►
From YouTube: Solo.io Sponsored Session - Idit Levine, Solo.io
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
Solo.io Sponsored Session - Idit Levine, Solo.io
A
A
So
why
we're
all
here
we're
all
here,
because
we're
excited
about
the
potential
we
excited
about
what
service
mesh
can
give
us
it
can
give
us
security
features,
can
give
us
observability
future
roarings
future,
it's
all
great
right,
but
what's
really
really
special
about
what
we
are
in
here
is
not
what
it's
actually
giving
you,
but
how
it's
actually
doing
it
anyway.
It's
actually
doing
it
if
you
think
about
it
actually
by
abstracting
the
network
right.
What
we're
actually
doing
is
we're
dating
there.
A
The
information
that
the
configuration
that
usually
we
put
inside
the
application
built
in
and
we
abstract
that
and
now
all
the
network
is
basically
abstract.
So
this
is
pretty
powerful
right
and
we're
great
power
comes
great
responsibility
right,
so
basically
think
about
what
we
just
did.
It
just
took
all
those
network
and
you're
giving
it
to
us
the
service
provider,
to
actually
end
all
that
what
will
happen
if
someone
will
push
their
own
configuration.
Your
data
center
is
done
right,
no
application.
Nothing
is
there.
A
So
the
real
question
that
we
need
to
ask
ourselves
as
a
community
is
how
do
we
make
sure
that
our
service
match
is
resilient,
because
that's
really
really
scary
right.
So
there
is
a
lot
of
things
that
we
can
do
about
it
right
like,
for
instance,
there's
some
stuff
that
pattern
that
already
exists
in
the
market,
for
instance.
First
of
all,
let's
make
it
simple
and
that's
what
we
did
with
SMI
right,
so
Michelle
talked
before
me.
A
This
is
something
that
we
helped
push:
SMI,
something
that
taking
the
configuration
of
the
mesh
and
just
basically
taking
80%
of
this
functionality
and
make
it
damn
simple
for
the
user
to
use
right
that
simple
API
that
was
our
driven.
So
that's
what
we
were
going
to
make
it
over
the
bridge
here,
because,
if
it's
simpler,
it
is
right,
therefore
less
error-prone.
The
second
thing
that
we
did,
we
specifically
in
solo,
announced
like
last
year,
a
service
mesh
hub
service.
A
We
want
to
make
sure
that
what
the
configuration
we're
letting
you
do,
we
kind
of
like
putting
and
well
God
will
write
me
making
sure
that
what
you're
doing
is
actually
make
sense.
Besides
that,
we
actually
also
giving
you
the
ability
to
troubleshoot
it
or
bring
you
all
the
locks
from
everywhere,
including
all
the
way
for
SD
for
invoice
and
actually
bring
it
back
to
you
right.
So
every
service
match
no
matter
which
one
we
can
group
in
together.
We
can
give
them
one.
A
Freedom
is
one,
and
then
we
can
also
make
sure
that
the
configuration
is
not
bad.
So
that's
the
number
second
that
can
actually
help
us.
The
third
one
is
gate,
ops
and
service.
Mashup
is
actually
come
build
in
with
this
out
of
the
box.
This
is
the
ability
of
actually
making
sure
that
the
configuration
of
use
of
your
service
match
it's
actively
being
treated
as
a
actually
code
right.
How
do
we
try
treating
code?
A
If
it's
that
important
to
us,
we
have
a
repository
that
we
storing
there
we
never
pushing
directly
to
MAME
to
a
to
master
right.
First,
we
actually
full
request
and
there
is
reviewers,
and
only
then
it
that's
giving
us
comes
like
the
compliance
right
to
make
sure
that
we're
not
going
crazy
that
nothing
slip.
So
all
of
this
is
actually
really
really
useful
right.
A
That's
to
make
your
mesh
resilient,
but
if
we're
looking
at
all
those
approach,
there's
something
common
between
all
of
them
and
that's
the
ability
that
they
all
driven
by
the
user
themselves
right,
the
user
itself
say
I
want
to
change
my
configuration
now.
