►
Description
Starts at 04:00. This week I chat with Betty Junod of Solo.io on #microservices, #servicemesh, #serverless, and more. Learn more about them at www.solo.io. Every week I host a YouTube Live to take your #container, #docker, #kubernetes, #swarm, and #devops questions.
You can also check out my Docker courses and free resources at https://www.bretfisher.com/docker
Listen to my new podcast of this show's past episodes at https://www.bretfisher.com/podcast
Signup for my newsletter to get short emails when I launch new things (videos, podcasts, open source, etc.) https://www.bretfisher.com/newsletter
www.solo.io/events
Meet the Solo.io team at an upcoming event or webinar and learn more about API Gateways and Service Mesh in talks, demos and workshops.
https://servicemeshhub.io
Try/evaluate service meshes with Service Mesh Hub to make installation easy and try some operations with a demo app.
https://www.solo.io/glooe
Try modernizing legacy apps with new features added as microservices and functions and create a "hybrid app" with Gloo Gateway.
https://slack.solo.io
Join the community to ask questions, start a discussion and compare notes with each other on APIs, Microservices and Service Mesh.
A
Welcome
to
the
show,
thank
you
so
much
for
being
here
on
YouTube
live
I'm,
Brett
and
your
host
today
and
we're
gonna
be
talking
about
service,
mesh,
kubernetes,
container,
orchestration,
everything,
docker
and
DevOps,
and
I'm
glad
you're
here.
So,
if
you're
not
watching
this
live
sorry,
but
you
can
not
get
into
the
chat.
But
if
you're
here
live,
you
get
to
ask
your
questions
and
I'll
introduce
our
guest
today
in
a
second
and
you
get
to
sort
of
have
a
conversation
with
us.
A
That'll,
give
you
a
notification
when
I
go
live,
you
can
come
back
every
week,
ask
questions
about
containers
and
everything
DevOps
and
now
we're
gonna
get
to
the
fun
part.
The
fun
part
is
introducing
my
good
friend
Betty
to
node,
who
you
may
know
from
docker,
but
now
she's
heading
up
a
great
company
at
solo,
IO,
she's
VP
of
Marketing
there,
and
let
me
just
tell
you
a
little
bit
about
Betty
Betty
spent
over
four
years
of
docker
running
product
and
partner
marketing.
A
During
that
time
she
helped
so
many
people,
including
myself,
and
launching
a
laundry
list
of
docker
product
and
projects
over
that
time.
She
also
helped
to
cultivate
the
dr.
captain's,
Thank
You
Betty
and
which
is
where
we
got
to
know
each
other
and
basically
had
a
lot
of
fun
at
conferences
and
working
on
docker
cool,
Duggar
stuff.
For
you,
the
audience
on
the
Internet
welcome
to
the
show
Betty
Thank.
A
So
glad
we've
been
talking
about
doing
this
for
awhile
and
I'm
glad
we
finally
made
the
time
to
do
it,
and
so
for
those
of
you
that
don't
know
Betty
if
you've
ever
seen
me
do
a
selfie
on
the
internet
with
docker,
if
you've
ever
seen
any
of
that
on
my
Twitter
stream.
That
is
me
trying
to
be
one
tenth
as
cool
as
Betty.
She.
C
C
A
Are
cool
I
want
one
so
yeah?
That's
it's
always
a
fun
icebreaker,
I!
Think
for
people
and
now
what
happens?
And
you
know
this
happened
with
you
for
years
and
years
was
you
know
people
would
show
up
and
you're
you're
now
like
they're,
coming
up
to
you
for
the
selfie?
Isn't
it
that's
the
goal
right
and
yeah.
B
A
So,
but
now
what
you're
focused
on
is
sort
of
the
next
shift
or
wave
in
container
problems
and
solutions
right.
So
you
join
solo
I/o
this
year.
We
we
knew
about
them
for
a
while
as
the
docker
captains,
because
docker
has
been
talking
about
this
as
sort
of
one
of
the
next
big
problems
to
enhancing
our
container
orchestration
and
fixing
a
lot
of
the
the
problems
that
we've
created
by
putting
so
many
different
containers,
all
in
a
bunch
of
servers,
so
maybe
starting
off
what?
What
is?
B
And
you
know
for
those
of
us,
like
you,
me,
who've,
been
in
the
container
space
for
a
while
kind
of
since
the
beginning.
It's
interesting
that
you
know
we
feel
like.
Oh,
it's
been
around
for
so
long
we're
so
deep
in
it
like
this
is
the
next
thing
service,
mesh
and
all
these
other
areas,
but
when
you
think
about
it,
docker
just
celebrated
its
sixth
birthday.
Kubernetes
is
celebrating
his
fifth
birthday,
so
they're
the
folks
that
have
been
deep
in
this
kind
of
cloud
native
and
micro
services
world
for
a
long
time.
B
But
you
know
there's
everybody
else.
That's
not
looking
at
kind
of
like
this
is
the
next
stage
to
do
this
and
I
was
doing
a
little
thinking
about
this.
You
know
and
part
of
my
reason,
reasoning
for
joining
solo
as
well
as
that,
when
I
even
look
at
my
own
career,
you
know
I
spent
a
long
time
at
VMware
and
then
I
went
to
doctor
and
now
I'm
here
and
so
in
a
nutshell.
The
statement
is
I
really
like
abstractions.
B
And
when
you
think
about
the
abstractions,
it's
it's
interesting,
because
it'll
kind
of
explain
why
mesh
is
a
thing
we
need.
We
need
to
look
at
honestly
today
is
that
when
we
were
when
we
first
started
with,
like
let's
say
VMware
in
virtualization,
we
had
a
high.
We
had
an
application
architecture.
That
was,
you
know,
monolithic
and
it
was
highly
dependent
right,
there's
a
high
level
of
dependencies
and
it
started
out
being
tied
to
you
know
you
have
like
the
physical
box
and
then
the
operating
system
and
then
the
applications.
B
It
then
said:
oh
what
about
this?
How
do
I
then
think
about
things
like
storage
or
security?
All
those
same
all
those
questions,
and
what
was
interesting
is
then
you
know
it
VMware's
been
around
for
20
years
and
over
those
20
years
vmware
and
there
that
ecosystem
has
been
solving
a
lot
of
those
problems
and
doing
more
interesting
things
right.
Once
things
are
separated,
you
can
do
interesting
things
fast-forward
to
the
you
know
about
five
years,
I
spent
at
docker.
B
We
just
what
we
did
was
create
yet
another
separation,
but
just
at
a
different
layer
right
and
then
developers
rejoiced
and
then
the
lot
of
the
architects
and
the
IT
kind
of
sysadmin
operator
said.
Oh,
my
gosh.
Now
what?
Because
it
changed
how
it
because
it
changes
how
they
you
know
how
they
were
orchestrating
those
applications
before
you
went
from
like
one.
You
know,
you
know
one
code
base
right
that
was
kind
of
deployed
as
one
thing
challenges
that
you
know
there
were
pluses
and
minuses
of
those.
