►
From YouTube: Kubernetes SIG Federation 20170424
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
B
C
A
Yes
shopping,
then
you
cleaned
up.
The
dog
has
really
said
last
time,
so
thanks
a
lot
for
that
sure
sea
level
of
la
légion
at
it.
So
let's
get
started.
The
first
is
the
cluster
version.
Skew
should
like
marui
broader.
You
can
double
click
on
the
link
so
yeah.
There
are
multiple
issues.
As
I
said
the
same.
The
main
one
was
how
to
determine
if
a
feel
disappointed
and
so
I
respect
seems
like
that
I
think
to
me
to
figure
that
out,
but
we
still
don't
have
a
clear
answer
for
what
to
do.
A
G
It
for
implementation
purposes,
I
think
those
issues
are
separable
like
detecting,
there's
a
field
difference
and
then
actually
dealing
with
it.
So
I
mean
if
that's
a
priority
for
17
I
would
say
we
could
start
work
on
getting
the
detection
mechanism
in
place
and
then
we'll
have
time
to
figure
out
the
best
strategy
or
best
configuration
options
for
the
user
to
deal
with
differences.
Yeah.
A
G
But
but
I'm
saying,
like
I,
think
that
we
could
split
it
up
rather
than
having
a
sort
of
pause
implementation.
Until
we
have
everything
done
just
focus
on
like
designing
for
detection
of
the
differences
and
implementing
that
and
then
separately,
because
I
think
the
other
ones
a
lot
more
big
like
how
I
should.
G
D
And
it
could
end
up
being
either
very
complex
because
we
can't
necessarily
an
overriding
behavior
for
all
objects,
and
you
could
imagine
a
world
in
which
the
user
wants
certain
objects
to
hold
a
writing.
Others
not
to
you
have
to
decide.
Is
that
something
that
we
want
to?
Let
the
user
do
or
it
doesn't
see
me
I,
agree,
it's
not
straightforward
to
do
what
what
you
will
feel
is
not
supported
and.
G
I
mean
this
kind
of
ties
into
the
the
need
for
I.
Don't
like
status
is
maybe
what
I've
been
thinking
about
it
as
but
I
mean
up
to
now.
The
goal
has
been
like:
I
want
to
be
able
to
propagate
objects
to
these
clusters
and
I
want
to
make
sure
that
that's
consistent
over
time
and
when
new
clusters
come
on,
but
I
think
that
there's
a
lack
of
visibility
into
the
system.
You
know
if
I'm
federating
resources,
how
do
I
know
which
resources
have
been
successfully
federated?
G
How
do
I
know
you
know
when
there's
a
problem,
what
the
problem
was
and-
and
that
doesn't
even
begin
to
touch
what
happens
when
we
talk
about
policy
and
how
we
was
all
policy
conflict.
So
this
is
kind
of
a
separate
issue
and
it's
not
on
the
agenda,
but
I
do
think
that
it's
important
to
raise
a
cycle
and
start
thinking
about
it's
not
into
Mary
yeah.
A
G
A
I
The
users
do
can
I
suggest
we've
got.
We
got
a
lot
to
get
through
to
this
week,
including
prioritizing
all
of
the
1.7
work.
So
I
would
suggest
that
that
we,
you
know
hive
of
a
separate
group
whoever's
interested
in
this
whole
topic.
I
think
all
of
the
issues
you
mentioned
LaRue
there
are
actually
solutions
to
so
I
would
suggest
that
we
put
put
a
design.
We
have
a
has
the
time
and
inclination
to
put
a
design
document
around
the
version.
A
I
Ultimately,
doesn't
have
to
happen
for
17
with
ya
ideal,
but
you
know
I,
don't
think
it's
it's.
You
know
a
train
smash.
If
we
just
say
you
know,
you're
FCP
must
be
the
same
version
at
the
moment.
You
know.
Most
of
these
things
are
not
even
in
production
anyway,
they're
alpha
and
beta,
and
until
we
graduate
out
of
there
you
know
the
versions
have
to
line
up
a
ballsy.
You
know
does
likely
weird
thing
if.
G
I
So
about
that
I
actually
personally,
quite
like
the
idea
where
we
identify
reasonably
well
defined,
features
that
optimism
I'm
reasonably
comfortable,
putting
them
in
the
features,
repo
and
labeling
them.
