think
this
is
not
even
the
first
time
that
you
got
to
see
like
the
Chris
Nova
behind
the
scenes.
Preeti
gik,
when
I
just
see
it
in
here
and
talked
to
myself
for
half
an
hour,
it's
really
quite
funny,
I
kind
of
want
to
start
recording
it
just
so
that
we
can.
They
have
like
the
TJ
I
can't
bloopers,
because
it's
just
me,
and
here
like
typing
and
just
talking
to
myself
without
the
other
camera
there,
which
is
funny
because
it
just
gets
me
like
the
mode
for
two
gik
but
yeah.
A
It's
probably
looking
saying
to
most
people
anyway.
I
have
to
go
so
getting
started.
This
is
Duffy.
It's
a
deep
hello
from
Poland
ttak
Martin,
hello
from
another
heff
do
says
this
is
the
hack
of
d
which
will
pull
that
up
in
a
second.
So
that's
the
hack,
IAM
d
for
folks
at
home,
who
don't
know
what
that
is
in
the
chat
and
in
github
COMSAT
do
/t
gik.
A
You
can
find
a
link
to
a
hack
indeed,
that
I
will
pull
up,
which
is
basically
a
live
readme
file
for
github
that
we
can
all
add
it
together,
so
feel
free
to
take
notes
or
write
things
down,
and
we
can
all
collaborate
on
that.
Why
we
do
RTG.
I
ka
live
suresh,
says:
hey
guys
and
also
girls
to
chat,
says
chat,
looks
good,
Louise,
San
Martin
says
good
night.
A
Everyone
greeting
from
Amsterdam
happy
to
be
here,
we're
happy
to
have
you
Louie
good,
to
see
you
all
so
I
apologize
if
I
mess
up
anyone's
name,
I
try
my
best
but
I'm,
not
always
the
best
at
pronouncing
people's
names.
Miguel
says
hello
from
Mexico
hola
Miguel
que
pasa
hello,
Fort,
San,
Francisco,
hi,
Aaron,
hello
from
Johannesburg
hi
Michael
L
chat,
esto
from
cuñado
Chad
has
up
and
running
Thanks
rasca
sube,
hello
from
India
hair
looks
great.
So,
okay,
so
funny
story
here:
Amy
the
N
Brautigan
she's.
A
She
just
commented
in
the
chat
wanted
to
give
her
a
big
shout-out.
Today
she's
been
helping
me
with
my
TGI
K
intro
and
outro
game.
So
we
really
want
to
try
to
step
up.
My
like
welcome
to
TGI
K
everybody
and
so
Amy's
been
giving
me
personal
guidance
on
that
and
she
joined
today.
Just
so
I
can
make
her
proud
and
also
I
guess
to
compliment
my
fabulous
hair.
Roy
says
Norway
Tron
him
says
hello,
Steve
sloka
says
happy
Friday.
A
Everyone
good
to
see
you
again,
Steve
thanks
for
joining
love,
to
see
you
here,
okay,
so
let's
pull
up
this
hack
MD
and
get
started
through.
What's
new
in
TG
I
came
this
week,
so
we
go
over
here
so
again.
This
is
there's
a
link
to
this
in
the
chat
if
you
want
to
go
and
pull
it
up,
but
it's
just
this
website
hacking
di,
oh
and
then
on
the
left
side.
Here
we
have,
which
is
basically
just
like
github
style,
markdown
and
then
on
the
right
side.
A
We
have
a
rendering
of
this,
and
after
we
are
done
with
the
episode
today,
we
will
merge
this
into
github.com.
Slash
have
do
/t,
tik
and
you
can
come
in.
You
can
see
that
we
actually
have
all
the
episodes
list
you
care
in
the
episodes
directory.
So
anything
you
type
will
get
merged
upstream
and
that's
how
we
like
to
do
things
around
here.
So,
let's
see
I
got
somebody
messaging
me
awesome.
It's
Amy's
telling
me
that
I'm
killing
it.
Thank
me.
So,
let's
see
you
go
back
to
that
chat,
room!
Okay!
A
So
here
on
the
Left,
we
have
table
of
contents.
So
we
try
to
do
like
a
what
point
in
the
video
you
can
do
whatever.
So,
if
folks,
if
there's
any
major
points
in
the
video
that
you
think
folks
would
want
to
jump
to
feel
free
to
add
a
line
here,
and
then
these,
if
you
just
use
this
format,
it'll
automatically
build
a
link
whenever
we
push
it
to
the
repo,
so
you
don't
have
to
worry
about
creating
a
link
or
anything
just
the
time
and
we'll
take
care
of
the
rest.
A
Yan
says
hello
from
Palo
Alto
Nikolas
is
here
I'm
here
and
I'm,
looking
Duffy
directly
in
his
space.
That's
awesome!
Where
are
mocoso
Nikolas
and
Duffy
both
work
here.
Obviously,
where
are
y'all
hanging
out
that
today
are
you.
Are
you
like,
together
at
the
office
somewhere,
cuz
I?
Don't
think
you're
here,
maybe
you
are
I,
don't
know,
there's
not
very
many
people
here
in
Seattle
today,
Christopher
says
hello
from
Germany,
yeah
I
think
that's
all
the
chat
messages.
So
anyway,
let's
see,
we've
got
two
things
to
look
at.
A
We
have
the
last
week
in
kubernetes,
and
then
we
have
Q
weekly
and
Duffy
was
kind
enough
to
put
some
of
the
stuff
together
for
us.
So
let's
go
through
and
check
some
of
this
out
and
then
also
I
just
want
to
mention,
while
I
eat
one
of
my
M&Ms
here
so
this
week
has
been
a
crazy
week
for
me.
I've
been
super
busy,
so
this
is
one
of
the
few
episodes
that
I'm
actually
going
to
be
going
like
I've
never
touched
glue
before
in
my
life.
A
I
know
enough
about
it
to
know
that
we
can
do
ingress
and
we're
gonna
be
figuring.
Everything
else
out,
live
so
you're,
actually
gonna
get
to
see
like
the
oddest,
like
me,
figuring
stuff
out
on
the
fly
without
any
work
setup.
All
I've
done
so
far
is
just
fire
up
a
kubernetes
cluster,
but
we'll
jump
on
my
terminal
and
look
at
all
that
stuff
later
Nicholas
says
bellevue
office.
Okay,
you
guys
are
in
bellevue
I'm.
Sorry,
you
folks
are
invaluable.
A
Bellevue,
Louisa
and
Martin
glow
sounds
a
lot
like
glow
Sully,
as
in
sigur
rós,
okay,
so
Louie.
You
win
the
contest
today.
So
every
week
for
kubernetes,
if
you
look
here,
we
have,
we
have
a
title
card
like
the
one.
I
just
showed
you
and
we
always
pick
out
an
image,
that's
relevant
to
like
whatever
we're
doing,
RTG
I
came
on,
and
it's
like
this
inside
thing
that
we
do
and
if
you've
been
watching
TGI
K,
you
may
or
may
not
have
picked
up
on
it.
A
really
great
example
was
when
we
did
spring
boots.
A
A
Which
is
a
picture
I
took
with
my
fancy
camera
while
I
was
in
Iceland
for
precisely
that
reason
that
Louie
mentioned
down
here
the
word
for
overjoyed
and
Icelandic.
If
you
look
over
here,
it's
actually
pronounced
blue
or
it's
a
closest
word
in
Icelandic
to
glue,
and
it
means
overjoyed
so
I
took
it
my
favourite
picture
from
Iceland.
That's
the
background
today,
because
it
means
overjoyed
in
Icelandic,
so
glued
means
overjoyed.
A
So
today
you
learned
some
Icelandic
and
the
band
sigur
rós
is
actually
from
Iceland,
so
yeah
Louise
was
the
first
one
to
get
the
the
reference
there
and
why
we
picked
that
picture
for
this
week.
So
I
wanted
to
call
that
out.
So
anyway,
we're
gonna
go
through
this
weekend.
Kubernetes
the
myrna
jump
into
glue
and
I've
got
a
bit
of
a
surprise
kind
of
twist
to
the
to
the
what
we're
going
to
be
doing
today.
But
we'll
talk
more
about
that
in
a
moment.
Okay,
so
first
things.
A
A
There
we
go
Steve
silica,
says:
glow
sounds
a
lot
like
glozell
and
then
Nicolas
Lane
says
give
Edelin
it
value
to
last
Shou
pauses
own,
so
I
think
that
means
they're
in
Bellevue
hanging
out
with
Paul,
which
is
exciting.
That's
really
cool
hi
to
everybody
in
Bellevue
and
thanks
for
planting
for
me,
I
appreciate
it.
Okay,
so
anyway
kubernetes
scope.
A
So
let's
see
there's
one
contributor,
so
Brian
grant
added
this
on
January,
30th,
hep
geo,
says
yep
I
agree,
president
says:
greetings
from
Las
Vegas
awesome
present
I
will
be
in
Las
Vegas
doing
the
kubernetes
meetup
in
a
couple
weeks:
I
just
tweeted
about
it,
but
I'm
gonna
tweet
about
it.
Some
more
so
follow
me
on
Twitter
for
more
but
I'm
gonna
be
doing
basically
I
have
the
whole
night
at
the
meetup
to
myself
so
they're
doing
like
an
evening
with
Chris
Nova.
A
So
we're
going
to
do
some
virtualization
demos,
we're
gonna,
write
some
containers
and
basically
just
do
like
a
live
Q&A.
So
if
you
want
to
come
and
hang
out
in
Vegas
I'll
be
there
for
an
evening,
and
it
should
be
a
lot
of
fun,
so
hopefully
you
can
make
it
out,
and
hopefully
you
can
tell
some
folks
that
we
will
be
down
there
having
a
party
okay,
so
Louie
says
yay
going
to
visit
Iceland,
for
the
first
time
in
may
also
may
is
the
fabulous
time
to
visit.
A
Iceland
make
sure
you
bring
sunglasses,
because
it
will
be
sunny
almost
the
entire
time.
Okay,
so
kubernetes
scope,
so
Brian
grant
merged
this
in
on
January
30th,
okay.
So
it's
relatively
new
purpose
of
the
stock
clarify
factors
affecting
decisions
regarding
what
is,
and
it's
not
in
scope
for
the
kubernetes
project:
oh
wow,
okay,
so
this
is
the
big
one.
A
This
scope
of
kubernetes
is
like
we've
I
feel
like
we've
been
talking
about
this
for
the
past
three
years,
and
this
light
starts
to
allude
to
the
conversations
that
we
used
to
have
about
what
is
core
and
what
does
corny.
Nor
what
is
the
first-class
feature
or
what
is
a
second-class
feature?
What
do
we
support?
What
do
we
not
support
and
there's
always
been
sort
of
these
blurry
lines
around?
A
What
is
where
kubernetes
stops
and
we're
like
the
rest
of
the
ecosystem,
picks
up,
because
the
whole
point
of
kubernetes
is
to
make
everything
modular
like
stick
pieces
together
to
build
a
more
complex
system,
but
like
what
are
the
boundaries
for
kubernetes
itself
and
what
what
actually
defines
what
is
kubernetes
and
then
what
is
everything
else
and
that
that
boundary
is
not
as
clear
as
you
would
think.
It
is
especially
when
you
start
looking
at
things
like
dns
or
container
runtimes.
Those
can
vary
from
cluster
to
cluster
and
from
environment
to
environment.
A
So
is:
does
that
count
as
kubernetes
or
does
it
not
so
anyway?
I
think
this
talk
is
gonna,
come
through
and
kind
of
call
this
out
line
by
line
yeah,
so
we've
got
scope
for
the
github
kubernetes
project,
okay,
so
it's
all
the
github
orgs
kubernetes
SIG's,
kubernetes
client,
other
github,
orgs
I'm.
Assuming
this
is
github
incubator.
We
have
release
artifacts,
so
all
of
the
the
kubernetes
releases
are
owned
by
the
project,
which
is
good
because
kubernetes
releases
are
actually
super,
awesome
and
surprisingly
stable.
Well,
the
release
process
is
the
release
itself.
A
A
Ok,
so
if
you're
interested
in
learning
more
about
what
exactly
is
Cooper
net
where
it
starts
and
where
it
stops,
there's
actually
a
doc
document
on
this
now
see
we
can
come
and
point
to
this
if
we
ever
have
any
confusion
about
what
is
and
isn't
or
out
of
scope
for
kubernetes
and
what
I'm
effectively
thinking
of
as
the
core
components
of
kubernetes,
which
again
these
are
things
like
the
API
server,
the
scheduler,
the
controller
or
the
cubelet.
You
know
at
CD
all
those
good
things
that
make
kubernetes
nice
and
happy.
