►
From YouTube: CNCF SIG Network 2020-06-04
Description
CNCF SIG Network 2020-06-04
A
A
B
C
C
C
B
C
C
C
C
All
right,
fair
enough:
who
else
do
we
have
here,
Luke
Nikolaj
we've
got
Watson
Ami's,
I'm,
Jonathan,
Matt,
David,
very
good,
all
right
we're
about
four
after
three
after
now.
If
the
meeting
minutes
are
a
community
effort
I'm,
so
please
don't
be
shy.
If
you're
on
the
call
today
and
your
fingers
still
work,
go
ahead
and
slap
your
name
into
the
attendees
list,
there.
C
And
we've
really
just
got
one
agenda
item
today,
which
will
give
I
think
you
the
present
today's
maintain,
errs
and
presenters
some
some
comfort
and
some
room
to
to
tell
us
about
ambassadors,
to
tell
us
about
their
project
and
hopefully
won't
overflow.
This
SIG's
plate,
so
so
I'm
pleased
that
we
just
have
one
one
item
on
that
agenda
today:
I'm
with
that
we're
a
few
minutes
we're
a
few
minutes
after
so
let's
go
ahead
and
get
going.
Daniel
is
here
with
us,
and
so
is
Richard
and
I.
C
C
B
Can
everyone
see
me
okay
there?
It
is
sweet,
oh
it's
prettier!
So
folks
on
the
call
that
hadn't
bump
into
me.
My
name
is
Daniel
brand
product
architect
at
data
wire,
Richard,
Lee,
co-founder
and
CEO
of
data
wise
also
on
the
call
as
well
I'll
be
presenting
today
and
then
Richard
and
I
can
take
questions
at
the
end.
B
They
do
iris
the
sort
of
the
founder
in
the
student-
if
you
like
of
ambassador
at
the
moment-
and
this
is
our
discussion
around
proposed
donation
to
the
CN
CF,
the
TLDR
ambassador-
is
an
open
source,
API
gateway
for
kubernetes
powered
by
envoy,
very
much
focusing
on
the
north/south
use
case.
Wimp
we've
tried
to
make
it
as
developer,
focused
and
developer
friendly
in
the
cloud
space
as
possible,
so
custom
resources
for
configuring,
your
endpoints,
your
routing
and
so
forth.
We
support
kubernetes
ingress
too,
very
much
into
open
standards
active
a
on
the
dado.
B
Our
team
he's
been
working
hard
on
the
community's
1.18
ingress
class
support
and
the
path
type
feels
as
well
and
there's
wide
adoption
in
the
industry.
We're
using
like
fans
of
orgs
lots
of
docker,
poles,
3,000
plus
folks.
In
our
community
slack
and
over
130
contribution
and
non-trivial
contributions,
I'll
run
through
some
of
those
later
on,
but
there's
some
fantastic
people
in
the
community
really
sort
of
contributing
to
interesting
discussions.
Interesting
bits
of
code,
that's
extending
ambassador
all
the
time,
so
we
are
a
effectively
a
control
plane
on
to
envoy.
B
We
are
focused
in
the
north/south
use
case
and
we
convert
kubernetes
config
or
your
mappings
and
so
forth
into
envoy
convict
under
the
hood
envoys.
The
data
plane
traffic
directed
through
I'm
sure
on
this
call
folks,
you've
seen
the
contour
stuff,
you've
seen
probably
other
gateways,
probably
nothing
new
to
yourself.
So
I'll
skip
over
that,
and
we
can.
You
know,
move
on
to
the
perhaps
more
interesting
content.
B
We've
come
a
long
way
in
three
years,
so
I
think
it
was
early
early,
March
2017
when
we
announced
the
project
where
the
1.0
release
early
this
year,
which
was
like
super
exciting,
really
pretty
pumped
to
see
that
get
up
stars
2.8
K,
because
you
know
we're
like
get
up
stance
the
way
to
measure
success
for
any
project
right.
