►
From YouTube: Kubernetes WG K8s Infra - 2021-06-23
Description
A
Hi
this
is
june
23
and
we
are
in
the
working
room,
kids
and
from
meetings.
I
will
be
your
host
today.
My
name
is
announcement.
This
meeting
run
under
the
code
of
content.
So
if
you
have
any
issue
with
this
meeting,
you
can
reach
out
the
code
of
content
to
committee
at
contact
at
kubernetes
dot
io,
I'm
dropping
the
link
of
the
dots
in
the
chat.
A
I
think
we
can
start
before.
We
start
god.
A
A
A
A
And
I
think
that
was
the
next
item.
Aaron
is
not
here
today,
so
I'm
gonna
give
just
a
quick
update
on
the
item
he
puts
into
meeting
now.
So
basically,
I'm
working
with
state
scalability
folks
to
identify
which
resource
code
we
need
to
raise
on
gcp
in
order
to
run
the
5k
snow,
the
project
that
needs
5k
notes.
A
A
A
So
that's
all
for
me.
I
think
next
item
is
ep.
C
I
wanted
to
to
to
give
us
another
more
time
to
work
on
this,
as
I'm
digging
into
it.
I
am
having
to
reach
out
to
to
more
people,
and
I
don't
have
anything
new
to
update,
but
that's
where
our
energy
is
going.
C
With
regards
to
my
team,
we
also
have
a
lot
of
updates
to
the
different
asn's
that
are
apart
on
the
manual
pieces
that
we've
looked
at
kind
of
a
separate
note,
and
I'm
also
looking
to
reach
out
to
I
reach
out
to
the
infra
leads
arnaud
and
his
crew
about
the
amazon
accounts,
because
it's
currently
I
I
spoke
with
our
friends,
my
technology
notes,
as
I'm
going
for
the
amazon
accounts,
that's
actually
an
entire
cncf
level
account,
and
so
billing
is
at
that
level,
and
so
I
need
to
like
find
a
way
to
request
that
at
least
my
team
has
access
to
that
and
let's
figure
out
how
we
can
pull
out
the
billing
information
for
this
project
separately,
so
that
we
can
review
that
together
or
if
we
want
to
punt
that
up
to
the
cncf
level
and
create
one
for
all.
C
A
Linus,
I'm
not
sure
about
that.
So,
let's
post
the
question
on
the
zoom
chat.
A
D
Questions
so
once
you
figure
out
the
aws
account,
I
I
tried
to
strike
up
a
thread
with
tim,
aaron
and
justin,
and
I
was
going
to
act
as
the
aws
representative
on
the
account,
but
that
thread
didn't
go
anywhere,
so
I
can
bump
that
again.
C
D
C
Justin
santa
barbara,
those
are
the
those
are
the
official
and
only
key
holders.
As
far
as
I'm
concerned,
outside
of
amazon.
A
Great
anything
else,
any
question.
A
E
So
yep,
oh
cool.
I
just
had
a
random
question.
You
know
I've
been
into
these
meetings
in
the
past,
but
I
never
bothered
to
ask
in
this
forum.
But
you
know
if
we
do
have
a
crazy
amount
of
billing
and
if
there's
any
issue
with
that
like
who
will
be
the
first
one
to
to
see
that
or
you
know
if.
E
B
That's
a
good
question.
I
mean,
I
think,
the
first
time
we
noticed
that
the
billing
report
was,
you
know
more
than
a
fraction
higher
than
previous
month
would
be
the
moment
we
would
go
and
investigate
so
I
mean
unless
we
spend
a
million
dollars
in
a
month,
we're
probably
okay.
A
Basically
is
just
check
what's
remaining
in
the
credit
we
have
that's
the
only
thing
we
can
really
check
so
what's
happening.
Most
of
the
time
is
people
people
come
to
us
and
raise
the
issue
like
we.
We
basically
exit
some
quarters
or
we're
running
out
of
I
mean
something
and
we
can
act
on
that,
but
I
don't
think
there's
an
american
instance
monitor.
D
A
You
need
to
contact
will
be
like
apc,
eo
and
justin
sambar
and
for
google.
I
think
you
can
contact
iran
and
team.
A
A
A
A
Problem
so
also
one
thing
we
we
don't
do,
and
this
is
something
like
so
just
as
a
back
practice.
We
we
don't
tag
on
level
braces.
We
have,
and
I
feel
like
we
can
do
something
like
that.