B
A
B
And
the
thing
is
it's
like
now:
we
have
this:
we
have
honey.
You
know
any
number
of
containers
as
you
want.
There
is
no
right
magic
number,
it's
whatever
the
application
team
wants
and
then
the
idea
is
that,
like
you,
can
now
kind
of
scale
up
and
down
and
change
those
things
all
independently,
like
that's
the
beauty
of
it.
So
that's
and
then,
when
you
look
at
that
it,
you
know,
as
we've
all
seen
in
kind
of
all
the
questions
in
the
ecosystem.
B
What
does
that
mean
for
things
like
state
and
storage
and
networking
when
it
used
to
be
like
I?
Have
a
physical
thing
here
and
a
physical
thing
to
here
I
can
just
build
it
in
to
my
business
logic
like
they.
Those
two
are
the
only
things
that
talk
to
each
other.
Those
two
things
are
now
like
a
hundred
things
that
are
sometimes
on
sometimes
off
changing
all
the
time
who
talks
to
what
and
that's
really
where
service
mesh
is
born.
B
A
B
A
It
started
the
the
problem
started
showing
up
for
me
when
suddenly
this
thing
that
was
normally
on
one
server
and
I'm
going
to
use
a
simple
example
like
a
PHP
in
a
patchy
app
or
something-
and
it
was
a
very
large
monolithic,
app
and
it
was
on.
We
had
it
on
a
couple
of
servers,
but
for
redundancy,
but
it
wasn't.
It
was
all
talking
to
itself
right
and
to
troubleshoot
that
was
largely
about
talking
to
you
know.
A
The
development
team
enable
the
debug
flag
in
that
in
the
app,
so
it
would
just
log
everything
and
then
eventually,
as
those
kind
of
apps,
became
more
distributed
and
more
focused
on
micro
service
architecture
type.
Suddenly,
I
had
to
care
so
much
more
about
network
tools.
You
know
because
now,
I
I
couldn't
just
go
debug
the
app
or
load
up
some
sort
of
performance
debugger
in
the
app
to
figure
out
what
was
my
performance
problem?
A
Why
was
this
thing
not
working
and
now
suddenly
I'm,
looking
at
all
sorts
of
system
level
tools
at
the
Linux
level
and
at
the
networking
level
and
I'm
trying
to
figure
out
how
to
plug
into
firewalls?
You
know:
is
this
port
even
open
and
getting
there
and
all
these
sort
of
networking
things,
and
on
this
show
we
talk
a
lot
about
actually
a
sort
of
a
tool
that
came
out
of
that
called
net
shoot
made
by
one
of
the
docker
engineers
originally.
But
it's
just
a
it's.
A
A
crap
ton
of
low-level
network
and
systems
trouble
assuring
tools
and
I,
and
it's
a
big
problem.
Once
you
get
to
canal
large
scale,
containers
right
when
you
have
a
couple
of
servers,
it's
fine,
but
once
you
really
get
to
you
know
dozens
and
hundreds
of
containers
across
dozens
and
hundreds
of
servers
it
it
becomes
really
difficult
to.
You
know
deal
with
the
security
between
these
things.
Right.
Do
you
really
want
to
plug?
In
you
know,
add
security
features
to
every
single
one,
your
apps,
you.
A
Do
we
have
to
do
how
do
I
make
sure
that
I'm
always
getting
to
the
right
version
of
my
app
and
the
right
on
the
right
location
and
all
these
other
little
nuanced
issues
that
you
probably
don't
even
realize
until
you
have
the
system
and
you're
like
ok,
this
is
really
hard
for
me
to
troubleshoot.
You
know
measure
performance,
you
know
deal
with
security
and
access
layers
and
all
that
stuff.
So
so
your
story
is
around.
Like
the
birth
of
service,
mesh
and
I
guess
we
had.
A
B
And
so
there's
the
mesh
providers
in
that
you
know,
there's
like
SEO,
Hasek
or
console
there
is
linker
D
by
buoyant
AWS,
recently
launched
at
mesh,
which
is
natively
inside.
You
know
the
AWS
cloud
and
then
you
know
I'd
imagine
like
I,
think
IBM's
working
on
IBM,
mayor,
Maori
hackers
working
on
one,
so
I
expect
that
you
know
when
you
look
at
something
like
service
mesh.
It's
really
part
of
like
the
infrastructure
plumbing
but
application
level
infrastructure
right.
So
there
should
be
things
that
are
well
integrated
into.
B
You
know
that
stack
that
infrastructure
stack,
but
right
now
it's
you
know
the
product.
Some
of
the
projects
like
sto
is
an
open
source
project
that
comes
out
of
Google
right,
the
same
people
behind
kubernetes,
and
you
know
there.
There
are
organizations
that
are
kind
of
taking
us
to
you
and
then
distributing
it
into
their
platforms
right.
You
know,
like
we
see
with
some
other
open
source
projects
and
then
there's
a
you
know
like
these
other
ones
here.
So
it's
there's
a
lot
of
providers.
Each
one
has
a
slightly
different
implementation.
B
B
When
you
design
your
app
you're,
not
starting
first
with
you
can
instead
of
starting
first
with
like
okay,
what
is
a
tech
stack,
that's
approved
by
my
company,
you
can
say
like
what
is
my
application
need
to
do
for
the
people
or
the
customer
that
I'm
building
for
right,
how
they
want
to
interact
with
it?
And
then
you
know
it
could
be.
You
know
hundreds
of
services,
it
could
be
ten
services
and
it
needs
to
do
X,
Y,
&
Z.
B
The
beauty
of
this
kind
of
diversity
in
the
marketplace
is
you
can
pick
the
tool
you
can
pick
that,
based
on
what
your
application
needs
to
do,
steel
is
has
like
tons
of
tons
of
features,
but
it
may
not
be
what
people
need
link
RDS
focused
on
being
very
fast
and
easy
to
use,
so
that
could
be
good
for
different
types
of
applications.
So
really,
you
know
back
to
like
developer
choice
and
using
the
right
tool
for
the
job.
This
allows
for
that.
B
But
you
know
the
problem
that
we
as
a
company
look
at
it
is
we
want
to
help,
make
that
process
easier
and
then
we're
also
like
you'll.
Probably
you
may
have
more
than
one
thing
in
your
environment,
especially
if
you're
large
large
team
and
then
how
do
we
make
that
easier
for
you
to
kind
of
operate
overtime,
yeah.
A
The
says
there's
so
much
diversity
me
so
I'm,
looking
at
this
and
I'm
thinking,
okay,
so
a
lot
of
us
one,
we
don't
even
know
what
service
misha's
or
we
didn't
even
know
that
it
was
a
thing
and
we
maybe
have
started
to
hear
about
it,
and
it's
following
the
similar
pattern
to
docker,
maybe
five
years
ago,
right
like
it's
the
six
years
old
as
a
project.
It
was
you
know,
after
about
being
about
a
year
old,
most
of
us
that
were
just
sort
of
average.
