►
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
Intro: Telepresence - Rafael Schloming, Datawire & Adnan Abdulhussein, Bitnami
How do you set up a productive development environment on Kubernetes? With microservices and cloud databases, creating a local development environment is very complex. In this talk, we’ll cover the different approaches to developing services on Kubernetes. We’ll then introduce Telepresence and show how it can be used to enable local development of Kubernetes services. We’ll cover how Telepresence is used for real-world development of the Kubeapps project. We’ll take a peek under the hood to show how Telepresence works. We’ll finally wrap with a discussion of the Telepresence roadmap.
To Learn More: https://sched.co/GrbM
A
Hey
everyone-
hopefully
everyone's
here
for
the
telepresence,
let's
official
name,
telepresence
intro
intro
session,
so
my
name
is
reka
Shlomi.
This
is
unknown.
Abdul,
Hussein
and
I
work
at
a
company
called
data
wire.
This
guy
works
at
binami
and
data
wire.
If
you
haven't
heard
of
us,
we
are
I
think
what's
called
a
brave
group
of
souls
who
decided
to
build
a
company
that
caters
to
developers,
so
we
love
application
developers.
We
want
to
make
their
lives
better.
We
want
to
make
it
easier
for
them
to
get
that
work
done
and
we
love
open
source.
A
We
do
this
in
two
ways:
we
have
ambassador,
it's
an
API
gateway
built
on
Envoy.
You
may
have
heard
about
this
and
we
also
make
telepresence,
which
actually
we're
happy
now
as
a
CFCF
project
and
come
actually
telepresence
alone.
Has
thousands
of
users
combined?
You
have
many
many
thousands
of
users,
so
in
fact,
including
bitNami,
and
so
since
we
like
it
better
when
our
users
say
good
things,
then,
when
you
know
we
say
good
things.
I'm
gonna
turn
this
over
to
Adnan,
to
give
you
guys
a
an
intro
into
telepresence.
Thank
you.
B
Cool
thanks
well
so
everyone,
as
my
fellow
said,
I
work
at
vid
Nami.
If
you
haven't
heard
of
us,
try
not
to
make
this
a
marketing
pitch,
but
we,
we
are
one
of
the
leaders
in
application.
Sorry
in
application,
packaging
in
VMs
containers,
cloud
images
and
stallers
a
whole
range
of
different
packaging
tools.
We
are
a
kubernetes
certified
service
provider,
which
means
we
sometimes
will
come
to
events
like
this
and
provide
training
and
other
things
we've
also.
B
So
kubernetes
is
we
kind
of
see
it
as
like,
a
really
great
production
environment.
A
lot
of
our
customers
will
will
tell
us
that
that
they
love
deploying
stuff
on
cuba,
Nettie's
and
it
just
makes
it
it's
just
a
really
great
environment.
To
run
production
workloads,
on
top
of
you
know,
has
a
lot
of
great
primitives
built
in
things
like
services
and
load
balancing
to
your
deployments.
Your
staple
sets
all
that
good
stuff
that
just
works.
B
The
problem
is
what
we
often
hear
is
that
the
development
experiences
is
not
a
great
and
we
found
out
ourselves
when
we.
When
do
we
do
it
of
developer
tools
on
that.
Well,
you
know
just
to
give
a
kind
of
an
example
if
you're
developing
an
application
on
top
of
Cuba
at
ease.
It
might
look
something
like
this.
Where
you,
you
know,
you
kind
of
have
these
different
services
that
connect
to
maybe
a
database.
That's
over
inside
the
cluster
or
external
to
the
cluster.
B
Develop
a
community
native
application
like
that.
You
might
have
something
like
this,
which
is
talking
post
file.
I,
actually
took
this
from
the
monocular
project,
which
is,
which
is
something
that
we
worked
on
previously
and
I.
Just
have
all
my
different
services
that
I'm
I'm
developing
here
in
the
Stockton
first
file
have
things
like
volumes,
environment,
variables
and
I'm
plugging
in
focal
local
volumes
that
I'm
trying
to
mount
N,
and
these
are
all
things
that
kubernetes
has
already
solved
for
me,
but
I'm
having
to
re-implement
it
into
compose.
B
Just
that
I
can
develop
my
application,
so
I
know
how
much
we
all
like.
