►
From YouTube: 013 What's New and Coming to Gloo Edge
Description
Including legacy workloads are a reality of today's cloud-native architectures. Technology like Envoy proxy can simplify application networking but what about for legacy technology or security protocols? In this talk we see how running Gloo Edge in large enterprises has influenced the roadmap and improved the evolution of application architectures.
A
Everyone
thanks
for
joining
us
at
solocon,
thanks
for
coming
to
our
talk
about
glue
edge,
we're
here
to
talk
about.
What's
new
in
glue
edge
1.7,
we
have
the
release
candidate
coming
out
and
then
we'll
have
the
ga
in
a
couple
weeks.
We're
soaking
it
for
a
little
bit
and
we'll
actually
talk
about
what
we're
doing
in
soaking.
A
A
You
can
find
me
on
the
community
slack
the
solo
io
community
slack,
if
you're
not
part
of
that,
yet
go
sign
up
I'll,
have
an
action
item
at
the
end.
So
don't
worry
about
it.
Give
you
the
address,
go
sign
up
and
you
can
find
me
at
gaundered
I'd
love
to
hear
about
what
you're
working
on
the
enhancements
that
you,
your
that
are
required
for
any
production
implementation
of
glue
edge
as
well
as
you
know,
feedback
on
the
talk.
A
If
you
had
something
that
you
thought
was
interesting
or
not
interesting
or
anything
in
between
just
please
please
reach
out
very
friendly.
You
could
also
find
me
on
twitter
at
gone.
Eddie's.
I've
been
working
in
kubernetes
for
over
six
years,
so
I've
picked
up
that
tag.
A
I'm
joined
today
by
kevin
derose
he's
our
engineering
lead
for
glue
edge.
We
have
a
bunch
of
different
teams
each
with
their
own
speciality
for
glue
edge.
Kevin
is
leading
that
up
and
he's
gonna
show
us
some
of
the
demos
of
the
new
features
that
we
have
in
glue
edge.
You
could
find
him
as
well
on
the
community
slack.
A
You
know,
don't
reach
out
as
much
yes
to
work
as
an
engineer
and
manage
a
team
and
get
things
out
the
door,
but
yeah,
say
hello,
and
if
you
see
his
name
pop
up
in
the
in
this,
in
slack,
you
know
who
it
is
now.
You
could
also
find
him
on
github.
So
again,
if
you
have
a
request,
an
enhancement
request
that
you
put
into
the
repo-
and
you
see
his
name
pop
up-
you
again
know
who
he
is.
A
First,
I
wanted
to
start
with
a
poll
I'll
give
you
a
few
seconds
to
fill
this
out.
I
just
want
to
get
a
sense
of
who's
in
the
audience.
So,
first,
if
you're,
currently
a
glue
enterprise
user,
you
know-
let
us
know
if
you're,
currently
an
open
source
user
like
to
know
that
as
well
or
if
you've
never
used
glue
before.
That's
all
right,
I'd
like
to
hear
that
as
well.
So
I'll
give
you
a
few
seconds-
and
this
is
pre-recorded,
so
my
surprise
is
going
to
be
completely.
A
A
What
are
we
here
to
talk
about?
Well,
we
want
to
introduce
you
to
glue
edge.
Some
of
you,
according
to
that
poll,
have
never
used
glue
edge
before
or
new
to
glue
edge
or
just
on
the
open
source
and
maybe
not
even
familiar
with
the
enterprise
version.
So
we
want
to
give
you
a
sense
of
what's
in
it,
what
we've
done?
That's
new
to
glue
edge
1.7
as
well.
A
A
You
want
to
know
that
it's
doing
well,
so
I
want
to
touch
on
that
and
then
what's
new
in
glue
edge,
1.7,
what's
coming
in
1.8
as
well
in
the
roadmap,
we'll
talk
about
those
two
topics
for
what's
new
in
1.7,
we'll
have
some
demos
on
soap,
which
is,
I
know
you
thought
soap
was
deprecated,
but
I'll
talk
about
that,
we'll
talk
about
enabling
soap
and
glue
edge
and
and
how
that's
going
to
be
coming
up.
A
We'll
talk
about
release
hardening
as
well
as
our
integration
with
blue
fed,
which
is
it's
a
tool
that
you
haven't
used
before.
You
should
check
it
out,
it's
something
if
you're
into
it.