A
C
A
Done
we're
hearing
about
this
thing
called
docker
and
we
don't
really
know
what
it
is.
Yet
we
read
a
quick
blog
article
and
we
get
a
generator.
We
get
a
basic
idea,
but
we're
still
very
confused
about
a
lot
of
the
things.
I
feel
like
we're
kind
of.
Are
we
in
that
stage
with
with
service
mesh,
or
we
were
we're
trying
to
figure
out
what
it
is,
what
it
isn't
and
trying
to
just
get
people
to
understand
the
concept
I
think.
B
In
some
respects,
maybe
because
I
think
it's
being
a
more
broadly
talked
about
I'm
kind
of
in
in,
like
you
know
the
last
year
18
months,
however,
I
think
it's
a
little
different,
because
things
like
docker
have
been
around
for
six
years.
Things
like
kubernetes
has
been
around
for
five,
and
so
yes,
some
of
the
bleeding
edge
folks
are
kind
of
deep
down
that
path
like
the
web
scale.
B
Companies
have
already
been
using
this
for
quite
some
time,
but
the
body
of
knowledge
that's
been
kind
of
from
all
those
kind
of
people's
experience
over
five
or
six
years.
You
know
that
lives
on
the
Internet
as
blogs.
You
know
engineering
blogs
from
a
lot
of
companies,
samples,
demos
and
so
in
that
anybody
who's
looking
at
I'm
gonna
do
containers
and
micro
services
today,
you're,
not
starting
at
the
same
place
where
you
and
I
were
starting
five
years
ago.
B
Do
you
know
what
I
mean
like
there's
already
been
like
a
lot
of
like
work
and
application,
architectures
that
have
been
built
and
from
that
they
can
look
at
it
with
the
you
know,
it's
like
hindsight
is
20/20.
You
can
now
look
at
a
net
new
journey
for
some
one
container,
izing
a
new
micro
service,
or
what
have
you
but
gaining
from
the
knowledge
of
everybody,
who's
already
gone
before
they're
gonna.
Ask
though,
and
there's
questions
on
like
how
do
you
handle
things
like
networking
and
storage
and
security?
A
lot
of
those
answers?
B
B
A
We
all
like
to
see
yeah
the
I
like
those
are
now
well
those
stories
because
they
it's
hard,
I,
think
one
of
the
things
that
you
know
once
you
learn
about
service
mesh
and
it's
it's
the
same
pattern
as
docker
and
kubernetes
right
like
once,
we
learned
about
docker,
then
we
had
to
make
decision,
do
I
care
and
then
eventually
you're
like
okay.
Everyone
around
me
is
doing.
This
I
need
I,
clear,
there's
clearly
something
in
the
water
and
I
need
to
go
check
this
out
and,
and
that
started
to
happen
with
that
has
happened.
A
I
think
for
a
lot
of
people
with
kubernetes,
of
course,
everyone's
on
their
own
path.
Right,
there's
still
be
people
just
getting
into
orchestration
in
five
years
right.
So
that's
a
thing,
but
with
with
service
mesh
for
solo
I/o,
how
do
you?
How
do
you
help
people
understand
if
they
even
need
it?
You
know
like
when?
Are
they
too
small
or
that's
yeah?.
B
B
We
may
get
to
that
point
that
you
know
that
could
be
interesting
to
see
how
the
ecosystem
evolved,
but
for
those
folks
who
are
kind
of
starting
they're
thinking
about
like
hey,
we're
ready
to
containerize,
we
have
new
applications
that
we're
going
to
build,
that's
going
to
use
containers
and
that's
going
to
be
from
more
of
a
micro
services
structure.
Those
folks
should
think
about
service
mesh,
and
it's
really
from
the
concept
of
like
when
I
map
out
my
app
architecture.
You
know
I'm
gonna
white
board
out,
like
here's
mine,
here's
this
new
app
project.
B
We
needed
to
do
this
and
you
start
talking
about
what
are
the?
What
are
the
components
that
we're
going
to
use
in
here
right?
So
then
they
start
to
think.
Like.
Oh,
we
have
polyglot.
We
have
different
things,
we're
gonna
try.
Then
you
have
containers
immediately.
The
next
question
should
be:
how
are
all
these
things
going
to
talk
to
each
other?
What
how
are
we
going
to
you
know
in
what
order?
What
kind
of
policies
do
we
want
to
have
for
it
if
we
start
to
update
a
container?
B
Even
if
you
have
five
containers
in
your
app,
you
know
you're
gonna
roll
out
an
update
or,
if
you're
gonna
scale
out.
If
you
need
to
scale
out
certain
services
differently
than
others,
you
know,
how
are
you
gonna
manage
routing
the
traffic
between
them
during
updates,
you
know
so
that
you
know
that
it's
good
before
you,
you
know,
send
100
percent
of
the
traffic.
Those
are
all
things
that
are.
Those
are
all
like
networking
questions
at
the
application
level
that
are
solved
with
service
mesh.
B
A
B
Know
and
everybody
rejoices,
and
then
everybody
also
kind
of
freaks
out,
but
the
idea
is
like
you
should
be
able
to
update
and
update
your
code
and
make
changes,
deploy
new
features
separate
from
you
know,
hey.
How
should
this
thing
talk
to
this
other
thing
and
you
know
I
have
a
hundred
people
coming
to
my
website
today.
Inbound,
like
you,
know,
I'm
just
rolling
out
this
update.
Do
I
want
all
hundred
people
to
roll
to
the
new
hit
the
new
thing
right.
A
B
Do
I
want
or
do
I
want,
like
10%
of
them
to
hit
it,
and
then
it
feels
good
I'm.
Not
seeing
any.
You
know,
like
you,
say,
degradation
and
performance
I'm,
not
getting
any
error,
so
I'll
start
to
then
at
some
rate
send
more
and
more
that
traffic
and
then
I
can
go
kill
the
other
container.
That
is
the
older
code
right.
A
B
A
It
sounds
a
little
bit
like
so
for
people
learning
in
my
courses
and
stuff.
They
learn
docker,
swarm
and
then
kubernetes
and
sort
of
in
that
order-
and
you
know
each
one
is
more
complex
and
more
abstractions
in
the
last
and
they
get
used
to
having
you
know,
DNS,
round-robin
or
load
balancing
in
there
as
the
way
that
one
part
of
their
app
finds
another
part
of
their
app.
Is
this
sort
of
simple
DNS
and
simple
load,
balancer
concept?
And
it's
it's
usually
out-of-the-box.
It's
very
basic
right.
It
doesn't
add
a
lot
of
functionality
there.
A
So
when
you
start
talking
about
10%
of
traffic
or
you
know
looking
at
the
requests,
real-time
and
the
traffic
or
you
know
controlling
which
one
can
talk
to
other
ones,
at
an
application
layer
level,
instead
of
just
firewalling
things
or
in
like
docker
and
docker
swarm,
where
you
just
create
little
virtual
networks
and
that's
sort
of
like
a
VLAN.