Yam,
all
right,
I,
don't
know:
I'm,
not
a
big
fan
of
having
to
manage
two
different
types
of
Yama
and
two
different
environments
entirely
so
like
manage
this
one
thing
like
it
can
quickly
become.
It
can
quickly
get
really
hairy
and
go
really
wrong
and
you
have
all
these
different
paths,
their
intertwining
of
each
other.
So
it'd
be
really
nice.
B
There
are
other
tools
like
scaffold
and
draft
that
we
looked
at
and
telephone
and
we
found
telepresence
as
well.
One
of
the
things
that
we
needed
in
in
our
case
for
key
wraps
is
we
use
things
like
ingress
jobs
and
cron
jobs
and
it
containers
we
have
web
building
controllers
as
well.
So
in
order
to
do
that,
we'd
have
to
like
in
duck
and
pose
we'd
have
to
mount
keep-keep
config
for
extensions,
so
obviously
get
again
going
out
to
the
point
we
don't.
We
didn't
want
to
have
to
re-implement
all
of
this
for
doc
compose.
B
So
we
really
needed
a
way
to
do
this
inside
of
kubernetes
another
way
of
doing
this,
and
this
is
similar
to
what
scaffold
and
draft
kind
of
do
is
they
go
into
just
like
development
cycle?
Where
you
know
you
build
an
image,
you
push
it
up
to
a
registry,
and
then
you
either
do
it
keeps
the
deal
set
image
or
you
do
you
know
draft
or
scaffold
can
kind
of
do
that,
for
you.
B
But
that
can
take
some
time,
and
especially
the
the
application
that
we're
working
wave
is
a
react
application
and
then
I've.
You
work
with
nodejs,
but
it
takes
a
bit
of
time
to
do
something.
Sometimes,
so
you
got
the
classic.
You
know
developer
slacking
wats
what
the
docker
is
building
or
whatever.
B
So
then
we
came
across
telepresence,
which
is
again
by
date.
Wire
and
the
way
telepresence
works
is
that
you
have
all
your
services
that
already
deployed
inside
of
your
kubernetes
cluster,
and
you
tell
telepresence
that
you
want
to
replace
a
certain
port
or
deployment
with
a
proxy
back
to
your
local
machine.
So,
with
this
set
up,
I
can
deploy
my
application
into
my
cluster
or
into
my
shared
cluster
or
cluster
running
locally.
B
That
I've
set
up
something
like
mini-keyboard
stalker
for
for
Mac,
given
at
ease
support,
and
it
just
works
the
same
way
just
picks
up
my
cube,
config
and
I
tell
it
what
deployment
I
want
to
point
it
to
and
all
go
and
replace
that
part
with
a
proxy
that
then
points
to
a
port
on
my
local
machine
and
additionally,
to
that
I
also
get
service
discovery.
So
now
I
can
talk
to
anything
in
a
cluster.
That's
using
s
shuttle
for
that
volumes.
B
It
mounts
so
it
mounts
the
volumes
that
are
mounted
into
that
into
your
pods.
Using
sshfs
and
I.
Have
the
exact
same
essentially
the
exact
same
environment.
I
would
have
running
that
thing
inside
that
portal,
if
I
did
keep
CTL
exec,
for
example,
but
it's
all
in
my
local
machine
so
before
I
go
into
a
demo
here,
there's
a
couple
different
proxying
methods
that
her
presence
supports
today.
I
know:
there's
some
changes
that
coming
as
well
around
that
and
Raphael
can
speak
to
that
later.
B
But
there's
kind
of
s,
the
main
method
or
like
the
the
default
method,
is
the
VPN
method
which
works
well
for
your
developing
go
services,
but
there
is
a
caveat
in
that.
It
won't
work
while
there
are
other
VPNs
or,
if
you're
connected
to
have
a
VPN
S
on
your
machine.
So
if
you're
connected
to
your
company
VPN
and
need
that
kind
of,
you
have
to
turn
it
off
and
then
and
then
use
the
VPN
eff
it.
So
it
kind
of
interferes
with
that.
B
The
second
method
is
inject.
Tcp
excuse
me,
which
basically
works
by
injecting
a
shared
library
into
the
process
to
basically
reroute
a
network
course
to
to
telepresence.
