►
From YouTube: OpenFaaS Cloud + Linkerd: A Secure, Multi-Tenant Serverless Platform - Charles Pretzer & Alex Ellis
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
OpenFaaS Cloud + Linkerd: A Secure, Multi-Tenant Serverless Platform - Charles Pretzer, Buoyant & Alex Ellis, OpenFaaS, LTD
In this talk, maintainers of the OSS projects OpenFaaS and Linkerd discuss and demonstrate combining the features of each to build a scalable and secure multitenant serverless platform. OpenFaaS is a lightweight serverless platform built on Docker and k8s. Linkerd is an ultralight service mesh for k8s. The projects share philosophy around simplicity, ease of use, speed, and low resource impact. OpenFaaS was designed for small teams and tackled multitenancy by creating OpenFaaS Cloud to layer on top. Scaling for multitenancy requires more robust networking, encryption, detailed metrics, and load-balancing strategies than Kubernetes L4 can offer. Linkerd fits because it provides the features listed above required for scaling multitenancy while remaining focused on simplicity, security, and performance as a part of a modular platform built around OpenFaaS.
https://sched.co/Uabn
A
So
we're
here
to
talk
about
open
fires
and
linka
D.
We
explore
them
a
little
bit
separately
and
then
why
we've
combined
them?
Why
we
think
they're
powerful
was
a
combination
and
then
I
want
to
show
you
open
fires
cloud
and
what
that
can
bring
to
open
VARs
into
teams,
but
we
also
had
to
do
a
bit
of
work
around
security
and
that's
involved
Cuba
Nettie's
primitives,
but
also
some
stuff
we
had
to
build
in
as
well
once
you've
been
through.
A
That
will
then
look
at
you
know
what
is
the
future
work
where
we
can
make
more
use
of
linka,
D
and
potentially
other
service
meshes,
and
what
do
we
want
to
achieve
through
open
vast
cloud
going
forward?
Ok,
so
it's
at
open
fires
and
at
Lincoln
II,
if
you
want
to
tweet
take
photos,
would
be
much
appreciated.
A
Now
open
fire
started
in
2016
I
don't
know.
Has
anybody
here
used
at
all
heard
of
it
before
yeah?
Oh
wow,
great,
it's
good
to
have
you
here
so
I
won't
bore
you
too
much
with
the
details,
but
I
wanted
to
kind
of
have
a
portable
serverless
code
and
function
experience.
I.
Could
run
on
containers,
and
that
was
really
the
catalyst
for
it
and
from
there
it's
grown
massively
I
mean
there's
2,000
people
on
slack
now
and
over
250
contributors.
A
A
So
that's
a
confess.
It's
a
kind
of
context
set
why
we
need
open
fires.
We
could
look
at
cloud
functions
and
you
know
I'm
calling
this
serve
less
1.0
the
original,
and
you
know
they
pioneered
some
things
and
there's
a
great
experience
here
and
if
you
use
this
and
it
suits
your
needs,
you
should
probably
stick
with
it.
But
the
problem
that
we
found
was
the
the
programming
language.
A
Let's
say
no
js',
the
signature
of
that
function
would
be
different
in
different
platforms,
and
so,
if
you're
going
to
migrate
something
you're
going
to
struggle
there,
the
zip
file
is
not
the
best
kind
of
packaging
format.
That's
out
there
there's
certain
things
that
we
can
do
better.
For
instance,
if
you
want
to
build
a
graphical
library,
you
can
have
to
download
the
Amazon
Linux,
build
against
it
and
upload
that
and,
as
we
all
know,
cube
con
docker
containers
kind
of
set
the
way.
A
You
then
also
going
to
run
into
different
problems
when
you
move
something
like
GCP
and
you
have
different
regions
available
a
different
API
and
there
are
some
efforts
to
try
and
consolidate
it.