So
it
sounds
like
this
is
like
once.
That's
not
enough
and
just
simple
robbing
load-balancing
isn't
there
is
that
when
you
should
start
considering
service
mash
is
that
kind
of
the
entry
gateway?
Drug
sorts
I
mean.
B
There's
a
level
of
for
folks,
there
could
be
like
a
level
of
scale
in
number
of
services
right
number
of
containers,
but
even
when
you
have
in
order
to
even
understand
the
concept
of
what's
built
in
coming
out
of
the
box
with
like
a
swamp
with
former
kubernetes,
is
to
say
like
based
on
based
on
the
application
based
on
who
it's
serving
based
on
the
SLA
s
and
whatever
I
need
to
provide
is
leaving
it
all
to
be
completely
automated
by
that
system
enough.
You
know
so
it's
it
is.
A
B
A
Cool,
so
if,
if
someone
was
someone
had
learned
swarm,
and/or
kubernetes,
it
sounds
like
at
least
it
looks
like
that
the
marketplace
is
really
centering
this
around
the
kubernetes
orchestration.
So
do
you?
Do
you
sort
consider
kubernetes
a
requirement
before
you
get
into
super
Matt
super
service
mesh
or
is
there
I
think
I
guess
I
guess
AWS
has
their
own
and
I
don't
even
know
if
that
is
on
kubernetes
or
I.
Guess
it's
on
top
of
eks
yeah.
B
B
A
B
So
from
an
architectural
standpoint,
if
you
want
a
picture,
it
is
this
concept
of
like
sidecar
right
and
that's
what
really
and
that
sidecar
container
runs
alongside
the
actual
service.
The
container
that
has
your
application
code,
the
business
logic
and
that's
what
I
meant
by
like
abstracting
that
out,
instead
of
having
that
built
into
like
your
actual,
like
business
logic
it
sits
alongside,
and
so
when
there
is
a
when
the
traffic's
going
through
it's
gonna
before
it
hits
the
container.
B
It's
always
kind
of
traveling
through
that
sidecar
in
and
out,
and
so
that's
where
what
gets
interesting
is
like
when
you
abstract
it
out
so
you're
managing
all
like
the
routing
and
the
rules
and
the
policy
separately
you
can.
It
just
allows
for
more
interesting
things
you
can
update,
but
you
know
you
can
update
the
code
or
update
the
mesh
things
separately
from
each
other,
so
think
of
another
separation
of
concerns
from
that
point
and
then
be
able
to
kind
of
manage
those
separately
right.
A
C
B
The
stuff
inside
your
cluster,
it
needs
to
talk
about
certain
way
to
each
other
to
make
sure
that
the
app
actually
works
and
the
user
has
the
experience
that
you
wanted
them
to
have
with
the
application
and
then
there's
there's
other
kinds
of
traffic.
Is
the
traffic
coming
into
your
cluster
from
the
outside?
That's
when
somebody
actually
accesses
your
service
yeah,
so
those
are
like
kind
of
the
two
ways
to
think
about
that.
Yeah.
A
One
of
the
first
source-
you
know,
people
that
are
typically
following
these
videos
are
very
hands-on
right
and
so
one
of
the
first
questions.
Okay,
that's
great!
How
do
I
install
it?
Well,
how
could
I?
How
could
I
look
at
it?
And
you
told
me
something
about
I?
Think
it's
a
server
smash
hub
is
that
right?
Yes,.
B
And
so
that's
the
thing
that
we
put
together,
because
we
know
that,
like
with
all
new
technology
we
it
can
be.
You
know
sometimes
scary.
If
you
haven't,
like
you,
know,
read
everything
on
the
internet
before
you
try
and
use
it.
So
what
we
wanted
to
do
is
make
that
whole
getting
started,
process
super
easy
and
making
it
easy
to
try
different
ones
right.
One
like
we
said,
there's
different
options
out
there
so
service
mesh.
B
It
is
a
what
Brett's
showing
right
now
that
is
a
read-only
website,
so
you
can
see
you
know
it's.
What
we're
trying
to
do
is
put
a
GUI
based
dashboard
in
front
of
all
the
things
that
are
meshes
so
from
here.
What
you
can
do
is,
if
you
click
on
that
click
here
right
there.
This
is
a
read-only
environment,
but
in
that
click
here
thing
it'll
do
a
little
pop
up.
Basically,
you
can
install
this
mesh
hub
on
your
own
kubernetes
cluster
and
then
from
there
there's
the
ability
to
once
it's
stood
up.
B
You
know
very
three
simple
commands
what
you
can.
What
it'll
do
is
little
one
automatically
discover
if
you
by
any
chance
happen
to
already
have
meshes
running
in
your
environment,
because
you
never
know
your
buddy
and
your
buddy
down.
The
hall
might
already
be
playing
with
this
yeah
separately
from
here.
If
you
go
back
to
the
GUI,
you
can
do
install
a
new
mesh
from
the
dashboard,
so
this
dashboard
is
really
kind
of
like
your
your
operations
hub
for
it.
You
can,
if
you
click,
install
a
new
mesh.
B
What
it'll
do
is
you
can
pick
like?
We
have
sto
and
link
or
deselect,
give
it
a
name
pick
your
namespace
and
it'll
install
it
for
you,
so
the
meshes
that
are
available,
like
all,
have
slightly
different
api's
they're.
You
know
provided
by
different
vendors.
What
we're
trying
to
do
here
is,
instead
of
you,
as
an
end
user,
having
to
learn
the
details
of
each
one.
B
A
So
it
it
eases
the
like
all
open
source
projects,
especially
at
the
beginning
right.
We
focus
on
feature
set
and
solving
problems
before
we
solve
the.
This
is
easy
to
install
and
deploy
and
in
use
problem,
so
this
is
sort
of
instead
of
waiting
for
any
one
standard
on
installing
this
stuff.
You
sort
of
solve
that
problem
for
people,
and
this
is
all
free
right,
free
to
use
this.
A
A
B
So
superglue
is
kind
of
the
Misses
service.
My
job
is
kind
of
like
an
evolution
of
super
glue,
and
super
glue
is
really
a
fundamental
building
block
around
it.
This
is
all
CLI
based.
This
is
a
mesh
Orchestrator
right.
So
it's
a
thing
that
kind
of
dud
the
translation
and
kind
of
orchestrate
multiple
meshes
in
a
consistent
way.
What
we've
done
is
taken
super
glue.
It's
been
in
it's
an
open,
its
a
project,
we
open
sourced.
B
B
This
thing
that
you
know
with
the
the
kind
of
the
point-and-click
UI
on
it
and
to
be
able
to
kind
of
like
view
all
status
in
a
nice
display,
it
I
think
it
makes
it
more
access,
so
it
makes
it
accessible
to
more
people,
not
everybody's
comfortable,
doing
all
command
line,
commands
and
trying
to
remember
all
of
them.
