►
From YouTube: Kubernetes Community Meeting 20200716
Description
The Kubernetes community meeting is intended to provide a holistic overview of community activities, critical release information, and governance updates. It also provides a forum for discussion of project-level concerns that might need a wider audience than a single special interest group (SIG).
See this page for more information! https://github.com/kubernetes/community/blob/master/events/community-meeting.md
Like what you see here? Continue the conversation on https://discuss.kubernetes.io
A
So
my
name
is
sue
Schmidt
and
I'll,
be
your
host
for
today's
community
meeting.
I
have
recently
joined
the
kubernetes
community
and
I've
been
a
part
of
SiC
enhancement,
so
I've
been
having
a
lot
of
fun
getting
to
know
a
lot
of
interesting
stuff.
Let's
get
started
before
we
get
started.
I
just
wanted
to
remind
everyone
that
this
is
a
community
meeting,
and
this
will
be
nice
streamed
and
and
posted
publicly
on
youtube.
A
B
B
Alright,
so
he's
not
here,
I
guess,
Bob
and
I
can
go
if
you
want
sure
okay
Bob
go
ahead
and
share
we'll
do
a
quick
sit,
contributor
experience,
update
and
then
kick
back
over
to
release
team
when
someone
shows
up
this
is
your
say:
contributor
experience
community
meeting
update
for
July,
we
haven't
done
this
one
in
a
while,
so
we've
got
a
lot
to
get
through
next
slide.
Please,
there
are
two
new
chairs
first
to
contribute
experience,
that's
Bob,
killin,
mr.
Bobby
tables
and
Castro
Jo.
B
That's
me
I'd
like
to
thank
Elsie
in
Paris
for
their
years
of
service
to
the
sig
and
they've
gone
on
to
volunteer
for
other
things.
The
major
themes
that
we've
been
concentrating
on
in
this
past
cycle
have
been
sustainability
around
the
projects
that
we
support,
automation
and
working
on
our
tech
debt.
So
things
like
we
finally
were
able
to
triage
all
the
issues
in
our
in
the
community
repo
as
an
example,
next
slide
for
the
next
upcoming
cycles,
we're
planning
on
sustaining
our
sustainability
efforts.
B
So
these
these
are
the
kind
of
little
things
that
are
annoyances
across
the
projects
that
we've
been
trying
to
streamline
and
automate
to
make
it
easier
for
SIG's
to
do
stuff.
So
things
like
fixing
mailing
list
permissions,
your
calendars,
that
kind
of
thing
and
we're
gonna
continue
to
invest
in
automation
like
removing
workloads
from
SIG's
and
I'll.
Go
into
that
in
a
little
bit
next
slide.
B
So
we've
got
a
lot
of
sub
projects,
so
the
first
one
is
a
community
management.
Sub
project
is
kind
of.
Is
the
catch-all
bucket
for
a
bunch
of
things
that
this
sake
does
for
the
rest
of
the
project,
so
things
like
managing
the
project,
mailing
lists
and
calendars,
and
things
like
that,
so
that,
when
a
contributor
comes
to
the
project,
they're
kind
of
getting
a
consistent
view
of
the
places
where
we
keep
stuff
so
like,
if
I
needed
to
join
a
sig
mailing
list?
Where
do
I
do
that?
How
do
I
do
that?
B
We're
set
up
a
very
long
time
ago
and
our
vestigial,
so
we've
been
slowly
moving
away
from
that
model,
as
we've
been
making
new
SIG's
to
doing
things
like
having
the
RO
mailing
lists,
and
things
like
that
owned
by
G
suite
accounts
under
the
control
of
the
project
themselves
in
the
past
and
currently
today,
there's
many
of
our
mailing
lists
and
calendar
entries
and
stuff
are
owned
by
you
know:
people's
old
Google
accounts
and
things
like
that.
So
we've
been
slowly
trying
to
move
all
those
things
over
over
to
a
more
managed
suite.
B
B
This
also
includes
some
of
our
work
that
we're
doing
on
YouTube,
which,
where
you
see
videos
like
this,
for
those
of
you
who
are
unfamiliar
with
this
all
the
SIG's
strive
to
record
their
YouTube
meetings,
and
then
we
put
them
on
YouTube.
We've
tried
to
automate
this
over
the
past
year
using
a
tool
called
splain,
but
recently
there's
been
issues
with
api's
between
zoom
and
youtube,
and
ideally,
what
we
want
is
for
a
sig
co-chair
to
click
a
button
in
zoom
and
then
have
the
meeting
automatically
publish
to
youtube.