But
you
know
in
the
enterprise
we
don't
really
have
an
answer
and
when
it
comes
to
micro
services,
you're
kind
of
stuck
with
that
as
well,
you
can't
lift
and
shift
a
big
PHP
laravel
application
into
one
of
these
easily
right,
and
so
that's
what
a
mning
to
do
is
kind
of
find.
A
solution
for
that
and
solution.
A
A
If
you've
got
a
binary
like
bash,
we
have
a
wrapper
called
the
watchdog
and
you
can
put
that
in
the
container
and
have
that
give
the
HTTP
interface
right,
and
this
is
the
important
thing
that
lets
us
have
that
portability
that
we
didn't
have
before.
We
also
can
run
that
anywhere.
Now
it's
important
for
debugging,
but
also
for
consistency
to
be
able
to
have
that
same
experience
on
this
laptop
as
I
have
on
Amazon
right
in
the
cloud,
and
that's
something
that
we
struggled
with
1.0.
A
A
B
Having
deployed
quite
a
few
applications,
not
necessarily
a
nodejs
application,
but
we
have
to
write
our
our
code
to
serve
the
pages.
So
here
we
see
an
index.html
page,
so
I
know
that's
30
lines
of
code.
Okay,
we
have
to
have
our
file
to
manage
our
dependencies
so
another
24
and
that
way,
okay
and
we
have
to
have
a
handler
to
handle
the
requests.
And
then
we
have
to
have
a
service
resource
to
deploy,
to
kubernetes
the
deployment
resource
to
deploy
to
kubernetes
and
wienies
deployment.
A
B
A
What
if
there
was
an
easier
way,
what
if
we
could
think
back
to
object-oriented
programming
and
this
idea
of
abstraction
and
do
not
repeat
itself,
and
that's
what
we've
done
here
now-
those
llamo
files
we'll
come
back
to
you,
but
the
docker
file
and
the
index
have
been
abstracted
away
into
a
template
and
that's
what
we
had
to
write
three
lines
of
code.
Okay,
now
you
can
have
a
docker
file
as
well.
We'll
get
to
that
and
this
yam
will
file
here
is
an
open
file
or
file.
It
can
be
a
CID.
A
So
if
you're,
using
git,
ops
or
flux-
that's
great-
you
don't
have
to
abandon
that.
But
here
you
can
just
set
your
memory
limits
and
secrets.
Environmental
variables
can
figure
it
very
easily,
so
docker
files
are
fair
game
like
I.
Want
you
to
take
this
home.
If
you
don't
know
this
about
open
fires,
you
can
get
a
ruby
on
rails
app
and
if
you
listen
on
the
certain
port
you
can
deploy
it
in
a
professor
you'd
be
good
to
go
now.
A
Some
of
the
customers
and
end
users
here
use
it
as
a
pure
service
functions
framework
and
others
use
it
as
a
container
platform
DX
a
company
in
Norway
and
they're,
one
of
the
leading
brands
for
e
ticket
sales.
What
DX
do
is
they
have
a
PHP
laravel
monolith
is
huge
and
they
really
didn't
want
to
break
it
apart.
So
they
wrote
a
docker
file
for
it
put
it
in
a
git
repo,
and
now
it's
on
open
for
some
cuban
Hetty's,
really
that's
simple.
A
Live
person
use
open
files
purely
as
a
farce
and
they
embedded
it
into
their
UI,
so
that
customers,
when
they're
chatting
to
operators
they
can
have
code
flows
execute.
So
if
you
are
buying
a
pair
of
Nike
shoes
and
the
customer
rap
is
rude
to
you
and
you
rate
them
bad,
it
can
send
an
email
to
their
boss
afterwards,
right,
for
example,
and
you
as
a
customer,
can
write
the
code
and
extend
their
platform
vision.
Banco
got
a
lot
of
stability
by
moving
to
Cuba
Nettie's,
and
it's
not
all
due
to
open
fires.
A
B
Lincolni
well,
thank
you.