If
you
use
glue,
fed
or
been
interested
in
sorry
and
kubernetes
federation
or
interested
in
kubernetes
federation,
solves
many
of
the
same
use.
Cases
but
is
is
completely
hardened
and
product
worthy.
A
But
what
is
glue
edge?
It's
a
next
generation
api
gateway,
an
english
controller
built
on
envoy
proxy
and,
as
you
learned
in
the
keynote
glueage
2.0,
also
built
directly
in
this
istio
native
to
enable
and
secure
traffic
at
the
edge.
A
It
comes
with
a
number
of
exciting
features:
advanced
traffic
management
delegation
federation,
which
we'll
learn
and
go
deeper
on
today,
observability
zero
trust
security,
so
you
have
waff
and
some
features
there:
authentication
and
authorization
integration
policies
like
advanced
rate,
limiting
which
are
key
to
an
api
gateway
and
it's
cloud
native.
So
all
those
skills
that
you've
learned
as
your
cloud
native
architect
or
devops
or
sre.
They
apply
to
glue
edge
as
well.
A
Okay,
let's
touch
on
the
glue
edge
momentum.
How
is
the
community
doing?
How
are
we
doing
in
general,
so
these
are
some
stats
that
you
saw
in
our
momentum
press
release
as
well
as
in
its
keynote,
but
I
wanted
to
address
them
here
as
well.
You
know
glue
edge,
is
our
first
breakout
product
at
solo
I
o,
and
so
a
lot
of
I'm
not
gonna
break
out
glue
edge
in
particular
base.
This
is
this
is
first
old
solo.
A
I
o,
but
I
don't
you
know,
we're
we're
a
private
company,
so
we're
not
going
to
break
out
blue
edge,
but
rest
assured
that
glue
edge
is,
is
you
know,
a
healthy
part
of
these
numbers,
so
we
had
550
bookings,
growth
in
solo
io
over
2019.
So
this
is
the
2020
numbers.
We
had
125
net
dollar
growth
in
renewals.
What
does
that
mean?
That
means
you
have
this
cohort
of
customers
that
joined
in
2019.
A
How
did
they
grow
as
far
as
dollar
spent?
Did
they
increase?
Did
they
decrease?
This
includes
everybody
by
the
way.
So,
if
someone
didn't
renew,
you
have
to
include
those
numbers
or
or
whatnot-
and
this
is
a
very
healthy
number.
This
means
that
the
customers
grew
by
25
percent
from
2019
to
2020..
As
far
as
industry
goes,
this
is
phenomenal.
You
see
numbers
anywhere
between
15
30
30
300
increase
in
new
customers.
Again,
this
is
over
2019.
Whatever
the
customer
count
was
there.
A
We
tripled
that
in
2020
and
I'll
tell
you
that
this
quarter
is
doing
phenomenal
as
well,
and
we
have
a
bunch
of
new
customers
coming
on
this
quarter.
We
have
five
new
fortune,
100
cut
businesses
in
2020
and
two
of
the
top
global
telecom
companies
again
very
exciting.
A
These
are
some
of
the
names
of
our
customers
right
now,
vonage
schneider,
electric
dc,
both
classical
enterprises,
as
well
as
new
companies,
as
well
as
companies
transitioning
small
large.
All
over.
We
run
the
gamut
there's
no
particular
industry.
Where,
where
it's
stand
out,
you
actually
see
blue
edge
across
the
board.
It's
big
small,
traditional
new
enterprise,
new
new
internet
company,
all
over
the
place.
A
So
growth
and
delivery
we've
seen
or
if
you
look
today,
there's
over
1600
people
in
the
very
active
edge
channel.
If
you
go
into
solo,
I
o
there's
3
600
plus
people
in
there
in
the
actual
just
edge
channel
alone,
there's
over
1600
and
it's
very
active.
We
get
a
lot
of
questions
in
there.
It's
almost
like
free
support.
If
you
love
money,
you
know
go
in
there.
If
you
hate
money
and
don't
want
free
support,
don't
go
in
there,
but
you
do
get
that.
A
You
know
help
from
other
users
help
from
sometimes
our
engineers
as
well
hope
for
me.
So
please
join!
Please
join
the
slack
channel.
We've
had
over
400
releases
for
glue
edge.
You
want
to
talk
about
agile,
that's
the
most
agile
thing.