Obviously
this
is
not
going
to
work
for
statically
linked
things,
because
because
because
it
requires
on
that
that
shared
library
right
as
a
result,
that's
not
going
to
work
well
with
go
because
often
you're
compiling
into
a
statically
linked
binary.
The
other
massive
method
which
is
really
cool,
is
the
docker
method.
So
this
works
well,
if
you're
already
developing.
B
If
your
development
environment
already
has
containers-
and
you
can
do
things
like
docker-
run-
to
set
up
a
container
to
run
and
it
will
just
run
essentially
a
sidecar
container,
with
with
the
network
stack
set
up
to
talk
to
you
to
the
president
proxy
and
then
and
and
yeah.
So
that's
just
dust
reason.
I
do
sorry
so
and
then
the
container
that
runs
for
you
will
share
the
same.
Networking
stack
as
the
the
other
container.
B
So
yeah,
that's
really
great
if
you,
if
you're
already
using
containers
in
your
divert
environment
today,
so
it's
it's
fun
to
talk
about,
but
it's
way
more
fun
to
do
an
actual
demo.
So,
let's
head
over
to
that,
hopefully
this
is
big
enough
for
all
of
you
at
the
back.
Can
you
can
you
see
it
out?
Cool,
okay,
so
I'm
just
gonna
make
it
a
little
bit
bigger
just
in
case
oops
I,
just
reset
it
Oh
City!
B
Okay,
so
I
have
a
cube
entities.
Cluster
set
up
right
now,
it's
actually
using
docker
for
Mac
kubernetes,
because
I
don't
want
to
worry
about
the
network
right
now,
but
I
have
a
if
I
just
do
you
keep
CT
I'll
get
pods
here.
I
have
some
services
running
and
discussed
already
the
ones
the
one
that
I
have
here?
B
Is
it's
just
a
basic
API
called
quixote
m,
which
is
quite
over
a
month
and
it
just
responds
and
Jason
with
a
quote
and
then
I
have
this
telepresence
test
container
as
well,
which
is
which
we
can
take
a
look
at
I
guess.
First,
let
me
just
show
you
the
service
here
so
I
have
this
cure
TM
service,
that's
running
on
port
5000
and
it's
a
cluster
IP
service,
so
I
don't
have
access
to
it
right
now
outside
of
the
cluster.
B
So
I
have
this
telepresence
test
file
or
deployment.
I
guess-
and
this
will
be
my
application-
that
I'm
that
I'm
kind
of
debugging
it
has
an
in
it
container
that
basically,
is
just
seating,
some
data,
so
you
can
see
it's
just
mounting
a
volume
which
is
an
emptied
a
volume
and
echoing
some
data
to
this,
this
file,
random
and
then
the
main
container
start
starts
up
and
it's
basically
just
doing
a
tale
of
of
that
file.
I
also
have
a
secret
environment
variable
here
coming
in
coming
in
from
here
as
well.
B
So
if
I
take
a
look
at
this
I
just
do
qtt
old
logs
here
I
should
see
that
yeah
it
works
as
expected.
I'm
getting
random
data
out
there,
but
let's
just
pretend
that
there's
something
wrong
with
this
service
and
I
want
to
kind
of
debug
it
and
I'm
gonna
use
telepresence
to
do
that.
So
I'm
gonna
do
a
watch
of
the
pods
here
just
to
see
what's
going
on
as
telepresence
starts
off,
but
what
I'm
gonna
do
is.
B
It's
gonna
start
out
with
the
VPN
proxy
method
here
and
you
can
see,
as
this
new
part
is
starting
up
here.
It's
still
going
through
that
init
containing
process.
So
it's
still
still
running
that
in
in
container
and
then
it
and
then
it
starts
up
with
the
proxy,
and
then
it
puts
me
into
a
bash
shell
because
I
didn't
provide
any
other
command
to
it.
It
just
puts
me
into
a
bash
shell
when
I
can
I
can
just
run
commands
in
here,
so
first
of
all,
I'm
gonna
run
LS
and
I'm
in
the
same
directory.
B
B
B
Oh,
the
other
thing
is
the
environment.