First
of
all,
for
letting
us
be
part
of
the
open,
faz
experience.
You're
welcome.
Linker
D
is
the
only
service
mesh
in
the
ciencia
as
a
project.
There
are
other
pieces
of
service
meshes
in
Open
de
ciencia,
but
linker
D
is
wholly
donated
to
the
CNC
F
and
that's
important
to
us,
because
we
appreciate
the
code
of
conduct
and
the
open
governance
that
goes
along
with
that
link
to
you
is
meant
to
be
transparent,
easy,
zero,
config
and
easy
to
install.
B
So
in
under
60
seconds,
you
can
install
the
linker
D
control
plane
begin
to
deploy
your
proxies
and
immediately
get
metrics
out
of
that
metrics
that
are
actionable,
that
you
can
get
valuable
information
from
it's
love
by
engineers.
We've
got
a
huge
community,
not
unlike
open
faz,
and
that's
why
I
think
these
two
projects
are
complementary,
not
only
technologically,
but
also
from
a
social
perspective
as
well
and
along
with
the
metrics
you
get
these
deep,
runtime
Diagnostics
so
that
you
can
debug
your
application
in
real
time.
A
And
it's
great
to
know
that
this
is
not
only
an
incubating
project,
but
there's
many
production
users
Bosa
both
of
linka,
d1
and
2.0,
and
the
team
are
absolutely
great
I,
really
love
getting
involved
in
the
community.
I've
been
to
some
of
the
online
calls,
as
super
organized
really
well-run,
very
welcoming
Thanks.
A
So
you
know
like
why.
Why
would
I
want
to
look
at
Lincoln
II?
There's
all
of
this
stuff
about
service
mesh
people
are
upset
about
it.
People
are
excited
about
it.
Well,
it's
look.
Linka
D
is
very
lightweight
and
easy
to
use
and
don't
get
me
wrong.
There
are
users
of
open
bars
that,
like
yes,
you
too,
but
I
just
found
it
easier
to
use
and
I,
particularly
like
the
traffic
splitting
capability.
This
is
where
we
can
do
Canaries
and
progressive
deployments
without
changing
open
files
and
I.
A
Think
that's
why
the
cloud
native
computing
foundation
is
so
great.
You
don't
have
to
rewrite
everything
yourself.
You
can
tap
into
this
existing
code
and
communities.
We're
actually
going
to
look
at
me
too
TLS
in
the
dashboard
today
and
they're
also
very
important,
particularly
for
customers
that
might
be
running
in
environments
where
they
need
compliance,
link,
ADIZ,
dashboard,
he's
really
great
I
mean
if
you
install
it's.
A
Probably
the
first
thing
you
see,
and
here
I've
got
a
topology
of
what's
in
the
open
files,
namespace
with
open,
fast
cloud
and
if
you've
deployed
open,
vag,
you'll
notice,
there's
some
extra
stuff
here
and
we'll
cover
that
in
a
second.
So
what's
the
difference
between
open
files
on
open
files?
Cloud
opens,
as
is
this
platform
I
just
introduced
to
you,
but
quite
often
somebody
like
the
person
that
sent
to
coop
con
like
Carson,
for
instance,
or
one
if
you
have
a
guy
or
girls,
will
be
really
into
containers,
really
get
it.
A
But
you
get
back
to
the
office
and
you
like
I've,
got
this
thing's
great
and
they're
like
what's
docker.
What's
kubernetes,
have
you
done
uck
ad
yet,
and
so
what
we
do
is
allow
open
fires
to
answer
that
question
and
allow
somebody,
it's
super
technical
platform
engineer
to
come
back
and
deploy
this
for
their
team.
A
Anybody
that's
familiar
with
guitars
are
going
or
flux
will
know
that.
There's
this
idea
of
pushing
code
into
a
github
repo
and
getting
an
endpoint
out
the
other
end.
And
that's
what
we've
got
here.