I've
ever
seen
over
400
releases.
That
includes
the
patches,
the
the
gas,
the
betas,
but
we
do
a
release
almost
every
other
day.
A
Yeah
there's
some
overlap
with
service
mesh
and
edge
gateways,
traffic
control,
traffic,
routing,
tls
and
mtls.
There
is
that,
but
to
go
from
istio,
ingress
or
envoy
itself
into
an
api
gateway
is
a
journey,
it's
something
where
you
need
rate
limiting
and
federation
and
security
and
advanced
rate
limiting
and
developer
portals
and
whatnot.
So
it's
not
just
about
the
ingress.
A
It's
not
just
about
what
you
could
do
with
this
dio
and
envoy
there's
a
lot
of
features
that
we
add
on
top
of
that,
and
this
is
just
a
list
of
the
ones
that
that
are
available
today
and
the
the
differences
from
just
using
the
skew
ingress
to
using
glue
edge.
A
Again,
you
see
that
advanced
rate
limiting
that
federation
that
developer
portal,
which,
if
you
haven't,
seen
my
talk
from
from
the
other
day,
I
highly
suggest
it
because
I
go
into
why
you
would
need
a
developer
portal
for
istio
and
then
blue
edge
itself
is
integrated.
Today,
with
this
do
and
glue
edge,
2.0
is
istio
native.
You
saw
that
in
the
keynote
we
have
a
tight
integration
that
we're
currently
working
on
with
with
glue
mesh.
That
is
very
exciting
and
you'll
see
in
the
upcoming
quarters.
A
So,
what's
new
in
glue
1.7,
well,
customers?
First,
we've
delivered
over
40
customer
requests
for
enhancements
in
glue
edge
in
this
release.
That
is
that
is
incredible.
I'm
so
proud
of
the
team
kevin
for
getting
all
this
at
the
door,
but
a
lot
of
those
are
blockers.
A
lot
of
those
were
highs
that
you
wanted
in
those
releases,
and
we
we
got
those
to
you
glue,
fed
integration.
It's
now
one
repo
one
download
for
enterprise.
A
One
platform
we'll
have
further
ux
integration
there,
but
right
now,
if
you're
using
glue
edge
enterprise,
you
could
use
glue,
fed
and
kevin
will
highlight
some
of
the
features
that
are
now
about
available
in
glue
fed
when
you
use
it
with
enterprise.
For
those
of
you
that
are
not
familiar
with
that
portal
enhancements
as
well
again,
if
you're
interested
in
the
deep
dive
on
portal
enhancements
join
my
other
presentation,
I
talk
about
why
you
need
a
portal
for
istio.
Why?
A
A
So,
first,
here's
all
the
features
that
we've
done
enhancements
over
the
past
past
three
months.
There's
a
few
more
that
we're
getting
out
the
door
before
we
actually
cut
the
rc
and
do
a
feature
freeze,
but
you
can
see
there's
a
lot
there
there's
many
blockers
that
you
have
come
to
us
with
there's
a
lot
of
highs
that
you
know
were
something
that
you
really
really
wanted
to
see
within
the
release
and
we
got
those
to
you
as
as
quickly
as
as
we
could.
A
Oh,
instead
of
going
through
each
one
of
those
40
different
line
items,
I
wanted
to
break
them
into
some
synthesis.
Some
analysis
of
what
we're
seeing
as
far
as
the
requests
that
are
coming
in
20
of
the
requests
are
related
to
help
change
this
in
helm.
Add
this
to
helm!
Why
isn't
this
in
the
helm
chart
enterprise
home
chart?
Can
it
be
more
in
line
with
the
community
helm
chart?
A
A
Another
five
percent
of
the
requests
are:
can
you
expose
this
an
envoy,
and
can
you
expose
that
in
envoy
and
so
we'd
like
to
address
that
as
well
in
in
upcoming
releases
and
then
four
percent
of
the
requests
that
are
coming
in
are
about
observability?
Can
you
see
licensing
expirations
in
prometheus
things
of
that
nature?
A
So
what
is
glue
federation?
And
why
is
it
important?
So
you
might
be
familiar
with
kubernetes
federation
and
if
you
are,
then
you
sort
of
get
it.
You
can
use
it
for
federating
of
configurations
across
duluth,
but
the
exciting
thing
is:
you
could
also
use
it
for
failover.