As
you
know,
some
future
version,
as
opposed
to
any
specific
one,
I
don't
have
anyone
else
disagrees
and
I,
don't
know
if
they're
contradicts
any
other
advice,
but
we
need
some
some
well-defined
place,
I
think
to
to
kind
of
put
these
things
down,
because
you
know
I
get
at
conferences
and
elsewhere.
I
get
questions
on.
Oh
you
know
when
you
guys,
are
you
going
to
handful?
I
You
know
security,
two
clusters
and
it's
really
handy
to
be
able
to
send
them
to
the
features.
Repo
and
say:
look,
that's
that's
where
it
is.
You
know:
here's
the
status
is
or
isn't
it
the
long
document
there
isn't
is
or
isn't
a
plan
to
release
it
in
a
given
version,
rather
than
having
that
stuff
scattered
in
everybody's
brains,
all
over
the
place.
Yeah.
A
Maybe
I
think
he
chose
rubbers
good,
I
generally
using
them
to
the
spreadsheet
that
we
have
which
has
links
to
other
issues.
Okay-
and
that
is
the
tradition
specific,
so
they
can
just
look
at
the
phoenician
things
winning,
but
we
can
do
that
same
with
the
repo
is
pointing
to
all
issues
with
the
Federation
label.
Ya,.
I
J
G
Don't
know
I'm
a
little
bit,
hesitant
to
say
spreadsheet
should
go
away,
because
github
is
really
terrible
at
providing
the
kind
of
view
that
a
spreadsheet
can
provide
like
arbitrary
and
being
able
to
like,
rather
because
I've
looked
at
the
list
of
of
say
issues
for
Federation
or
you
know,
work
in
progress,
and
it's
just
it's
amorphous
to
the
uninitiated.
I
mean
I
agree
that
it
needs
to
be
visible,
but
spreadsheet
gives
you
a
tabular
like
summary
and
until
github
has
a
similar
capability.
I
Anybody
can,
you
know,
keep
their
own
spreadsheets
or
whatever,
but
at
some
point
as
long
as
soon
as
you
have
month,
copies
are
the
same
information
they
get
out
of
sync,
you
know
one
is
true
and
the
other
one
is
false
and,
and
it's
ambiguous
as
to
where
to
send
people
to
so
yeah
I
mean
I,
totally
agree
with
all
the
issues
you
raised.
I'm
not
sure
that
having
multiple
copies
of
the
same
amount
of
sync
with
each
other
is
the
best
solution.
I
K
K
Forever,
this
is
david
erotic,
so
from
that
perspective,
that's
done.
I
think
there
was
speaking
earlier.
A
Christian
if
you
in
Quentin
would
like
to
be
I'll,
get
a
got
some
folks
on
the
line
to
do
it.
I
wiII
that
said
it
we're
not
married.
You
know
when
we
went
to
the
feature
repo,
never
suggested
that
there
was
one
and
only
one
tool
out
there.
K
The
problem
we
had
is
we
needed
a
single
place
that
tracked
versions
over
time,
allowed
assignments
back
and
forth
to
people
and
that
that
was
free
flowing
to
allow
comments
that
persisted
indefinitely.
That's
that's,
basically
why
we
said
you
know
what
spreadsheets
are
good
is
a
summary
but
they're,
not
an
ongoing
tool,
and
and
so
again,
if
there's
other
things
that
we
want
to
choose
as
a
cig
that
makes
total
sense.
You
know
Trello
or
I,
don't
know
we
can
figure
out
a
way
to
make
spreadsheet
support
in
that
way.
I
Yeah,
I
think
the
features
we
brief.
It
makes
sense
me
if
we
can
just
build
a
slightly
more
convenient
view
onto
it,
like
Maru
said,
which
seems
totally
doable
and
like
a
kind
of
a
fun,
a
fun,
a
part-time
project.
For
somebody
who
wants
to
water
around.
With
these
things,
it
would
seemingly
be
very
valuable
to
multiple
cigs
rather
than
yeah,
no.
K
K
L
However,
I
just
caution
as
we
come
up
to
deadlines,
that
the
single
pane
of
glass
that
we
have
right
now
for
tracking
its
Federation
features
and
Federation
issues
that
don't
fall
under
the
features
repo
is
in
the
spreadsheet
and
as
deadlines
start
coming
up,
we'll
need
to
paint
people
to
make
sure
everything
gets
skinny.
Waiting
for
these
days.
I
would.