A
Okay,
so
that's
really
cool.
We
have
kubernetes
scope,
so
this
next
one
is
it's
a
bit
wordy,
but
basically
I
think
what's
happening
here
is
we
we
made
a
guarantee
that
for
readiness
and
live
meet
luminous
probes,
which
these
are
how
you
would
go
through
and
validate
that
your
pod
is
running
and
behaving
how
you
want
it
to
so.
A
So,
basically,
all
we're
changing
is
the
way
that
the
301
and
302
redirects
behave
and
how
the
the
cubelet
responds
to
that.
So
there's
a
little
bit
of
a
consideration
here.
If
we
pull
this
up
in
some
rare
cases,
this
might
actually
change
the
behavior
of
what
your
systems
are
doing,
but
for
the
majority
of
folks
this
is
just
going
to
be
a
nice
to
know
and
not
it
needs
to
know
so.
A
You
can
come
through
and
read
this
if
you
want
to
actually
see
how
these
HTTP
redirects
will
behave,
that
it's
going
to
be
based
on
the
hostname
that
you're
using
so
redirects
to
the
same
hostname
as
you
used
originally
you'll,
still
not
be
followed,
but
a
redirect
to
a
new
host
name
will
be
considered
a
success.
So
that's
what's
going
on
with
liveliness
probes
and
excuse
me,
readiness
probes.
A
A
A
Cube
weekly
has
been
around
since
I
first
made
my
first
commit
to
kubernetes
forever
ago.
I
think
I
was
like
I.
Remember
the
11th
one
of
these
I
want
to
say,
but
anyway
we're
up
to
165
now
and
every
week.
This
is
a
really
great
blog
that
puts
together
all
of
what's
kind
of
happened
in
kubernetes,
and
it
just
consolidates
that
and
it
just
sort
of
acts
as
an
index
for
the
week.
A
So
this
week
we
had
a
very
exciting
post
the
what
and
the
why
of
the
cluster
API-
and
this
is
by
Tim
Sinclair,
he's
one
of
our
engineers
here
at
VMware.
He,
it
was
one
of
the
chairs
for
SiC
cluster
lifecycle.
He's
a
really
great
engineer.
I've
been
working
with
him
for
a
while
he's
on
the
steering
committee
Tim's
awesome,
and
he
wrote
this
really
great,
very
detailed
definition
of
the
cluster
API,
what
it
is
where
it
came
from
and
why
it's
here
and
he's
even
got.
A
We
there's
a
blog
on
that
I
think
they
have
to
go
blog.
That
tells
the
history
a
little
bit,
but
basically
all
the
cluster
API
is.
Is
it's
a
declarative
way
of
managing
and
mutating
kubernetes
infrastructure,
so
sort
of
doing
things
the
kubernetes
way,
so,
instead
of
saying
I'm
gonna
run
a
script
to
get
me
a
new
kubernetes
cluster,
you
say:
I'm
gonna
create
a
CRD,
that's
going
to
define
my
cluster
and
then
they'll
be
some
controller
that
runs
somewhere,
probably
in
a
cluster
or
in
its
same
cluster.
A
That's
going
to
then
go
and
reconcile
whatever
that
definition
was
and
continue
to
reconcile
that
over
time.
That's
like
the
big
win
here
right
is
we
get
that
continual
reconciliation?
So
if
something
does
happen
or
something
does
break
or
something
does
go
wrong,
the
goal
of
cluster
API
is
to
have
your
infrastructure
sort
of
refix
itself
to
the
way
you
originally
declared
it.
A
So
that
was
like
the
big
win
of
clustering
API,
and
this
talks
a
lot
about
that
in
a
lot
more
detail
than
what
I
was
able
to
kind
of
spit
off
here,
but
that
is
cluster
API
on
the
VMware
blog,
so
make
sure
you
go
and
check
that
one
out
alright.
So
coming
up
next,
we
have
o
here
to
shout
out
to
some
Tim
Sinclair,
shout
out
to
Marko
madrina,
okay,
so
Marko
another
really
good
friend
of
mine.
A
He
is
one
of
the
maintainer
zon
cubic
horn,
which
is
a
project
I
started
two
years
ago.
I
forget
maybe
one
year,
I,
don't
know
it's
been
a
minute,
though
it's
been
around
for
a
while
and
Markos
been
nice
enough
to
help
out
with
the
project,
especially
after
I
got
so
busy
here
in
the
last
year
and
a
half
or
so
so
anyway,
Marko
wrote,
this
blog
I
think
he
works
for
our
friends
over
at
Lucy
who
are
based
out
of
Berlin.
A
So
this
is
probably
a
similar
concept,
probably
handled
a
bit
differently
and
detailed
step
by
step
instructions
on
how
to
make
that
work.
Okay,
let's
see
so
Khalid,
says
hi
guys
from
Dubai
hi
good
to
see
you
and
then
we
have
Michael
who's
waving,
hi
Michael
good
to
see
you.
Thank
you
again.
Everyone
for
joining
I
know
it's
it's
a
weird
time
for
some
folks
in
other
parts
of
the
world
and
I
think
the
majority
of
our
audience
here
is
is
not
from
the
US.
A
So
it's
always
great
to
see
folks
who
are
spending
their
Friday
nights
or
Saturday
mornings
or
whatever
time
it
is
for
you
hanging
out
here,
live
with
us
at
VMware,
so
things
joining
so
anyway,
there's
Markos
blog
and
then
we
want
to
give
a
shout
out
to
Paris
Pitman
who's.
The
community
manager
for
Google.
Another
good
friend
of
mine,
I,
seriously
feel
like
Duffy
picked,
all
of
my
all
of
my
home
homies
from
kubernetes
for
the
key
weekly
this
week,
but
Paris
did
a
retro
and
look
forward
to
two
on
the
contributor
summit.
A
That's
coming
up
so
Paris
and
George
here
at
VMware.
Work
really
hard
on
these
kubernetes
contributor
summits.
They've
evolved
quite
a
bit
over
the
past
couple
of
years.
That
started
out
originally
as
like
a
couple
of
us
sitting
in
a
room
who
got
some
free
t-shirts,
hacking
on
kubernetes
and
cutting
a
release
live
at
cube
con
to
now.
This
way
more
advanced
way,
more
formal.
A
We
actually
can
go
and
be
productive
and
it
gets
recorded
and
we
take
notes-
and
it's
like
a
very
exciting
part
of
our
year
when
everybody
gets
to
get
together
and
sit
in
a
room
and
talk
about
kubernetes
in
the
direction
the
project
is
going.
So
none
of
this
would
be
possible
without
Paris.
So
a
big
thank
you
to
Paris,
we're
all
very
grateful,
and
then
you
can
also
see
here:
here's
George
Paris,
Bob
Jeff
and
a
handful
of
others.
A
Josh
burkas,
oh
he's
over
at
Red,
Hat
he's
another
great
one:
Jason
D,
Tiberias,
I,
Tim
pepper,
a
lot,
a
lot
of
really
great
people
here,
a
lot
of
really
great
names.
So
thank
you
so
much
for
everything
y'all.
Do
we
really
appreciate
it?
Let's
see,
what's
going
on
in
chat,
shameless
plug,
I'm
demoing,
the
cluster
API,
a
debut
of
provider
I
had
to
meet
up
in
Seattle
on
the
27th
okay,
so
this
is
Ashish.
He
just
mentioned
this
in
chat.
Hi
Ashish,
it's
good
to
see
you
all
so
I
think
I'm,
pretty
sure.
A
I'm
pronouncing
your
name
right,
but
correct
me.
If
I'm
not
I
apologize,
but
yeah
Ashish
is
doing
a
demonstration
on
the
cluster.
Api
he's
been
working
hard
in
that
project
for
quite
some
time
now
and
I
guess
he's
doing
a
demo
of
the
cluster
aid
API
AWS
provider,
which
is
something
that
we
were
working
hard
on
at
a
hefty.
A
Oh,
so
it'll
be
exciting
to
see
how
that's
up
and
going,
we
did
a
TTI
K
on
the
cluster
API
for
AWS
quite
some
time
ago,
so
I'm
excited
to
see
how
it's
evolved
since
the
last
time,
I
looked
at
it
Ashish
do
you
know
if
that's
gonna
be
recorded
anywhere,
and
if
so,
can
you
share
a
link
in
the
document
here
so
that
folks
can
find
it
later
after
the
fact,
because
I
know
a
lot
of
people
come
back
and
watch
this
over
the
sorry
I
have
bad
bones
over
the
coming
weeks
after
we,
we
really
said
so
anyway,
shall
up
to
Paris
things
Ashish.
A
We
learned
about
the
word
glow
and
Icelandic,
and
then
now,
let's
start
talking
about
the
reason
we're
all
here
which
is
glue.
So
Patrick
he's
one
of
our
engineers
here
at
VMware,
I
typed,
this
up
in
slack
and
I,
don't
think
he
realized.
I
was
going
to
quote
him
on
this,
but
it
really
was
a
fantastic
little
snippet
that
he
did
as
she
says.
Yes,
I'm
pronouncing
his
name
right.
So
that's
good!
A
So
I
just
want
to
read
this
really
quick,
so
basically
sto
ambassador
at
contour,
etc,
are
all
just
ways
of
configuring
envoy
on
kubernetes.
We're
gonna
talk
a
little
bit
about
what
that
means
more
concretely
in
just
a
moment,
but
glue
decomposes.
All
of
that
and
that
process
down
into
more
fundamental
types-
and
this
is
the
same
thing
that
kubernetes
itself
does-
is
take
this
complexity
and
break
it
down
into
like
small,
convenient
little
interfaces
that
you
can
replace
with
the
various
other
pieces
of
logic.
But
the
interface
is
always
stay.
The
same.
A
The
Linux
kernel
itself
also
behaving
the
same
way.
So
Glu
did
this,
but
with
networking
built
on
top
of
the
Envoy
service
proxy
from
our
friends
over
at
lyft
and
our
friend
Matt
Klein.
So
you
can
come
here
and
you
can
actually
see
this
glue.
Dot.
Solo
IO,
dot,
dev
and
you
can
see
how
it
breaks
all
of
this
down
to
buy
no
dresses
domain
sets
and
then
routes
and
TLS
configs
for
you
to
your
domain
sets.
So
it's
sort
of
a
logical
way
of
breaking
down
how
you're
gonna
configure
your
layer,
seven
networking.
A
A
So
anyway,
going
back
here
that
was
sort
of
the
TLDR,
quick
and
dirty
explanation
of
glue
and
how
it
relates
to
envoy-
and
we
all
know
contour-
is
one
of
the
the
projects
that
Dave
Cheney,
one
of
our
star
engineers
here
based
out
of
Australia,
has
been
working
on
he's
at
VMware
now
as
well
and
Condor.
Does
this
similar
thing
so
this
in
a
weird
way
sort
of
a
compare
and
contrast
between
contour
and
glue?
A
And
we
want
to
see
some
of
the
differences
and
similarities
between
the
both
of
them,
but
yeah
basically
makes
it
really
easy
way
to
build
envoy
configuration
services.
In
other
words,
it
provides
an
abstraction
on
top
of
envoy.
That
makes
it
easy
for
engineers
to
interact
with
so
it
sort
of
boils
it
down
and
gives
you
a
toolkit
for
working
with
and
around
envoy
and
contour,
of
course,
does
the
same
thing.
A
Okay,
so
we're
gonna
pause
there
with
envoy
and
glue
fur
in
contour
for
just
a
moment
and
we're
gonna
go
on
this
other
little
tangent
and
then
we're
gonna
tie
the
two
back
together.
So
this
other
tangent
here
is
on
Twitter
I
forget
who
it
was,
but
it's
on
my
Twitter
and
I'll
retweet
it
after
after
the
show
somebody
had
mentioned
hey.
A
Will
you
check
out
the
new
kubernetes
dashboard
and
I'm
like
yeah,
totally
I
love
a
good
kubernetes,
dashboard
and
I,
always
like
think
it's
very
rewarding
at
the
end
of
a
TTI
K,
to
get
something
up
and
running
in
kubernetes
and
actually
be
able
to
go
and
use
it.
So
the
kate's
dashboard
here
you
can
find
it
at
docker
hub
/hr,
Branson's
last
kate's,
you
can
see
here's
a
screenshot
of
it.
A
It
looks
pretty
similar
to
the
old
one,
although
it
looks
like
a
little
bit
more
like
Java,
Script
II
like
fancy
or
whatever,
like
chardy
graphi.
I
don't
know
if
there's
an
official
ux
term
for
that
I'm,
a
colonel
engineers.
Don't
don't
get
mad
at
me
if
there's
better
words
to
be
using
there.