A
You
could
do
a
request
to
one
glue
edge
on
on
one
on
one
cluster
for
an
upstream
service,
but
if
that
service
is
not
there,
it
actually
fails
over
to
the
other
glue
edge
that
is
registered
in
federation
and
gets
you
the
like
service
on
that
other
cluster.
So,
as
a
user
from
your
perspective,
you
never
have
an
outage.
A
You
could
also
use
gluffed
in
order
to
configure
the
r
back
on
clusters
as
well
again,
if
you
are
familiar
with
kubernetes
federation,
then
you're
familiar
with
blue
edge
federation
very
similar
concepts.
I'm
really
going
to
save
you
a
lot
of
time.
Instead
of
federating
or
sorry
configuring,
each
cluster
individually,
you
could
just
use
federation
in
order
to
do
that.
A
It
saves
you,
mishaps
and
configuring,
one
one
way
versus
configuring,
another
another
way
which
you
don't
want
either
portal
enhancements.
What
what's
going
on
in
the
portal
and
again,
if
you're
interested
in
this,
please
join
my
other
talk
from
the
other
day
we
added
grpc
api
types
into
the
portal.
We,
you
know
the
portal
used
with
blue
edge,
as
well
as
the
only
only
portal
available
for
istio
on
the
market.
A
It
is
also
the
only
portal
available
with
grpc
as
an
enterprise
product,
there's
other
open
source
efforts
on
grpc
and
getting
those
into
a
developer
portal,
but.
A
A
We've
done
a
lot
of
work
on
authentication,
so
authentication
to
the
portal
itself,
which
was
in
the
last
release,
but
also
now
authentication
to
the
apis.
You
could
actually
use
your
idp,
not
just
api
keys,
but
user
idp
in
order
to
start
using
the
apis.
A
Super
cool
really
helps
with
self-service,
and
then
we
made
the
apis
available
publicly.
If
you
wanted
to
this,
is
you
know
if
you're
a
user-
and
you
want
to
just
read
about
the
api
documentation
without
signing
in
you
can
think
of
this
as
like?
If
you're,
like
you
know,
you
want
to
do
like
the
spotify
or
or
twitter
and
just
list
your
apis
in
case,
like
you
know,
partner
or
some
user
is
interested
in
in
using
them
and
then
in
order
to
use
them,
you
actually
sign
in
and
you
get
access
same
method
here.
A
Okay
release
hardening.
What
do
we
do
here?
Well,
we've
had
the
release
aren't
anything
in
our
all
our
releases,
but
now
we
have
it
within
cdn.
We
did
a
lot
of
improvements.
One
is
that
we
do
long-term
testing
soaking
above
betas,
as
well
as
the
least
candidates
for
ga.
A
So
basically
every
release
we're
soaking
right
now
that
helps
us
find
memory
leaks
that
helps
us
find.
You
know
if
logs
fill
up
over
time-
and
you
run
out
of
storage
things
of
that
nature-
things
that
you
can't
find
in
unit
testing
or
or
anything
like
that.
So
we
let
just
let's
sit
around,
we
give
it
a
customer
load,
a
workload,
so
it
simulates
how
you
use
it
directly
and
we
see.
If
anything
happens,
we
also
have
vulnerability
scanning
which
kevin
will
show
as
well.
A
This
is
now
in
our
cd
and
it
allows
you
to
or
allows
us
to
see
if
a
pull
request
actually
added
a
vulnerability
and
address
it
right
away
and
then
license
scanning
very
similar,
we're
scanning
our
you
know,
licensing.
So
every
pull
request.
We
we
see
if
there's
a
gpl
violation,
so
that
you
could
rest
assured
that
our
software
is
compliant
with
the
open
source,
licensing
of
the
components
that
we're
using
and
in
large
scale
testing.
This
is
something
that
we're
going
to
do
in
q2.
Not
doing
yet.
A
Obviously,
envoy
itself
is
large
scale
tested,
glue,
edge
is
used
at
large
scales
within
the
enterprise
and
we're
seeing
some
of
our
customers
in
this
conference.
Talk
about
that,
but
we'd
like
to
test
it
in
every
single
release
at
the
scales
that
our
customers
run
at,
and
so
this
is
something
that
we're
working
on.
A
Okay,
so
wait
isn't
so
deprecated?