We
can
make
our
code
into
a
git
repo
and
there's
an
integration
between
the
get
provider
and
some
code
web
book
receiver
it
receives.
The
web
hook
validates
it
clones
to
set
repo,
builds
a
docker
image.
A
That's
our
artifact
on
kubernetes
puts
it
in
a
registry
and
then,
in
this
instance
we're
going
to
deploy
it
on
the
open,
faced
gateway
for
you.
At
that
point,
it
will
show
up
on
your
dashboard.
So
this
is
a
conceptual
overview
and
then
we're
going
to
dig
in
a
bit
more
depth
at
each
part
of
that
pipeline.
A
But
the
features
that
you'll
get
is
the
CIN
CD.
So
you
commit
to
master,
it
will
build
the
code,
it
will
deliver
an
artifact.
You
have
an
endpoint
just
like
that.
Github
commie
support
it
and
get
labs.
So
if
you're
self
hosting
get
get
lab
is
another
option
for
you,
you
can
have
runtime
secrets
encrypted
with
bit
Nami
sort
secrets
and
you
can
also
get
a
dashboard.
A
I
will
fend
off
means
that
we
can
have
these
permissions
for
teams
so
that
you
can't
see
my
functions
but
for
in
the
same
organization
we
can
share,
and
then
you
both
have
logs
for
builds.
But
recently
a
chap
called
Alistair
in
the
UK
came
up
in
the
github
and
it
slack
and
it's
like
I
really
want
to
get
involved.
So
we've
got
all
this
stuff.
Can
you
add
logs,
run
I
must
confess
cloud,
and
he
did
it
in
about
a
week.
A
Source
yeah,
so
you
can
try
the
community
cluster
where
you
can
run
the
bootstrap.
This
is
what
it's
going
to
look
like
when
we
see
the
commits
detected
on
get
absolutely
integrated
and
then
we'll
get
a
dashboard
with
what
we've
deployed
now.
What
do
we
want
to
do
today?
Do
we
feel
like
a
micro
service
with
a
docker
file,
we'll
run
a
function
so
docker
file?
No
Jess
come
on
I've
wanted
to
do
that.
One
Oh
Taylor
actually
is
had
it
already.
Okay,
so
we're
doing
a
function
right.
A
And
you
notice
I,
don't
have
docker
running
on
my
machine.
Just
yet
Fasil
I
store
less,
we
got
a
template
store
and
that's
the
function
store.
We
want
the
template,
store
list
and
we'll
see
all
the
different
options
there
I'm
going
to
pick
nodejs
12
fast,
you
light
em
plate
store,
pull
no
12
and
that's
going
to
come
from
a
github
repo
template
store
whatever
got
wrong.
Oh.
A
A
A
So
all
up
and
running
now
and
I
can
look
at
the
code.
Remember
the
stack
yeah
Mel
is
that
minified
version
of
what
you'd
normally
be
deploying
you
can
have
a
CID.
If
you
want
it
and
then
I've
got
my
code
there,
it's
just
going
to
say
what
the
input
was
received.
Taco
instead
and
remember,
we've
don't
have
anything
running
on
a
local
computers
and
no
docker
and
I'm
just
going
to
rename
this
to
stack
yeah
mall,
that's
how
we
like
it
initial
origin/master,
okay.
A
So
in
a
short
period
of
time,
we'll
see
a
git
commit
status
appearing
here
and
it's
gonna
tell
us
what
is
discovered.
So
you
have
more
than
one
in
that
repo.
If
you
want-
and
this
says
github
actions
but
its
github
checks
and
statuses,
we're
going
to
get
the
locks
and
I
can
click
on
this,
and
if
I
had
a
unit
test,
for
instance,
that
failed,
it
will
block
a
builder.
Tell
me
Warren
wrong.
The
logs
are
now
here
and
that
was
11
seconds
from
git
push.