So,
basically
you
need
a
running
kubernetes
cluster.
A
You
know
metric
server
installed
it's
optional,
but
strongly
recommended
I
think
we
might
skip
this
step
today
because
we're
just
going
to
show
a
demo
in
a
kubernetes
cluster
configured
for
Open,
ID,
Connect,
auth
also
optional,
and
also
we
need-
and
this
is
where
things
start
to
tie
together.
If
you
look
down
here,
there
is
an
ingress
defined
and
we
all
know
that
if
you
define
an
ingress,
you
of
course
need
an
ingress
controller
in
order
to
make
this
run.
A
So
what
better
ingress
controller
to
demo
the
kate's
dashboard
with
other
than
glue
the
controller
built
on
Envoy.
So,
as
I
was
looking
at
checking
out
the
kate's
dashboard
here,
I
was
like
trying
to
also
come
up
with
a
project
that
we
could
use
glue
to
run
on
and
on
these
two
just
sort
of
fit
magically
together.
So
we're
gonna.
A
Do
both
we're
gonna
check
out
the
new
kubernetes
dashboard
here
to
see
what
it's
all
about
explorer
kubernetes
cluster
and
we're
gonna
get
this
ingress
down
here
at
the
bottom
up
and
reconciling
using
glue
sound
like
a
plan,
everyone
and
so
Louis-
hey
Martin,
wait!
No!
That's
from
earlier
okay!
So
that's
the
game
plan.
So
let's
go
here:
doo
doo,
doo
jump
in
my
terminal.
A
Can
everybody
see
this
okay
I?
Think
I
can
turn
this
okay
CD
down?
Let's
do
that
really
quick?
How
do
I
do
this
profiles?
Open
profiles
default?
Maybe
no,
oh
close
that
one!
Let's
do
edit
I,
don't
spend
too
much
time
on
this,
but
maybe
not
I
was
gonna.
Try
to
turn
the
opacity
down,
but
it's
fine
I'll
just
make
sure
that
I
go
slow
so
that
it
doesn't
bleed
through
too
terribly
much
moans
uber
says:
hi
Chris
hi
good.
To
see
you!
Oh
and
happiest
as
the
metric
server
is
up
and
running
already.
A
A
One
point
ten,
and
if
we
do
our
other
alias
here,
I
always
do
Kade
dump
is
equal
to
K,
will
type
it
out
explicitly.
Q
Bechtel
get
all
all
namespaces
this
we
can
just
actually
just
do
a
dump
of
the
majority
of
the
objects
that
are
currently
declared
in
our
kubernetes
data
set
here
eric
says:
good
case
is
much
cooler
with
the
metric
server.
Okay,
cool
and
vault
says
hi
from
Finland
high-salt.
Welcome
we're
in
Finland.
Are
you
from
I've
always
wanted
to
go
to
Finland
I
have
some
really
good
friends
who
are
finished?
A
Ok,
so
we
can
see
here
we
have
a
couple
of
pods
that
are
running
on
the
nodes.
I'm
assuming
this
is
like
a
they're
running
in
cube
system.
This
has
got
to
be
an
e
KS.
Something
like
this
has
got
to
be
a
demon
set
or
some
like
that
running
on
each
of
our
nodes
running
and
keep
system.
That's
doing
magic,
AWS
things
behind
the
scenes.
A
For
us,
we
have
our
kubernetes
service,
cube,
DNS
and
AHA
service
metric
server,
yay
already
up
and
running
thanks,
stuffy
I
should
ask
you
to
spin
up
clusters
for
us
more.
This
is
really
convenient
for
today
you
can
see.
Basically,
the
only
thing
we
have
running
in
default
is
just
our
kubernetes
service.
That's
always
there,
and
everything
else
is
over
here
on.
The
left
side
in
this
left
column
is
running
in
the
cube
system
namespace,
which
is
effectively
route
for
kubernetes.
A
Let's
see
what
folks
are
saying:
ok,
so
mows
offer
says:
is
it
too
late
here
I
will
watch
the
recording
tomorrow?
Okay
have
a
good
night,
most
effort.
Thank
you
for
joining
and
then
sult
says:
I
live
in
I'm
gonna.
Try
to
pronounce
for
this.
What
is
it
Lulu,
Oh,
Lu
I
think
maybe
the
first
one
but
we'll
see
and
Josh
says
cue,
bechtel,
new
fav
pronunciation,
so
the
story
behind
québec
duel
is
when
folks
were
arguing
over
C
cube.
C
T
L
keep
cuddle
cute
control.
A
At
the
same
time
me
and
Tim
Hawken,
both
I
think,
came
to
the
same
conclusion
based
on
high
octal,
which
is
ioctl
that
Q
Bechdel
would
be
kind
of
funny
to
say
so.
I
started
saying
it
as
a
gag
and
then
I
just
got
so
used
to
saying
it
over
the
past
two
years,
I've
actually
stuck
with
it
and
now
I
just
say,
and
that's
just
what
I
call
it
and
yeah
I
call
it
Q,
Bechdel
and
I
think
that's
totally
rad.
A
A
I
have
a
coop
cuddle,
t-shirt,
I,
know,
there's
QB,
cuddles
and
folks
say
that
as
well.
It's
just
you
know
you
do
you
as
long
as
you're
happy
and
as
long
as
you're
using
kubernetes.
That's
all
we
want
so
OBS
config,
let's
get
out
here
and
go
to
my
go
path
and
that'll,
be
our
home
base
for
now
cool
and
let's
go
figure
out
how
to
install
glue
which
again
I
have
not
looked
at
yet.
So
this
is
gonna
be
fun.
So
here
on
the
glue
documentation
site,
which
is
glue,
dot,
solo
IO.
A
We
have
this
awesome
little
person
monster
thing
here.
That's
really
cute
I
think
it's
just
the
glue
thing.
I,
don't
know
if
they
have
a
name,
but
I
really
hope
they
do.
This
is
a
Hugo
site
which
Hugo
is
a
go
program
that
allows
you
to
build
and
run
static
websites
and
that's
what
we
use
for
advocacy,
site,
hep,
tio,
now
VMware,
and
so
this
is
the
the
doc
site
that
glue
has
built
for
us.
A
So
if
you
come
here
on,
the
Left
I
wanted
to
check
out
installing
glue,
and
maybe
even
the
user
guide,
so
it'll
probably
be
the
the
things
that
I
want
to
check
out.
Oh,
we
have
a
question
in
the
chat:
Yong
Chris.
Could
you
please
briefly
describe
what
is
ingress
and
why
it's
needed
yeah,
totally,
okay,
so
ingress
TLDR
is
basically
how
do
I
get
traffic,
particularly
HTTP
HTTP
traffic
into
my
cluster
and
then
more
importantly,
what
do
I
do
with
my
traffic
after
that?
A
So
when
kubernetes
was
very
young,
we
had
this
capability
of
setting
up
our
services
to
type
load.
Balancer
and
then
we
can
map
a
service
to
various
deployments
or
pods
based
on
labels,
and
that's
how
we
did
all
of
our
interesting
mapping
between
the
scenes-
and
there
were
some
concerns
about
that
and
we
found
out
that
wasn't
as
dynamic
and
composable
enough
as
we
would
have
liked.
So
a
really
basic
here,
you
know
what
I
can
do
this
really
quick
I'll
just
draw
a
picture
for
you
camera.
A
Let's
see
a
really
basic
example.
Here
is:
if
we
had
a
domain
name,
let's
say
we
had
T
gik
dot-com
and
we
had
TJ
com,
slash
auth
and
we
had
t
GI,
akom,
slash,
let's
say
API,
and
then
we
had
another
one.
I
was
like.
What's
call
this
one
had
been
right,
so
we
have
all
these
endpoints
that
we
want
to
do
various
things
like.
Maybe
the
authoring
point
just
needs
to
go
to
an
authentication
service
that
all
it
does
is
off,
but
the
API
needs
to
run
whatever
version
of
the
API
ever
currently.
A
So
an
admin
is
like
we're
all
like
the
admin.
Ii
stuff
happens
that
you
have
to
have
the
secret
password
to
that
all
the
orchestrator
operators
can
come
in
and
play
with.
The
point
I'm
trying
to
make
here,
though,
is
if
we
had
kubernetes
running,
and
we
had
a
namespace
here
called
the
auth
namespace.
So
we'll
put
this
in
a
box
here
and
in
that
namespace
we
had
like.
A
We
actually
are
able
to
come
in
and
define
slash,
auth
and
we're
able
to
take
this,
and
we
were
able
to
map
that
over
here
to
this
service
and
if
they
take
API,
we
can
map
that
over
here
to
this
service
and
so
on
and
so
forth,
so
that
we're
able
to
actually
do
this,
this
complex
routing
between
various
services.
Now
this
is
exciting,
because
this
helps
us
break
down
the
monolith
right
if
we're
doing
our
services.
A
This
way
we
now,
if
we
want
to
change
out
our
service,
we
don't
have
to
touch
the
API
or
the
admin
service.
So
this
gives
us
a
lot
of
flexibility
as
architects
and
engineers
of
keeping
our
services
up
and
making
things
much
more
resilient.
So
the
way
universe
works
is-
and
this
is
the
controller
pattern
that
we
see
all
over
kubernetes.
So
this
is
like
literally
the
sort
of
blood
of
kubernetes
here,
the
way
that
ingress
works
I'll
draw
lines.
This
is
a
new
picture
down
here.
A
This
is
an
example
of
one
here,
so
you
would
define
one
of
these
and
you
would
go
and
you
would
actually
save
that
in
kubernetes.
You
know
came
back
to
apply
that
and
then
let's
go
back
my
dog
camera
and
you
would
apply
that
and
that
would
go
over
and
that
would
get
stored
in
at
CD.
So
a
CD
is
here
and
then
we
would
have
this
other
piece
of
software
here
that
we
call
a
controller.
A
Sorry,
my
handwriting
is
sloppy.
That's
how
you
know
I'm
a
good
engineer,
controller
and
then
that's
going
to
go
and
it's
going
to
do
a
couple
of
things.
That's
going
to
run
this
thing.
We
call
a
control
loop,
which
basically
is
this
ongoing:
never-ending
reconcile
the
state
of
the
world
loop
that
runs
it
over
and
over
again.
So
if
you
check
out
my
book
cloud
native
infrastructure,
we
have
this
thing
that
we
call
the
reckons.
Oh
that
annoying.
We
have
this
thing
that
we
call
the
reconciler
pattern.
A
Let's
see
if
I
can
fix
this
and
that
talks
a
lot
about
the
patterns
behind
kubernetes
and
the
patterns
behind
reconciling
infrastructure
in
general.
And
that's
what's
going
on
here
with
ingress,
and
so
the
ingress
controller
would
read
with
a
CMO
file
that
you
defined
and
would
actually
go
and
handled
routing
the
traffic
itself
to
wherever
it
needs
to
go
up
here
here
and
here.
So
this
is
what
contour
does.
This
is
what
all
of
the
ingress
controllers
do.
A
A
We
can
handle
this
and
we
can
mutate
this
or
set
this
up,
and
a
lot
of
different,
interesting,
fashions
and
envoy
is
exactly
what
we
need
to
have
all
these
really
cool
capabilities
like,
for
instance,
what
if
our
off
service
also
was
running
in
a
second
kubernetes
cluster,
behind
the
scenes
right
and
that's
the
second
kubernetes
cluster.
We
want
it
to
not
only
load
balanced
internally
across
that
kubernetes
cluster,
but
also
on
the
second
one,
and
this
is
what
gimble
does
and
there's
a
handful
of
other
projects.
A
Okay.
So
nothing,
this
that's
a
ingress
controllers
and
that's
what's
going
on
here
behind
the
scenes.
Does
anybody
have
any
questions
all
right
and
on
that
note,
I'm
gonna
grab
a
sip
of
my
a
my
happy
Friday
beer
here
that
somebody
left
in
the
fridge
that
I
was
nice
enough
to
leave
for
me.
So
thank
you
for
that
in
yon.
Again,
I
hope
this.
This
answers
your
question
here
and
then
we're
gonna
go
back
and
we're
gonna
jump
into
the
terminal
in
here.
Just
a
second.
A
Happy
Friday:
that's
how
you
know
it's
the
weekend
right
there,
okay,
cool
so
back
into
our
terminal.
We
want
to
well
actually
no,
let's
go
back
to
the
docks,
that's
what
we
want
to
do.
We
want
to
install
glue
and
we
all
know
I
am
a
tab
closer.
So
let
me
close
all
these
tabs,
and
so
now
we
have
our
markdown
file.
Here
is
the
dashboard
that
we're
going
to
get
up
and
running,
and
then
here
is
glue
so
using
glue.