No,
it's
not
deprecated.
These
are
some
of
the
statistics.
I've
seen
out
there
about
soap
in
the
enterprise.
Now,
if
you
look
at
the
larger
api
usage,
so
rest
is
about
61,
but
this
is
just
within
the
enterprise
about
80
percent
of
apis
are
a
rest,
but
that
leaves
about
15
that
are
still
using
soap
and
in
some
organization
in
some
teams
it's
almost
entirely
soap.
A
We
wanted
to
give
that
path
that
that
way
to
get
from
where
you
are
to
cloud
native
to
you
know
transitioning
to
to
not
just
rest
but
grpc
and
graphql
all
within
the
same
product.
All
within
the
same
portal,
so
embracing
soap
is
part
of
that.
We're
not
saying
oh
it's
all
or
nothing.
You
could
actually
take
this
stepping
stone
to
that
cloud.
A
Native
journey,
okay,
so
what's
in
glue
edge,
what's
in
the
enterprise
I
just
wanted
to
highlight
here
that
portal
is
is
now
in
the
the
enterprise
glue
edge,
so
that
is
part
of
the
product
as
well
as
glue
federation
and
we'll
see
that
quite
shortly,
kevin
would
demo
that,
and
so,
with
that
I'll
hand
it
over
to
kevin
to
give
us
a
quick
demo.
B
Awesome,
thank
you.
Thank
you,
chris
very
excited
to
show
you
show
you
all
some
of
the
things
that
we've
been
working
on
here
at
solo
on
the
ledge
team.
I
think
you
know
a
lot
of
it
should
be
a
lot
of
fun,
really
really
exciting
stuff.
B
So
there's
kind
of
two
different
demos
I
wanted
to
to
run
through
today,
the
first
one
kind
of
the
namesake
of
the
talk,
or
it's
part
of
it-
we're
talking
about
xslt,
bringing
soap
to
cloud
native
and
bringing
so
up
to
cloud
native
by
that
we
mean
xslt
and
envoy
so
xslt,
if
you're
not
familiar
as
a
transformation.
B
Language
for
xml
and
soap
requests
are
built
on
xml
and
that's
not
natively,
provided
in
envoy,
where
you
build
our
own
builds
of
envoy,
so
we're
building
xslt3,
so
the
newest
xslt
built
on
saxon
and
we're
supporting
that.
So
I
want
to
quickly
run
our
build
of
envoy
and
show
off
some
of
that
functionality.
B
B
This
is
probably
going
to
look
pretty
familiar
and,
if
not
I'll,
try
to
run
through
it
here
with
you
real
quick
first
thing
here:
we've
got
a
static
listener,
so
we
can
see
that
we're
listening
on
zero,
zero,
zero,
zero
port,
eight
thousand
listening
for
requests
and
forwarding
them
onward.
B
You
can
see
that
we've
got
your
standard,
http
connection
manager,
filter
here,
so
we're
receiving
http
requests.
The
part
that's
we're
doing
here
is
we're
matching
every
single
domain
star
domain,
we're
matching
routes.
That
begin
with
the
forward,
slash
prefix,
so
all
requests,
and
we
send
basically
all
requests
to
this
echo
service,
which
the
config
lives
at
the
bottom
of
this
file.
I'll
show
you,
but
it's
an
extreme
service
that
just
responds
with
what
I
got
so
an
echo
service.
B
The
more
interesting
part
here
that
the
xslt
related
configuration
lives
here
in
the
I
o
solo
transformation
filter.
This
is
something
we
run
and
we
support.
I
know
we
asked
earlier.
Some
of
us
are,
probably
you
know:
glued
users,
glitch
enterprise
users,
and
some
of
us
have
never
heard
of
gledge
before
we
here
at
solo.
Do
our
own
filters
and
envoy,
in
addition
to
the
ones
that
envoy
typically
shifts
with
ships
with
one
of
the
most
popular,
is
our
transformation
filter.
B
This
is
one
that
allows
you
to
transform
from
json
and
rewrite
requests
and
the
request
body,
the
headers
things
of
that
nature.
Take
information
from
onboard
dynamic
metadata,
it's
one
of
the
most
popular
powerful
filters.
We
have
today
that
lives
in
open
source,
we're
adding
to
the
enterprise
the
ability
to
transform
xslt.