G
Say
that
there's
like
a
general
just
need
for
reporting
capability
to
come
out
of
github
I
mean
I'm
kind
of
involved,
with
the
release
side
of
things
I'm
like
being
able
to
look
at
bugs
and
know
which
cigs
you
know,
Ron
them
and
what's
the
status
of
those
issues
like
an
issue
tracker?
Does
that
any
good
issue?
Tracker
and
github
is
just
a
terrible
issue
tracker
and
we
need
to
smell,
make
it
work
for
a
really
large
project.
As
an
issue
tracker.
K
I
Pretty
good
spreadsheets
linked
here
somewhere
I
think
anyway,
so
we're
we're
20
minutes
into
a
60
minute
meeting
and
we
have
too
far
it
has
all
of
the
1.7
features
and
get
them
captured.
An
issue
today,
I
believe,
is
the
plan.
Should
we
limit
the
time
we
spend
on
the
other
things
in
this
gender
just
so
that
we
can
get
through
that
or
perhaps
Jeff
inches
and
prioritize
them
and
then
come
back
to
to
this
agenda
once
we
have
completed
that.
L
I
I
D
I
I
I
Yeah
a
lot
of
them
well,
so
then
that
is
all
of
the
features
that
are
currently
there
that
are
open,
and
so
there
are
a
couple
that
I
think
we
can
trivially
close
and
or
so
we
there's
a
bit
of
a
process
here
that
I
don't
know
whether
it's
been
agreed
upon.
Yet
there
are
milestones,
15,
16
17,
then
one
called
next
milestone,
which
I'm
not
sure
what
the
intention
for
that
is,
and
we
have
a
bunch
here
that
are
labeled
17,
which
I
just
went
through
them.
I
A
I
I
But
it's
not
clear
to
me
how
what
we
should
do
to
track
that
this
feature
was
alpha
in
release
1.5
and
went
to
beta
in
16
and
went
to
GA
in
17,
for
example:
I
guess
there
is.
You
will
actually
see
that
if
you
go
dig
through
the
particular
feature,
you
will
see
that
the
you
know
release
was
changed
from
those
to
that
and
the
label
was
changed
from
alpha
to
beta.
So
you
could
go.
I
I
Features
repo:
we
have
there's
a
release
well
as
a
milestone
field
which
can
be
16,
17
or
next
milestone,
and
then
there
is,
but
there
are
a
set
of
labels
stage,
alpha
stage
beta
stage
release,
and
then
there
are
a
bunch
of
deprecated
labels
which
are
like
alpha
1
point,
5,
18,
1.6,
etc.
I
So
my
the
way
I
have
used
this
right
now
is
to
take
the
current
state
of
the
feature
and
reflected
in
its
milestone
and
it's
stage
and
then,
as
the
feature
moves
from
alpha
to
beta
2
will
never
be
paint
another
milestone
and
we
change
the
label
so
that
we
can,
you
know
see
where
we
are
is
that
is
that
the
intended
use
of
these
labels
and
things?
Yes.
K
K
I
So
in
it's
more
like
some
future
milestone,
as
in
the
next
milestone,
presumably
is
17
like
what
anyway
ok
cool,
so
we
will
use
it
that
way.
I
mean
the
one
downside
of
this.
Is
that
there's
no
way
that
you
can
put
a
single
place
and
say,
oh
I,
can
see
this
feature
was
alpha
in
15
feet
in
16
and
GA
in
17.
You'd
have
to
dig
through
the
history
of
all
each
one
of
these
features
to
find
out.
I
K
Is
an
alternative
that
I
have
seen
some
groups
use
and
that-
and
this
is
totally
an
alternative-
your
feature
is
federated
resource
quota
alpha
and
then
you
have
the
label
stage
alpha
just
for
the
purposes
of
sorting.
But
when
that
feature
is
done,
you
close
it
and
you
create
a
federated
resource
quota.
Beta,
ok,.
I
And
you
some
out
channel
movies
together,
so
that
you
can
see
them
yeah
I,
guess
you
still
not
kind
of
lose
some
Thunder
sample,
so
I
don't
really
care
either
way.
I
just
want
to
be
consistent
with
everybody
else
is
using
it.
Otherwise,
his
features
Reaper
becomes
useless
because
you
can't
you
know,
filter
if
each
sig
is
doing
it
differently.
So
if
you
can
just
tell
us
which
of
those
two
approaches
to
use,
we
will
I
think
use
it.