That's
the
stand
of
cloud
native
metric,
but,
but
you
know
it
shows
interest.
It's
like
the
appreciation
of
the
sort
of
the
supporters
that
ends
the
international
project.
B
We've
had
130
contributors,
I
can't
leave
anyone
today
actually
1700
and
pull
requests.
If
you
want
to
read
more
back
to
the
later
journey,
about
1.0
release
and
so
forth,
I
put
the
blog
blog
post
for
you
there.
It's
been
a
fantastic
ride.
I've
been
involved
in
a
whole
bunch
of
this
time,
richest,
obviously
been
there
from
day
0.
It's
been
just
amazing,
working
with
the
community.
We're
looking
forward
to
driving
this
forward
on
in
the
future
to
call
features
again.
B
I
probably
won't
labor
this
too
much
because
I'm,
assuming
this
crowd,
are
pretty
happy
with
sort
of
envoy
features
and
we
layer
on.
On
top.
We
provide
that
ease
of
use.
We
provide
that
north/south
use
case
and
we
provide
that
ease
of
use
for
developers,
so
many
things
I
will
point
out
Evans
sort
of
resilient
side
there.
If
you
look
to
the
left
and
per
example,
we
did
all
your
standard
resilience
features,
but
we've
also
had
off
from
a
very
early,
very
early
development
initiative
in
ambassador.
Initially
it
was
a
custom
extension
to
envoy.
B
We
then
worked
with
the
upstream
envoy,
a
community
help
form
the
XT
or
Z
interface,
and
we've
worked
through
that
now
and
that's
actually,
you
know
once
it
was
up
streamed
in
the
Envoy.
We
then
changed
ambassador
to
use
that
well
established
their
well
agreed
API
as
well.
So
we
love
working
with
the
upstream
envoy
community,
fantastic
community,
all
around
there.
Rate-Limiting,
for
example,
is
done
by
the
an
hour,
less
proto
as
well
I've
got.
B
For
me,
as
a
hybrid
use
case
situation,
if
you're
pod,
the
Ambassador
is
deployed
in,
can
write
to
somewhere
on
the
network,
you
can
route
out
of
the
kubernetes
cluster,
so
people
in
Google
are
gke
and
GCP
use,
IP
aliasing
in
their
network,
and
they
have
VMs
running.
They
have
gke
running
ambassador
running
in
there
with
console
and
we
can
use
console
for
service
discovery
for
IPS
and
ports
and
we
can
route
to
not
only
to
kubernetes
services.
B
We
can
rank
two
VMs
outside
of
the
cluster,
so
that
kind
of
supports
their
lift
and
shift
the
hybrid
model.
That's
been
a
super
interesting
journey
and
a
hat
tipped
of
hash
eagle
folks.
We
worked
with
quite
closely
they're
a
fantastic
community,
as
we
all
know,
and
that's
been
a
really
interesting
sort
of
enabling
use
case
for
getting
folks.
You
perhaps
stuck
with
the
VMS
or
not
even
stuck
they
like
the
VMS,
but
they
want
to
dabble
with
cuban
at
ease
to
all
your
the
staff.
We
have
good
stuff,
zero,
downtime
config.
B
We
use
kubernetes
all
that
all
the
bridges
and
communities
to
manage
our
state,
so
ambassador
pods
can
come
up
and
down.
The
state
is
stored
within
kubernetes
itself.
L7
support
again
had
tipped
envoy.
This
is
basically
building
on
all
the
great
stuff
in
envoy
there.
Nothing
too
exciting
on
that
one
and
from
our
perspective,
but
obviously
super
useful.
The
l7
routing
is
becoming
really
a
big
thing,
as
folks
move
more
towards
micro
services.
B
Our
primary
use
case
is
in
the
api
gateway
space,
traffic
man
and
security.
App
development.
You
know
allowing
folks
to
sort
of
move
at
their
own
pace
do
different
releases.
They
can
contribute
to
their
own
keep
Nettie's
files
with
the
mappings
in
so
different
teams
can
go
at
different
paces.