B
So
here
you
can
see
that
we're
matching
every
single
prefix
gets
a
transformation.
That
transformation
is
going
to
be
an
xslt
transformation
which
lives
here,
and
this
is
if
you're
familiar
with
xslt
just
a
standard
or
like
you
can
put
any
xslt
string
here
and
then
this
will
be
transformed
on
the
request
path.
B
What
we're
doing
here
is
if
you're
unfamiliar
with
xslt
we're
just
doing
xml
to
json,
so
we're
going
to
take
an
xml
request
on
the
request
body
and
rewrite
that
as
json
and
send
that
his
new
request
body
is
json
to
the
upstream
and
then
before
we
talk
about
the
echo
service,
we're
just
going
to
route
to
this
website
http.org,
and
this
is
going
to
show
us
what
I
got
and
so
just
quickly
demonstrate
we'll
go
ahead
and
I'll
run.
B
So
we're
going
to
run
our
build
of
envoy
here.
This
is
the
xslt
demo
we're
going
to
expose
port
8000
on
the
local
machine.
That's
where
it's
listening
and
we're
going
to
run
with
the
bootstrap
that
I
just
showed
and
we're
going
to
run
on
trace
level
debug
trace
level
logs.
B
So
this
should
look
familiar
if
you've
used
android
before
looked
at
on
boy
logs,
you
can
see
all
you
know
real
live
envoy
here,
and
then
this
is
the
request
that
I
want
to
make.
So
here
I'm
just
gonna,
I'm
echoing
the
xml
request
and
then
we're
gonna
post
it
to
that
to
envoy.
So
looking
at
the
xml
we're,
basically,
if
we're
talking
about
car
prices,
we
got
four
door
cars
that
cost
six
l
whatever
that
means,
and
five
door
or
hatchback
cars
that
cost
13
l.
B
This
is
centered
xml
and
we're
going
to
hit
envoy
on
localhost,
and
if
we
make
this
request,
then
it
hits
http.org.
We
can
see
in
the
form
that
this
was
transformed
with
xslt
to
json,
and
so
you
can
see
this
is
val
json.
If
you,
you
know,
render
out
the
new
lines
and
some
of
the
formatting
here.
B
And
that's
that's
it
for
the
xslt
demo,
like
I
said
this
will
be
live.
You
know
hitting
one
of
the
eight
1.8
betas
very
soon
and
live
in
in
1.8.
B
Additionally,
I'd
like
to
demonstrate
some
of
the
new
glu
fed
features
that
we've
been
working
on,
because
there's
a
bit
of
setup
involved
and
blue
federation
involving
several
clusters
and
glue
installations.
I've
taken
the
liberty
of
bootstrapping
a
lot
of
that
environment
for
us
all
I
did
is.
I
ran
this
command
glue,
cuddle
demo
federation,
and
this
will
set
up
the
environment
for
you
and
then
you
can
do
what
I'm
doing
right
here.
B
I'll
walk
through
what
the
environment
looks
like
right
now.
The
first
thing
that
I
wanted
to
show
is
that
we
have
two
clusters,
so
we
have
a
local
cluster
and
a
remote
cluster,
and
if
we
want
to
go
ahead
and
get
the
pods
in
both
clusters,
we
can
see
that
we've
been
bootstrapped
with
in
the
remote
cluster.
B
B
The
words
green
pod-
that's
it
so
we're
doing
like
a
blue
green
deployment
if
we
get
the
same
pods
but
in
the
local
we'll
see
that
we
have
the
blue
deployment
as
well
as
a
glue
installation,
so
the
blue
deployment,
we
have
our
blue
installation
and
then
we
have
blue
fed
itself,
the
glue
fed
controller,
and
so
we
want
to
talk
a
little
bit
about
what
bluefed
is
doing
and
how
it
pieces
together
here,
but
so
the
environment's
been
set
up
and
bootstrapped
already.
B
In
addition,
we
have
already
registered
the
clusters,
so
by
that
I
mean
gluefed
is
controller.
Its
job
is
to
take
federated
resources
and
write
them
and
place
them
on
all
of
your
glue
clusters.
So
if
you
have
you
write
a
federated
virtual
service
and
you
place
that
on
one
cluster
or
both
clusters,
it
gets
written
by
the
fuku
fed
controller.
To
both
to
do
that,
you
need
to
register
the
clusters.