Some
people
want
to
be
able
to
do
a
quick
snapshot
and
see,
like
you
know,
what's
the
red,
yellow,
green
in
my
environment,
right,
yeah,
plus.
B
A
A
B
Yes,
so
you
know
at
the
end
of
the
day
when
you
look
at
what
we
as
a
company
are
doing,
is
we're
at
we're
working
with
the
application
network
level,
we're
working
at
that
layer
and
we're
trying
to
connect
applications
so
that
you
know
people
can
do
interesting
things
with
them,
and
so,
when
we
look
at
that
that
we
there
are
two
sides.
When
you
look
at
the
application
landscape,
you
know
what
kind
of
apps
could
you
have?
B
One
is
well,
you
know,
there's
a
lot
of
stuff
you
already
have
and
then
there's
stuff
you
want
to
build,
and
so
almost
we
just
talked
about
what
service
mashup
is
people
when
you're
looking
to
build
new
applications?
Everyone
is
looking
at.
You
know,
containers,
micro-services
and
kubernetes,
and
that
means
a
new
ecosystem
of
stuff
around
it
and
which
service
mesh
is
an
enabling
kind
of
architectural
component.
And
then
there
is,
you
know
the
loads
and
loads
of
applications
that
have
been
running.
You
know
your
business
for
a
really
long
time.
B
A
lot
of
them
might
be
monoliths
and
those
need
to
you
know
you
still
need
to
iterate
on
them.
You
still
need
to
deliver
new
features,
so
with
that
side
is
on
that
side
we
have
an
API
gateway
and
so
we're
helping
kind
of
connect
evolve.
Those
applications
on
their
journey
to
becoming
you
know
more
cloud
native
by
using
the
API
gateway
to
help
add
new
features
to
an
existing
app.
You
know
that's
outside
of
the
monolithic
code
base
and
using
an
API
gateway
to
help
route.
A
And
I
guess
we
should
reiterate
that
this
is
about
technically
I
guess
would
be
called
greenfield
greenfield
apps
micro
service
apps
and
that
it's
not
unlike
just
containers
in
general.
That
can
run
everything
because
that's
a
question
we
get
a
lot
is
what
do
I
not
put
in
containers
but
I.
Think
at
this
point
generally
every
workload
it
should
default
to
a
container
and
if
you
can
figure
out
how
to
you
know
it's
usually
just
an
environment,
variable
thing
or
something
like
that:
it's
not
very
little
code
that
has
to
be
changed.
A
I
think
PayPal,
talks
about
and
they're
there's
their
container
story
about,
replacing
or
updating
a
hundreds
and
hundreds
of
apps
to
run
in
containers
that
they
didn't
have
to
change
any
code
for
them.
So
so
so
legacy
apps
can
still
run
in
containers
and
can
still
run
on
kubernetes.
But
it
seems
like
you're
talking
about
that.
Really
it's
it's
the
distributed
micro
service
architecture.
That
really
needs
something
like
a
service
mesh
tube
to
help
manage
it
right.
Yeah.
B
And
so
the
service
mesh
is
great
for
Greenfield
right
because
it's
you
know,
that's
the
one
where
you'll
have
like
more
than
one
service
right
and
you'll
start,
and
then
they
will
scale
it
in
different
ways
and
you'll
want
to
be
more
granular
on
how
you
kind
of
control
that
right
and
then
the
when
we,
when
you
kind
of,
went
over
to
the
the
glue
aspect.
That's
like
you
know.
B
If
you
convert
your,
you
know,
convert
your
legacy
workload
into
like
a
big
container
and
say
you
know:
you're,
secure,
you're
gonna,
keep
it
around
right,
but
you
want
to
add
a
feature
to
it.
You
could
add
a
feature
new
feature
as
a
function
or
you
can
turn
that
one
container
into
two
and
then
you
need
to
start
being
able
to
manage
those
connections
between
them
right.
A
And
so
one
of
the
questions
Andrew
asks
how
to
set
up
service
mesh
on
docker,
swarm
and
I.
Don't
know
if
Betty
has
a
better
answer
than
I
do
but
I
think
right
now,
I,
don't
know
of
any
service
mesh
that
works
out
of
the
box
with
swarm
and
I.
Think
one
of
the
more
challenging
issues
with
it
is
that
the
service
met
all
the
service.
Most
features
aren't
built
into
your
app
they're
built
as
a
sidecar
or
a
separate
container.
A
Now
with
kubernetes
you
get
pods
out
of
the
out
of
the
architecture
dot
you
know
of
kubernetes.
You
get
this
functionality
where
a
pod
can
be
multiple
containers
all
deployed
into
the
same
server.
You
know
they
both
they
all
have
the
same
networking
and
volume
space,
but
in
swarm
we
don't
have
that
pot
concept.
We
it
sort
of
is
removed
from
the
stack
to
increase
simplicity,
and
so
you
just
really
have
the
load
balancer
with
built-in
DNS,
round-robin
and
load
balancing
so
service
mesh
I.
B
C
B
He
was
playing
around
with
with
swarm
and
some
of
the
some
of
our
open
source
technology
to
see
if
he
can
like
build
the
concept
of
like
a
service
mesh
into
it,
but
yeah.
This
warm
comes
with,
like
you
said,
that's
probably
the
the
built-in
kind
of
like
just
the
routing
mesh,
which
is
kind
of
like
a
you
know
like
the
round-robin
II
I'm,
just
gonna
make
sure
the
traffic
gets
the
right
place,
but
that's
also
yeah.
It's
not
separate
yeah.
A
B
A
Yeah,
in
fact,
I'm
I
remember
this
reading
this
recently
about
the
people
asking
for
dr.
swarm,
support
in
SEO
and
I.
Think
there's
just
part
of
the
challenges.
I
think
that
people
generally,
you
know,
swarm
ISM,
more
enclosed,
it's
a
single
binary
right
so
to
add
functionality,
unlike
with
kubernetes,
where
there's
all
these
moving
parts
and
all
these
different
layers
that
increases
flexibility
and
allows
you
to
add
these
different
layers
of
technology
into
your
solutions,
whether
with
swarm
it's
unless
it's
supported
out
of
the
box
or
there's
a
specific
plugin
or
something
for
it.
A
And
we
don't
really
have
specifics
on
that
yet,
but
you
know
I
think
if
anyone,
if,
if
anyone
wants
you
know
service
mission,
swarm
hard
enough,
go
find
these
issues
like
just
Google
around
and
find
these
issues
and
give
your
thumbs
up
to
show
that
you
would
like
to
have
the
simplicity
of
swarm
with
also
the
benefits
of
service
mesh,
because
I
I
think
it's
a
right.
Now,
it's
a
big
tens
of
being
a
maturity
model
right.
You
you.
C
A
Dot
doc
or
swarm,
and
then
you
get
to
a
level
where
you're
like
we
want
invoice.
We
want
service
mesh,
we
want
aisle,
you
know
we
want
all
these.
We
want
helm
charts.
We
want
all
these
other
things
that
swarm
doesn't
get
support
for
and
so
that
forces
people
to
try
out
kubernetes.