A
A
If
we
get
time,
I'll
show
you
the
docket
that
sorry,
okay,
someone's
taking
a
photo
there,
you
go.
If
we
get
time,
I'll
show
you
the
dockerfile
version
as
well,
which
I
which
I
really
like.
Now
it
doesn't
end
there,
because
you
thought
that
was
the
end
of
the
demo.
No
I'm
gonna
show
you
my
dashboard.
I
saw
your
dashboard,
and
this
is
an
organization.
Now,
if
I
open
an
incognito
tab
and
try
to
log
in
it
will
use
two
factor
auth.
A
A
A
He
would
have
used
lambda
if
it
was
suitable
for
he
wanted
he's
using
open
fire
and
eks,
but
he
needed
support
for
ECR
and
so
through
open
files
limited
being
able
to
implement
that
for
him,
and
now
we
can
push
into
ECR
they're,
really
great
customer
and
learning
a
lot
getting
a
lot
of
feedback
as
we
go
along
and
if
you're
building
open
source
I
think
it's
really
important
to
get
that
kind
of
feedback.
Now
the
other
side
of
open
fosters
community
centricity.
If
someone
can
take
a
photo
of
roger
at
will
be
very
happy.
A
Yes,
he
asked
for
that.
Roger
works
for
Proofpoint
and
he
has
his
bot
and
we
heard
from
Kelsey
about
this
idea
of
inclusivity
and
being
able
to
contribute.
There's
this
theme
of
everyone,
being,
you
know,
accepted
a
coupon.
Well,
what
better
way
to
get
involved
in
finding
a
good
first
issue
to
work
on
Microsoft,
Google,
Doc
jet
stack
opened,
perhaps
even
buoyant
at
some
point-
are
using
this
bottle.
A
You
need
for
secrets,
was
Fewster
and
yeah
you're
right,
and
you
really
don't
want
to
put
that
in
there
get
repo,
even
if
it's
private
right
so
we've
using
sealed
secrets
from
bitten
ami
and
that
allows
us
to
encrypt
it
one
way
when
it's
attached
in
the
cluster,
you
can
get
read
back
again:
ok,
so
a
little
bit
more
technical.
Now,
let's
look
at
the
pipeline.
There's
two
things
that
github
will
do
here.
There's
no
API
for
open
fast
clouds
at
all.
You
can't
delete
a
function.
You
can't
create
one.
A
A
So
that's
the
CIC
DV
bit
of
it.
The
second
part
of
it
is
the
dashboard,
and
you
see
these
are
actually
functions
hanging
off
the
open
fire.
So
someone
said
to
me:
if
open
fires
are
so
great,
why
don't
you
build
open
fires
with
open
fires
and
I
said?
Ok,
a
will-
and
this
is
what
we
got.
The
dashboard
does
a
get
to
the
list
functions
function
which
has
a
user
parameter
the
ghosts
that
open
fires
API
filters
according
to
metadata,
we've
got
there
and
draws
the
dashboard
for
you
where.
A
A
This
is
an
example
of
how
the
sealed
secrets
works,
because
you
know
we
wanted
to
look
a
bit
into
security
in
this
talk.
You
have
your
plaintext
secrets.
You
run
fast
CLI
cloud
seal,
it
will
generate
this
encrypted
version
and
then,
when
it's
uploaded
it
effectively
gets
unsealed
and
attached
again.
We
only
need
to
give
you
the
public
key
which
is
non
confidential
and
we
keep
the
private
key
in
the
cluster.
A
Now
we've
in
cuba,
Nettie's
there
are
a
number
of
ways
we
can
get
multi-tenancy
and
enhance
security
and
I'm
sure
people
will
come
and
tell
me
afterwards
that
I've
missed
a
couple
and
I
know
it's
confusing.
It's
very
hard
to
get
it
right,
but
we're
playing
memory
and
CPU
limits
across
the
cluster
to
try
and
help
reduce
people,
exhausting
the
resources.