So
if
I,
if
I
run
and
here
I,
can
see
all
these
environment
variables
that
they've
come
from
the
pod,
but
also
the
ones
that
are
clearly
from
my
local
machine
as
well,
so
I
have
I've
kind
of
both
mixed
in
here
and
I
should
be
able
to
see
and
I
can
that
secret
environment
variable
here
at
AI
that
I
put
into
that
that
deployment.
B
B
B
What
I
want
to
do
is
when
I'm
developing
this
I
want
to
replace
the
dashboard
service
which
I'm
currently
connecting
to
virus
this
port
forward
here
over
on
running
and
if
I
go
to
chrome
here,
I'm
logged
in
to
keep
apps
so
cube.
Apps
is
an
application
dashboard
for
your
cluster.
It
allows
me
to
basically
deploy
and
manage
helm
charts
inside
my
cluster
and
also
manage
the
Service
Catalog
provision.
B
Services
from
the
Service
Catalog
I
haven't
set
up
right
now,
but
it's
something
I
can
do,
but
all
this
needs
to
use
the
cube
Nettie's
API,
so
that
it
can
actually
request
information
about
Resort.
What
resources
are
running
and
give
you
in
output,
like
you
know
what
pods
and
sorry
what
IPS
imports
or
things
are
running
on,
so
that
you
can
actually
access
that.
B
So
what
we
do
is
we
run
telepresence
and
we
pass
in
once
again
we
pass
in
the
swapped
appointment.
So
the
point
we're
going
to
swap
is
this
one
to
keep
out
the
internal
dashboard
and
certainly
in
space
here
perhaps,
and
then
we
also
are
going
to
expose
oops
expose
the
port
from
my
local
machine.
I'm
gonna
use
3,000
and
we're
gonna,
basically
replace
it
with
port
8080
here.
B
And
this
is
again
why
we
chose
not
to
build
containers
and
push
them
up,
because
it
can
take
just
that
long
to
start
the
development
server.
So
you
can
only
imagine
how
long
it
takes
to
build,
but
now
that's
running
I
can
take
refresh
here
and
once
again,
I
have
access
to
the
same
thing.
It
looks
the
same
except
that
it's
actually
running
off
of
my
local
machine
from
this
and
I
can
even
you
can
see
if
I
go
to
this
address
here
on
my
local
machine.
B
B
B
And
once
again,
this
is
all
happening.
If
I
refresh
able
to
see
its
recompiling
here,
it's
picked
that
up
once
it
gets
recompiled
it'll
refresh
the
browser
with
that
change.
Now,
if
I
were
to
exit
telepresence,
let's
just
go
and
control
see
here,
I
thought
I've
done
developing
what
it's
going
to
do.
If
we
go
back
to.
B
A
So
yeah,
so
this
is
that
that's
a
great
demo
of
where
we
are
sort
of
today
I
think
we're
at
what
point
nine
something
for
94.9.
Four.
We
try
to
release
you
know
frequently,
so
it's
hard
to
track
the
number,
but
the
things
we're
really
looking
to
get
to
before
sort
of
we
consider
at
1o
are
a
bunch
more
things
on
robustness
and
sort
of
speed
of
the
UX
right.
So,
if
you're
using
this
in
a
sort
of
a
network,
flaky
environment,
you
might
appreciate
reconnect
that's
something
right
now.
A
If
your
network
connection
dies,
you
know
you
sort
of
have
to
set
up
everything
again
and
it's
a
little
bit
of
a
pain,
and
sometimes
when
that
happens,
because
the
client
depends
on
the
client
does
cleanup
in
the
cluster.
That
cleanup
doesn't
happen
so
we'd
like
to
make
that
cleanup
happen
automatically.
A
It's
it's
getting
less
and
less
sort
of
useful.
So
we'd
like
to
we'd
like
to
eliminate
that
before
we
hit
one
else
so
that
you
know
we
don't
have
to
support
it
going
forward.
So
we're
actually
pretty
much
95%
of
the
way
there
on
this
stuff,
which
I
guess
it
makes
sense
since
we're
about
version
point
9
for
that
of
course
worked
out
on
purpose,
so
yeah,
and
so
the
we're
at
this
sort
of
place
where,
where
we
have
sort
of
made
a
bunch
of
architectural
changes
to
do
pretty
much
all
this
stuff.
A
But
we're
in
this
interesting
situation
with
telepresence,
because
it
is,
it
is
a
command
line
program
it
runs
in
it
runs
in.