A
We
can
even
just
go
really
quickly,
read
the
the
introduction
here
so
earlier
when
I
drew
the
graph
I
talked
about
all
the
different
configuration
that
you
would
care
about
as
you're
looking
at
routing
various
pieces
of
networking
traffic
to
different
services
and
here's
a
list
of
all
the
features
that
glue
can
do
for
us,
so
we
can
handle
TSL
termination.
We
can
handle
monitoring,
we
could
have
cut
up
to
open
tracing
so
that
we
can
actually
see
where
all
these
packets
are
going.
A
So
it
gives
us
a
lot
of
introspection
and
a
lot
of
observability
into
what's
going
on
in
the
networking
stack
using
kubernetes
ingress.
So
that's
really
exciting
and
then
you
know
we
got
some
buzz
words
in
here
scalability
and
performance,
but
glue
leverages
envoy
for
its
high
performance
and
low
footprint
and
if
we
go
and
we
look
at
envoy,
so
here's
a
really
good
blog.
I
should
have
brought
this
up
earlier.
A
That
sort
of
explains
what
envoy
is
and
why
the
api
is
important,
but
if
we
actually
just
go
to
the
Envoy
docs
on
left
envoy
proxy
home
come
down
here.
Why
envoy
features
right
here
on
boy?
Is
a
self-contained
high-performance
server?
That's
an
important
word
there
with
a
small
memory
footprint
and
runs
alongside
any
application
language
or
framework
because
of
the
API
that
is
defined
here.
So
there
should
be
yeah.
A
So
here's
like
you
can
scroll
down
and
you
can
see
the
different
API
is
and
what
they
do
and
all
the
different
endpoints
and
this
common
API
then
plugs
into
this
underlying
Envoy
code,
which
is
all
written
in
C++
again
thanks
to
Matt
Klein
and
our
friends
at
lyft,
and
this
actually
makes
all
that
happen
and
brings
all
that
to
life.
But
it
gives
developers
this
really
really
convenient
API.
A
So
again
receive
this
other
pattern
of
taking
complexity,
providing
a
meaningful
abstraction
on
it
and
then
making
that
whole
thing
lightweight,
and
that's
really
why
envoy
has
been
so
successful
and
why
it's
such
a
great
tool
that
engineers
choose
to
work
with
and
why
going
back
to
Patrick's
original
quote
here:
misty
Oh,
ambassador
contour
are
all
just
ways
of
configuring
envoy
on
kubernetes,
which
he's
absolutely
right,
but
that
was
a
really
really
short
way
of
saying
everything.
I
just
said
so:
I
kind
of
broke
it
down
a
little
bit
there
for
you.
A
So
anyway,
here's
on
boy,
you
can
come
check
out
the
code
if
you're
a
C++
person
I
have
in
written
C++
since
I
worked
at
SolidFire
in
Boulder
years
and
years
ago.
So
I'm
not
even
gonna,
try
to
go
look
through
this
code
right
now,
but
it
is
because
it's
written
in
C
C++,
it's
super
lightweight
and
it
gets
very
low-level
very
quickly.
So
we're
very
grateful
for
folks
who
have
made
the
wonderful
API
on
top
of
this,
to
make
it
easy
for
us
to
interact
with
very
handy
okay.
A
So
let's
go
back
to
Glu,
okay,
so
yeah
Glu
leverages
on
boy
for
its
high
performance
and
footprint.
You
can
build
plug-ins
for
it.
There's
tooling,
you
can
even
do
event
based
events
using
cloud
events.
You
can
do
pub/sub
json
to
j
RPC
transcoding,
but
I
think
probably
the
most
exciting
thing
about
glue
that
I
noticed.
Let's
do
glue
kubernetes.
A
I
guess
this
is
the
site.
Maybe
it's
not
here
the
exciting
thing
that
I
noticed
about
glue?
Was
they
called
itself
a
hybrid
load?
Balancer,
let's
see
if
it's
here
at
the
github
page,
maybe
high
here
we
go.
Yeah
glue
is
uniquely
designed
to
support
hybrid
application
in
which
multiple
technologies,
architectures
protocols
and
clouds
can
exist
so
in
other
words,
glue,
allows
you
to
not
only
route
in
the
same
way.
We
we
routed
here
internally
in
kubernetes,
but
it
also
allows
you
to
like.
A
So
you
can
start
to
like
look
at
how
we're
gonna
migrate
to
and
from
the
cloud
and
how
we're
going
to
start
looking
at
true
hybrid
cloud
infrastructure,
which
is
really
exciting,
because
this
gives
us
that
flexibility
and
declarative
nature
of
kubernetes
and
we
can
come
in
and
now
start
to
define
our
baterry
services
using
tools
like
glue
and
contour,
which
is
really
rad.
Ok.
So
let's
go
back
to
my
screen
hybrid
apps.
Here
it
is
here
so
yeah
grab
for
the
word
hybrid
and
you
can
see
that
it
does
a
lot
of
things.
A
I
think
one
of
the
things
folks
are
most
excited
about.
Is
this
new
fancy
thing
called
serverless,
which
is
a
great
example
of
a
really
unfortunate
name,
because
there
are
actually
a
lot
of
servers
involved,
but
serverless
is
just
another
way
of
saying
functions
as
a
service.
So
if
you
just
have
a
very
simple
use
case
for
a
very
simple
micro
service,
you
could
write
one
of
these
small
lightweight
service
functions
and
you
could
use
glue
to
map
your
traffic
from
legacy
apps
to
micro
services
and
kubernetes
to
these
service
service
functions
and
managing
mutate.
A
All
of
that
using
glue,
which
is
exciting,
we're
going
to
be
using
blue
very
mentally.
If
that's
a
word
today,
which
is
basically
just
going
to
be
as
a
load
balancer
and
an
ingress
controller
in
kubernetes,
ok,
so
back
to
glue
installing
glue,
there's
a
glue
command
line
tool
like
all
glue
users
should
have
to
make
their
lives
easier.
Well,
we
should
install
that
unpacked
the
glue
my
gosh,
the
Google
octal
glue
octal.
Oh
my
gosh.
What
are
we
gonna
call?
A
A
Updating
homebrew
again
I'm
on
my
personal
laptop,
so
I
don't
use
this
thing
very
often,
I
usually
just
use
my
work
one,
so
there's
going
there
might
be
some
like
stuff
on
here,
like
I,
think
I
don't
even
have
W
get
installed,
so
we
might
be
spending
some
time
today
installing
a
few
tools.
If
we,
if
we
need
them,
okay,
so
glue
octal
is
installed.
A
Oh
yeah
and
it's
written
go
I,
know
what's
written
and
go
because
this
is
clearly
the
cobra
command
library
that
we
see
all
over
kubernetes
and
all
over
pretty
much
every
go
project
everywhere,
and
this
is
one
of
my
favorite
command
line
tool,
flag,
parsing,
libraries
in
the
world.
It
looks
like
we
have
focusing
stuff
in
chat.
Let's
see,
what's
going
on,
yon
says:
well
thanks
a
lot
for
that
explanation.
A
Currently
retire
services
to
AWS,
load,
balancer
or
use
cluster
DNS
name
yeah,
so
the
AWS
load
balancers
do
a
lot
of
the
same
things
that
cost
a
lot
of
money.
This
is
a
way
of
you
know,
taking
some
of
that
and
running
it
in
kubernetes,
which
requires
a
little
more
overhead,
but
there's
some
flexibility
and
control
that
you
would
get
it
otherwise.
Michael
says:
glue,
CTL,
we're
going
with
glue
octal
today
on
t
gik,
happy
Friday,
love,
Chris,
Nova,
so
yeah
octal.
So
we
got
that
installed.
A
Let's
see
what
is
next
next
step
install
glue,
so
when
you
either
kubernetes
cluster
set
up
an
installation
for
glue
on
a
kubernetes
cluster,
that's
what
we
wanted
to
get
so
kubernetes
prep.
In
this
document
we
will
review
how
to
prep
kubernetes
environments
before
installing
glue
gke
okay.
So
this
is
like
where
we
need
to
set
up
like
our
back
and
stuff
like
that.
So
so
far,
I'm
thinking
GK
is
gonna,
be
the
one
we're
gonna
follow.
Oh,
he
casts
perfect.
We
have
an
e
KS
cluster
cuz
dub.
A
A
In
my
opinion,
yeah
I
just
I,
just
like
glue
octal,
that's
just
some
I'm
gonna,
say
it
so
yeah
here
it
says
we
suggest
using
eks
cuddle,
which
I'm
off
on
a
tangent,
but
if
you
want
to
set
up
EK
s,
mu,
KS
cluster
in
Amazon,
this
tool
makes
your
life
astoundingly
easy,
and
the
irony
of
it
is
this
tool-
and
this
was
so
funny.
This
happened
earlier
today,
this
tool
vendors,
a
lot
of
go
programs
that
I
wrote
like
forever
ago,
particularly
one
that
generates
a
random
name
for
you
and
the
kubernetes
cluster.
A
We're
using
today
uses
this
tool.
Uk
s,
cuddle
about
vendors,
a
lot
of
my
old
code
and
created
a
kubernetes
cluster
named
after
my
old
dog
Charlie,
and
that
just
made
me
really
happy
when
I
saw
that
today.
So
that
was
exciting,
so
eks
cuddle
is
an
exciting
tool,
so
it
says:
eks
called
create
named
your
cluster
named
your
region.
I
think
Duffy
did
a
similar
thing
here,
but
probably
configured
it
a
little
bit
better
for
us
again,
Thank
You
defi.
That
was
very
much
appreciated.
A
Okay,
so
the
next
thing
it
says
is
verify
that
your
cubital
context
is
curly,
pointing
out
your
eks
clusters.
So
we
should
have
that
already.
So,
let's
go
here
and
let's
do
K
get
nodes.
We
already
did
this,
but
yes,
we
are
in
fact
talking
to
AWS.
We
can
tell
from
the
cname
here.
So
let's
go
back
and
it
says,
use
your
current
context.
Okay
and
you're
all
set
glue,
install
guide
here
so
we're
good.
A
We
didn't
really
have
to
do
anything
options
to
install
glue
via
the
command
line,
interface,
CLI
recommended
or
via
a
helm.
I'm
gonna
go
with
the
the
CLI
here.
So,
let's
see
what
it
says
says:
choosing
a
deployment
option
for
installing
glue
into
your
kubernetes
cluster
gateway,
ingress.
Let's
gonna
do
ingress
so
install
the
glue
ingress
controller
to
your
kubernetes
cluster,
using
glue
octal,
no
way,
no
wow.
This
is
really
cool.
So
apparently
all
you
have
to
do
is
type
glue,
octal
install
ingress,
let's
see
how
accurate
this
is.
A
They
said
this
will
be
really
cool.
This
is
all
you
have
to
do
so
it
creates
okay,
let's
see
what's
going
on
here,
so
it
created
some
CR
DS
for
us.
So
we
have
settings,
we
have
gateways,
we
have
virtual
services,
we
have
proxies.
We
have
up
streams
so
I'm
guessing
these
are
the
actual
concrete
abstractions
on
top
of
the
Envoy
service
proxy.
That
glue
is
going
to
be
using
for
us
and
it
looks
like
they're
pretty
self-explanatory.
A
We
have
proxies.
We
have
up
streams.
We
have
virtual
services,
which
I'm
assuming
are
just
like
ways
of
doing
dynamic.
Routing
no
resource
is
found,
I,
don't
know
what
that
line
is
from
Google,
so
we
created
the
namespace
glue
system
and
then
we
installed
the
config
map
ingress
envoy,
config.
Okay,
so
then
we
did
a
cluster
role,
glue,
role,
ingress
and
then
we
did
a
cluster
role
binding,
which
again
I
know.
A
So
it's
just
a
way
of
managing
who
can
do
what
and
the
cholesterol
binding
is
just
this
thing
in
the
middle
that
binds
one
to
the
other.
So
that's
all!
It
is
okay,
so
back
here
so
then,
when
we
created
the
ingress
service
proxy,
which
okay
so
the
way
that
ingress
works,
is
you
basically
create
a
regular
kubernetes
service?
And
then
you
map
that
to
your
ingress
controller
and
then
your
you
guest
controller,
then
maps
that
internally,
so
there's
actually
like
two
layers
of
networking
going
on
there.
A
We
created
the
service
glue
so
yeah,
so
this
is
where
the
traffic
comes
in
and
then
this
is
where
we're
going
to
distribute
that
traffic
according
to
what
ingress
rules
were
about
to
define
deployment
ingress
created.
So
this
is
now
I
bet.