All
the
good
patterns
we
see
with
cloud
native
all
the
good
patterns
we
see
with
microservices
decoupling
and
independent
release.
Would
you
see
some
folks
running
multiple
ambassadors?
B
That's
totally
a
bad
use
case,
often
with
an
internal
and
an
external
ambassador,
so
like
internal
devs
and
then
external
API,
offering
on
top
they
can
change
it
different
paces
and
so
forth.
We
do.
We
don't
really
encourage
it,
but
we
do
see
that
this
hub-and-spoke
model-
it's
like
a
bit
of
a
service
meshing,
light
type
thing
where,
if
you're
running
a
very
shallow
graph
of
services,
maybe
you've
got
the
monolith
and
you're
breaking
out
services.
Norman's
you're,
comfortable
routing
traffic
rounds
with
the
outside.
B
You
can
use
Ambassador
as
that
service
discovery,
sort
of
mechanism,
if
you
like,
all
that
routing
mechanism
config
wise,
so
customer
sources
and
ingress
and
config.
So
we
have
mappings
and
hosts
again
folks
used
to
sort
of
all
the
traditional
routing
like
nginx,
actually
proxy.
This
is
probably
nothing
new
there.
You
have
your
endpoints,
your
back-end
services
host.
You
can
configure
your
hosts,
your
TLS
config,
that
kind
of
good
stuff
and
then
from
I
think
was
September.
Last
year
we
added
support
for
ingress
as
well.
B
So
it's
been
a
fantastic
journey
and
Alex
is
thoroughly
enjoyed.
Working
with
the
wider
community
to
test
that
these
ideas
give
his
opinions
and
so
forth
in
terms
of
being
sort
of
proven
and
growing
rapidly.
We
have
many
production
deployments,
just
a
handful.
We
picked
out
here,
mainly
because
they've
done
great
blogs
and
they've
done
coop
con
presentations
and
then
there's
been
as
Li
an
hour
talking
sort
of
anyone
off
mic
noise,
leather,
Keith
car-
and
this
is
a
highlight
for
myself,
Richard
the
whole
team.
B
Actually
people
coming
up
just
at
the
booth
come
out
so
from
where
doing
talks
and
saying
hey,
we're
using
bass
today,
we're
doing
this
app
direct,
fantastic
story,
Ticketmaster
sure
needs
no
introduction.
Chick-Fil-A
blew
my
mind
because
chick-fil-a
run
a
committees
cluster
in
each
of
their
restaurants
around
the
u.s.,
so
I've
chatted
as
person
who
things
are
armed
chick-fil-a.
B
We
run
like
hundreds
of
ambassadors
because
we
run
an
ambassador
in
every
kubernetes
cluster
in
each
of
our
restaurants,
kind
of
like
at
the
edge
use
case,
and
that
was
just
hearing
his
story
of
how
they
managed
his
stuff
was
just
fascinating
right,
so
loved
learning
from
you
know,
production
use
cases
and
it
feeds
back
into
our
design
goals,
feeds
back
into
things
we
put
back
into
the
community
too
in
terms
of
community
contributions
and
just
sort
of
highlighting
this.
Our
latest
release.
B
It's
just
one
release
I
wanted
to
pull
out
a
few
interesting
things,
a
hat
tip
to
procure
Joshi
and
contributed
and
preserving
extra
quest
IDs.
Their
mortal
blonde
voice
got
water
blond
boys
in
the
stack
and
great
work
like
a
chat
to
procore
and
the
slack
he
worked
with
Flynn
on
the
dev
side.
He
works
at
a
hot
star
like
the
Indian,
say
Netflix,
and
it's
where
Disney
Plus
is
available
over
in
India.
He
was
loving,
the
community
loving
get
involved,
and
we
really
appreciated
this
kind
of
tweet
shouting
out.