It
runs
on
people's
laptops,
people's
laptop
and
run
on
different
networks.
What
it
is
doing,
you
know,
is
sort
of
working
at
a
network
level
at
multiple
network
levels,
it
messes
with
DMS
it
messes
with
your
firewalls
and
stuff
like
that,
and
so
that
those
things
are
sensitive
to
all
aspects
of
your
environment.
You
know
as
a
little
example.
You
know
we
had.
A
So
I'm
max,
you
know
I
think
max
III
until
until
I
got
involved
in
this
I
didn't
learn
the
max
actually
I
didn't
realize,
but
max
actually
cache
negative
DNS
lookups.
So
if
you
get
a
DNS
lookup
failure-
and
you
you
know-
swap
your
DNS
you'll
still
get
that
negative-
that
NX
domain
result
until
the
cache
decides
to
flush
and
there
are
actually
ways
you
can
manually
flush
this.
A
In
fact,
we
have
incorporated
code
to
do
that,
but
the
the
way
to
do
that
I
think
it's
changed
four
or
five
times
over
the
over
sort
of
the
recent
set
of
Macra.
In
fact,
it
went
what
it's
gone
back
and
forth
between
between
different
versions.
So
I
have,
unfortunately,
don't
have
a
slide,
but
there's
a
great
snippet
of
code
that
has
less
than
greater
than
on
minor
version
numbers
to
figure
out
the
right
way
to
flush,
DMS
caches.
So
well.
A
We
have
architectural
II
sort
of
put
the
stuff
in
place
and
we
have
a
component.
That's
sort
of
it's
almost
a
new
method
to
to
do
this.
Proxying
that
we
believe
is
sort
of
is
more
flexible,
robust.
It
could
more
robust
and
can
do
all
these
things.
Well,
we
have
that
that
we're
in
this
place,
where
we
don't
want
to
just
roll
it
out,
because
we
don't
want
to
you-
know
a
lot
of
people
use
telepresence.
A
In
the
future
part
of
the
architectural
changes
we've
made,
we
want
to
enable
sort
of
some
more
flexible
use
cases,
so
the
swab
deployment
scenario
saw
was
kind
of
combining
three
three
sort
of
categories
of
use:
cases
right:
it's
replicating
the
execution
environment,
the
file
system,
the
environment
variables
all
the
stuff
that
the
unit
containers
are
doing.
So
it's
replicating
the
execution
environment
in
kubernetes
on
your
laptop.
A
It's
also
taking
outbound
network
traffic
from
your
laptop
and
directing
it
back
to
the
cluster,
and
it's
capturing
inbound
traffic
from
the
cluster
and
bringing
it
down
to
your
network,
and
it
turns
out,
if
you
we
think,
if
you
actually
break
these
into
a
little
bit
more
of
a
toolkit.
You
can
put
these
together
in
some
interesting
ways
like
you
can
maybe
shadow
traffic
from
a
live
cluster
that
you
wouldn't
want
to
necessarily
send
run
to
telepresence
on
directly
or
send
or
send
traffic
back
into,
and
then
you
can
actually
take
that
traffic.
A
A
So-
and
you
know,
one
of
the
one
is
I
recently
learned
that
that
turns
out
choices,
even
though
we
think
are
there
they're
a
good
thing.
Actually,
when
people
have
more
and
more
choices,
it
turns
out
beyond
a
certain
point
that
actually
makes
people
less
happy.
So
we'd
like
to
we'd
like
to
remove
some
of
the
choices
right,
you
saw
how
complicated
that
that
method
slide
was
and
the
sort
of
the
matrix
of
understanding
what
what
works
with
what
that's?
A
Not
something
we'd
like
you
to
have
to
think
about
so
simpler
UX
and
you
know
maybe
it's
sort
of
more
first-class
support
for
higher
level
developer
tools
like
IDs
service
mushrooms.
That
kind
of
thing.
So
that's
that's
where
we'd
like
to
go,
we're
doing
a
we're
doing
a
sort
of
more
in-depth
talking
some
of
that
stuff
on
Thursday.
So
hopefully
we'll
get
some
feedback
and
see
you.
C
A
A
A
D
D
I
was
really
unable
to
use
telepresence,
but
not
because
of
anything
wrong
with
telepresence.