Then
they
adopt
and.
A
And
now
they
have
these
new
features
available
to
them,
rather
than
you
know.
If
people
just
want
to
get
work
done
so
they're
not
going
I'm
gonna
go
make
a
service
match
that
works
with
swarm,
because
I
have
a
need
in
my
job.
Not
everyone
can
have.
Has
that
luxury?
Of
course
some
people
do
and
that's
how
these
projects
get
started.
But
but
not
everybody
can
do
that.
So
I
think
we're
early
I
think
I'm
not
saying
that
service
mesh
would
never
be
in
swarm.
A
I
think
we're
just
too
early
days
with
service
mesh
force
forum,
because
the
the
people
that
adopts
for
more
and
my
for
my
consulting
and
like
talking
to
people
they
just
want
to
get
work
done
rather
than
play
with
leading-edge
technology
and
so
swarm
out
of
the
box.
That's
one
of
its
key
features:
is
it
just
easy
to
get
started,
and
but
that
also
means
it's
not
going
to
have
every
feature
that
you
could
possibly
imagine,
which
is
I,
think
we're
the
realm
of
kubernetes.
A
You've
got
80,
plus
projects
now
on
how
just
to
deploy
apps
on
kubernetes.
So
we've
got
a
lot
of
experimentation
and
I.
Think
that
you
know
when
I
hear
about
a
service
mesh
for
swarm,
you
will
definitely
be
hearing
about
it
from
me
because
I'll
be
talking
about
it,
we'll
be
showing
it
on.
This
show
we'll
have
Betty
back
in.
B
Yeah
cuz
I'm
I,
want
to
say
at
darker
Con
I
did
see.
They
did
announce
that
they
were
gonna,
be
supporting
sto
in
the
enterprise
suite,
but
I.
Don't
I
haven't
seen
any
more
details
on
that
either.
So
it'd
be
interesting
to
see
how
that's
gonna
play
out
with
the
fact
that
they
support
both
swarm
and
kubernetes
inside
the
platform
together
like
side
by
side.
How
that
will
play
with
sto
support,
yeah.
A
And
I
think
I
thought
the
exact
same
thing
when
I
saw
it
on
the
slide
and
announced
the
docker
kind
of
thought,
okay,
because
by
default
you're
still
deploying
swarm
out
of
the
box
with
Ducker
enterprise,
although
it
doesn't
really
have
in
a
strong
opinion,
you
can
turn
on.
Kubernetes
cannot
turn
on
swarm,
have
both
both
of
those
work
with
enterprise,
but
I
immediately
thought:
okay.
Does
that
mean
we're
gonna
get
some
like
bonus?
A
Plugin
features
for
swarm
to
you
know
to
get
support
for
it,
CEO,
because
it
would
be
great
for
this
warrant
that
you
know
swarm
team
would
be
the
best
people
to
work
on
that
because
they
know
swarm
so
well.
The
I
think
one
of
the
things,
though,
with
swarm,
is
even
though
I
don't
see
swarm
ever
getting
pods,
at
least
not.
There's.
No
there's
no
discussion
on
the
roadmap
of
adding
pod
support.
We
could
probably
figure
out
something
around
constraints
so
that
you
could
ensure
that
you're.
A
Basically,
you
can
ensure
that
that
service,
mesh
sidecar
is
always
running
with
your
app
container,
and
we
just
need
a
couple
of
extra
features
in
swarm
mode
and,
if
you're,
if
it
been
around
long
enough,
if
there's
anyone
on
the
chat,
that's
been
around
long
enough,
there
was
this
old
swarm
called
swarm
classic
it.
It
actually
supported
that
it
had
a
feature
in
it
that
allowed
you
to
run
a
container
and
basically
it
would
find
the
right
other
container
to
run
side-by-side
with
it,
and
it
wasn't
technically
a
pod.
It
just
meant
they.
C
A
We
got
that
support
back.
We
might
build
together,
something
like
that.
I
have
to
bring
that
up
with
the
captains
and
see
if
we
can't
hack
away,
isn't
like
that
and
I
and
I'm,
not
sure
it
Chen
wits
I
couldn't
find
his
tweet,
but
I
know
he's
been
talking
about
it
lately
and
let
me
just
send
people
to
that
on
Twitter.
A
B
In
order
to
kind
of
like
set
up
the
rest
of
the
environment.
We
are
gonna,
go
into
mash,
and
so
you
know
they
always
say
like
we're
playing
with
both
this
tea,
only
Kerdi,
okay,
great
and
so
we're
playing
around,
and
then
it
really
comes
down
to,
like
you
know,
SEO
does
a
ton
of
things,
that's
kind
of
like
kubernetes.
It
does
a
ton
of
things
so
how
much
of
it
do
you
need?
Link
or
D
does
what
it
does
really
really
well,
and
it's
like
is
that
is
that.
B
Is
that
what
you
need,
or
do
you
need
a
few
other
things
from
over
here?
You
know,
that's
really
the
question
so
the
the
questions
kind
of
come
into,
like
hey.
We
do
a
bunch
of
this
stuff,
so
they
all
have
either
like
deep
mesh
related
questions
or
it's
a
what
is
a
mesh.
You
know
or
I'm
looking
at
its
do
and
then
it's
like
wow,
oh
for
what
we're
like
I,
don't
know.
I
heard
I
need
to
look
at
mesh,
so
we're
definitely
there's
a
there's.
B
A
lot
of
people
like
it's,
it's
a
buzzword,
kind
of
you
know
we
had
the
whole
container
buzzword
like
years
ago,
so
it's
really
kind
of
those
two
camps
and
then,
when
it
comes
to
like
I
need
X
I'm
looking
at
X,
because
I
heard
I'll,
it
need
to
look
at
it.
It's
really
kind
of
going
back
to
like
hey.
What
are
the.
What
are
the
things
that
you
guys
are
doing?
Are
you
doing
micro
services?
Do
you
have
one
implant?
Have
you
already
implemented
containerization?
You
know
at
some
point.
A
B
A
A
A
Yeah
and
because
it
it
it
seems,
like
you
know,
you
all
are
find
like
finding
a
rough
I
get
a
thing
that
service
mesh
either
is
not
doing
well
out
of
the
box,
and
it
needs
help
with
it
or
something
cool
that
you
can
do
on
top
of
service
mesh
to
add
even
more
functionality
or
insight
into
your
apps.
So
that
seems
we're
either
that
are
you're
just
wanting
to
make
up
really
cool
characters,
and
so
you
making
open.
A
B
B
The
land
of,
like
you
know,
I,
have
these
apps
that
I
can
tilt
play
faster,
like
the
promise
of
micro
services
is
to
be
more
agile
to
be
to
be
faster
and
then
to
have
more
freedom
for
your
developers
right
in
their
creativity
on
how
they
build
these
applications,
build
your
app
with
the
best
things
for
the
job,
so
we
want
to
help
people
get
there,
and
so,
when
you
look
at
that,
glue
is
on
the
Left.