B
You
know
yes,
like
we
recognized
that
the
community,
the
contributing
experience,
is
super
important
to
an
open
source
project,
other
couple
of
hat
tips
to
fill
pebble
there
and
contributing
the
they're
setting
that
envoy
shared
memory
base
ID
to
like
multiple
envoy
proxies
in
a
pod.
So
if
you
got
sto,
there
was
his
use
case
and
you're
an
ambassador
to
that
was
a
you
know,
great
bit
of
work,
lots
of
fun
chatting
to
feel
on
the
slack
really
big,
an
interesting
piece
of
work
and
share
it
out
to
newer
font.
B
Is
there
from
puppet
and
puppet
have
got
real,
a
sh
which
is
their
take
on
github
actions
or
Argo?
That
kind
of
thing
and
Noah
is
been
using
fantastic
work
with
the
K
native
support
with
ambassador,
so
he's
been
doing
some
performance
improvements,
support
for
path
and
time
out
options,
and
then
the
K
Native
gateway
and
he's
working
on
some
blog
posts,
which
is
a
fantastic
story
all
round.
So
we
really
enjoyed
seeing
more
more
advanced
contributions
to
ambassador
from
the
community
robat
wise,
rich
and
I
thought
about
breaking
it
down
into
two
strands
here.
B
Let's
get
some
guidance
from
the
CN
CF
on
how
to
make
this
even
better.
Because,
again,
if
we're
all
trying
to
drive
forward
the
innovation
of
the
North
South
use
case
for
envoy,
we
think
we're
in
a
great
position
to
do
that
or
to
help
do
that.
But
we
love
the
input
from
the
CN
CF
to
make
that
make
that
happen,
features
wise,
wise
and
support.
Obviously,
in
the
envoy
space
wise,
a
miss
is
super
hot
he's
in
general,
but
we'd
love
to
and
look
at
this
and
as
well
as
our
future
roadmap.
B
So
since
you
have
donations,
we're
going
for
an
information
proposal,
it's
a
mature
project,
ambassador
we'd
love
to
advance
the
north/south
use
case
for
envoy
again,
I.
Think
I'm.
You
know
I
don't
need
to
really
preach
too
hard
here
we
want
to
know
and
love
envoy
and
we
think
the
north/south
use
case.
Is
it
really
important
part
for
that?
The
edge
proxy,
the
ingress
part
of
your
cluster
we've
got
production
and
you're
approved
in
production
and
xur
deployments?
Lots
of
interesting
use
cases
we've
linked
in
the
deck
we
can.
B
We
can
share
more
we're
all
like
driving
this
climb
native
best
practices
for
kubernetes
ingress
and
again.
That's
me
I
think
I
chose
nicely
with
the
CN
CF
goal
of
promoting
this
cloud
native
experience.
All
the
architectures
all
the
operational
models
and
how
we
can
figure
is
a
big
part
of
this.
We
really
invested
into
clarity
of
config
love
to
get
ups.
We
preached
at
the
we
works
team
around
these.
These
things
a
self-service
comprehend
integration
is
the
CN
CF
ecosystem.
For
me.
B
Yes,
all
these
good
things
too,
and
while
committing
we
are
focusing
on
the
north-south
traffic
management
use
case,
but
where
we're
making
it
easy
to
integrate
with
east-west
too,
we
have
integrations
yet
with
an
link,
addy
was
console,
and
with
this
do
we
often
find
that-
and
you
know
folks
get
on
board.
We
say
the
north-south
use
case.
They
spin
up
kubernetes,
they
get
some
CI
in
some
CD
in
there
just
to
deploy
their
containers
in
the
next
thing
you
need
to
do
is
get
traffic
into
that
cluster.
That's
the
ingress!
B
When
they
start,
you
know
doing
more
micro
versus
they
often
want
to
move
towards
service
mesh.
So
we
weren't
we're
looking
to
make
it
as
easy
that
the
on-ramp
is
easy
as
possible
for
them
the
communities
have
been
fantastic.
It's
been
humbling
being
part
of
this
and
we'd
like
to
make
it
better
again.