A
A
Finally,
CRI
container
runtime
interface
can
allow
us
to
you
something
like
cata
containers
or
VMs,
to
have
additional
isolation
so
that
if
somebody
found
a
vulnerability
in
container
D,
they
may
not
be
able
to
break
out
now.
What
we've
done
is
by
default,
applied
a
read-only
file
system
to
everything.
That's
deployed,
that's
not
the
case
in
open
fires
by
default,
but
you
can
opt
into
it
in
open
vise
cloud
is
policy
we're
using
non-root
so
that
you
can't
escalate
privileges?
A
wolf
means
that
you
only
see
your
staff
will
stuff
in
your
team.
A
H
Mac
is
used
to
validate
the
web
books
that
come
from
github
so
that
we
know
that
it
really
is
github
and
we
can
delegate
your
user
authority
between
them
and
the
public
TLS
is
built-in,
so
let's
encrypt
can
be
quite
tricky
to
configure,
but
we
do
that
all
for
you,
you
say
well,
your
domain
name
is
and
we'll
generate
it.
So
the
secrets
we
covered
in
this
a
bit
an
army
project
and
also
you
know,
there's
a
default
configuration
that
apply
some
basic
rate
limiting.
A
B
Right
so
linker
lickity
out-of-the-box
provides
mutual
TLS
to
all
to
all
the
proxies
which
are
injected
with
the
linker
D
proxy.
Sorry
to
all
the
services
that
are
Jack.
Why
would
I
want
to
do
that?
Well,
you
want
encryption.
You
need
to
be
safe,
especially
if
you're
working
in
a
multi-tenant
environment.
So,
looking
at
this
slide
here,
we
see
we've
got
multiple
services
in
the
open,
faz
environment
that
communicate
with
each
other.
B
Okay,
alongside
those
are
underneath,
those
I
should
say,
is
the
linker
deep
proxy
linka
deep
proxy
has
the
certificates,
and
now
the
services
no
longer
communicate
directly
with
each
other
link.
Rd
is
managing
all
that
traffic
in
and
out
and
securing
it
encrypting
it
along
the
way,
with
a
trust
or
certificate
from
the
trust
or
in
the
control
point.
Well,.
A
So
this
is
what
we
get
and
I
really
love
the
the
CLI
with
link
ID,
and
this
is
the
edges
command.
What
it
allows
me
to
see
is
what
are
the
edges
for
the
proxy,
how
they
talking
to
each
other
and
I
know
you
probably
can't
see
that
so
I'm
going
to
help
you.
The
first
thing
we've
got
is
nginx
talking
to
the
edge
router.
This
is
the
bit
that
decides
what
user
to
serve
and
what
function
we're.
Looking
at,
you
can
see
at
the
foreign
that
that's
secured.
A
Next,
we
go
from
the
edge
Rooter
to
the
Gateway,
and
the
edge
rate
is
a
new
bit
the
Gateway
you
guys
probably
already
have.
Then
the
Gateway
goes
to
my
pod.
Alex
Alice
run
on
Kate's
and
we
get
so
a
TLS
served
on
ingress
controller
and
we've
got
two.
There
got
TLS
all
the
way
down
into
our
code
during
auto
scaling.
A
You
will
see
that
link
ADIZ,
also
picking
up
metrics
and
there
was
a
spike
in
traffic
and
then
the
latency
went
up
because
we
needed
more
capacity
that
was
added
into
the
cluster
and
the
latency
in
the
bottom
corner.
There
dropped
back
down
again-
and
this
is
all
built
into
open,
vas
and
then
again,
there's
a
policy
in
open
first
cloud
to
keep
it
within
sane
limits.
A
If
you
want
to
get
open
fast,
open,
fast
cloud,
rather,
then
there's
a
bootstrap
tool
that
you
can
run
now.
This
will
install
a
whole
bunch
of
things.