If
we
do
K
get
P,
oh
namespace
glue
system,
I
bet
we
can
see,
we
have
pods
for
all
of
these
different
deployments
or
anything
here,
let's
see
yep
and
they
of
course,
we
do
discovery,
glue,
ingress,
ingress,
proxy
discovery,
glue,
ingress,
ingress
proxy,
so
those
math
tubs
just
in
a
different
order
there.
A
So
we
are
running
pods.
It
looks
like
our
pods
are
nice
and
happy
and
healthy,
so
Wow.
That
was
really
awesome.
One
command
and
we
have
numerous
controller
up
and
running
from
a
statically
link
to
go
binary.
So
that's
really
exciting.
So
let's
go
back
to
the
docks
and
it
says
yeah
after
you
verify
your
installation.
Let's
do
that
check
that
glue,
pod
and
services
have
been
created
depending
on
your
install
option.
A
You
may
see
some
differences
from
the
following
example
and
if
you
choose
to
install
glue
into
a
different
name,
space
you'll
then
need
to
query
that
chosen
names
face
instead.
Okay,
so
it
says
just
get
all
four
glee
system
and
we
kind
of
already
did
this.
We
ready
listed
our
pods
here
just
naturally
anyway,
and
then
you
can
see.
We
also
have
deployments
and
replica
sets
and
services
that
came
along
with
it,
which
we
just
talked
about
a
moment
ago,
so
yeah.
A
So
now
we're
going
to
need
to
do
is:
let's
actually
get
our
services
here
and
figure
out
where
our
cluster
is
listening
on
the
interwebs.
So
we
want
to
do
a
git,
SVC
namespace
glue
system,
because
this
is
where
English
is
deployed
and
we
want
to
do
oh
yeah
nope.
We
want
to
do
Oh
long
or
wide.
What
is
it
I
think
is
why
so
wide
might
be
go
ahead?
Let's
try
it
away
there.
We
go
okay,
so
let's
clear
a
screen
run
this
again,
so
we
get
a
nice
easy
output
and
bam.
A
A
I
did
one
on
traffic,
so
you
feel
free
to
go
check
those
out
and
those
will
explain
in
a
little
bit
more
detail
how
I
knew
to
instantly
run
this
command
here
and
basically
do
the
O
ID
to
get
the
name
of
our
AWS
load
balancers.
So
let's
go
jump
into
my
AWS
console
and
see
what
actually
happen
behind
the
scenes
here.
So
let's
go
here
and
we
want
to
go
to
Amazon
management
console.
A
Do
do
log
in
and
then
we
want
to
go
to
services.
It's
got
easy
to
make
sure
that
we
have
we're
using
the
right
user
here
nope.
We
need
to
switch
over
to
dev,
nova
or
Canova.
I
don't
know
if
it's
a
silent,
K
or
not-
and
you
can
come
in
here.
Oh
this
is
running
in
duffy's,
amazon.
Isn't
it
okay,
so
usually
I
would
have
my
kubernetes
cluster
running
in
my
own.
A
Ada
abuse
account
we're
running
in
Duffy's
right
now,
so
we
actually
were
just
able
to
interact
with
Duffy's
AWS
account
using
the
magic
of
kubernetes,
which
is
pretty
exciting.
If
you
think
about
what
actually
just
happened,
I
was
able
to
create
a
load
balancer
which,
if
it
was
in
my
AWS
account,
would
be
showing
up
here
under
the
load
balancers
tab,
but
you
can
see
these
were
created
march.
Aids
and
I
should
probably
delete
those
because
they're
not
cheap,
so.
A
Anyway,
we
would
have
seen
the
load
balancer
created
here
and
that
would
have
been
mapped
to
the
instances
that
we
have
running
in
our
kubernetes
cluster.
So
we
actually
would
have
gotten
the
real
public-facing
ELB
load,
balancer
in
front
of
our
kubernetes
cluster,
mapped
to
a
specific
service
and
then
again,
services
can
be
mapped
to
other
things
based
on
labels.
That's
defined
in
a
lot
of
our
other
ingress
episodes.
A
A
This
page
isn't
working
and
so
I
think
what
what's
happening
here
is
we
have
ingress
up
and
running.
We
have
an
ingress
controller
up
and
running,
but
we
don't
have
any
ingress
routes
to
find.
So
doesn't
there's
nothing
to
do
this,
nothing
for
to
listen
on.
So
what
we
want
to
do
is
and
I'm
gonna
jump
off
the
screen
because
we're
gonna
use
my
personal
AWS
account.
Actually,
let's
not
do
that.
I
think
I
bought
a
domain
when
I
was
that
when
we
were
happy,
oh
I
might
be
able
to
use
that.
A
A
A
We
can
delete
mailing
and
we
want
to
change
our
TGI
Kate's.
Nobody
want
to
add
an
a
record
okay,
so
this
is
awesome.
So
this
is
a
very
simple
TGI,
Kate's,
calm,
a
record
and
we're
gonna
do
actually
look
up
and
resolve
what
this
this
address
resolves
to.
So
we
can
dig,
do
I
have
dig
I
do
have
dig
here
and
you
can
see
I
I
think.
A
Fernan
says
hi
Chris,
you
know
how
one
would
set
up
multiple
universe
controllers,
ie,
serving
tea
viacom
using
multiple
addresses,
and
how
would
your
failure
recovery
look
like
when
a
controller
dies
for
fern,
and
can
you
explain
more
or
elaborate
a
little
bit
more
on
your
use
case
for
why
you
would
be
using
multiple
ingress
controllers
like?
Why
is
one
ingress
controller?
Not
enough,
for
you
would
be
my
question.
A
Roy
says:
does
it
have
integration
with
certain
manager
for
wiring
up
TLS,
so
it's
automatically
from
lit
scripts,
so
it
can
terminate
each
gtps.
It
did
mention
TLS
on
there,
whether
or
not
that
uses
a
certain
manager
and
let's
encrypt
I,
don't
know.
Yet
we
can
look
at
that
in
a
second
though,
but
I
know,
TLS
will
be
something
that
we
should.
We
should
go
at
least
figure
out
what
it's
advertising
to
do
so
anyway,
let's
try
to
ping
this
here,
there's
my
mouth,
so
we
can
get
it's
real
true
public
address
here.
A
So
where
was
this
thing?
Svc
grab
this
ping
this
BAM-
and
here
is
our
public
IP
address
yay
hacking
right,
okay!
So
let's
go
here.
Let's
paste
this
in
here
and
we'll
create
that
and
bam
poof.
We
have
an
a
record
for
TGI
Kate's
comm,
that's
pointed
to
our
public
IP
address
that
should
be
broken,
didn't
send
any
data,
it's
not
resolving
just
yet,
and
if
we
refresh
this,
let's
see
what
again
BAM
and
then
Duffy.
A
If
you're
still
with
us,
can
you
do
me
a
favor
just
cuz
there
might
be
a
concern
with
the
AWS
cloud
provider.
Can
you
check
and
see?
If
you
go
to
your
Amazon
console,
can
you
check
and
see
your
load
balancer?
If
you
go
to
ec2
and
then
scroll
all
the
way
down
on
the
left
to
load
balancers,
you
should
be
able
to
click
on
a
load
balancer
that
you
have
here
and
then
down
here
in
this
bottom
pane.
You
should
be
able
to
see
if
it's
healthy
or
not.
A
Can
you
verify
that
the
load
balancer
is
in
fact
healthy,
and
it
says
up
next
to
it
when
you're
looking
for
that
magic
word
QP.
That
would
be
great
if
you
could
help
out.
So
anyway,
let's
go
back
and
let's
now
install
there's
to
be
so
cool
that
worked
out
of
the
box
listens
to
all
the
cakes
dashboard.
A
So
Eric
says
I
thought
you
were
supposed
to
use
cname
records
instead
of
a
records.
2.90
lb,
see
that's
why
I
wasn't
wanting
Duffy
to
check
if
it
was
up
I
couldn't
remember
if
it
was
if
a
name
or
the
cname
worked.
So
we,
if
you're
suggesting
we
do
cname
we
can.
We
can
set
that
up.
I
haven't
really
done
this
in
a
while.
So
let's
see
what
we
got
going
on
here,
let's
go
back
to
route
53
2.
A
Eric
says:
yeah:
oh
I,
see
okay,
so
backing
up
josh
says
alias
the
a
record
in
New.
York
says:
I
believe
that
the
eye
piece
can
change
at
any
time.
So,
let's
keep
that
the
same,
and
then
we
want
to
do
a
record
set
here
and
see.
If
this
will
work,
I
am
NOT.
The
DNS
person
I,
don't
do
DNS
anymore.
Dns
is
hard.
Is
the
DNS?
It's
always
DNS?
A
Okay,
so
we
want
to
create.
Let's
see,
you
know
why,
but
we
can
just
Google
this
route,
53
hosted
zone
to
yell,
be
routing
traffic
to
an
e
lb
load.
Balancer
Amazon
rel,
53
prereqs
configure
to
route
traffic
to
an
e
lb
load
balancer.
If
you
created
a
route
53
zone
and
yell
be
load
balanced,
we're
using
the
same
account
skip
to
step
2,
alias
a
record
rule
will
work
tango.
A
A
So
it
would
be
really
helpful
to
be
like
full,
complete
sentence
so
that
I
understand
the
context,
because
sometimes
I'll
say
something
and
you'll
respond
to
me,
but
I've
already
moved
on,
because
my
brain
works
super
fast
like
that,
unfortunately,
so
if
you
could
be
a
little
more
explicit,
that
would
help
me
understand
in
which
context
I
you're,
exactly
speaking,
alias
checkbox,
got
it
Oh,
Hayley's
checkbox.
Thank
you,
wonderful,
create
the
record
set
cannot
be
saved
because
a
Lia
target
contains
an
invalid.
Without
you
enter
target
name,
you'll
be
classic
load,
balancers.
A
Let's
see,
click
and
then
select
radio
for
alias,
oh
I,
see
I,
think
I,
see
what
folks
are
saying.
Okay,
so
come
here,
select,
alias,
say:
record,
set,
enter
target,
name,
I,
guess
to
Jake
it's
calm.
Let's
try
that
one
and
see
what
happens
so
I
think
because
it's
not
the
ELB
is
not
in
my
AWS
account.
We
might
need
Duffy's
to
do
this
for
us,
but
regardless
we
should
still
be
able
to
hit
it
here.
We're
just
gonna
have
to
use
this
as
our
host
name
for
it
for
ingress
right.
A
Let's
see
what
folks
are
setting
okay,
so
Michael
says
glue,
TGI,
Kate's,
calm,
okay,
I'm,
all
about
doing
that.
That
seems
like
it
would
make
the
most
sense
to
me.
So
what
Michael
suggested
is
we
do
glue,
got
TGI,
Kate's,
calm,
we
paste
in
our
a
record
here
and
we
give
it
a
cname,
and
this
should
work
yeah.
So
now
we
can
use
glue,
TGI
Kate's
calm,
and
we
should
get
our
record
here
glued
to
gie
Kate's.
A
And
we
got
a
course
wait
for
a
TTL
on
this,
so
we'll
give
that
a
minute
to
do
it's
internet
magic
behind
the
scenes,
and
then
we
can
go
move
forward
with
installing,
via
the
dashboard
hats
off
to
Michael,
for
the
magic
tip.
Thanks
for
the
reminding
me
that
we
could
do
that,
that
was
really
really
helpful.
A
Thank
you
for
everyone
else,
for
helping
with
the
alias
stuff,
I
appreciate
it
as
well,
like
I
said
that
we're
doing
this
one
completely
on
the
fly
so
yellow
already.
So,
let's
go
back
to
we
can
close
this
blog.
We
can
close
envoy.
We
can
close.
The
github
will
keep
all
this
other
stuff
open
and
let's
go
here
and
let's
do
getting
started.
I
love!
This
note
that
they
put
in
here
never
trust
a
file
downloaded
from
the
internet.
A
If
we've
said
that
once
on
TV
I,
K,
we've
said
it
a
thousand
times,
and
we've
also
said
if
it's,
if
it's
we're
not
looking
at
yeah
Monty
gik,
we're
doing
it
wrong
that
week.
So
here
we
are
gonna,
be
doing
a
classic
Nova
and
Joe
download
and
look
at
yam
all
in
the
command
line
that
we
all
know
and
love
and
I
like
ice
mentioned
earlier.
I
don't
think
I
have
W
get
so
yeah.
Let's
do
it
grew.
Install
W
get
sorry
I'm
a
w
get
user.
A
A
Okay,
cool
installing
W,
good
dududu
awesome.
So
now
we
can
W
get
this
thing
here
and
now
we
can
I.
Think
I
have
my
Emacs.