That's
what
we'd
love
the
CNCs
guidance
s,
helping
this
space,
but
plenty
of
slack
uses
lots
of
contributions
and
and
multiple
coupon
talks.
We
don't
even
know.
Sometimes
a
few
years
ago
we
heard
the
Canadian
about
ambassador.
B
We
were
like
hang
on
your
ambassador.
This
is
fantastic
and
that's
what
then
triggered
even
more
work
around
ke
native
to
so
yeah.
She
spent
this
humbling
to
see
folks
building
on
our
technologies,
we'd
like
to
help
them.
Do
that
you
more
the
asks
from
the
CN
CF
we're
looking
for
the
vendor-neutral
home
to
grow
the
Ambassador
Croom
community
to
grow
the
North
South
use
cases
for
envoy
and
kubernetes.
We
love
help
the
CIC
and
CD,
or
more
folks,
probably
on
CD,
continuous
delivery
infrastructure
and
assistance,
improving
the
docs.
B
How
to
the
general
kind
of
onboarding
experience
that
you
know
contribution
experience,
because
we
recognize
if
we're
trying
to
all
these
goals,
are
trying
to
drive
forward
the
north/south
use
case
for
envoy
humanities
for
all
the
kindness
of
tech,
it's
all
about
making
it
easy
for
developers.
Particular
you
know
late
majority
that
adopters
this
is
a
big
hurdle
for
them,
so
we'd
really
love
some
guidance
and
some
help
from
the
CN
CF
forum
on
how
to
improve
this
experience
for
folks
looking
to
get
traffic
into
their
Cuba
Nettie's
clusters.
At
that
point,
actually,
thanks.
B
C
B
D
When
you
say
that
wasn't
well
you're
not
currently
building
on
armed,
although
we've
had
starting
to
see
a
couple
requests
for
that,
so
it's
something
we'd
like
to
do,
but
we
haven't
actually
looked
at
it
beyond
sort
of
theoretical
sort
of
a
theoretical
perspective.
It
doesn't
seem
like
it's
super
hard
to
do,
but
we
just
haven't
done
it.
D
B
C
C
Yeah
I'm
leaving
space
for
other
something
very
hard
practice
for
me.
That
was
there's
a
lot
of
strength
in
that
silence.
Just
then,
the
of
the
so
good
so
pleased
to
see
and
can't
confirm
that
that
that
the
proposal
here
is
an
incubation
level
which
is
appropriate,
the
thousands
of
deployments
or
kind
of
the
the
numbers
around
that
how
are
those
measured?
How
are
those
confirmed,
I
think.
D
It's
a
little
bit
of
triangulation
I
mean
we.
You
know
we
based
on
Twitter
the
number
of
folks.
We
have
in
a
slap
back,
and
you
know
we
get
some
data
from
dr.
pols
I.
You
know
people
pull
all
the
time,
so
you
know
it's
a
little
hard
to
say
so,
but
you
know
we.
We
actually
shifted
from
keyed
IO
to
docker
hub
because
of
the
persistent
outages
and
and
with
in
a
week
we
had
half
a
million
poles.
D
D
Think
when
people
want
to
extend
their
ingress
for
a
sort
of
more
broader
use
cases,
that's
it
that's
a
big
one.
So
we
we've
actually
had
a
lot
of
users
migrate
from
contour
to
ambassador,
because
contour
hasn't
supported,
authentication
and
authentication
is
a
pretty
common
requirement
at
the
edge.
D
So
we've
worked
pretty
hard
to
try
to
make
it
because
there's
so
many
things
that
you
have
to
install
to
get
kubernetes
working.
The
way
you
want
we've
tried
to
make
it
easy
to
just
integrate
with
you
know,
all
the
other
stuff
you
need
so
I'd
say:
that's,
probably
the
so
there's
a
ton
of
how
to's
on
our
website,
around
sto
link
or
D
console
yeah.
