►
From YouTube: Kubernetes WG K8s Infra - 2021-04-28
Description
A
A
The
first
thing
is
the
billy
report,
but
I'm
not
sure
I
want
to
go
over
the
billing
report
or
anyone.
I
have
a
question
about
the
billing
required
for
this
week.
A
A
D
A
A
A
Blocking
so
before
we
go
over
the
pressure
pump,
I
think
you
had
a
question:
is
it
better
or
if
we
had.
D
I
can
go.
Am
I
yes,
I
just
wanted
to
see
if
there's
anything
external
holding
that
merge
happening.
If
there's
anything
more,
we
we
need
or
that
you
need
from
us,
or
is
this
1469
just
a
hold
on.
A
Basically,
the
4069
is
the
umbrella
mobile
roadmap
for
this
working
group.
So
there
are
many
things
I
mean.
There
are
many
details
depending
of
the
issue
we
we're
facing.
D
So
yeah,
my
question
is
only
specific
to
1469
this
or
to
the
very
bottom.
A
What
I'm
gonna
do
about
this
is
I'm
gonna.
I'm
gonna
apply
the
chain,
so
you
can
basically
get
rid
access
to
the
to
to
the
bracket.
Maybe
I'll
run
our
team
with
gtm
the
pull
request.
Okay,
so
it's
merged
in
the
full
request.
A
D
Thank
you,
yeah
and
that's
expected
to
give
up
to
get
access.
I
believe.
A
Okay,
I
see
the
command
from
ap,
so
basically
the
bigquery
access.
I
would
do
that
as
a
follow-up
okay,
so
because
it
was
not
clear
to
me
what's
the
eye
analysis,
you
need
from
basically
you're
running
your
analysis,
so
I
prefer
to
just
give
you
access
to
the
to
the
logs.
But
if
you
need
request
to
do
analysis,
I
will
I
will
work
on
that
in
the
next.
I
think
next
week.
D
E
Okay,
is
that
all
that's
the
way
I
just
wanna,
I
just
wanna
confirm
so
one
1966,
the
only
thing
you
just
need
to
apply
that
small
net
that
was
pointed
out
by
them
and
after
that,
it's
all
to
cheat
him
and
approve.
A
A
D
The
sandbox
I
have
access
to
it
and
so
far
I've
not
had
problems
with
permissions
on
it.
Okay,
caleb
have
you.
C
A
A
Oh
wait:
the
1469
because,
like
I
said
this
is
basically
the
umbrella
issue
that
would
that
guide
the
direction
of
this
group,
where
we
still
have
a
lot
of
unfinished
business
about
the
issue.
So
I
feel
like
we
should
basically
follow
this
myself
focus
the
effort
on
finishing
the
migration
of
all
the
basically
project
hosting.
D
And
caleb
has
a
sub
set
of
these
tickets
in
his
section
that
is
specifically
in
here
is
gubernator
as
a
one
of
the
projects
that
need
to
move.
I
mean
he
has
some
questions
for
moving
that
forward,
because
we're
looking
at
kubernetes
and
starting
there
as
a
project
to
move
all
the
parts.
A
I
will
quickly
say
something
about
giving
it
a
basic
argument:
it's
a
user
interface
you're,
helping
contributors
who
have
a
unified
view
of
the
request.
They
have
been
tied,
it's
maintained
by
the
sick
testing,
so
I
will,
I
think
you
can
basically
go
directly
to
scene
testing,
ask
questions
if
you
need
to
if
you
need
some
answer,
because
I'm
not
very
aware
about
how
kubernetes
is
running
right.
Okay,
so
again,
I
think
you
saw
the
best
advice
I
can
give.
You
is
basically
playing
around
in
sick
testings
like
channel
and
asked
about
this.
The
issue.
A
A
That
needs
basically
input
from
other
cities
and
somewhere.
We
can
basically
focus
on
that.
I
will
say
for.
A
F
I
think
I'll
for
myself
I'll
reach
out
to
the
test
infrastructure
as
you
recommended
yeah,
so
we
we
won't
worry
about
that
in
this
meeting.
A
Okay,
I
don't
want.
I
don't
want
to
spend
too
many
time
on
this
dashboard
on
this
spot,
because
there
are
too
many
things
that
need
to
be
discussed
and
there
is
there
are
some
access,
I'm
not
really
aware.