B
There
are
some
issues
there,
and
sometimes
we
just
have
to
do
these
by
hand.
So
a
co-chairs
like
hey
I,
don't
have
time
to
figure
out
this
thing.
Here's
a
bucket
with
my
videos
and
then
we
have
a
team
that
uploads
this
we're
also
looking
for
people
to
join
our
stream
team,
especially
those
of
you
in
the
West
Coast
time
zone.
So
we
have
a
lot
of
programs
like
meet
our
contributors,
our
office
hours
and
there's
things
that
have
meetings.
B
But
we
are
always
looking
for
people,
especially
in
the
West
Coast
us,
where
we
currently
have
a
whole
next
slide
slack
our
slack.
The
kubernetes
slack
now
has
over
a
hundred
thousand
users.
That
is
incredible
to
me,
so
that
is
just
a
recent
milestone
that
we've
achieved
zoom
maintenance,
the
only
the
one
thing
we're
asking.
If
everybody
is
please
check
that
zoom
is
on
your
computer,
is
that
is
on
the
latest
version.
Over
the
past
few
months,
there's
been
lots
of
issues,
lots
of
updates,
especially
those
of
you
on
linux.
B
They
do
provide
binaries,
but
a
lot
of
those
aren't
self
updating.
So
please
check
your
your
PC's
when
you're
going
to
the
zoom
meetings
for
kubernetes
and
ensure
that
you're
staying
safe
with
that
discuss
the
doc
Cooper
9000
is
our
internet
forum.
It's
sitting
in
about
6,000
users
right
now,
and
the
only
thing
to
really
report
there
is.
We
have
had
instances
of
people
trying
to
use
the
forums
to
cheat
on
their
certification
exams,
so
we've
been
working
with
the
CNC.
B
You
have
to
kind
of
clean
that
stuff
up
so
that
it's
not
getting
posted
so
that
we
can
keep
the
integrity
of
that
process
and
if
any
of
you
are
listening
out,
there,
computers
have
logs
of
stuff.
So
we
know
when
you're
cheating
and
so
will
the
CNC
F
just
a
pro
tip,
if
you're
pursuing
a
C
CAD
and
we're
always
looking
for
moderators
to
moderate
our
properties.
Things
like
zoom,
YouTube,
slack,
all
the
things
that
we
do.
B
That's
all
run
by
volunteers,
so
we're
always
looking
for
people
who
can
join
our
contributor
process
and
help
us
out,
especially
if
you're
in
the
Asia
Pacific
time
zone
next
slide,
share
and
technically
meaning.
This
is
a
meeting
that
we
have
we're
the
co-chairs
and
tiel's
from
around
the
project.
Come
they
saw
tips,
we
give
them
announcements
and
then
we
send
them
on
the
little
newsletter.
Generally
speaking,
that's
all
transparent
to
most
of
you.
Next
up
stream
marketing
met,
bro
berg
has
been
taking
unis
on.
B
This
is
a
new
team
and
said
contributor
experience
that
has
been
handling
a
lot
of
our
things
like
our
new
twitter
account,
which
strongly
suggest
you
follow.
That's
at
Cates
contributors
and
we'll
be
posting
things
that
are
important
to
you
too
contributors
on
this
account
on
a
regular
basis.
So
this
is
a
way
that
we're
planning
on
helping
you
find
more
information.
B
Next
slide,
please
and
contributors
documentation
is
something
we've
been
working
on
a
while
is
a
hugo
front-end
to
the
community,
the
community
repo,
where
we
have
things
like
how
do
I,
you
know,
make
my
first
contribution
how
to
labels
work.
All
that
kind
of
stuff
is
all
being
worked
on
right
now
and
the
hugo
site
is
almost
done
and
it's
gonna
be
nice
and
awesome.
We
also
have
Eric,
Anderson
and
Joel
Barker
have
dived
in
and
started
to
tackle
the
developers
guide
to
kubernetes,
which
we've
needed
help
here
for
a
long
time.
B
So,
if
you're
interested
in
helping
them
out,
that's
how
you
contact
them
next
slide.
Please
and
then
deaf
stats,
Laurie
Apple
has
taken
this
on.
We
know
that
there's
a
bunch
of
dashboards
and
there
are
challenges
and
trying
to
get
the
information
that
you
want
out
of
it.
So
we're
gonna
be
doing
a
review
of
all
of
those
and
seeing
what
we
can
do
to
make
that
better.
Next
slide,
please
and
then
events
a
view.