A
B
B
Just
you
know,
keeping
an
eye
on
all
these
things,
this
sort
of
checkbox
stuff
in
some
ways
the
l7
support
l
for
support
s
ni,
you
know,
check,
check,
check
and
then
we
always
add
sort
of
it.
Comments
on
on
top
in
terms
of
what
the
as
rich,
as
mentioned,
for
integrations
being
super
important.
There's
additional
value
over
the
checkbox
stuff
of
the
checkbox
stuff
is
covered
quite
nicely
by
external
folks.
C
E
Hi
Jenn
Ian,
hey,
you
can
hear
me,
we
can
yeah
Daniel
I
was
just
questioned.
That's
it
for
Derek
and
I
have
been
in
the
boss,
seeing
the
control
damning
right,
telling
the
things
they
have
done.
So
they
have
a
very
good
useful
of
use
case
of
them
and
when
you
have
an
English
controller
for
that,
and
you
have
a
fully
quite
qualified
domain
name
that
have
a
vet's
that
is
so
similar
to
that
so
kind.
E
In
a
in
a
native,
kubernetes
and
ingress,
you
can't
do
the
same:
philic
qualified,
a
name
name
for
separate
two
ingress
and
egress
services.
So
but
using
contour
you
can
achieve
that.
I
have
some
very
unique
use
cases,
but
people
often
talk
about
that.
We
have
an
English
controller
and,
let's
say
I've
at
wwe.com,
and
you
have
a
same
in
this
services
that
have
a
same
WW,
not
test
dot-com
and
/.
There
was
other
product,
we
have
a
limitation,
but
they
have
the
D
equals
self
domain.
So
in
using
contour
you
can
do
that.
E
D
D
So
so
the
answer
is
yes,
you
can
have
multiple
different,
fully
qualified
domain
names.
You
know
the
reason
why
I
think
a
lot
of
ingress
controllers
don't
necessarily
support.
This
is
because
it's
not
really
part
of
the
ingress,
spec
and
so
common
use
case,
though,
would
be
you
know
you
want.
You
have
two
different
domains.
You
know
a.com
and
b.com
that
you
want
to
host
on
the
same.
Kubernetes
cluster
expose
the
same
ingress
controller
and
you
also
want
TLS
with
them.
D
D
Mean
I
think
it's
like
so
we've
had
a
positive
experience
with
CN
CF
telepresence
is
a
sandbox
project,
so
I
would
say
that
the
engagement
level
with
the
CMC
F
for
telepresence
has
been
I'd
characterize
a
little
more
arm's
length
and
which
is
which
is
perfectly
appropriate
given
in
sandbox
nature,
and
you
know
and
I
think
we've
also
been
caught
up
in
the
the
vortex
of
what
is
sandbox,
which
seems
to
be
a
present
RIA
merges
on
the
TOC
every
few
months.
So
so
you
know
we
have
I
would
say
we
have
higher
expectations.
D
Around
incubation
as-
and
you
know,
ambassador
I
would
say
just
from
a
community
and
breadth
of
adoption.
Standpoint
is
more
mature
than
telepresence,
so
we
think
incubations
appropriate.
So
we
would
expect
sort
of
just
more
more
engagement
with
the
CN
CF
community.
Around
ambassador
and
I.
Would
I
would
hope
that
the
community
kind
of
pushes
us
in
directions
we
may
not
have
considered
and-
and
we
consider
that
to
be
a
good
thing
so.
B
Definitely
one
thing
like
that's
that
it
is
been
the
sort
of
facilitation
aspect
ly
like
going
to
the
coop
cons,
doing
the
presentation
in
the
maintain
his
track.
You
know
just
meeting
sort
of
folks
has
been
fantastic
because
in
my
I
did
the
m1a1
talk
at
us
in
San,
Diego
and
I've
done
the
other
ones
as
well,
and
that
visibility,
even
at
the
sandbox
level,
was
great
to
Rich's.
Point
I.