The
first
thing
is
the
functions
that
make
up
the
build
pipeline
and
the
CI
CD
in
the
dashboard.
Next
installs
build
kit
and
if
there's
an
on
route
configuration
for
build
kitten,
privileged
configuration
as
well
Kiros
done
a
great
job.
A
Maintaining
that
and
helping
me
out
on
the
project,
and
so
that
means
you
can
have
these
secure,
builds
helm
can
be
used
with
or
without
tiller
to
then
install
the
open
face,
gateway
itself
using
its
chart.
Nginx
using
its
chart,
cert
manager
will
give
you
the
certificates,
sealed
secrets,
and
then
we
add
Mineo
by
default.
But
if
you
want
to
use
your
own
s3,
you
can
just
configure
that
too
right.
So
you'll
fill
out
a
Yama
file
with
your
domain
name,
your
email
address
for
your
TLS
anything
else
you
want.
A
If
you
want
the
read-only
policy
across
the
board
or
not,
and
then
finally
you'll
probably
take
about
100
seconds.
You
have
all
of
that
deployed.
So
what
I
demo
to
you
and
get
on
your
own
environment?
And
if
that's
a
bit
too
much-
and
you
don't
want
the
TLS
and
the
DNS
you
like
Alex
I,
just
want
to
try
it
that's
okay,
too,
and
you
can
look
on
my
blog
and
have
a
quick
start
that
just
turns
those
bits
off
you
can
just
play
around
with
it
and
see
if
it
fits
your
needs.
A
So
the
future.
What
we
have
in
mind,
as
we
just
come
up
to
close
the
talk,
is
this
idea
of
feature
branches,
and
this
again
this
is
me
saying:
open-source
needs
to
listen
to
real
users,
and
our
customer
@dx
needs
a
preview
branch.
So
right
now
you
can
install
open,
fast
cloud
and
point
to
a
branch,
so
one
that
does
master
one
that
is
staging.
A
We
want
to
combine
it
into
one
cluster
build
time
secrets
are
something
we're
trying
to
work
on
if
you've
got
ideas
of
how
we
could
do
this,
please
come
and
talk
to
me
inputs,
welcome
the
quotas
and
also
I,
think
static.
Analysis
is
interesting.
I
think
of
tools
like
snick
they're,
able
to
look
for
from
room
abilities
in
your
code
within
link,
addy
I,
know
Oliver's,
here
really
looking
forward
to
mutual
TLS
for
TCP.
There's
part
of
open
fast
clouds
that
isn't
meshed
at
the
moment,
which
is
the
asynchronous
bus
for
Nats
streaming.
A
We
use
that
to
run
the
builds.
Canaries
would
be
great
if
they
were
built
in
and
using
the
SMI
spec.
We
don't
have
to
change
open
fires
chargeback.
This
is
something
that
Enterprise
really
keen
on
the
ability
to
do
that
is
enhanced
by
licking
at
the
bandwidth
metrics
in
Lincoln,
Lee,
then
being
able
to
bubble
that
up
through
open
vests
and
maybe
even
like
we
trying
automatically,
as
people
have
asked
for
that
too.
A
So
just
want
to
sum
up:
the
new
will
almost
at
a
time
open
files
cloud
is
open
fires,
but
automated
the
build
pipeline.
It's
done
for
you,
the
security
is
handled
for
you
and
the
management
of
that
I
love
you
go
and
try
it
out,
and
this
is
how
you
can
get
involved.
We've
got
a
github
repo
feel
free
to
take
a
photo
of
this.
A
This
is
the
main
repo
of
all
the
code
and
there's
a
lab
if
you're
just
interested
in
open
files.
At
the
moment,
you've
got
your
own
pipeline,
you
can
bind
it
with
link
ID
and
you
can
try
all
this
cool
stuff
feel
free
to
send
pull
requests
of
that
too,
and
both
of
our
slacks
are
looking
forward
to
seeing
you
there
so
okay.