B
As
you
know,
we
we
didn't
have
any
contributor
events
here
with
lack
of
real
cube
cons,
so
we're
currently
looking
at
a
virtual
contributor
summit
at
the
later
half
of
the
year
and
a
new
contributor
workshop
is
moved
under
the
mentoring,
sub
project
and
we'll
talk
about
that
during
the
announcement.
So
that
is
still
on
next
slide
and
bob
is
gonna.
Take
it
from
here.
Okay,.
C
Sir,
just
like
powering
through
all
these,
because
we
try
and
get
this
done
in
ten
minutes
becoming
github
management.
The
big
thing
is
that
the
github
or
cleanup
is
underway.
Most
of
you
have
probably
seen
the
email
that
went
out
regarding
the
new
policy
on
off-boarding
members.
Essentially,
if
you
there
are
no
contributions
within
the
past
18
months,
we'll
go
ahead
and
remove
people
from
the
various
owners
files
and
then
from
the
they
can
of
orgs.
C
C
Next,
big
one
is
mentoring,
so
we
have
a
couple
different
mentoring
initiatives.
One
is
we
have
started
to
ramp
up
and
build
programs
not
targeted
towards
new
contributors,
but
by
current
or
active
contributors
to
try
and
grow
the
amount
of
reviewers
improvers
in
the
SIG's.
Paris
is
sort
of
leading
this
effort.
C
If
this
is
something
you're
interested
in
for
your
cig
or
you'd
like
to
get
involved
or
you
know,
you're
looking
for
a
mentor
ping
Paris
or
reach
out
in
the
state
contribute
slack
channel.
The
other
thing
is
we
also
hold
our
monthly
meet
our
contributors,
which
is
sort
of
the
partner
to
our
office
hours
or
you
sort
of
have
a
bunch
of
you
know,
contributors
on
we
answer,
questions
live
and
that
gets
a
pretty
decent
mix
of
people
that
are
both
new
to
the
project
and
current
contributors
that
you
know
ie.
C
You
know
even
reviewers
and
approvers
are
looking
for
to
you,
help
like
navigate
new
areas
and
right
now
it's
at
7:30
a.m.
Pacific
time
and
a
second
session
is
being
scheduled
in
the
in
the
evening,
just
to
reach
more
people
less
ultimate,
just
sort
of
to
book.
A
book
and
the
other
one
is
the
new
contributor
workshop
has
been
moved
under
the
mentoring,
some
project
and,
as
George
said,
we'll
hear
more
about
that
under
announcements.
C
C
It's
it's
mostly
been
like
giving
done
too
much
with
it
for
a
while,
but
we
do
have
some
upcoming
stuff
that
we're
gonna
try
and
get
done,
namely
we
want
to
connect
github
or
like
a
reap
or
something
that
to
announcements.
Just
so,
you
can
PR
like
you
PR
something
in
there,
and
it
goes
out
to
multiple
notification
sources
like
the
announcements
channel
and
slack
and
then
the
next
big
one
is
channel
level
moderation.
C
Right
now
we
have
a
team
of
like
I,
think
20
moderators
that
help
moderate
all
hundred
thousand
users
in
slack
and
we'd
really
like
to
be
able
to
delegate
moderation
down
to
you
know
other
groups
that
can
support
that
sort
of
thing,
especially
for,
like
the
other
projects
that
aren't
necessarily
like
upstream
kudos
related
and
with
that
we're
sort
of
through
our
big
sub
projects.
A
couple
little
things
to
be
aware
of.
C
We
have
a
cap
that
is
in
flight
regarding
changing
the
issue,
triage
workflow
and
some
of
the
automation
around
there,
namely
we
are
removing
some
of
the
triage
labels,
so
like
duplicate,
not
reproducible
unresolved,
so
the
other
ones
are
getting
renamed
so
tree
are
supports,
becoming
kind
support
and
eats
information
is
lifecycle,
needs
information
and
we
are
adding
two
of
them.
This
is
still
provisional
by
the
way.
So
it's
like
I
think
it
might
be
a
little
double
check,
but
we
are
adding
needs
triage,
which
is
essentially
every
thing
coming
into
you.
C
C
Another
big
thing
that
I'm
sure
a
lot
of
people
saw
is:
we
are
now
working
with
a
another
working
group,
the
naming
working
group.
This
is
mostly
about
changing
some
of
the
language
that
we
use
throughout
the
project,
with
a
long-term
goal
of
renaming
the
master
branch
to
main
that
one
will
take
some
time
and
we'll
be
working
with
github
and
a
few
other
groups
on
on
that.
One
and
there'll
be
plenty
of
notifications
that
dictated
for
that.