Why
does
he
need
to
change
the
configuration
you
need
to
change
the
configuration
because
of
the
environment
right
so
either
it
is
someone
pushed
something
together
or
someone.
You
know
some
met
with
change
in
your
environment
or
some
poor
diet
right.
So
what
do
you
do?
A
You
need
to
change
the
configuration
to
actually
adapt
right
to
the
new
situation.
So,
if
you
think
about
it,
if
you
really
wanted
to
actually
make
your
service
mesh
resilient,
what
you
really
need
to
do
is
to
make
it
adaptive.
That's
what
you
will
in
it,
because
too
be
fair.
Use
are
great
and
I'm
a
user,
but
sometimes
we
making
a
mistake
and
that's
too
dangerous
right
now
with
the
responsibility
that
we
got
okay.
So
how
do
you
make
your
service
measured
up
right
and
to
be
fair?
A
This
is
a
pattern
that
already
exists:
let's
create
operators
right
so
we'll
create
operators
and
actually
going
to
make
sure
that
those
driven
that
it's
going
to
snip
your
environment
and
based
different
of
different
events
that
happening
in
the
environment
is
going
to
change
the
service
match
configuration.
So
there
is
a
lot.
A
lot
of
operator
frameworks
out
there
and
there's
a
lot
and
coming
more
incoming
mode,
but
because
they
are
so
low
level,
because
they're
kind
of
like
all-purpose
operators,
they
make
it
a
little
bit
harder
to
actually
build
them.
A
So
I'm
really
really
proud
to
announce.
Today,
a
new
open
source
called
a
project
called
autopilot,
and
the
idea
with
autopilot
is
basically
a
damn
simple
SDK
that
will
help
you
to
auto,
generate
all
your
code
and
Dragon,
and
basically,
it's
coming
with
hook
up
with
its
very
opinionated
right,
it's
open
yet
for
service
match,
which
means
that
it's
coming
hook
up
to
the
metrics
of
your
environment
to
actually
configure
your
match
and
therefore
it's
built
in
all
the
library
inside.
A
It's
Auto
generate
your
code
and
basically
the
only
thing
that
you,
as
a
user
need
to
actually
write
is
the
brain
is
the
rule
is
what
to
do
one.
So
this
is
an
open
source
project.
We
just
open
source,
it
look
at
it.
It's
really
can
go.
I
think
it's
just
you
know
in
the
nut
shells
operator
is
not
new,
but
we're
always
building
more
and
more
and
everybody's.
A
I
think
it's
it's
something
that
will
help
everybody
to
drive
their
meshes
so
again,
just
that.
Well
summarize,
because
they're
five
minutes
what's
actually
autopilot
shabbat
autopilot.
So
the
first
thing
it's
already
come
building
with
snooping
all
your
matrix
and
an
opinion
metrics
for
the
service
mesh
itself,
it's
going
to
snow
web
works
and
deployment
and
CR
this
basically
right.
The
second
thing
that
is
going
to
do
is
going
to
drive
your
configuration.
So,
for
instance,
for
s,
do
we
build
a
builder
pattern?
We
thought
that
it's
easier.
A
If
you
have
a
builder
button
to
actually
configure
the
mesh,
the
other
evening,
gator
said,
is
either
expose
or
invoke
aware
books.
It's
provided
out-of-the-box
geeks
experience.
So
when
you
actually
write
in
your
rule,
you
can't
you
have
a
flag
and
you
can
say:
do
I
actually
want
it
to
go
all
the
way
and
configure
the
match
or
what
I'm
actually
want
is
first
to
check
in
to
get
and
get
some
approval
for
it
so
kind
of
like
it's
building
with
the
flip
of.
A
If
you
flip
over
of
a
flag
and
that's
what
I
will
say
again,
it's
all
about
the
workflow
right
you're
doing
something
you're
moving
to
the
next
in
the
pneumothorax
in
the
document
and
and
you
making
sure
that
all
these
process
that
operate
or
need
to
do
is
being
done
so
go
check
it
out,
we're
so
low
love
to
get
your
feedback
mainly
and
hear
about
your
use
case
for
building
operator
thanks.
So
much.