Let's
see
if
we
have
my
Emacs
installed
on
here.
Oh,
we
just
have
regular
old,
vanilla
and
Emacs.
Okay,
that'll
work,
so
kind
deployment,
this
running
and
cube
system.
So
here
is
okay,
so
this
image
tells
us
that
it's
gonna
pull
from
that
docker
repo
docker
hub
repo.
We
were
just
looking
at
so
this
is
good.
A
A
Name:
Kate's
out
Kate's
there,
so
that's
how
that
mapping
between
the
service
matches
back
to
the
deployment
here.
So
this
looks
good
to
me.
I
think
this
is
go
ahead.
We
can
go
ahead
and
install
this
lets
que
apply
f,
kubernetes,
Kate's
da
animal
poof,
so
deployment
Kate's
and
service
Kate's.
So
now,
let's
check
out
our
ingress
and
see
what's
going
on
here,
okay,
get
in
all
namespaces
no
resource
is
found
good.
A
So
let's
go
back
to
our
Docs
and
it
says
to
access
Kate's.
You
must
make
it
publicly
visible
if
you
have
an
ingress
server
setup
which
we're
in
the
process
of
getting
this
up
and
running
with
glue,
you
can
accomplish
that
by
adding
a
route
like
the
following.
So
we're
gonna.
Do
this
we'll
copy
this
and
let's
do
can
our
terminal
emacs
we're
gonna
call
this
in
dot.
Yeah
mole
paste
this
in
here
look
my
ammo
is
all
messed
up.
Actually
wait!
That's
fine
right!.
A
A
Okay,
so
we
want
to
change
this
to
glue
dot,
T
gik,
it's
calm
and
then,
if
you
look
here
in
there
yeah
mol,
this
is
where
we
start
to
see
that-
and
this
is
the
kubernetes
ingress,
which
only
concerns
itself
with
HTTP
HTTPS
and
glue
itself
is
actually
more
advanced,
and
that
adds
some
Envoy
itself
and
you
could
do
more
with
that.
But
in
this
case
we're
gonna
be
using
glue
to
route
to
various
community
services.
A
So
we're
gonna
just
basically
change
our
host
glue,
TG
iCade
calm-
and
you
can
see
here
that
that's
just
going
to
be
on
the
root
path,
but
we
could
actually
come
through
and
define
various
other
paths
here
and
we
can
add
multiple
paths
under
the
path.
This
plural
declaration
here,
so
that
we
can
map
different
different
paths
to
different
service
names.
Remember
the
service
name:
kate's!
Wasn't
it
yeah?
No,
let
me
just
like
that
a
moment
ago.
So
that's
how
we're
gluing
glue
together.
A
Okay,
so
let's
save
this
and
kay
apply
f
in
Guiana
ingress
created
now,
let's
run
our
kay
get
in
command,
so
ing
is
of
course
short
for
ingress
and
then
we're
going
to
do
all
namespaces
and
you
can
see
here
and
keep
system.
We
have
glue,
TGI,
Kate's,
calm
and
now
is
when
we
get
to
start
debugging
or
if
we're
super
lucky,
it's
gonna
work
and
we
all
get
to
go
home,
we'll
see
nope.
It
doesn't
work
okay,
so
now
I
get
to
start
debugging,
which
I
know
is
everybody's
favorite
part
of
T
tik.
A
So
let's
see
what
we
have
going
on
here,
so
the
first
way
we
can
start
to
be
bugged
glue
is
we
want
to
get
the
glue,
ingress
controller
and
actually
see
if
any
happening,
wherever
we
hit,
if
we're
even
getting
traffic
to
our
cluster
or
not.
So,
let's
like
verify
that
we're
actually
hitting
the
cluster
and
then
we
can
diagnose
what's
going
on
with
the
cluster
first.
So
this
is
the
way
I
like
to
approach.
Debugging
networking
in
kubernetes
is
start
from
the
tippy
top
and
then
work
your
way
down.
A
Oh
Roy
said
you
removed
blank
from
path.
Isn't
it
required?
Oh,
it
is
required,
did
I
really
remove
that?
Let's
try!
Oh
thank
you.
Okay,
so
Roy
with
eyes
like
a
fox
Kay
applied
f
king
camel
okay,
so
we
reconfigured
that
and
let's
do
our
get
in
and
then
we'll
do.
Oh
yeah
mal,
just
to
verify
kate's
pass
is
set
to
root.
Okay.
So
that's
what
we
want
there
so
clear
my
screen,
let's
see,
did
that
work.
That
would
be
crazy.
If
that
was
it,
no
I
had
a
feeling.
A
It
was
gonna,
be
too
easy.
Okay,
as
I
still
think
something's
going
on
with
DNS.
That's
just
my
suspicion,
but
we're
gonna
start
to
keep
on
well.
Okay,
so
how
we
want
to
do
that?
Is
we
coming
here
to
a
terminal
kay,
get
queue,
Bechdel
get
Pio,
pods,
name
space,
glue
system
and
before
I,
move
on
I'm?
Actually
going
to
do
this
alias
will
do
kg
for
K
glue
is
equal
to
Q
Bechtel
in
glue
system.
A
A
So,
let's
see
Jorn
moanin
says
it
late,
belgium
hi
good
to
see
you
and
finances
right.
If
you
only
have
one
ingress
and
it
dives
in
what
the
glue
art
pitch
you
shown
has
multiple
envoys
wondered
if
you
know
what
that
setup
would
look
like
Rex,
don't
end
up
on
the
floor,
so
requirements
don't
end
up
on
the
floor.
Okay,
so
for
it,
Fernanda
I'll
take
a
stab
at
answering
that
a
second,
let's
finish,
debugging
this
and
get
it
up
and
running,
and
then
what?
A
At
the
end,
we'll
talk
about
sort
of
the
pros
and
cons
and
all
the
different
ways
we
could
configure
glue
or
contour
envoy
and
how
we
could
start
to
solve
the
multi
cluster
and
multi
molar
issue
and
we'll
talk
about
at
the
end.
Please
remind
me
in
fact,
I'm
gonna
write
a
note,
really
quick,
just
to
remind
myself
from
Nan's
question:
cuz,
that's
gonna
be
a
bit
of
it'd.
A
You
telling
answer
and
I
want
to
stay
focused
on
this
right
now,
if
that's,
okay,
okay,
so
if
this
worked
correctly,
we
should
be
able
to
come
here
and
do
a
refresh
or
even
more
importantly,
we
should
be
able
to
come
here
and
hit
this
thing
or
last
but
not
least,
we
should
be
able
to
come
and
hit
the
a
record
directly
which
again,
let's
do
a
recursive
search
for
SBC
grab.
This
thing
again
tailor
our
logs
again
put
some
white
space
in
there
and.
A
Empty
response,
so
we
should
at
least
see-
or
at
least
hope
to
see
that
something
would
be
happening
here
in
the
logs
denoting
a
failure.
Sorry,
equal
Rex
equals
requests
touch
day,
typing
sorry,
so
yeah
again
for
nan
I'll,
bring
up
different
requests
or
I'll
bring
up
your
requests
for
me
to
talk
about
multiple
requests
and
how
not
to
drop
them
in
a
moment
and
there's
some
interesting
patterns.
We
can
explore
that.
But
again
that's
gonna
be
a
bit
involved.
So
let's
get
this
thing
up
and
running
first.
A
A
Gws
node
cube
products,
e
ingress
proxy
discovery,
glue,
I'm
wondering
if
it's
this
glue
on
here.
So
let's
do
kg
get
P.
Oh,
let's
see
what
glue
itself
is
kg
logs,
minus
F
for
NAND
I.
Believe
multiple
envoys
on
the
picture
equals
multiple
instances
of
the
same
ingress
controller.
Yes,
I
think
yon
is
absolutely
correct.
A
Okay,
so,
let's
see
we
have
some
logs
here
looks
like
we
had
some
event
something
EDS
plugin
failed,
that's
not
good
I,
don't
know
what
ets
plugin
is,
but
we
should
be
able
to
come
here
and
resolve
this,
and
we
can
also
just
just
for
grins
here.
Http
colon,
slash,
slash,
interesting
yon
says
have
leave
now.
Thanks
for
the
episode,
yeah
have
a
good
good
one.
Yon
we're
gonna
keep
playing
with
this
whoa.
Where
are
you
an
hour
and
15
minutes
into
the
episode?
A
I
did
not
realize
it
had
gotten
so
late,
so
quickly,
ping
shows
pakka
lost,
sure
resolves,
probably
DNS,
client
and
Max
are
nutrias
for
caching
dns
answers
as
well
good
tips
again
we're
just
trying
to
stay
to
debug
here.
I
want
to
see
what
I'm
looking
for
now
is.
I
want
to
see
if
we
can
actually
find
the
pod.
A
That's
actually
running
the
server,
that's
supposed
to
be
forwarding
traffic
to
us
and
see
if
we
can
actually
demonstrate
that
it's
getting
traffic
does
it
require
the
path
and
the
requests
or
is
it
assumed
route
I
think
it
assumes
the
route
by
a
default.
So
if
you
don't
specify
it
should
just
go
straight
to
route
and
that's
sort
of
like
the
catch-all
or
the
asterisk
of
ingress
so
yeah,
let's
do
our
cage
Ekpo.
A
Maybe
ingress
proxy
kg
logs
English
proxy
e
minus
F
cluster
manager,
kate's
complete,
ok,
so
yeah,
ok.
So
this
looks
like
this
actually
detected
our
route
and
set
it
up
here.
So,
let's,
let's
just
start
out
our
services,
so
k
get
kg,
get
SVC
and
see
what
folks
in
the
chat
are
saying.
Fernandez
thanks,
Chris
no
deal
if
no
time
I
see
TX,
Jung,
okay,
so
yeah,
so
our
load
balancer.
Here
it's
called
ingress
proxy.
So
we
can
actually
do
kg,
get
SVC
ingress
proxy.
Oh
yeah
Mille!
We
can
see
here
it's
a
service.
A
A
Let's
try
to
dig
this
up
again:
Kate
glued
TG,
ie
Kate
stock,
okay
and
here's
our
C
name
again.
So
that's
happy
as
well.
So
let's
go
Keiji
get
p!
Oh
and
let's
do
our
service
one
more
time
and
let's
see
our
names
here,
her
label.
Sorry.
So
it's
going
to
eaten
rest
proxy.
So
that's
where
this
is
going
so
now
we
want
to
do
kg
get
Pio,
ingress
proxy.
We
already
looked
at
this,
but
let's
look
at
this
again
kg
logs
ingress
proxy
minus
F,
so
this
should
be
the
the
layer
of
software.
A
That's
actually
going
to
be
14
on
our
networking
traffic
and
basically
behaving
as
a
proxy,
which
would
make
sense
because
it's
called
ingress
proxy,
so
it
updated.
The
cluster
did
some
things.
This
looks
like
it's
configuring
envoy.
If
you
know
that
the
way
envoy
is
set
up,
it's
all
API
driven.
So
when
you
see
things
like
this,
this
to
me
is
a
very,
very
obvious
line
from
envoy.
So
this
is
the
thing:
that's
actually
an
abstraction
on
top
of
envoy
itself,
so
this
is
sort
of
the
meat
and
bones
of
glue
here.
A
So
we
really
should
be
I,
don't
know
if
Logging's
just
not
turned
on
or
what's
going
on
here,
but
we
should
be
able
to
see
it
up
and
running
and
I'm
really
frustrated
that
this
thing
is
not
wanting
to
route
traffic
for
us
interesting.
So,
let's
do
K
getting
all
namespaces
and
let's
just
check
this
one
more
time:
ports,
80,
cube
system
kate's,
so
that
makes
sense
glue.
Tgi
Kate's,
calm
I
mean
so
we
know
the
traffic
is
going
it's
resolved
incorrectly.
A
She
says:
Cates
ash
service
end
points
we'll
check
in
in
points
in
a
second
Ashish.
So
we
know
the
traffic
is
getting
routed
correctly
via
DNS
to
Glu
TGI
Kate's
calm.
Let's
just
sanity
check
that
really
quick
glued
TGI,
Kate's,
calm,
14
to
art
a
record
here.
Do
we
need
to
do
alias?
No,
let's
just
do
this,
and
actually
you
know
we
can
just
refresh
a
page
to
cancel
that.
Ok,
so
we
know
that's
happy.
A
We
know,
what's
going
to
a
be
seven
four
three,
and
if
we
dig
again
a
be
seven
four
three
here,
so
that's
happy
so
something's
going
on
and
it's
not
wanting
to.
Are
you
sure
the
EOB
is
routing
traffic
to
the
cluster
Erik?