Those
have
been
done
already.
B
Let's
take
a
look
at
some
of
the
federated
configuration
that
already
exists,
so
first
we
have
a
federated
upstream.
We
can
see
that
this
federated
upstream
resembles
very
much
a
glue
upstream.
If
we
look
at
the
metadata,
this
is
our
our
blue
service
on
port
10
000.
It's
got
envoy
health
checks,
which
will
become
relevant
later
for
this
demo.
We
want
to
fail
over
from
the
blue
service
to
the
green
service,
pretending
that
the
blue
service
is
unhealthy
and
then
the
relevant
for
blue
fed
is.
B
B
B
So
we're
matching
all
domains
matching
to
the
forward,
slash,
prefix
and
routing
to
the
blue
pod,
and
we
can
see
that
this
was
also
placed
placed
on
the
kind
local
cluster
and
then
likewise,
we
could
get
virtual
services
and
simple
route
does
indeed
exist
here,
and
it
looks
a
lot
just
like
what
we
had
written
before
and
the
last
thing
we
we
have
for
setup
already
is
a
service
failover.
So
we
have
defined
a
failover
scheme
in
blue
fed
already,
and
this
failover
scheme
is
talking
telling
us
how
to
do
failover.
B
So
we're
saying
if
we
we
have
in
our
primary
cluster
kind,
local,
which
is
where
the
blue
pod
lives
and
also
where
blue
fed
lives
right
now
in
our
demo
environment,
we're
routing
to
the
the
blue
service,
the
blue
system
blue
service,
but
if
we
were
to
fail
over
so
this
were
to
start
failing
health
checks
for
whatever
reason,
then
we
would
route
to
the
green
service
and
this
green
service
lives
in
kind,
remote,
so
a
totally
separate
kind
cluster
altogether,
and
just
to
demonstrate
that
all
this
behavior
real
quick,
I
just
want
to
run
through
first,
let's
port
forward
to
this
gateway
proxy
service.
B
B
So
if
we
do
that
we're
routing
everything
to
blue
pod,
you
can
see
that
we're
always
getting
blue
pod
because
it's
healthy
and
it's
passing
health
checks.
B
We
can
see
the
glue
once
registered,
like
registering
the
ability
to
read
those
those
clusters
you
register,
fed
with
clusters.
It
has
the
ability
to
discover
glue
instances.
So
if
you
had
more
than
one
glue
instance
say
you
installed
in
namespace
mode,
woofed
is
responsible
and
has
these
blue
instances
crds?
We
discovered
all
those
for
you
automatically.
B
The
last
thing
that
that
we
want
to
run
by
for
the
1.7
release
is
to
show
some
of
the
things
we've
made
in
release
hardening.
So
first,
we've
mentioned
before
we're
doing
open
source
attribution.
Now
we
can
look
at
the
licenses
that
we've
in
all
of
our
dependencies.
B
This
is
integrated
in
auto,
generate
as
part
of
our
ci
cd.
This
is
important
because
we
want
to
avoid
using
things
like
gpl
licenses.
You
know
those
copy
forward
and
you
know,
as
you
can
see,
what
we're
using
today
and
we
get
notified
as
part
of
csed.
If
we
were
to
try
to
bring
one
of
those
dependencies
in
likewise,
we
are
doing
security
updates.
So,
for
example,
we
do
scans.
B
Nobleman
is
found
on
the
latest,
but
if
we
were
to
look
at
say,
release
1.6
like
an
older
one,
we
could
see
cves.
All
of
our
assets
are
being
scanned
for
all
known,
cves
part
of
independent
third-party
tools,
and
this
is
what
it
would
look
like
here.
We're
also
additionally
doing
soak
testing
a
lot
of
that's
internal,
but
I
don't
take
too
much
more
of
our
time
here
or
your
time
here.
We
just
thank
you
all
for
joining
and
with
that
I'll
hand,
it
back
to
chris.
A
Thanks
kevin
okay,
I'm
gonna
cover
the
road
map
a
little
bit
into
question
time,
but
you
can
obviously
find
us
on
slack
and
ask
questions
there,
and
I
just
want
to
leave
a
few
minutes
for
questions
at
the
very
least.
So,
let's
cover
the
road
map
for
1.8
and
beyond.