K
I
Until
we
announced
that
they
will
just
use
a
single
issue,
/
feature
and
we
will
just
roll
it
forward
from
alpha
to
beta
2
GA,
and
we
will
change
that
if
we
hear
otherwise
from
the
PM
group
excellent,
should
we
just
work
that
we've
got
17
of
them
here?
Should
we
just
work
our
way
through
the
one
at
a
time
and
agree
who's
working
on
them?
I
A
I
A
And
then
you
are
clearly
the
spreadsheet.
We
also
made
sure
that
for
each
feature,
there's
a
feature
Luna
and
there's
also
reviewed
to
make
sure
there
is
enough
review
bandwidth,
because
what
happened
in
one
dot
six
was
like
with
jobs.
There
was
someone
who
did
it,
but
we
didn't
have
enough
time
to
review
that.
You
don't
want
to
do
in
197
sure.
I
M
A
M
Yes,
plus
about
the
design
I
think
I
would
want
somebody
from
this
to
review
the
design
per
se,
because
I
have
received
many
comments
on
it,
but
probably
nobody
from
this
petition
sing
shaykh
husain,
indeed
review
it.
We
left
yes,
yes,
yeah
I
saw
12
comments
only
from
you,
Emily
I
thought
you
have
just
gone
through
like
an
overview.
I
need.
C
M
G
C
Good,
you
should
definitely
get
someone
from
the
gaps
to
review
the
proposals
because
they
have
a
lot
of
experience
working
through
this
shoes
yeah
that
already
yeah
okay,
I
saw
item
left
number
of
comments,
yeah.
M
C
M
I
J
G
I'm
I'm
a
little
bit
concerned
at
the
number
of
things
that
are
depending
on
how
we
think
about
policy
and
how
we
apply
it,
and
this
is
one
of
them
and
we're
just
kind
of
going
on.
Yes,
anyway,
I
don't
sort
of
devolved
into
unrelated,
but,
like
I,
just
think.
We
really
do
need
to
have
a
discussion
as
a
group
about
this
stuff
before
we
move
forward.
G
A
A
G
I
guess
my
concern,
though,
is
that
the
current
design
of
this
feature
is
predicated
on
a
lot
of
assumptions
like
number
one:
the
fact
that
we're
federating
resources
and
applying
metadata,
/
resource
and
I'm
not
sure
that's,
realistic
and
I'd
like
to
at
least
have
a
discussion
about.
You
know
that
issue
before
we
say.
Yes,
we
need
this
specific
implementation
based
on
these
assumptions
anyway,
I'll
stop.
A
Yeah-
and
it
did
have
some
other
discussion
in
the
issue
as
well.
I
agree
with
the
cherry
point.
I.
Don't
think
that
the
current
implementation
is
too
tied
to
the
annotation
stuff.
It
does
assume
that
there'll
be
something
on
the
resource
which
it
can
use
to
specify
which
all
just
as
it
should
go
to
and
I
know
it's
using
the
rice
election
right.
G
I
Lets
say,
comment
on
the
design
there
and
just
to
remind
everyone.
This
is
an
alpha
feature.
So
so
the
intention
is,
you
know,
publish
something
which
has
not
yet
decided
upon
api
to
invite
feedback
from
users.
That
is
the
intention
of
an
alpha
feature.
I
mean
clearly
we
shouldn't
completely
screwed
up
and
build
something
that
we
think
it's
ridiculous,
but
it
doesn't
have
to
be
correct
and
it
can
be
changed
afterwards.
That's
the
intention
of
an
alpha
APR.
M
A
M
M
And
module
not
enough
for
particle
auto
scanner,
probably
is
I'm,
not
sure
we
need
to
see
the
spreadsheet
yeah
I
see
marching
yeah
they're
in
volusia
yeah.
What
I
mean
to
say
is
much.
It
might
not
be
able
to
take
up
jobs
from
this
same
right.
So
then,
where
it
falls
into
your
coat.
So
maybe
you
need
to
assign
it
to
him.
Christian.
L
I
I
A
You
don't
really
have
any
known
issues
for
config
maps:
I
mean
I'm,
fine,
calling
it
beta
I.
Don't
there
was
I.
Think
one
issue.
We
didn't
have
any
e
2
e
test
at
all
for
content
like
this,
that
I
remember.
No,
so
we
should
fix
that
ensure
we
have
enough
testing.