C
Okay,
through
it
all
last
little
thing,
we're
gonna
find
us,
you
can
ping
George
or
myself
or
Christopher
Nikita,
as
the
the
tech
leads
in
slack
on
github,
we
have
our
select
channel,
secure
mix
and
our
bin
list
is
there
at
the
bottom.
Gracie
key
drugs
just
feel
free
to
reach
out
and
ping
us
at
any
time
Thanks.
A
D
Thank
you
so
I'm
trying.
First
of
all,
let
me
truth,
myself
on
fabrics
upon
DNA
I'm,
tackling
in
sequencer
a
cycle
I
work
at
in
firmware
and
today
I'm,
giving
you
an
update,
customize
icon,
and
what
we
did
in
the
recent
cycle
so
see.
Costigan
cycle
is
a
is
a
big
see.
We
have
a
lot
of
sub
projects
I'm
in
dealing
in.
It
is
like
a
very
quick
summary
but
below
in
the
presentation,
which
is
also
touches.
The
meeting
me
2
minutes.
You
can
find
more
detail
about
some
of
our
projects.
D
So
during
this
cycle
we
had
around
of
improvements
on
our
lifecycle
management
tools
and,
for
instance,
in
custardy
ie.
We
just
release
the
beast:
0
3,
7
ready,
and
these
are
releases
a
lot
of
improvement.
Even
if
it
is,
it
is
a
minor
one.
It
is
a
patch
1
and
I,
encourage
everyone
to
test
it
and
send
feedback.
D
The
mini
project
was
a
very
active
and
they
they
added
docker
and
portman
drivers,
so
you
can
run
a
more
easily
mini
tube.
Luckily,
and
also
the
long-awaited
feature
for
multi
not
support.
It's
not
available
in
YouTube
also
copes
with
a
lot
of
work
during
this
cycle
and
the
the
new
cops
controller,
which
managed
not
not
liberating
for
now,
but
in
future.
D
Probably,
we
managed
also
more
at
aspect
of
of
a
customer
cycle
and
also
the
copper
container
did
a
great
job
in
improving
and
to
add
coverage,
and
now
they
have
an
impressive
matrix
of
operative
system
and
CNI.
They
are
testing
against.
To
finally
govern
mean
incumbent
mean
there
was
some
work
to
improve
resilience
for
Kabam
enjoy
process,
especially
for
the
less
performant
machines.
D
D
We
start
moving
the
coop,
it
could
be
a
double
us
from
connects
incubator
to
186
and
we
also
removed
the
ownership,
from
our
sake
to
the
rest
clusters
directly
implicate.
This
is
a
legacy
directory
and
in
the
next
slide,
I
will
talk
about
some
work
to
to
basically
get
away
from
what
is
in
this
in
this
territory,
and
also
we
introduced,
we
added
a
new
project
in
kubernetes
6,
which
is
the
cluster
API
provider
for
pocket
and.
D
Last
to
note,
we
murdered
a
cab
for
define
standard
for
communicating
class
the
repository.
This
is
really
useful
for
people
using
Mini,
Cooper
kind
or
such
type
of
the
local
developer,
environment
and
they're,
aligning
on
on
a
local
image
repository,
and
it
was
a
really
interesting
effort
that
joining
different
company,
including
tilt,
which
is
a
provider
of
developer
tools
that
many
people
are
using
and
also
we
started
planning
out
to
remove
language
considered
effective
for
our
project
and
and
some
work
is
already
undergoing.
D
So
what
are
the
plans
for
for
the
next
cycle?
We
want
to
the
city
committed
to
complete
the
work
of
removing
offensive
language
and,
and
we
and
we
will
synchronize
with
the
workgroup.
They
may
be
more
group
as
soon
as
possible
and
the
in
terms
of
features
of
new
feature
where
the
main
teams
we
are
committed
to
is
to
improve
the
the
story
about
intones.
There
is
already
an
added
on
project
and,
and
we
want
to
disability,
to
to
to
gain
a
new
traction
and
to
and
to
deliver
co-creators
art
as
soon
as
possible.
D
D
The
last
main
team
for
our
sig
is
to
continue
working
directly
to
replace
the
kuba
Escape,
which
is
under
the
the
famous
Refresh
trust
the
directory,
because
basically,
this
discreet
is
is
used
only
in
our
in
kubernetes
and
west
end
to
end
jobs,
and
this
is
a
really
strange
because,
basically,
we
are
testing
kubernetes
with
script
that
is
not
used
out
in
the
wild,
so
we
want
to
replace
the
script
with
cluster
API
end
to
end
test
jobs.