There
is
the
API
gateway.
B
So
that's
really
about
like
helping
transition
that
monolith
to
be
more
cloud
native,
because
you
can
start
to
add
you
can
kind
of
iteratively,
add
capabilities
to
it,
and
the
new
capabilities
can
come
in
the
form
of
micro
services
or
like
service
functions.
We
call
that
concept
a
hybrid
app,
so
it's
really
about
you
know.
How
do
we
help
you
route?
B
You
know
ad
capabilities
and
route
traffic
on
things
that
are
not
using
mesh
today.
Squash
is
this
is
another
thing
on
like
you
know,
microservices
are
great,
but
when
you
kind
of
change
something
dramatically
like
that,
the
whole
tool
set
around
how
you
deal
with
micro-services
hasn't
evolved.
So
people
are
like
when
you
have
a
when
you
have
a
bug
in
your
let's
say:
100
service
micro
service.
How
do
you
know?
How
do
you
debug
it
which.
B
If
you
scroll
down
super
glue,
we
talked
about
that.
That's
really
service
mesh
orchestration,
so
we
see
like
glue
and
super
glue.
Super
super
glue
is
the
open-source
project
behind
service
mesh
hub,
but
we
see
the
glue
thing
and
the
servo
service
mesh,
helping
being
like
two
sides
of
the
coin:
right,
yeah,
yeah
and
then
glue
shot.
This
was
something
we
announced
just
at
Kuk
on
Europe,
but
that's
really
starting
to
think
about.
B
If
you
have
a
mesh,
if
you
have
your
application
network
abstracted
away
from
your
code,
what
kind
of
interesting
things
can
you
do?
This
concept
of
chaos
engineering
is
that
if
you
have
start
to
have
applications
that
are
lots
of
services,
the
network,
like
you
mentioned,
becomes
more
important
right.
How
those
things
talk
to
each
other
is
really
critical
to
troubleshooting
the
app
finding
out.
What's
going
on
so
once
you
start
having
a
lot
of
services
and
those
services
are
being
updated
by
different
people
at
different
times.
B
How
will
you
know
where
weak
points
are
so
this
is
about.
You
know
almost
thinking
about
it
from
like
proactive
testing
like
the
whole
idea
of
testing
applications
changes
when
you
have
micro
services,
so
cushaw.
What
that
does
it
injects
failures?
You
can
control
decks
bare.
You
can
inject
failures
or
latency
through
the
side,
cars
and
then
be
able
to
see
how
what
happens
and
then,
if
you
start
to
see
things
like
it's
controlled
and
that
you
could
turn
it
off
as
you're
monitoring
like
what's
happening
in
that
environment.
B
C
A
Pretty
cool
that
I
mean
that's
the
kind
of
thing
that
I
feel
like
the
only
if
you
didn't
have
a
service
match.
The
only
other
way
to
do
something
that,
at
that
level
of
your
app
is
to
either
I
guess,
have
completely
Software
Defined
Networking
or
put
this
in
your
app
directly
like
custom
code
it
so
this
so
the
glue
shot
is
for
it's,
it's
augmenting
the
service
mesh
or
is
it
augmenting
its?
It's
not
in
your
app
right?
It's
in
it's
in
the
service
man.
B
It's
augmenting
its
using
the
service
mesh
to
do
this
and
you
you
explain
it
really
well
in
that.
If
you
didn't
have
that
you'd
be
like
importing,
libraries
are
actually
making
code
changes.
So
then
you
have
problems
with
like
you're
building
more
dependencies
in
there
right
and
then,
if
you're,
you
know,
if
you,
if,
if
how
you
do
the
experiment,
things
are
the
tools
you're
using
those
change,
then
do
you
have
to
update
your
whole
application?
B
A
Very
cool
and
we
have
a
question
in
chat
a
little
technical
one
from
biker
biker
asks
about.
He
says
it
concerned
about
chatty
protocols
at
his
office,
so
you
know
I've,
never
measured
this
the
service
mesh
traffic
to
actually
look
at
what
it's
generating
I
don't
actually
know
what
is
the
back
end
of
service
mesh?
Is
it
storing
data
just
a
net
CD?
Do
you
know
yeah.
B
B
And
you
know
one
if
you
anytime
you're
doing
this
you,
you
should
always
test,
make
sure
you
always
test
test.
It
I'll
test
the
performance
and
see
how
it
behaves,
because
some
of
it
will
make
some
of
it
might
be
traffic
related.
Some
of
it
might
be
just
it
could
be.
It
could
point
to
like
your
network
infrastructure,
not
the
application
network
layer
and
then
also
it
could
point
out
some
other
areas
and
how
your
application
business
logic
might
be
handling
some
things.
B
So
it's
it's
worth
testing,
however
I'm
gonna
say,
like
you
know,
major,
like
things
that
you
use
every
day,
use
this
type
of
architecture,
things
that
you
as
a
consumer
expect
to
be
using
on
your
phone
and
it's
processing
a
lot
of
data
to
get
you
a
car.
You
know
like
to
your
doorstep
to
get
you
things
deliver
those
things,
and
those
are
really
happening
very
very
quickly.
B
A
B
So
when
you
look
at
this
and
you
look
at
some,
you
know
some
of
the
some
of
the
meshes
like
you
know,
there's
varying
levels
of
complexity
and
there's
varying
things
that
you
can
use
within
them.
You
don't
need
to
have
all
of
them
on.
If
that's
not,
those
are
not
capabilities
that
you
need
for
the
application
environment
right.
A
But
the
what's
crazy
is
that
we
are
we're
uncomfortable
with
the
idea
generally,
and
this
is
a
good
thing.
I
mean
engineers
are
generally
uncomfortable,
the
idea
of
adding
another
layer
of
abstraction
or
another
layer
of
networking
or
hop
until
we
start
doing
it,
and
then
we
realize
oh
we're
fine
99%
of
the
time
like
if
you
looked
at
docker
swarm
with
the
overlay.
Networking
I
never
talked
about
this
publicly
because
it's
changed
lately
and
I
wanted
to
so
I'm
gonna.
Throw
this
in
there
as
a
soapbox,
we
we
didn't.
A
So
when
you
start
telling
this
to
people,
they
start
to
they're
really
uncomfortable
at
first,
because
they're
like
what
you're
putting
my
TCP
packets
and
UDP
packets
and
sending
around
the
network,
and
then
they
use
it
and
the
benefits
are
so
outweighing
that
5%,
you
know
performance
degradation
that
they're
like
okay,
well,
I'm,
just
gonna
make
faster
networking,
I'll
use,
10
gigabit
instead
of
one
gigabit
or
something
like
that,
and
in
fact,
on
Microsoft's
website.
This
is
the
thing
that
I
figured
out
recently
for
swarm.
Microsoft
has
Azure
documentation
for
implementing
all
that.