Maybe
we
don't
have
any
known
issues.
I
A
Garage,
it
will
be
and
I'm
for
feature
so
I
don't
want
to
tie
that
to
like
all
those
resources.
The
when
I
say
like
a
config
map
is
Vida.
I
would
say,
like
the
use
case,
of
spreading
config
map
around
to
all
the
clusters
that
works
and
like
its
federated
across
or
clusters,
having
the
ability
to
restrict
it
to
a
specific
set
of
clusters.
Yes,
we'll
have
an
alpha
implementation
in
one
word,
seven,
but
I'm
not
sure
we
need
to
block
confirm
our
beta
on
that.
I
A
G
You
can
fig
map.
In
any
case,
I
mean
really
you're
going
to
be
distributing
applications,
you're
not
going
to
distribute
individual
resources,
and
there
needs
to
be
a
way
to
decide
where
those
application
resources
go.
As
a
group,
yeah
I
would
be
on
board
of
not
blocking
config
map,
because
it
is
a
larger
issue.
Yeah.
I
G
And
it
is
handling
right.
Oh
we're
talking
about
the
difference
between
a
developer,
I'm,
implementing
Federation
of
resources
versus
how
I
do
it
as
a
user.
It
may
be
fine
if
I'm
dealing
with
an
individual
cluster
to
say
I'm
just
been
dealing
with
a
paris'
or
so
here's
my
secret
of
my
config
nap,
but
realistically,
especially
when
I
think
about
how
I
want
to
move
like
how
I
want
to
actually
Feder
ate
like
an
application
like
I'm
gonna,
be
thinking
in
a
larger
scale.
It's
not
the
same
thing
as
Cooper,
Nettie's
and
I.
I
With
that
but
yeah,
let's,
let's
have
a
separate
discussion
about
that.
I,
don't
want
to
bring
everything
to
grinding
hope,
because
we
don't
have
a
you
know,
an
application,
abstraction
in
communities
or
all
Federation
APR
I
mean
we've
got
helm
and
all
of
these
things-
and
we
can
certainly
start
looking
and
thinking
about
them
and
I
think
we
need
to
provide.
I
You
know
the
basic
building
blocks
that
you
can
do
the
kinds
of
things
you're
describing,
for
example,
by
choosing
a
specific
cluster
or
set
of
clusters
to
put
things
in
it,
generation,
API
level
and-
and
yes,
I
agree,
people
will
use
those
to
you
know,
make
sure
that
their
application
ends
up
in
the
right
place
and
etc,
and
that
their
config
map
ends
up
in
the
same
clusters
as
they
replica
set
or
whatever
is
using
it.
But
I
don't
think
we
need
to
solve
all
of
those
problems
right
now.
G
G
A
B
Actually,
yeah
yeah
muah
way
we
had
put
up
our
design,
but
it's
like
okay,
so
we
are
doing
the
small
small
pieces
actually
so
later
election
feature
was
of
one
part
of
it
and
I
also,
I
think
I
think
it
should
be
viewed
by
the
much
bigger
audience.
I
believe
I
think
federation.
I
think
many
people
have
not
seen
that
design
document.
I
believe
so.
I
think
we
can
also
work
on
that
design
going
forward.
It's
definitely
not
in
there
for
1.7,
so
maybe
we
can
target
it
to
the
next
Malstrom
that.
I
N
I
I
Can
we
you
know
I,
don't
want
us
to
end
up
playing
catch
up
with
all
of
these
types,
because
there
are
constantly
things
being
added
to
many
of
the
API
types
and
I.
Don't
want
us
not
to
be
able
to
graduate
out
of
alpha
until
we
support
the
latest
greatest
features
of
any
given
type.
Another
example
came
up
recently
yesterday,
with
with
jobs
they
now
they're
busy,
adding
a
like
a
job
queue
thing
and
whatever,
and
we
obviously
don't
have
that
in
Federation.
A
Yeah
I
was
going
to
say
when
we
added
support
for
rolling
up
day
to
give
insights
and
coordinators.
There
were
API
changes
as
well,
and
since
we
share
the
same
API,
we
do
have
those
fields
in
federated
demon
sets
and
like
as
a
user
I
could
lead
the
API
documentation
of
fidelity
when
such
to
support
these
fields,
I
can
set
them
against
submitted
even
said,
which
says:
I
won't
rolling
update,
and
then
it
doesn't
happen.