This
is
a
long-term
work.
D
There
is
a
lot
of
work
to
do
this
to
make
this
happening,
but
we
are
coordinating
with
many
other
six
to
make
it
to
make
this
happen,
and
also
another
work
that
we
see
coming
is
to
start
work
with
signal
to
manage
this
potential
speed
of
docker
team
from
cooperate
and
to
figure
out
how
this
will
affect
user
installation
of
the
cluster,
and
things
like
that.
So.
D
That's
not
how
you
can
contribute.
We
first
see.
Of
course
we
have
a
set
of
data
collection
and
we
are
trying
to
to
be
to
respond
quickly
and
be
a
really
fun
day
so
reach
out
please.
Otherwise.
You
can
go
to
search
for
one
cat
and
code
first
issue.
We
are
doing
our
best
to
triage
everything
and
to
get
all
the
issue
property
level
and
the
only
promise
that
I
wish.
We
are
spread
in
a
huge
amount
of
repositories.
C
I
know
the
doc
doc
open
in
the
month,
but
if
you
want,
you
can
just
read
off
the
updates:
I
think
he
put
them
in
there.
You.
C
A
A
A
A
E
All
right
so
I
want
to
start
by
talk
about
kind
of
where
we
are
within
providers.
So,
as
many
people
are
aware,
one
of
the
ongoing
and
long-term
goals
of
our
sick
is
to
remove
the
internal
Hellfighters
that
are
compiled
into
kubernetes
today,
and
so
it's
Davies
cloud
providers
are
AWS
as
your
GP
OpenStack
and
a
sphere,
and
so
our
sake
has
been
working
on
on
this
effort.
Quite
some
time
now,
I
believe
this
whole
thing
started
around
kubernetes
1
e
1.8.
E
So
it's
been,
you
know
over
10
releases
now,
and
so
we've
made
some
pretty
good
progress
over
the
last
10
releases,
but
it
hasn't
been
enough
and
we're
we're
still
slowly
but
surely
chipping
away
away
at
this
effort.
So
currently
we
have
3
of
the
5
natively
compiled
entry,
clinical
providers
who
have
GA
versions
of
the
auditory
provider
and
for
those,
oh
sorry,
sorry,
two
of
them
are
GA
one
of
those
beta
but
of
those
3.
They
have
like
actual
users
running
production
coverage
clusters
using
a
provider,
and
then
we
have
two
providers.
Aws.
E
Good,
so
yeah,
so
within
our
larger
ecosystem,
we
have
quite
a
quite
a
lot
of
the
external
or
other
true-color
butters,
as
you
can
see,
and
we
actually
have
significantly
more
external
providers
at
this
point
than
the
you
know.
The
original
entry
providers
in
kubernetes-
and
so
this
is
this-
is
the
you
know.
This
is
the
direction
that
we
are
to
go
really
on
and
clearly
this
out
of
three
external
models
working
and
so
in
this
list.
E
I've
also
included
the
auditory
versions
of
the
intro
providers
that
have
actual
users
using
it,
and
this
isn't
even
the
fullest,
I'm
sure
I
missed
a
bunch
of
cloud
providers
I.
This
was
just
kind
of
a
quick
github
search.
I
did
the
other
day,
so
I'm
sure
there's
also
to
consider
there's,
probably
a
bunch
of
proprietary
or
closed
source
implementations
of
cloud
provider
which,
which
is
all
good
and
and
what
we're
hoping
for
out
of
out
of
this
pluggable
model
of
the
Kunichika
provider.
E
E
It
was
not
something
we
really
support,
so
I'm
CC,
who
is
an
engineer
at
Google,
has
been
doing
some
great
work
to
bundle
all
the
cloud
provider
packages
in
the
way
where
we
don't
have
to
import
the
main
kubernetes
tree,
which
makes
it
which
makes
kind
of
like
the
go
modules
importing
all
the
go
modules,
a
much
cleaner
and
easier
easier
to
tell
vendor,
and
so
this
was
mostly
done
in
119
there.
If
there
are
a
few
peers
that
slipped,
but
by
120
they
should
be,
we
should
cut.
E
Next
there
was
a
cap
approved
for
by
our
sake
and
signal
in
zakat
for
an
external
registry
provider
and
the
couplet,
and
so
today
the
cubit
has
these
native
integrations
with
AWS
gtp
and
Azure
to
authenticate
to
their
cloud-based
registry
imagery
container
registry.
And
so
we
want
to
remove
those
native
integrations
and
have
a
pluggable
model
like
everything
else,
so
that
anyone
can
any
other
vendor
can
tell
the
cubelet
how
to
how
to
authenticate
to
their
container
registries.