So
I
feel
like
during
the
till
next
week,
I'm
gonna
review.
I
will
discuss
three
hours
and
basically
review
some
issue
and
see
which
one
can.
A
So
just
being
me
on
songs
like
I
wish,
I
would
see
how
I
can
help
you
to
move
forward.
Thank
you
basically
and
identify
some
more
basic
of
the
brokers.
D
Do
you
have
time
of
day
that
works
better
for
you,
I
kind
of
straddle
between
south
africa
and
new
zealand
time,
so
I
can
move
closer
mornings
are
easier
for
you.
A
G
Hey
everyone.
I
should
start
by
apologizing.
I
have
been
relatively
unavailable
the
last
several
weeks.
I've
got
a
bunch
of
crazy
stuff
going
on,
and
I've
just
not
been
able
to
make
all
the
time
that
I
was
committing
to
making.
So
I
I
apologize
for
that.
G
I
want
to
jail
tonight,
I'm
not
ignoring
you,
I'm
I'm
watching,
but
not
able
to
be
as
active
as
I
wanted
to
be
this
popped
up
in
my
mailbox
this
week
I
got
an
email
that
said:
vulnerability
scanning
will
be
billing
will
start
being
enabled
in
july.
G
We
knew
this
was
coming
all
along.
This
was
this
was
published,
but
they
sent
us
a
nice
reminder.
So
I
went
off
and
took
a
look
at
what
the
actual
billing
was,
and
so
I
threw
some
ballpark
numbers
in
here.
It's
not
a
huge
amount
of
money,
but
it's
not
a
trivial
amount
of
money
either
right.
G
So
sorry,
since
you
switch
screens,
I
don't
have
the
numbers
anymore,
but
anyways
is
on
the
order
of
a
few
thousand
dollars
a
month
which
isn't
that
much
in
the
overall
picture
of
what
we're
dealing
with.
So
maybe
we
simply
don't
care,
but
I
thought
it
was
worth
bringing
up.
What
I
thought
was
interesting
was
all
of
the
the
bills
currently
are
around
the
staging
repos.
G
G
So
we
made
a
decision
a
while
back
to
turn
on
scanning
for
staging
repos,
because
we
thought
it
was
useful
to
get
those
reports
before
they
get
promoted
into
or
into
the
main
repository.
G
What
I
think
we
should
ask
is:
are
all
staging
repositories,
promotion
path
or
are
some
of
them
going
to
be
not,
and
if
not,
if
they're
not
going
to
be
promoted,
do
we
really
need
to
enable
scanning
for
them,
or
is
this
just
a
distinction?
We
don't
care
about
because
it's
just
not
worth
that
much
energy.
A
A
A
So
my
personal
opinion,
it
would
be
to
deactivate
the
scanning
images
for
this
specific
project,
because
I
don't
think
it's
really
useful
right
now
they
are
nightly,
builds,
I'm
not
sure
anyone
check
those
those
reports.
G
Yeah,
well
I
mean:
does
anybody
check
their
vulnerability?
Scanning
is
an
entirely
different
question,
but
if
we
know
for
a
fact
that
these
images
are
not
sensitive
in
this
way
like
they
are
ephemeral
or
or
they're
short-lived,
then
you're,
I
agree.
Maybe
we
should
turn
it
off
as
we
I'm
thinking
about
all
the
shell
scripts,
that
control
all
the
staging
repositories
and
it
doesn't
sound
like
a
pleasant
change,
but
it
might
be
worth
doing.
G
G
We
sure
we
can
reach
out
to
every
staging
group.
They
all
have
mailing
lists,
so
we
can
actually
literally
email
them
and
see
if
they
actually
check
these
vulnerability
reports.
I'm
gonna
guess
the
answer
is
almost
nobody
actively
checks
them
and
I'm
not
even
sure
we
did
all
the
work
that
we
needed
to
make
it
possible
for
them
to
check
them
regularly.
G
That
said,
two
thousand
dollars
a
month
isn't
gonna
break
us
right
and
given
the
highly
constrained
bandwidth
that
we
have
from
people
on
this
call,
you
know
all
the
other.
Things
that
are
in
progress
are
certainly
more
urgent
than
this.