This
is
what
I
was
really
hoping.
Duffy
was
gonna
help
us
with
if
he
could
check
out
the
load
balancer
for
us,
but
I.
A
There
was
that
bug
I
mentioned
earlier
that
sometimes
in
Yale
B
will
be
created
in
the
wrong
availability
zone
and
won't
be
able
to
communicate
to
the
instance
that
he
needs
to
communicate
with,
and
that
very
well
could
be
what's
happening
here
and
that's
just
a
concern
with
running
any
kubernetes
and
amazon
and
there's
some
ways
you
can
go
about
ensuring
that
but
requires
a
little
bit
more
configuration
to
prevent
that
from
happening.
So
if
Duffy
last
call
I'm
gonna
do
a
shout
out
here.
In
fact,
I'll
pin
you
on
slack,
hey
Duffy.
A
A
Do
you
do
and
let's
go
in
here
and
see
all
of
our
kubernetes
resources
one
more
time,
my
oh
I
guess
Nicolas
isn't
gonna,
say
anything,
but
either
way
we
can
keep
looking
around
and
checking
and
seeing
what
glue
is
doing
so
he
Oh
Nicolas,
oh
ok,
so
Nicolas
says
in
our
slack,
which
you
can't
see
right
now,
I
had
it
on
a
different
screen.
He
said
he's
on
in
the
mention
to
me,
so
if
Duffy
is
checking
it
out
now,
so
hopefully
you
will
we'll
find
out.
A
What's
going
on
with
our
ELB
and
nicolas
is
typing
again.
Thank
you
so
much
folks.
This
is
why
we
there's
a
whole
crew
of
people
that
make
TDI
k
possible
really
every
every
week
we
have
like
a
whole
line
of
documentation
that
we
go
through.
So
it's
really
a
lot
of
work
here.
Okay,
so
Nicolas
wants
to
know
which
LP,
so
it
should
be
the
one
that's
created
for
his
cluster,
the
e
lb
name,
I.
Think
I
can
get
here.
K,
get
SVC
name
space
or
will
do
all
namespaces.
Oh.
A
Yeah,
no
ok!
So,
let's
see
so
load
balancer
kubernetes
cube
system.
There's
the
metric
server.
Let's
get
rid
of,
though
yeah
Mel
we're
trying
to
find
the
name
of
this
lb.
It's
going
to
be
the
one
that
it's
going
to
be
the
only
one
created
that
has
this
a
record
attached
to
it.
Nicholas
so
I'm
gonna
put
this
in
chat
here.
It
has.
A
Ok,
perfect,
so
Duffy
comes
back
and
says:
what's
up
zero,
five
instances
in
service
aha,
so
we
found
the
problem,
which
is
unfortunate,
might
need
a
security
group
rule
to
a
lot
traffic,
a
lot
of
traffic
between
ELB
I.
Have
it
and
it's
not
healthy,
ok,
so
Duffy
I
think
what
Eric
suggested
here
is
precisely
what
needs
to
happen
in
the
security
group.
We
need
to
poke
a
hole
in
the
firewall
on
port
80
between
the
ELB
and
whatever
security
group.
The
instances
are
running
in
if
you
could
check
that
that
should
do
it.
A
I
think
Nicholas
swing
says
got
it
I'm,
just
like
the
broadcaster
here,
Nicholas
has
got
it.
I've
been
sooo
mean
that
means
that
they
were
able
to
poke
a
hole
in
the
firewall
for
us,
which
is
totally
happy
pandas
and
let's
just
like
I'm,
so
excited
I,
can't
wait,
let's
see,
maybe
here
so
even
after
they
opened
the
hole
in
the
firewall.
Aws
was
eventually
consistent,
so
it's
gonna
like
take
some
time
for
it
to
do
its
thing.
So
why
we're
here?
Why
are
you
here
we're
taking
a
break?
A
We're
gonna,
wait
for
it
for
them
to
get
the
firewall
stuff
figure
it
out.
I
can
answer
your
question
now
for
nan
that
you
had
earlier
and
I
can
I
can
take
any
other
questions.
That
folks
have
questions
so
far,
we'll
try
to
get
this
up
and
running,
and
then
we
can
close
out
the
episode,
but
if
you
have
any
questions
feel
free
to
drop
them
in
there.
Okay,
hep
tio
says
on
looking
into
the
firewall
thing
now.
One
second
take
your
time:
Duffy
I'm
gonna
rant
for
a
bit
and
answer
finance
question
here.
A
A
Fernan
hi
crust:
do
you
know
how
one
would
set
up
multiple
ingress
controllers,
ie
serving
TGI
Kate's
comm
using
multiple
ingresses,
and
how
would
failure
recover?
You
look
like
when
a
controller
dies
so
to
be
clear
in
this
example.
Make
sure
we
get
everything
here.
In
this
example,
we
only
have
one
ingress
controller,
but
we
have
multiple
ingress
reps.
So
the
universe
controller
is
down
here
in
this
part
of
the
diagram
and
just
reconciles
all
of
these
different
yamo
files,
and
all
of
these
up
here
are
just
different
ingress
routes.
A
So
in
this
example,
were
only
using
one
ingress
controller
and
if
that's
controller
did
go
down,
we
would
run
into
some
trouble,
which
is
why
we
run
our
universe
controller
as
a
deployment
and
in
some
cases
we
want
to
actually
load
balanced
across
our
load
balancers
so
that
we
have
that
double
redundancy
as
well
looks:
okay,
so
Nikolas
just
messaged
me
and
shot
and
says:
ok,
they're
fixing
the
firewall
issue.
So
that
sounds
good
anyway.
Fernand.
Does
that
answer
your
question
about?
A
What's
going
on
up
here
and
the
differences
between
the
routes
and
the
actual
controller
itself
sample
here,
let's
see
what's
going
on
Oh
beautiful.
Thank
you
look
what
we
have
here.
So
let's
do
a
refresh.
Let's
see
what
suresh
says
or
michael
says:
ok,
I'm
doing
the
same
on
gke
and
I
had
to
add
annotation.
Ok,
so
Michael
here
brought
up
a
good
thing.
He
says
he
had
to
have
the
annotation
kubernetes
slash
ingress
class
glue.
It
is
possible
to
expose
the
app
on
a
different
port
rather
than
80
or
443
using
glue
yeah.
A
You
can
do
some
port
for
Dean's
type
stuff
with
kubernetes,
but
that's
a
little
bit
out
of
scope
with
this
episode,
but
we
can
talk
about
that.
One
later
I
want
to
dig
into
what
Michael
saying
about
this
annotation
Michael.
Where
did
you
add
the
and
which
object?
Did
you
have
that
annotation
to
exactly
coop
nineties,
dot,
io,
ingress
class
cool,
so
I'm,
assuming
he
had
to
add
that
too?
Let's
pull
up
our
terminal
here.
Well,
we're
gonna
get
two
things
fixed
at
once:
ok,
so,
and
let's
go
and
K
get
P.
A
A
A
Namespace
cube
system
educates
that
should
do
it
Oh
deploy
gates,
so
we
gave
it
the
name
another
type,
so
the
type
is
deployment
or
deploy
for
short
on
the
service
rush
says
on
the
ingress
object
on
which
ingress
object,
so
our
use,
I,
guess
I
would
need
more
context
on
why
we're
even
adding
this
annotation
in
the
first
place,
which
u.s.
object?
Are
you
talking
about
the
one
for
the
ingress
controller,
the
one
for
the
ingress
itself
again?
A
A
I
was
gonna
edit
this
resource,
but
now
they're
saying
we
want
to
edit
the
ingress
which,
what
for
like
the
two
questions
are
one:
where
did
we
discover
about
this
mysterious
annotation
so
that
I
can
get
the
full
context
here,
instead
of
just
asking
you
for
this
magic,
a
notation
that
I
have
no
idea
about
and
then
second
of
all
which
specific
ingress
isn't
on
so
Michael
says
for
case
okay.
So
let's
try
this
so
we're
gonna
do
K
getting
all
namespaces
cube
system
gates,
okay,
so
let's
edit
this
one
so
K
edit
ingress.
A
Gates
not
not
found
kan
namespace
cube
system,
Oh,
Kate,
interesting,
okay,
so
here
in
the
annotations,
we
want
to
add
this
annotation
here:
okay,
so
yeah,
so
Michael
I'll
show
yours,
says
in
there:
okay.
So
let's
back
up-
and
let's
talk
about
what's
going
on
here-
because
it's
got
a
little
bit
out
of
a
little
bit
confusing
for
folks,
okay,
so
do
cunning
glue
all
of
this
back
together,
we
have
done
two
things.
A
Number
one
Duffy
and
Nicholas
are
kind
enough
to
check
out
the
the
the
firewall
configuration
in
the
AWS
account
on
their
end,
so
they
should
be
fixing
that
for
us
now
number
two.
The
other
thing
that
we're
also
working
on
is
the
secondary
issue,
which
can
be
described
here
in
basic
ingress.
So
I'm
gonna
drop
this
in
our
doc
here
and
we're
just
going
to
call
this
weird
and
herbs
gonna
call
this
the
annotation
documentation
so
that
folks
at
home,
if
you're
following
along,
you
can
see
this
as
well,
so
there's
basic
ingress.
A
So
this
explains
where
Michael
got
that
from-
and
this
also
explains,
probably
one
of
the
many
reasons
why
glue
hasn't
been
resolving
for
us
yet
so
hopefully
that
kind
of
paves
the
big
picture
for
folks.
There
is
a
tutorial
here
that
we
skipped
over
and
then
we're
going
back
now
and
adding
our
annotation
here
as
well.
Thank
you
so
much
for
your
help
and,
let's
add
it
here:
kubernetes
dot,
io
ingress
class
glue.
So
that
looks
happy.
So
you
should
be
able
to
save
that.
It
was
edited.
A
Let's
see,
okay,
let's
go
back
here
and
let's
go
here
still
nothing
still!
Nothing!
That's!
Okay,
I'm
hoping
that
there
is
they're
just
working
on
the
the
ELB
right
now
which,
by
the
way
duffy
or
nicolas,
if
either
of
you,
could
give
us
an
update
on
the
ELB.
That
would
be
super
handy
dandy,
but
if
not
it's
totally
okay.
A
So
let's
go
back
and
let's
look
at
this-
this
ingress
routing
here,
so
it
says
we
need
to
create
an
ingress
object.
Here's
an
example:
it
says
path
is
slash.
Dot
star
for
nancy
is
awesome.
Iran
thanks
Chris,
makes
sense.
Now
I
was
confused
with
traffic
suresh
says:
there's
no
default
ingress
controller
in
case
that's
needed
to
have
this
imitation.
In
my
opinion,
Thank
You
Suresh.
We
might
want
to
steal
this
little
sin
taxi
bit
here
and
add
that
to
our
our
ingress.
A
Let's
try
that
505
in
service
woo,
awesome,
okay,
Thank,
You
defi
and
thank
you
Nicholas,
and
so
now.
Why
we're
here,
let's
edit
our
ingress
again
and
we're
gonna,
do
we're
gonna,
go
all
the
way
down
here
to
the
bottom
two
paths
and
we're
gonna
paste
in
that
magical
line
there
and
we're
gonna
close
that
and
now
we're
going
to
come
up
here
to
glue.
A
Yes,
it
works.
Sorry
I
get
it
very
excited:
okay,
so
yeah
a
whole
team
of
people
working
together
to
get
glue,
TGI,
Kate's,
calm
and
it
says,
enter
your
auth
token
here
so
yeah.
Now
everybody
on
the
internet,
it's
gonna
be
blowing
up
glued
at
TGI,
Kate's,
calm
and
our
AWS
account.
Please
be
nice
and
don't
do
what
I
did
one
time
to
Joe
and
shut
your
company's
internet
down
by
spamming,
the
TGI
Kate's
broadcasting
stream
with
all
kinds
of
traffic
which
that
was
a
really
fun
episode
as
well.
Okay.
A
So,
let's
back
up
and
explain
what's
going
on
here,
so
we
created
a
route
53
DNS
record.
Here
we
created
a
cname
record
and
we
gave
it
the
prefix
glue,
TTI
Kate's
calm
and
we
wanted
to
create
a
cname
so
that
we
could
point
it
directly
to
the
a
record
for
the
ELB.
The
elastic
load
balancer
in
kubernetes
so
after
we
had
that
we
were
then
resolving
public
traffic,
but
the
ELB
that
was
created
in
an
AWS
account
that
I
don't
have
access
to
was
not
resolving
to
the
traffic
to
the
instances.
A
It
was
supposed
to
be
used
a
token
from
your
cube
config.
Thank
you
very
much
Duffy,
and
so
anyway.