E
Next,
we
heard
quite
a
quite
a
bit
quite
a
lot
of
feedback
from
from
various
cloud
providers
and
because
I'm
saying
that
our
controllers
can
be
very
aggressive
when
it
comes
to
reconciling
state
and
a
lot
of
those
controllers
when
they
reconcile
state,
they
reconcile
the
Cabrini's
resource.
To
you
know
a
cloud
resource
which
involves
you
know
sometimes
many
api.
Many
cloud
API
calls,
and
so
we've
heard
feedback
that
sometimes
it
gets
really
aggressive
and
a
lot
of
customers
end
up
hitting
their
API
quota
limits.
E
Because
of
that,
and
so
we
made
a
few
optimizations
in
our
controllers
to
try
to
be
moral
or
optimal
about
when
to
make
a
cloud.
Api
requests
and
then
whatnot
so
made
some
good
progress
on
that
for
119
and
then.
Lastly,
we
standardized
on
semantic
versioning
for
out
of
providers,
and
so
the
new
policy
we
have
in
place
is
that
a
tagged
version
of
a
external
cloud
provider
has
to
match
the
major
and
minor
version
of
the
kubernetes
release
that
it
supports.
E
So
when
we
release
119
we're
going
to
expect
cloud
providers
to
also
you
know,
tagger
release
119
X,
so
that
it's
super
clear
to
users
that
that
version
of
a
cloud
provider
is
supporting.
You
know
kubernetes,
190
and
so
on
all
right
plans
for
upcoming
cycles.
So
we
want
to
continue
and
get
broader
adoption
of
the
other
providers.
So
that's
especially
the
out
of
tree
providers
that
are
going
to
be
replacing
the
the
legacy
entry
providers,
so
I
mean
you
want
to
get.
E
You
know
AWS
as
your
in
DCP
the
external
cloud
providers
for
those
TGA
status.
So
we
can
have
really
good
confidence
to
finally
remove
remove
the
entry
and
create
providers
and
that's
the
yeah.
We
have
more
ongoing
work
to
make
it
really
simple:
to
implement
your
own
cloud
provider
right
now.
There's
an
API
like
there's
like
an
interface
you
have
to
implement
and
there's
some
code
that
you
might
have
to
copy
and
paste
from
the
Kaveri
book
to
get
it
working.
E
So
we
want
to
make
the
Cloud
Controller
manager,
package
and
library
super
easy
for
anyone
to
pull
down
and
implement
so
that'll
be
upcoming
for
120
121
all
right.
How
do
these
plans
back
to
you?
So
if
you,
if
you
manage
a
kubernetes
a
product
or
offering
and
you're
still
using
the
entry
cloud
providers,
most
likely
it'll
be
either
AWS
Azure
GCP,
please
plan
for
a
migration
we
are
targeting
to
remove
those
native
terminates
integrations
in
1.21,
so
please
no
upgrade
or
migrate
to
you
see
provider
nd.
E
So
things
we
need
from
you,
we
have
a
lot
of
cross-cutting
initiatives
requiring
you
know,
reviews
from
different
SIG's
for
kept
sinjar.
So
you
know
if
we
reach
out,
please
help
us
review
kept
some
PRS
and
yeah.
Lastly,
please
help
us
promote
the
auditory
providers.
I
think
this
is
really
challenging
for
us,
because
we
don't
we
don't
want
kubernetes
favouring
any
specific
vendor
by
having
it
compiled
Natalie,
but
at
the
same
time
you
know
we're
aware
that
many
users
are
using
this.
E
The
native
cloud
provider
integrations
in
production
and
you
want
to
be
mindful
about
not
breaking
users
as
they
upgrade
so
getting
us
to
that
finish
line
where
we
can
finally
remove
the
injury
cloud
providers.
It's
going
to
be,
you
know
a
community-wide
effort
and
we
all
have
to
be
on
board
with
that
and
get
there
together.
So
yeah,
please
help
us
build
those
out
of
tree
integrations
and
help
us
promote
those
and
get
users
on
those.
Instead
of
the
you
know,
the
the
entry
ones
that
most
people
are
using
today
and
yeah.
E
F
Hey
everyone:
can
you
see
this
great?
This
is
a
signet
work.
Community
meeting
update
for
July,
2020
and
yeah
I'm
Rob
Scott
I've
got
a
lot
to
get
through.
Signet
work
has
been
incredibly
busy
over
the
past
quarter
and
beyond,
and
we've
got
lots
going
on.