A
A
Rest
and
soap
came
on
on
on
the
on
into
the
market
in
the
basically
the
dot
com
error
and
got
quick
adoption
on
rest,
and
that
still
continues
to
be.
You
know
very
popular,
that's
not
going
anywhere.
All
aws
services
have
a
restful
api,
for
example,
but
as
happens
in
technology,
what
what
the
use
cases
you
start
to
see?
Larger
companies
figure
out
that
maybe
they
can
do
something
a
little
bit
differently
to
do
this
better
at
scale
or
to
address
this
edge
use
case
or
to
use
just.
A
Do
it
a
little
bit
differently,
so
it
accelerates
adoption,
and
so
that's
exactly
what
happened,
and
when
this
has
happened
in
the
past,
you
see
a
real
jump
in
adoption,
like
you
could
think
of
a
containers
when
kubernetes
came
out
the
door
and
you
just
see
multiple
organizations
start
to
adopt
it.
You
can
think
of
this
as
big
data
or
with
machine
learning
with
tensorflow
when
those
tools
are
released.
A
Not
only
do
you
start,
you
continue
to
use
the
the
existing
tools,
but
you
have
new
ways
of
addressing
your
issues,
your
business
case,
and
so
you
really
see
accelerated
growth
in
that
adoption
curve,
and
so
with
with
api
types
recently
grpc,
which
came
out
of
google
and
is
now
managed
by
the
cncf,
the
cloud
native
computing
foundation
and
graphql,
which
came
out
of
facebook.
These
addressed
some
of
the
issues
that
they
saw
in
in
in
rest,
especially
graphql,
and
so
you
see
that
jump.
A
What
are
the
other
initiatives
that
we
should
talk
about
that
we're
going
to
be
working
on
glue
edge?
The
majority
of
our
time
is
going
to
be
still
on
customer
and
community
requests
things
that
come
in
that
you've
asked
for.
This
is
something
that
things
that
we
know
we
need
to
address
in
the
platform.
So
we
treat
those
very
seriously
and
they
gravitate
towards
the
up
of
the
priority
list.
A
We
also
want
to
increase
the
api
types,
and
I
talked
a
little
bit
about
how
we're
working
on
graphql
and
we're
working
on
grpc
and
adding
that
to
the
portal
in
in
latest
release
glue
fed
integration.
You
saw
you
saw
that
today
we
it's
almost
100
there,
there's
some
ux
things
that
we'd
like
to
do
in
the
future:
around
glue,
fed
and
glue
edge
and
the
integration
helm
changes.
A
This
is
doing
hell-
maybe
a
little
bit
differently,
so
that
we
could
have
the
team
work
on
some
more
interesting
things
and
adding
things
to
a
helm,
chart
aha
and
other
architectures
and
then
exposing
bass
tech.
So
by
exposing
bass,
tech
I
mean.
Can
I
expose
this
in
envoy
and
eventually
istio
versus
you
know
doing
this
directly?
A
A
Okay,
what
are
the
next
steps?
What
you
should?
What
should
you
do
now?
You,
like
this
talk,
go
to
download
glue
edge
1.7
the
release
candidate
should
be
available
when
you
get
home
we're
all
at
home,
but
when
you
get
home
after
the
conference
go
download
the
1.7
and
then
in
a
couple
weeks
that
will
go
ga
when
we
feel
that
we've
addressed
the
number
of
the
bugs
or
some
of
the
feedback
that
we
got
from
you,
the
user
join
the
slack
channel
at
solo.
I
o
not
just
the
slack
for
solo.
A
I
o,
but
the
actual
glue
edge
channel
go
join.
It
say
hello,
I'd
love
to
see
you
there
and
then,
if
you're
new,
to
glue
edge
if
you've
never
seen
it
before,
and
you
want
a
more
basic
introduction
to
glue
edge,
go
sign
up
for
the
workshop.
It's
hands-on
it's
the
last
day
of
sologon,
it
might
be
filled
up,
but
join
the
waiting
list
and
we'll
have
future
glue
edge
workshops.
I
I
promise
you
okay
with
that.
I
thank
you
for
coming.
Thank
you
for
coming
to
solocon.
A
Thank
you
for
coming
to
this
talk
we're
still
in
a
pandemic.
You
can
see
I'm
working
from
my
home
in
in
lovely
new
york
city,
so
stay
safe,
get
vaccinated!
Please
and
we'll
now
take
your.