B
Definitely
think
these
have
with
incubation
of
even
more
more
that,
but
that
alone
it's
been
fantastic,
gives
us
that
platforms
talk
about
what
we're
doing
to
get
the
input
either
have
those
kind
of
hallway
breakout
sessions
where
we
can
share
and
learn.
That's
just
been
for
me,
fantastic
just
seeing
that
part
of
the
CNC
yeah.
C
D
You
know
we
basically
have
a
a
non
open-source
sidecar
that
you
can
deploy
in
the
same
pod
as
ambassador
that
provides
additional
enterprise
features.
Sort
of
our
general
belief
is
that
we
want
to
make
sure
that's
and
that's
also
why
the
the
open-source
component
has
just
public
api's
that
are
all
documented,
and
we
are
our
side.
D
Cart
just
uses
the
exact
same
api's,
so
so
folks
can
choose
to
replicate
all
the
functionality
or
a
sidecar
which
includes
you,
know,
open
ID,
connect
and
ooofff,
and
all
this
other
stuff
or
they
can
or
we
have
lots
of
users
who
don't
and
just
just
use
our
api's
and
the
open
source
and
and
build
their
own
authentication
service.
And
so
so
that's
that's
sort
of
how
our
business
operates.
Oh
very.
C
Good
of
of
that
proprietary
sidecar
is
that
and
in
the
presence
of
that
is
there
disabling
of
our
the
yeah.
You
know:
I've
been
in
service
smash
land
focused,
there's
so
much
that
I'm
yeah,
no
okay,
very
good,
see
you
got
you
had
on
the
site.
Does
that
add
a
network
hop
in
the
way
in
which
the
sidecar
works,
or
is
that
a.
D
G
D
C
A
A
D
Think
so
so
a
few
things
one
is
not
sure
what
CN
CF
is
planning
and
doing
in
terms
of
supporting
ephemeral,
kubernetes
clusters
right.
So
we
have,
we
have
a
pretty
comprehensive
regression
and
we
want
we've
run
the
regression
all
against
kubernetes
cluster
is
right,
so
being
a
support
that
particularly
a
case
where
you
spin
up
an
ephemeral,
kubernetes
cluster
run
our
full
regression.
D
D
D
A
Okay,
thanks
I
guess
there
are
some
provisions
in
CNCs
could
probably
get
I
mean
if
you
guys
were
probably
would
need
to
do
the
work,
but
as
far
as
packet
clusters
packet
resources,
so
you
could
have
arm
machines
as
well
as
regular
AMD
machines,
and
you
can
install
your
own
kubernetes
clusters
there
and
there's
plenty
of
resources.
There
I
mean
should
be
able
to
get
some
type
of
permission
for
that
as
far
as
assistance
with
porting
things
to
arm
I,
don't
know,
yeah.
D
I
think
we
would
I
mean,
of
course,
how
porting
to
arm
would
be
great,
I,
think
it's
more
in
so
much
as
theirs,
because
we
can't
be
the
only
project
that
runs
regression:
tests
of
software
that
gets
deployed
in
kubernetes
clusters
right,
so
in
so
much
as
there's
infrastructure
or
projects.
In
that
vein,
those
are
things
that
we
would
definitely
want
to
take
advantage
of.
So.
C
Good,
given
that
we
have
a
little
time,
maybe
some
additional
details
around
well
we're
talking
about
the
number
of
contributors
and
kind
of
the
the
diversity
of
the
maintainer
ship,
tether,
the
project,
governance
and
I.
Guess
what
what?
What
else
can
you
use?
You
guys
say
there
so
I'm,
just
kind
of
looking
through
now
is
the
the
roadmap
is:
is
public
facing
or
kind
of
a
how?
How
are
the?
How
does
the
community
run,
and
you
know,
I
I'd.
D
Say
you
know
it's
it's
very
much
a
work
in
progress.
You
know,
I
think
a
lot
of
the
things
we've
added
are
people
typically
file
github
issues
and
we
get
people.