I'm
assuming
and
Duffy
correct
me
if
I'm
wrong,
but
it
folks
at
home
were
probably
a
bit
curious
about
what
was
going
on
with
the
ELP
but
I'm,
assuming
it
was
either
the
security
group
or
it
was
in
the
wrong
availability
zone.
Any
any
of
those
would
be
helpful,
so
Roy
says
it
supports
certain
integer,
okay,
so
Ryan.
A
If
there's
a
zone
question
so
then
after
we
got
the
ill
be
up
and
running,
we
were
then
traffic,
so
then
route
53
kicked
in.
Then
we
passed
the
traffic
off
to
let's
go
here
back
in
the
terminal.
We
passed
it
off
to
take
it
P,
o
name
or
we
can
do
kg,
get
P
o.
We
passed
it
off
to
the
ingress
proxy
in
this
namespace.
Ok
get
logs
bla
minus
F,
and
this
is
the
actual
abstraction
layer
on
top
of
envoy
that
is
glue
itself
and
now
we're
resolving
traffic
to
the
kubernetes
dashboard.
A
So
we
want
to
get
the
token
from
our
cube,
config
config,
then
I
suspect
your
annotation
also
helps
so
yeah
I
think
this
is
totally
an
example
of
like
one.
It's
ok,
kubernetes
is
hard
for
everyone,
but
that's
why
we
do
these
TTI
cases,
so
we
can
learn
together
and
we
learn
more
about
kubernetes
along
the
way
and
then
so.
We
can
show
folks
exactly
how
we
are
able
to
solve
a
lot
of
these
issues
and
hopefully
speed
things
up
for
both
people
internally
and
for
all
of
our
followers
at
home
and
across
the
world.
A
So,
let's
see
what
we
have
now,
so
we
want
to
grab
the
token
from
my
cube
config.
So
I
want
to
do
that
by
creating
in
a
new
tab
and
getting
that
off.
The
screen
and
I
want
to
come
over
here
and
I
want
to
do.
Okay,
I'm,
just
gonna
top
type
out
what
I'm
typing,
even
though
I'm
doing
it
off
the
screen
here.
I
want
to
do
a
cat
home
directory
q,
/
config
and
we're
gonna
copy.
My
token
here,
BAM
minimize
this.
Let's
see
what
folks
are
saying
in
chat.
A
Let
me
just
really
quick
say
thank
you
to
Nikolas
and
everyone
else,
and
let's
go
in
here
to
deflect
or
not
the
slack
I'm,
sorry
to
the
dashboard
here
and
let's
paste
our
token
and
see
what
happens:
never
not
now,
okay
cool.
So
we
have
the
new
kubernetes
dashboard
up
and
running
in
kubernetes,
with
glue
as
an
ingress
controller,
serving
traffic
distributed
across
our
deployments
all
running
in
amazon,
and
it
only
took
us
an
hour
and
a
half
so
not
so
bad.
But
again.
A
This
is
why
I
usually
like
to
go
through
things
before
the
episode,
but
I
also
really
like
the
episodes
to
where
we
get
to
do
some
live
debugging
together
as
a
team,
because
it's
always
hard
to
see
how
the
stuff
comes
together
and
to
work
together
as
a
team.
So,
anyway,
let's
check
out
the
bored
and
see
what
we
got
going
on
here.
So
we
can
see,
we've
got
five
nodes,
which
is
cool.
We
have
18
pods.
So
this
is
like
the
cluster
overview
like
what's
going
on
in
my
cluster.
A
Oh,
this
is
really
cool,
so
these
are
like
the
oddity
logs
right.
So
we
can
see
like
things
that
have
happened.
Somebody
pulled
the
cube
system,
kate's,
which
this
was
me
and
then
you
can
see
we
created
a
container
called
Kate
again.
This
was
me,
so
this
is
like
effectively
a
record
of
everything
that
I've
been
doing
on
this
episode
of
tea
gik
to
get
glue
up
and
running.
A
So
this
is
really
cool
and
if
we
wanted
to
come
through
and
actually
see
some
of
the
steps
we
did
earlier,
we
can
see
like
here's,
where
we
created
the
glue
system,
ingress
and
so
on
and
so
forth.
So
that's
exciting
we're
getting
to
see
a
bit
of
history
in
our
cluster.
Let's
go
over
here
to
the
left
and
let's
click
on
nodes,
ooh,
I,
love,
graphs,
we'll
see
what
folks
are
saying
in
the
chat:
victory
awesome,
debugging
thanks
everyone.
We
really
appreciate
it.
A
Roy
made
his
comment
about
certain
manager,
which
will
look
at
TLS
in
a
moment.
I
want
to
go
through
this
dashboard,
really
quick
and
just
see
how
this
thing
totally
works
and
see
what
we
have
going
on
in
here.
Okay,
so
we
it'll
show
us
our
labels
and
our
nodes.
It'll
look
at
top
things,
so
we
can
see
if
it's
ready,
which
this
is
an
important
part
of
like
what
the
cluster
API
does.
A
Is
it
what
we
always
want
our
nodes
to
be
ready,
we're
using
six
point,
seven
percent
of
our
compute,
only
5%
of
our
ram
and
we're
going
to
get
to
see
like
oh
look
at
this
node
over
here
this
one's
actually
working
harder
than
either
others.
I
dentists
is
doing
the
metric
server
or
something
exciting.
So
we
can
come
in
here
to
namespaces.
Here
we
have
the
default
namespace.
We
have
cube
system.
Q
public,
as
we
all
know,
is
a
newer
one
and
glue
system.
A
The
one
we
just
created,
let's
check
out
our
workloads,
cube
X,
will
get
events.
The
graphs
are
brought
to
you
by
the
integration
with
the
metric
server.
Yes,
so
this
is
why
folks
were
saying
it
looks
way,
cool
or
if
you
have
the
metric
server
up
and
running
so
this
is
really
exciting
that
we
can
come
through
and
see
all
these
graphs
and
get
all
of
this
wonderful
mud.
A
All
these
wonderful
metrics,
because
the
metric
server
standardized
is
that
for
us
once
it's
up
and
running
and
we
can
just
gobble
those
up
and
just
regurgitate
them
on
the
page,
which
is
what
this
new
and
improved
dashboard
is
doing
here.
So
yeah
I'm
not
gonna,
spend
too
much
time
you
can
come,
you
can
see,
we
have
services.
Well,
you
can
see
our
replica
sets
pods,
here's
all
our
by
various
ingresses,
and
this
will
tell
you
the
path
that
the
ingress
is
running
on.
So
this
just
gives
you
like
everything.
A
I
was
doing
here,
pulled
up
all
of
the
command
line,
hacking
stuff
I
was
doing
here.
We
can
now
get
all
of
that
same
information
here
from
the
dashboard.
It's
a
lot
easier
for
us
and
we
can
come
through
and
we
can
like
copy
and
paste
stuff
directly
out
of
the
dashboard.
We
can
see
our
various
configs.
We
can
look
if
we
have
any
storage,
we
can
even
create
various
profiles
and
it
counts
and
here's
all
of
our
backe
stuff
as
well.
So
can
we
actually
edit
this
stuff?
A
Oh
you
can
okay,
so
this
is
really
cool.
This
is
like
an
inline
editor,
so
you
can
really
come
through
and
edit
and
configure
and
play
with
your
cluster
here
using
this
open
source
dashboard.
So
this
is
really
cool
and
glue
is
bringing
all
of
this
in
grassy
stuff
for
us
here.
So
this
is
quite
exciting.
Okay,
so
real,
quick
right
at
the
end,
let's
go
back
to
glue
then
to
dumb
and
I
bet.
If
we
go
to
user
guides,
is
there
setting
up
TLS
BAM?
A
Let's
just
see
what
this
is
here,
so
you
would
you
glue
octo,
get
upstream
default
pet
store.
You
would
add
a
route
you
can
use
virtual
service
okay,
so
this
is.
This
is
gonna,
be
a
lot
to
setup
in
this
episode.
Maybe
we
can
do
a
secondary
episode
of
folks
want
to
see
it,
but
it
looks
like
you
can
totally
come
to
and
set
up
TLS,
so
you
can
get
TLS
running
with
glue
as
well.
Michael
says
very
clean
dashboard.
A
So
anyway,
it's
been
a
great
episode
again
shout
out
to
everybody
in
kubernetes
who
make
all
of
this
possible
all
the
hard
work
that
went
into
the
dashboard
glue
itself
as
a
server
shout-out
to
Nicholas
and
all
the
folks
in
the
Bellevue
office
for
helping
me
out
today,
y'all
really
saved
the
day.
Thank
you
to
Ashish
and
Michael,
and
a
handful
of
other
folks
who
helped
us
find
me
a
notation
we
needed
to
add
and
helped
us
with
the
cname.
Earlier.
It's
been
a
really
fun
episode.
A
Well,
let's
see
what
access
just
to
be
clear,
the
dashboard
isn't
the
new
kubernetes
dashboard.
This
is
a
project
I've
been
working
on
for
a
couple
of
weeks,
so
I
love
to
hurt
to
become
the
new
dashboard.
Sorry,
that's
what
I
was
implying
Eric
when
I
said
the
new
dashboard,
so
this
is
I'm.
Gonna
go
on
a
bit
of
a
tangent
here.
Why
were
at
the
end
of
the
episode,
because
it's
Friday
afternoon
and
I'll
have
my
second
sip
of
beer.
So
what
Eric
is
saying
is
this
is
a
project
he
built
on
his
own.
A
So
it's
not
like
official
remember
earlier,
we
looked
at
the
kubernetes
core
scope.
What
isn't
out
of
scope?
So
all
Eric
is
saying
is
like
he
built
this
and
he
hopes
folks
can
use
it.
It's
really
cool
and
we
really
like
it
as
well,
but
this
button
by
no
means
when
I
say
the
new
kubernetes
dashboard
does
the
word
V
imply
anything
official.
This
is
just
like
a
project
right,
just
like
Cuba
Korn
has
nothing
to
do
with
kubernetes.
It's
just
a
project.
I
created
in
my
free
time
could
well
free
time.
A
It's
a
project
that
I
obsessed
with
like
over
a
year,
but
it
wasn't
officially
a
part
of
the
kubernetes
core
or
the
kubernetes
sig
code
responsibility.
So
anyway,
just
clarifying
that
for
Eric
Suresh
says
thanks.
Chris
joins
us.
Thank
you
for
teaching
us
Zolt
says
awesome.
As
always.
Thanks
for
nan
says.
Thank
you.
Chris
said
deep.
Thank
you.
Chris
y'all
I
really
do
love
doing
T,
GI
k,
I,
really,
love
hanging
out;
I,
love,
hacking,
I,
love
that
folks
come
and
get
involved
and
help
me
out.
It's
always
a
ton
of
fun.
A
Thanks
for
making
this
so
successful.
If
you're
going
to
be
a
cube,
con
in
Barcelona
come
check
out
my
name
Jose
talk
we
were
talking
about
earlier
today.
We're
really
excited
we're.
Gonna
have
a
few
surprises
for
you
now
that
we
review,
where
we're
gonna
be
growing
tea,
tik
a
lot
more
and
really
excited
about
what
the
future
of
tgia
is
gonna
look
like
and
how
we're
gonna
be
able
to
make
this
content
go
even
further,
make
it
even
more
awesome
for
folks
as
we
grow
and
as
we
become
this
new
great
entity.
A
So
we're
really
excited
about
that.
So
Martin
says
great
walk
through
Chris
Bjorn
says
thanks:
I'm
gonna
watch
the
whole
recording
tomorrow
have
a
nice
weekend.
Ok,
so
last
chance
questions
for
me,
questions
for
ingress.
Anything
else,
folks
want
to
say
I'm
gonna
grab
a
sip
of
beer
and
it
looks
like
while,
since
we
started
the
episode,
Mount
Rainier
aka,
Mount
Charlie,
a
k-8
Homa
has
decided
to
come
out
and
so
now
I'm
live
on
the
air
staring
at
my
favorite
mountain
with
all
of
my
favorite
people
drinking
a
horrible
beer.
A
So
anyway,
thank
you
from
Michael
and
as
she
says,
thanks
for
the
episode,
okay,
well,
happy
Friday!
Everyone
enjoy
your
weekend
pack
yourself
on
the
back.
I'm
sure
you
all
did
really
great
hard
work
this
week,
keep
it
up
and
we
will
see
everyone
next
week
with
another
exciting
episode
of
TGI,
K
I,
think
Joe
is
gonna,
be
doing
it
so
we'll
make
sure
to
have
a
good
time
with
Joe
and
make
him
feel
welcome,
because
we
don't
get
to
see
him
very
much
anymore.