A
You
know
the
orchestrators
and
stuff
like
that,
and
for
the
longest
time
their
swarm
page
talked
about
not
using
overlay
in
production,
because
the
overlay
networking
style,
where
it
encapsulates
these
packets
has
a
performance
hit,
and
so
they
would
caution
you
against.
Using
a
production,
and
instead
of
I,
would
argue
that
they
need
more
people
and
even
need
more
information
to
make
that
decision
they
need
to
know
like
does
my
it
doesn't
matter.
A
They
themselves
have
to
get
comfortable
with
a
layer
of
abstraction
in
that
idea,
and
this
is
I
think
just
the
next
level
right.
This
is
the
next
step
of
that,
where
we're
now
having
to
get
comfortable
with
the
idea
that
we're
going
through
a
bunch
of
proxies,
but
those
proxies
adds
so
much
richness
to
our
systems
that
it's
worth
the
hit.
Yeah
yeah.
B
A
Yeah
exactly
like,
and
it's
also
the
whole
there's
a
there's,
a
quote
around:
you
don't
have
a
performance
problem
until
you
have
a
performance
problem,
so
you
know
people
people
will
chew
up
networking
all
day
long
until
they
hit
a
limit
and
then
suddenly.
Okay,
now
we
have
to
troubleshoot
performance,
because
we
have
this,
you
know,
and
that
makes
it
makes
you
upgrade
servers
or
make
better
code
or
find
a
you
know,
fix
a
bug
in
someone
else's
code
or
whatever
and
I
think
that's
largely
what
this
will
also
do
right.
A
It's
like
like
Betty's
saying
it.
It
probably
will
expose
some
things.
You
didn't
even
realize
in
your
app
impact
in
our
workshops
and
especially
in
Jerome
pedos
oneís
workshops,
you
do
a
little
bit
of
basic
network
troubleshooting
to
understand
latency
between
containers
and
he
actually
throws
in
little
latency
bugs
into
the
demo
so
that
you
see
like.
Oh,
this
only
took
100
and
let's
took
him
100
milliseconds,
and
this
other
one
took
one
millisecond,
that's
weird
and
when
you're
using
it,
it
acts
fine.
C
A
B
We
just
chatted
the
other
day
I'm.
Actually
he
has
a
he
put
out
an
interesting
blog
post
talking
about
because
he
was
you
know,
he's
been
doing
like
darker
and
kubernetes
for
so
long
and
he
was
exploring
until
like
just
match.
Has
a
cotton.
You
know
in
general,
and
his
blog
post
recently
talked
about
like
in
the
days
that
before
docker
was
docker
and
it
was
dot
cloud
as
app
as
all
the
things
that
they
built
themselves
yeah.
B
It
was
basically
they
built
the
concept
servus
mesh,
they
custom-built
it
and
he's
kind
of
like,
oh
and
then,
and
then
let
me
you
know,
because
you
needed
it
for
what
they
were
building
and
then
he's
like
okay.
Now,
if
we
like,
look
at
that
concept
and
see
like
what's
coming
out
on
the
market
today,
yeah.
C
B
So
it's
back
to
like
you
know
he
got
micro
services,
you
need
a
start.
You
know
you
want
to
have
more
than
one
service.
You
need
to
start
thinking
about
how
those
things
are
talking
and
how
you
want
to
control
that,
and
that
is
basically
how
you
know
when
you
think
about.
Is
your.
You
know,
app
network,
which
is
more
like
mesh
yeah.
A
Well:
okay,
so
we're
hitting
an
hour
on
the
stream
and,
let's
take
the
last
question.
So
friend
gets
the
the
the
winner
of
the
last
question
award.
Do
you
think
that
is
a
good
idea
to
migrate
a
production
docker
swarm
to
K
eights,
using
a
glue,
API
gateway
to
make
the
transition,
or
is
it
not
possible
to
send
traffic
to
a
swarm?
You.
A
B
B
A
That
all
the
things
container
native
yeah
well,
this
has
been
like
I,
there's
so
much
more
for
me
to
learn
about
service
mesh
and
I'm
I'm
sure
we're
gonna
have
more
conversations
about
it.
We,
if
you,
if
you
follow
this
channel
by
the
way
we
both
have
a
friendly
cow
code,
who
was
on
the
show
weeks
ago,
maybe
a
month
ago,
and
he
was
at
kubernetes
a
coop
kampf
cloud
native
conf
and
he
did
a
session
on
talking
about
all
the
different
service
meshes.
A
You
know,
basically,
there's
courses
coming
from
me
and
I'm
not
saying
that
there
will
or
will
not
be
service
mesh
here
at
anytime
soon,
but
follow
follow,
follow
us
on
the
Twitter's
and
you
will
hear
about
that
over
there.
That's
over
there
and
because
I
think
that,
as
you
start
to
adopt
more
micro
service,
stuff
I
think
it's.
This
is
gonna,
be
more
and
more
important
for
us
and
Betty
and
solo
IO
are
helping
us
understand.
What
is
service
mesh?
A
When
do
we
need
it
like,
and
what
you
know
solving
the
problems
of
getting
it
installed,
that's
kind
of
stuff
and
I
love
it
I
love!
That's
why
one
of
the
reasons
I
rant
about
swarm
all
the
times,
because
it
just
makes
getting
started
so
much
easier
than
than
having
to
learn
a
huge,
complicated
products
and
so
I
like
easy,
because
we
we
all
we
all
want
to
do
the
new
stuff
and
when
you
make
the
new
stuff
easy,
that's
I
think
how
a
lot
of
people
really
want
to
use
it.
So
so
Betty.
B
A
Yep
there
there's
no
bed
service
mesh
question
on
Twitter,
but
in
case
you
don't
want
it
to
be
in
public
Dean.
We
both
have
open
DM
so
that
you
can
ask
those
questions
that
you're
not
quite
wanting
to
share
with
the
world
and
Betty.
Where
will
you
be
next?
I
know
you
go
to
conferences,
and
here
you
have
a
travel
schedule.
Yeah.
B
Yeah
yeah,
so
actually
everybody
on
the
team
has
been
doing
tons
of
we've
been,
you
know,
doing
webinars
we're
gonna
start
doing
some
of
these.
Like
live
types
of
you
know
unboxing
and
you
know
Q&A
type
stuff,
and
we
have
people
speaking
at
various
meetups
and
conferences.
So
next
week
myself
and
Scott
Granton
will
be
out
in
New,
York
City,
so
we're
gonna
be
a
coupon
and
then
we'll
also
be
at
the
New
York
City,
kubernetes
Meetup
and
then
Christian
posta
who's.
B
C
B
A
And
that's
how
to
how
to
get
the
real
info
on
service
mesh.
So
thank
you
so
much
Betty
for
joining.
This
has
been
a
great
learning
experience
for
me
just
meeting
with
your
team
this
week
and
talking
about
service
mesh
in
general
and
how
solo
IO
is
trying
to
help
make
that
easier
and
I
will
certainly
have
you
back
on
this
show,
because
we've
got
to
get
our
selfie
by
the
way.
I
don't
think
we
got
a
live
selfie!
No.