It
was
because
what
tow
presence
was
actually
good
for,
in
that
case,
physics,
exposing
the
fact
that
we
have
to
do
place
fell
pretty
short
of
the
hit
ball
back
draft
and
telepresence
assumes
that
your
apps
are
get
in
some
serious
waves
like,
for
example,
if
you
are,
if
you
have
a
legacy
application
that
uses
positive
spent
like
say,
user
file,
uploads
per
temple,
those.
Of
course
those
uploads
are
not
going
to
be
in
your
local,
cleaner.
A
A
You
know
I,
it's
a
little
bit
trickier
dealing
right.
Just
because
of
you
know
the
we're
using
sshfs
under
the
covers
s.
Aphasia
sshfs
is
most
almost
read-only
as
a
as
a
file
system,
and
so
when
you
get
into
shared
file
systems
yeah,
you
basically
need
a
custom
sort
of
a
custom
proxy
thing
and
you
know
I
think
it.
B
So
it's
used
by
our
team
on
cue,
bats,
I,
don't
think
it's
been
adopted
in
every
team,
but
certainly
within
the
Cuba
nations
team
we
were
using
it
we're
pretty
much
one
of
the
only
teams
are
developing,
or
that
requires
a
lot
of
do
cube,
Nettie's
api's
as
well.
It
will
keep
that
so
that's
kind
of
the
reason
why
we
look
for
something
like
that,
whereas
other
teams
don't
necessarily
have
that
problem
but
I
know
some.
B
B
We
actually
do
it's
a
good
question,
so
the
question
was
whether
we
use
shared
environments
or
if
we
spin
up
our
own
clusters.
We
actually
do
a
little
bit
of
both
I
prefer
to
use
dock
compose.
Sorry
talk
docker
for
Mac
I,
some
other
folks
tend
to
use
mini
cube
and
then
some
folks
to
use
our
shared
cluster
and
run
cube
apps
in
their
own
namespace
and
then
swap
it
out
on
there,
but
it
but
it
the
telepresence
works
across
all
of
those,
so
that
workflow
is
the
same.
B
But
but
yeah
we
do
have
these
kind
of
different
environments
in
some
ways
it's
useful
because
it
allows
us
to
test
cube
apps
in
all
those
different
environments
as
well,
but
at
some
point
I
think
it
would
be
nice
to
just
have
one
shared
environment
shared
cluster,
where
we
all
we're
all
developing
on
and
we're
able
to
just
swap
out
things
there.
Yeah.
B
Yeah
I
think
so
in
our
development
clusters,
they're
not
exposed
to
the
Internet
Turpin's,
just
just
works
over
keep
config
and
and
configures
the
just
yeah.
It's
just
working
kind
of
as
a
local,
proxy
or
tunnel
I.
Guess
it's
not
actually
exposed
to
the
internet
for
our
use
case.
But
if
you
word
and
yeah
you'd
probably
want
to
use
some
sort
of
internal
lobe
answer,
I
noticed
some
work
going
on
as
well
to
route
to
different
endpoints
based
on
an
authorization
header
or
something.
A
Yes,
so
one
of
the
one
of
the
areas
we're
working
on
one
of
the
reasons
we
want
to
separate
that
sort
of
inbound
execution,
outbound
stuff
is
yeah.
It's
because
yeah
you
can
only
swap
one.
Only
one
person
can
swap
a
deployment
at
a
time
which
means
to
benefit
from
telepresence.
You
need
a
way
to
sort
of
make
your
cluster
multi-tenant
with
respect
to
developers,
whether
you
sort
of
have
a
cluster
per
developer
or
whether
you
have
a
namespace
for
developer
one
way
or
another.
A
You
need
to
set
that
up,
and
so
there's
you
know
other
strategies
to
do
sort
of
you
know
per
request
routing
where,
where
it
would
actually
be
much
easier
to
roll
out,
you
wouldn't
have
to
do
that,
and
so
we'd
like
to
be
able
to
support
those
kind
of
things-
and
you
know
in
in
that
kind
of
scenario
that
gets
much
more
realistic
to
put
into
a
production
set
up
and
integrate
with
authentication
so
that
you
know
only
authenticated.
Requests
are
allowed
to
be
allowed
to
be
directed.
You
know
to
to
a
laptop.