So
this
is
just
going
to
give
a
high-level
overview
of
some
of
the
bigger
items
were
working
on.
I
am
missing
things
that
are
still
significant,
but
I've
tried
to
just
highlight
some
of
the
bigger
things
that
we
have
achieved
and
are
working
on
right
now.
F
So
first
off
of
big
news
is
that
ingress
has
gone
GA
and
119
ingress
is
now
a
v1
API.
It
was
a
long
road
to
get
there.
We
made
a
number
of
updates
along
the
way,
but
it
is
finally
a
v1
API
with
that
and
119.
We
are
seeing
a
few
new
teachers
go
to
beta
endpoint
slices,
the
API
and
some
functionality
was
already
beta,
but
the
big
news
in
119
is
the
core
functionality
in
coop.
Proxy
is
now
beta,
which
means
people
will
be
using
it
by
default.
F
Sctp
is
support,
form.
No
new
protocol
and
hack
protocol
allows
you
to
specify
an
application
protocol
on
a
service
port
for
alpha.
We've
made
huge
progress
on
dual
stack:
implementation
updates
and
I'll
get
into
a
lot
of
these
in
just
a
bit.
So
first
we'll
start
with
ingress
as
part
of
the
v1
API.
There's
three
relatively
new
changes
that
we
made
to
get
there.
We
added
a
new
ingress
class
resource
to
replace
the
class
annotation,
so
you
can
reference
it
controller
and
a
parameter
C
Rd
with
that
ingress
class
resource.
F
F
F
F
Mirrored
to
endpoint
slices
with
this
controller
and
also
we
have
cout
proxy
windows
support
and
alpha
now
for
endpoint
slices,
alright.
So
next
up
and
and
the
last
thing
that
I
wanted
to
cover
for
things
that
we
have
accomplished
our
dual
stack.
So
dual
stack
is
already
in
alpha,
but
we
identified
some
significant
changes
that
we
needed
to
make
and
I
wanted
to
highlight
a
couple
of
them
here.
D
F
There
is
a
new
IP
family
policy
field
which
allows
you
to
specify
single
stack,
whether
you
prefer
dual
stack
or
require
dual
stack
for
a
service
and
also
an
IP
family's
field
which
allows
you
to
specify
the
IP
families
that
should
be
provided
for
a
given
service
and
every
dual
stack
service
will
have
to
cluster
our
keys
corresponding
to
that
IP
families
field.
So
if
ipv4
is
first
in
that
order,
you'll
see
your
ipv4
address
first
as
well,
and
I
can't
emphasize
this
enough.
This
is
an
absolutely
massive
project.
F
It
spans
large
portions
of
codebase
odds
requiring
changes
everywhere.
It
feels
like
so
this
will
likely
span
to
other
SIG's
and
will
need
some
help.
Reviewing
the
PR,
and
on
that
note
there
is
already
a
huge
implementation.
Pr,
that's
ready
for
review
as
of
I.
Think
yesterday,
so
we're
hoping
to
get
this
really
early
and
be
120
release
cycle,
but
that's
just
the
big
things
that
have
been
happening,
but
there's
also
a
lot
that
we're
excited
about
coming
up
so
service.
F
Api's
is
a
group
that's
working
on
a
more
advanced
alternative
to
the
ingress
api
network
policy
plus
plus.
Is
a
working
group
exploring
ways
to
improve
B
network
policy
API
we're
doing
a
lot
of
work
towards
topology,
aware
routing
and
I
have
to
at
least
mention
multi
cluster
services.
A
cig
network
does
not
actually
own
multi
cluster
services,
but
we're
very
involved
in
the
implementation.
F
F
So
this
is
a
new
set
of
api's
that
provides
a
more
advanced
alternative
to
ingress,
as
so
ingress
is
great
as
a
very
simple
way
to
configure
ingress
for
your
cluster,
but
it's
missing
some
advanced
functionality
that
a
lot
of
people
wanted,
so
things
like,
say,
traffic
splitting
or
traffic
mirroring
or
advanced
load
balancer
configuration
that
have
traditionally
been
rep
rather
difficult
to
achieve.
Are
things
were
trying
to
make
very
straightforward
with
service
api's
HOD?
F
This
API
has
three
core
resources:
gateway
class,
which
is
roughly
equivalent
to
ingress
class
and
ingress
API
and
then
gateway
and
route
are
kind
of
a
more
advanced
concept
of
ingress,
roughly
split
things
out
and
added
a
lot
more
configuration
so
we're
working
on
an
alpha
API
design
as
a
cig,
Network
sub
project
we've
had
lots
of
great
press
participation,
especially
from
maintainer
z--,
ingress,
controller,
maintainer,
z--
and
other
people
who
are
in
the
community.