H
I
vote
we
put
this
on
the
board
with
a
lower
priority
and
when
we
have
extra
resources
and
time
it's
not
forgotten,
but
it's
not
time
spent
well
times
well
spent.
Yet.
B
B
B
C
H
World
I
added
that,
because
on
the
the
meeting
is
tomorrow-
and
we
wanted
to
not
take
too
much
time
here,
just
to
discuss
it,
but
there's
a
lot
of
overlap
between
that
kept
and
the
302
or
split
dns
redirect
implementation
so
making
sure
that
we're
meeting
the
needs
there
and
and
and
and
using
all
the
brains
in
the
community
to
go
the
right
direction
on
that.
H
That's
cross
cross
six.
So
I
have
no
idea
what
meeting
you're
talking
about.
Can
we
click
on
the
link
for
that
that
takes
us
to
slack
that's
a
slack
thing
in
sigin,
sorry
kate's
in
for
working
group
and
then
the
next
one
is
the
artifact
management
board,
which
is
underlying
that
cup.
A
So
lori
abbas
and
a
daughter
to
decide
which
day
we
we
can
meet
disco
about
the
artifact
management
cap.
I
think
it
was
two
weeks
ago
and
based
on
the
results
we
have
in
the
doubles.
The
meeting
is
supposed
to
happen
april.
A
G
F
Yeah,
so
I
made
some
progress
on
making
a
proof
of
concept
using
envoy
and
I
I've
only
got
that
the
mirroring
working.
Well,
it's
not
it's
proxying
kind
of
thing:
it's
not!
Actually
it's
not
even
proxy.
It's
redirecting.
F
So
I
got
that
working
for
container
image
artifacts,
but
I
haven't
I've
been
able
to
do
anything
with
just
any
kind
of
artifact
yet,
but
getting
it
so
does
a
302
redirect
when
so
you
have
your
your
host,
which
is
registrator
case
that
I
o
takes
your
source
ip
and
then
spits
out
a
host
for
you
to
reach
out
to
pull
a
container
in
which
from
so
I
did
a
basic
demo
of.
F
If
the
ip
address
is
a
local
one,
then
I'm
going
to
give
you,
I
think
it's
registry.kate
studio,
but
if
you're
a
remote
address,
then
I'll
give
you
docker
hub.
So
I.
H
Think
I
just
dropped
a
link
into
kate's
and
for
working
group
channel,
which
is
a
blog
that
we
wrote
kind
of
based
on
this.
We
have
it's.
This
is
just
a
demo
written
in
what.
F
Was
written
it
it's
a
envoy,
lua
filter,
there
you
go,
but
that
could
be
a
go,
wesm,
filter
or
anything
else.
F
G
H
Can
we
we've
merged
into
the
kate's
scio
repository
somewhere,
a
list
of
asn's
and
that's
per
company,
so
the
google
asn's,
the
single
amazon
asn,
the
multiple
microsoft
asn's?
H
It
would
be
good
to
get
feedback
from
those
companies
that
we
suspect
they
haven't
have
higher
usage
to
confirm
that
those
asn's
accurately
represent
their
company
so
that
we
can
there's
not
going
to
be
company
specific
lookup.
The
database
we
use
for
that
is
is
the
bgp
and
the
and
the
autonomous
system
groups
that
are
owned
by
those
companies.
H
If
they
are
accurately
representing,
I
think,
that's
the
biggest
the
best
stab
in
the
dark
as
far
as
existing,
not
geoip,
but
company
ownership,
ip
that
we
have.
G
So
that's
a
good
good
point.
I
had
been
thinking
about
geoip
as
geoip,
it's
probably
better
to
first
start
with
the.
If
we
know
you're
coming
from
amazon,
send
you
to
amazon
yeah
and
then,
if
we
fall
off
the
bottom
of
that
list,
that
we
have
no
idea
which
company
you're
coming
from
is
there
a
mechanism
for
us
to
figure
out?
G
H
We
haven't
done
a
lot
of
research
into
that
yet,
but
this
is
going
to
be
a
solved
problem
and
there
will
be
something
there.
We
did
this
current
research
at
the
envoy
level.
We
did
look
a
little
bit
into
what
was
the
next
one,
the
stuff
on
top
of
envoy
contour
yeah.