That
can
help
us
track
down,
basically
do
different
resources
and
we
can
easily
identify
the
different
I
mean
different,
let's
say
big
assets
for
resource
consumption,
so
I
feel
like
the
first
thing
we
we
should
do
is
basically
flag
and
label
every
reasons
we
have
and
trying
to
break
down.
Where
are
the
costs?
C
I
know
that
early
on
we
were
looking
for
a
way
to
to
do
billing
per
name
space
so
that
we
could
name
space
things
and
and
find
a
way
to
identify
stuff
so
that
within
our
proud
configurations
and
and
all
of
our
work,
even
if
things
are
running
in
the
same
cluster,
we
could
separate
out
the
billing.
C
C
At
that
point
it
would
be
the
time
to
start
identifying
costs
per
identifying
cost
per
namespace,
and
if
somebody
wants
to
dig
into
that
preemptively
who
has
experience
with
that,
particularly
within
the
the
way
that
we
use
gke,
that
would
be
lovely.
We
will
probably
circle
back
that
if
someone
doesn't
find
that
what
to
work
and
also
for
amazon.
But
I'm
I'm
unaware
of
a
of
a
mechanism
for
that.
A
Okay,
next
person
is
tyler.
H
Nope,
this
is
just
sort
of
a
verbal
update
on
the
work
I
presented.
I
believe
two
meetings
ago.
This
was
basal
deprecation
in
the
kubernetes
container,
the
cip
tool.
So
basically
right
now
we're
at
the
halfway
point
in
deprecation.
So
it
looks
like
there's
not
going
to
be
any
change
for
consumers
of
this
tool.
They
already
required
docker
we're
just
relying
on
it
a
little
more.
If
anything,
you
don't
need
basil,
which
is
awesome.
H
So
additional
information
you'll
see
a
cap
more
of
a
retroactive
kept
being
released
within
the
next
week
or
so,
and
this
is
just
sort
of
a
wrap
up
of
what
happened.
H
Sort
of
after
implementation,
but
most
of
the
information
found
in
the
design
dock
is
pretty
much
going
to
be
in
the
cap,
but
it's
more
for
documentation's
sake.
So
that's.
G
A
Okay,
so
next
item
was,
I
mean:
is
it
suggested
to
cancel
the
next
meeting,
which
is
it's
july
7.?
I
don't
know
if
any
the
same
problem
is,
I
think
it's
isn't
it's
the
holiday
us,
so
it's
possible
most
of
the
people.
Us
can
attend
this
meeting.
So
I'm
just
dropping
the
question
and.
C
Maybe
ai
is
a
email,
email
to
list
and
channel
yep.
A
A
July
7th
8
a.m.
Central
time.
A
A
B
I'm
sure
we
could
ask,
given
that
it's
funny
money
moving
from
you
know,
left
pocket
to
right
pocket.
I
don't
know
what
the
appetite
will
be.
I
don't
know
I've
never
done
the
committed
use
stuff
as
a
consumer,
so
I
don't
know
if
there's
a
contract
involved
or
you
know
what
would
be
necessary
to
get
that
signed,
certainly
even
more
than
triple
a
I'm
sure
we
could
commit
to
some
base
level
usage
across
all
the
ci
stuff
too
right,
so
it
may
not
be.
B
A
B
B
What
that
is,
I
can
I
can
ask,
or
I
can
see
I
can
find
someone
to
ask,
I'm
not
sure
who
to
ask,
but
I
can
go
look
around.
A
No
russian,
that's
just
I'm
just
kind
of
curious
and
see
we
can
basically,
because
when,
when
I
took
a
look
at
the
doc
I
was
like.
Oh,
this
is
just
a
chick-like
command,
so
it
can
be
easy
to
to
wait.
A
Infinite
wisdom
say
we
should
use
it
to
instance
type.
I
think
it
was
in
the
poor
request
somewhere
about
something
I
don't
remember,
but
I
remember
these
switches.
We
should
use
a2
and
some
style
and
looking.
I
did
some
simulation
with
the
gcp
calculator
and
I
think
if
we
we
right
now
for
triple
a
f
n
one
sense
type
and
if
we
switch,
we
can
basically
save,
I
think
at
least
thirty
percent,
in
the
cost
of
one
year
where
we
use
the
basically
for
the
pre-cluster
and
tuberculosis.
B
A
I
will,
I
would
think,
with
iron
about
this
and
we
I
will
try
to
do
something.
Maybe
next
month
I,
like
tim,
said
we
should
focus
on
the
migration
trust.