We
would
really
appreciate
more
participation,
especially
from
end
users.
When
we've
had
feedback
from
nine
users,
it's
been
really
helpful.
F
We
could
use
more
so
if
you're,
using
the
ingress,
API
or
interested
in
networking
in
kubernetes,
we
would
really
love
your
feedback.
We
have
meetings
every
week
on
thursdays
and
the
meeting
today
is
later
this
afternoon,
hoping
to
release
an
initial
alpha
api
in
august,
and
hopefully
alpha
implementations
will
follow
shortly.
After
that,
some
network
policy
plus
plus,
is
another
working
group
in
cig
network
and
that,
as
the
name
suggests,
is
trying
to
explore
ways
we
can
expand,
prove
the
network
policy
api.
The
first
step
was
collecting
a
whole
bunch
of
user
stories.
F
I
think
there's
a
little
over
20
now
for
network
policy,
implementation
and
potential
updates,
and
most
of
those
have
been
related
to
no
policy
reach
ability,
graphing
service
level,
policy,
external
policy
or
even
proxy
or
massive
mesh
related
policies.
There's
weekly
meetings,
every
Monday,
where
they
discuss
these
issues
and
there's
also
a
couple
dots
that
I've
linked
in
these
slides.
F
So
if
you're
interested
at
all
in
network
policy,
there's
some
really
interesting
work
going
on
to
improve
that
and
finally,
the
last
thing
I
wanted
to
cover
was
topology,
aware
routing.
So
this
is
another
feature
that
is
already
in
alpha,
but
it
requires
manual
configuration
there's.
The
initial
approach
was
really
not
default.
Anything
you
had
to
specify
every
bit
of
topology
we're
routing
it
wanted
and,
as
we
looked
at
it
more,
we
thought
well
what
most
people
want
is
going
to
be
identical.
F
Most
people
would
prefer
that
traffic
stays
as
local
has
just
try
to
keep
things
as
efficient
as
it
makes
sense.
So,
instead
of
I'm
going
further
down
that
path,
we
may
revisit
it.
We
wanted
to
see
how
much
we
could
accomplish
with
some
kind
of
automatic
algorithm.
So
users
would
not
need
to
configure
anything
there's
a
few
steps.
Here.
We
have
a
cap
to
standardized
topology
labels.
F
F
We
would
love
contributions,
as
you
can
tell.
We
have
a
lot
going
on
in
cig
network.
If
you're
just
getting
started,
I
recommend
you
come
to
cig
network
meetings,
we
have
them
every
other
Thursday,
including
next
week.
We
do
issue
and
PR
triage
to
start
each
meeting
and
that's
a
really
great
way
to
get
involved
as
an
example
with
issue
triage.
We
are
really
just
looking
for
people
to
help
understand
issues,
ask
for
more
information
and
help
that
help
understand
the
real
bug
behind
it.
You
don't
need
to
solve
the
issue.
F
You
just
need
to
help
gather
more
information.
It's
a
great
way
to
get
to
know
the
rest
of
the
Signet
work
community
and
to
get
started
on
some
issues.
You
can
also
get
involved
in
one
of
our
working
groups
if
either
service,
api's
or
network
policy
are
interesting
to
you.
I
highly
encourage
you
to
come
to
one
of
our
meetings
and
parts,
and
with
that
here's
where
you
can
find
us
on
slack
on
our
mailing
list,
our
chairs
are
Casey's
Davenport,
Dan,
Williams
and
Tim
Aachen
and
I.
Think
I
may
have
gone
over
time.
A
So
I
wanted
to
thank
all
our
presenters
today
we
got
to
know
a
lot
of
interesting
stuff,
so
thank
you
for
sharing
it
with
us.
Moving
on
to
the
announcement
we
have.
The
new
contributor
workshop
kick
off
to
be
announced
in
the
mailing
list
today.
So
keep
a
lookout
for
that
and
we
don't
have
regarding
the
host
for
the
next
community
meeting.
A
A
Okay,
then
I
think
like
the
next
would
be
the
shoutout,
so
it
can
be
read
on
the
agenda
and
again,
thank
you
all
for
your
time.
Oh-Hoh,
please
hope
to
see
you
in
the
next
community
meeting
I'll
take
some
time
to
try
it
out,
write
down
the
notes
and
post
it
so
I
oppose
that
soon
as
soon
as
optical
collars
done.
So.
Thank
you
all
for
your
time.