People
tend
to
vote
up
properly,
github
issues
and
those
that
tends
to
be
a
big
sort
of
source
of
prioritization
for
us
and
then
you
know,
people
sometimes
just
show
up
and
there's
this
pull
request,
and
you
know
and
and
that's
great
right.
D
So
that's
how
we
learned
about
Noah
working
on
K
native,
because
he
opens
pork
west
and
it
was
a
it's
a
pretty
sophisticated
pull
request
and
we
had
to
get
on.
You
know
a
couple
calls
with
them
go
back
and
forth
to
before
we
get
to
the
point
where
you
know
we
could
land
land
has
changes
right,
so
we
have
a.
We
have
a
developer's
channel.
Almost
all
the
chatter,
from
both
our
internal
and
external
developers
are
all
in
a
public
public
channel.
So
we
have
we.
D
We
have
that
and
then
and
basically
and
we'll
probably
new
formalize
this,
but
essentially
the
folks
who
have
become
maintainer
is
not
from
data
where
they
basically
just
jump
in,
and
then
it
turns
out
that
they
know
more
about
some
part
of
the
system
than
we
do
so
we're
just
like
okay,
well,
we're
not
qualified
to
to
we're,
not
the
best
person
to
actually
view
the
code.
So
do
you
mind
you
know
overseeing
this
part
of
the
code
and
and
that's
what
happens
so.
D
So
we
don't
have
formal
community
calls
right.
So
this
is
so
it's
a
little
more
ad
hoc
and
there's
just
in
most
of
it
just
happens
through
our
developer
slack
channel
and
then
for
complicated
things.
We
just
hop
on
the
phone
with
a
contributor
just
work
through
issues
that
take
too
long
for
own
github.
We.
B
Do
good
together,
cube
comms,
that's
a
fantastic
resource
that
again
the
hallway
track.
We
had
a
dinner
last
year
at
San,
Diego
bunch
of
folks.
Together,
let's
go
spot-check
folks
came
along,
we
were
chatting
to
them,
so
that's
kind
of
a
touch
point
in
real
life
which
obviously
missing
at
the
moment,
but
we're
more
than
happy
so
yeah
extending
onto
that
virtual
world
as
well
resume
nice.
C
And
then,
to
clarify
of
the
of
all
of
that,
ambassador
is
said:
there
there's
an
ambassador
operator,
and
maybe
maybe
a
few
other
things
as
well,
that
I'm
I'm
not
familiar
with,
but
to
clarify
the
of
what
we're
looking
at
I'm
donating
here
is,
is,
is
which
repositories
and
in
which?
What
all
is
this
I
think.
D
D
So
we
maintain
a
couple
patches
to
envoy,
usually
at
any
moment,
I'm
that
we're
still
sort
of
the
process
upstream.
It's
also
where
we
it's
actually
not
a
public
repository
or
some
parts
of
it
are
not
public,
because
when
they're
embargo
on
which
security
patches,
that's
where
we
do
our
work,
so
it
would
not
actually
necessarily
be
public.
So
because
we
we
test,
we
test
with
embargo
and
envoy
and
then
we'll
generally
release
more
or
less
the
day
the
embargo
is
lifted
so
make.
C
H
C
B
The
guns
for
myself
and
Amy
there
early
on
that
as
in
we
definitely
that
crypt
from
some
of
the
other
examples
share
to
appreciate
the
input
they're
super.
We
wanted
to
deliver
what
was
most
most
useful
as
possible
to
yourselves,
so
that
was
really
useful.
Getting
that
insight
from
yourself
anemia.
Thank
you.
C
With
that
will
the
waltz
of
Matt,
myself
and
and
others
in
the
community
with
it
will
be
in
in
touch
to
do
that.
You
know
begin
that.
Go
through
due
diligence,
I
mean
you
know,
go
through
nice,
great
okay,
yeah
I
was
confused
at
first
when
we,
when,
just
before
the
call
I
thought
that
you
guys
were
filing
for
sandbox
and
I
thought
how
inappropriate
that
would
be.