So
that's
not.
The
usual
expectation
is
not
good
yeah.
We.
I
A
I
think
it's
an
extensions
we
won't
be
done
in
common,
it
is
I'll
have
to
check,
but
I
think
it
is
an
extension
beta
1,
so
it
might
not
be
GA.
Ok,
I
do
it.
We
will
have
the
problem
of
catching
up,
but
I
think
once
we
I'll
beta
is
well
with
the
same
resource.
Then,
when
we
change
the
area
in
common
it
is,
we
can
require
that
the
same
change
should
be
made
in
Federation
as
well
and.
I
I
B
I
I
think
that's
going
to
be
a
continually
moving
target
and
we,
you
know
in
the
in
the
extreme
case,
we'll
never
be
able
to
do
anything
because
we'll
never
be
up
to
the
latest
greatest
thing
and
I
don't
want
us
to
get
into
that
situation.
It
tons
of
people
out
there
using
daemon
sets,
as
they
were
at
least
version
1.5
or
sex
or
whatever,
and
they
want
to
Fenner
ate
them,
and
they
should
be
able
to
do
that
and
the
ones
who
I
mean.
I
A
To
do
one
time,
I
was
just
going
to
say
to
do
what
Daniel
and
I
nanny
both
of
you
didn't,
have
our
own
API
version
and
locking
locking
to
it
so
that,
if
govern,
it
is
add
some
features
we
are
not
impacted.
We
can
do
that
battle
to
do
that.
We'll
have
to
maintain
our
own,
like
API
machinery
code
as
well.
Right
now,
we
just
reuse
whatever
there
is
in
coitus.
If
you
want
to
sort
of
diverge
or
for
Kate
will
have
to
maintain
our
own
versions.
Yeah.
I
I
mean
that
sounds
like
a
big
problem.
It
sounds
like
we
need
to
think
more
about
this,
but
let
me
put
it
on
the
table
and
have
a
discussion
and
the
pros
and
cons
of
you
know
the
two
different.
Essentially
the
two
different
approaches
both
have
pros
and
cons,
and
we
need
to
wear
them
up
in
the
side
which
ones
you
change
them.
So.
A
A
A
I
C
D
G
D
A
I
love
did
the
issue
aside
into
Jonathan
and
I
wondered
sir
cool
next
is
generalized
cluster
selection
and
waiting.
So
the
first
part
of
a
generalized
cluster
selection,
danville
cell,
is
working
on
it,
which
we
talked
about
earlier.
It
would
just
be
an
elf
I,
it's
based
on
annotations
right
now,
but
the
aim
is
to
get
it
in
one
doctor.
G
A
I
We
should
just
make
sure
we
will
talk
about
the
same
thing
here
so
so,
as
far
as
I'm
concerned,
we
have
a
generalized
cluster
selection
waiting
in
annotations
at
the
moment,
and
it's
used
by
a
multiple
resource
types.
You
know
the
implementation
itself
might
be
not
perfectly
generalized,
but
certainly
the
API
is
not
specific
to
any
particular
type,
though
so.
I
M
A
M
Yes,
so
preferences
are
specified
during
and
me
using
annotations,
and
that
is
what
I'm
talking
about
the.
A
G
But
but
this
is
tied
into
like,
should
we
be
using
annotations?
Are
there
other
mechanisms
and
also
the
issue
of
you-
know
resource
versus
application,
Federation
so
I
mean
there's
just
things
to
talk
about,
I
certainly
have
opinions,
but
I'm,
not
saying
they're,
correct,
I
think
you've
really
gone
through.
G
We
haven't
really
gone
through
a
process
of
codifying
like
what
are
the
use
cases
and
what's
the
best
solution
to
me,
it's
like
we've
gotten
good
solution
for
some
of
the
use
cases
based
on
like
what's
quickest
to
implement
what
can
we
get
out
in
front
of
users
and
I?
Think
at
this
point
we
just
need
a
little
bit
more
thought
to
go
into
it.
So
Reuben.
G
A
I
was
going
to
say
I'm
fine,
adding
another
meeting
everyone
who's
interested.
We
can
have
discussion
based
on
this,
because
this
topic
is
come
with
many
times
and
we
just
say
like
this
meeting.
We
only
have
an
hour
and
we
have
so
many
topics
to
discuss
so
I'm,
fine,
cheering
a
specific
one
for
this
specific
topic.
I.
A
It
may
five
minutes
late,
so
the
next
one
was
yeah
for
userlist
authentication,
authorization
and
orbit
lobby.
So
I
did.
I
did
send
out
a
talk
with
more
details
and
I
did
present
it
at
sig
earth
as
well,
and
the
feedback
I
got
there
was
a
position,
gives
a
lot
more
power
than
users
one
or
what
Federation
should
get
so
I
got
a
lot
of
pushback
against
using
impersonation
palace.
A
F
G
Something
that
wasn't
really
discussed,
but
maybe
I'll,
follow
up
with
you
offline
is
the
possibility
that
maybe
you
just
restrict
Federation
to
you,
know
a
set
of
namespaces
or
a
certain
criteria,
so
that
you
know
maybe
give
whatever
off
mechanism
do
given
cluster
full
access
to
these
namespaces,
but
not
access
to
the
whole
cluster,
because
that
would
at
least
alleviate
the
concern
you're
getting
route.
Basically
I
think
that
is
possible.
A
Today
and
I
did
file
a
separate
issue
with
another
I
give
a
rather
than
using
a
single
service
account
for
holo
filtration.
We
can
use
namespace
service
accounts.
So
if
you
want
to
fit
rate,
you
use
Federation
only
for
let's
say
two
new
spaces.
You
can
create
two
service
accounts
for
each
of
those
namespaces
and
Alec
actually
replied
on
that
issue.
Saying
even
today
are
the
single
identity
that
you're
using
even
get
use
our
back
and
get
that
identity,
access
to
only
two
names
faces
and
all
that
you're
providing
to
finish
in
control
games.
Thank.
I
A
Cool
yeah
I
see
21
and
you
have
21
minutes.
Someone
was
private,
federated
services.
Again,
it's
mod
mix,
menstrual
I.
Think
it's
fine
to
have
us
on
the
backlog.
Yeah.
The
last
one
is
interesting:
federated
read,
access
to
cluster
field
objects,
we've
had
a
lot
of
discussions
and
I
were
designed
in
mine
and
I
can
I
did
send
I
had
send
out
one
earlier,
but
we've
had
some
discussions
after
that.
A
If
someone
is
ready
to
implement
it
again
clean
that
up
and
send
it
again
work
with
cindy,
I
believe
it
was
to
implement
it
in
spin.
Yes,
yes,
that's
correct
awesome,
so
I
will
probably
send
you
links
to
the
existing
one.
So
you
can
go
through
this
link
and
see
where
we
are
right
now
and
happy
do
when
you
up
to
speed
yeah.
E
A
I
Just
gonna
mention
to
kill
that
that
I
have
actually
given
you
all
the
background
and
send
Cindy
to
all
of
the
previous
documents
and
issues
and
things
and
she's
undertaken.
You
essentially
consolidate
those
into
a
new.
You
know
refreshed
version
of
the
design
that
would
that
we're
actually
going
to
implement
now
cool
awesome.
So.
E
Basically,
I
will
clone
an
eco
design,
doc
and
refresh
it
22
like
I.
The
opening
issues
are
too
wise,
the
design
on
using
proxy
API
versus
caching.
The
second
was
like
if
they
need
to
change
the
API,
how
it's
called
adding
the
clusters
or
not
I'm
planning,
to
make
it
step
by
step.
Like
say
we
can
do
in
stages.
E
A
A
I
Y'all
go
ahead.
Painting
now
it's
going
to
say:
is
there
anything
that's
too
urgent
to
wait
until
next
Monday
and
we
just
pick
it
up,
then
yeah.
I
think
that
time
what
what
the
main
you
need-
incidentally,
just
before
we
disband-
is
that
for
some
reason
and
I
don't
understand
how
these
things
work.
Some
people
can
have
issues
assigned
to
them
and
other
people
cannot
and
that's
causing
quite
a
lot
of
pain
and
suffering.
Yeah.
A
I
think
that's
a
known
issue,
not
everyone!
So
the
way
github
works
you
need
to
be
added
to
the
all,
and
once
you
are
added
to
the
oh,
it's
like
all
or
nothing
permission
model.
So
there's
some
restriction
of
when
you
can
be
added
to
all
so
for
new
members.
They
generally
need
like
a
proxy,
so
you
can
assign
I
see.
This
is
what
you
did
already.
You
can
assign
it
to
yourself
as
a
proxy
for
someone
else.
Okay,.