F
The
implementation
didn't
appear
as
possible
for
something
like
the
contour
ingress
controller
setting
some
kind
of
inline
code
snippet.
But
if
there's
anything
that
is
like
that,
I
think
would
be
cool
to
use
as
well
to
get
another
poc
across.
G
If
it
makes
more
sense
to
lean
into
the
lua
or
wasm
for
envoy
or
if
we
should
just
write
this
as
a
simple
go
program
like
it,
doesn't
feel
like
that
complicated
of
a
go
program
to
listen
on
http,
look
at
the
client
ip
look
it
up
in
a
table
of
vsn's
issue.
The
redirect
like
it
feels
like
you
know,
order
hundreds
of
lines
of
code
and
we
actually
have
a
working
demo.
F
G
Mean
the
redirect
part
like
one
of
the
things
that
go
does
really
well,
is
http
serving
right.
So
the
redirect
part's
super
easy,
we'll
have
to
think
about
the
opera
operationalization
of
this.
But
it
sounds
like
even
in
the
envoy
path,
you're
going
to
end
up
funking
out
to
some
code
to
do
the
database
lookup.
So
I
wonder,
then,
if
it
makes
sense
to
even
involve
envoy
or
nginx
or
anything
or
if
we
should
just
do
ourselves
and
deal
with
the
operationalization.
H
I
think
some
of
my
initial
thoughts
around
that
were
trying
to
use
some
of
the
the
existing
cncf
community
things
as
well
to
bring
them
in
and
get
their
opinion
and
feedback.
Ii
has
another
area
where
we're
working
directly
with
the
primary
envoy
api
shepherds
and
the
api
the
community
team
there.
G
Yeah,
I
think
what
whatever
we
do
here,
we're
going
to
definitely
set
trend
yeah,
and
so
we
should.
We
should
be
super
careful
about
what
it
is
that
we
do.
Staying
within
the
family
seems
totally
reasonable,
but
if
it
was
a
simple
go
program
that
said
look
we
run
this
on
kubernetes.
We
set
up
with
these
these
auto
scaler
configs
and
these
memory
limits,
and
if
you
wanted
to
mirror
your
registry,
you
could
take
this
go
program
and
run
it
yourself,
like
that's
a
that's
a
pretty
cool
story,
honestly,.
H
I'm
trying
to
fix
to
get
it
some.
I
need
to
think
about
it
a
bit
away.
I
I
we
have
caleb's
primary
on
it
right
now,
and
so,
if
we
want
to
shift
that
direction,
do
a
poc
over
the
next
week
or
two
we
can.
We
can
head
that
direction.
G
I
think
it
would
be
very
interesting.
I
just
think
it
would
be
interesting
to
compare
the
the
overall
complexity
like
it
feels
like
it
would
be
a
small
program.
Maybe
I'm
wrong,
like
maybe
maybe
actually
ingesting
the
asms.
We
need
to
make
it
plugable
and
externalized,
and
so
now
we're
talking
about
a
database
and
like
maybe
that
spirals
out
of
control,
or
maybe
it's
simply
a
config
map
that
we
update.
G
You
know
monthly
or
something
and-
and-
and
it
really
is
that
simple,
so
I
was
tempted
at
some
point
to
start
the
poc
myself
and
then
I
was
like
nope.
You
know
I'm
kind
of
busy
right
now
and
caleb
caleb's
actively
working
on
this.
So
why
would
I
go?
Do
that
you?
You
feel.
H
Free,
I
think
we
will
get
to
a
configmap
at
some
point.
We
have
to
get
the
asns
accurate
and
then
we
might
just
as
a
team
look
like
they're
already
merged,
but
it
would
be
great,
for
example,
for
google
to
at
least
step
in
and
say
yep.
That's
our
asms
get
some
new
microsoft
yeah,
that's
ours
and
amazon,
those
at
least
those
three,
because
we
have
a
starting
point
and
then
we
can
use
that
to
have
a
program
that
takes
those
as
sense
and
dumps
it
to
a
config
map.
G
It's
I
don't
know
if,
but
I
mean
like,
I
don't
know
if
we
actively
publish
it
in
a
place
where
you
can
like
refresh
it
daily
and
get
new
numbers
or
whether
we
would
be
willing
to
say
well,
you
figured
it
out
confirmed
or
whether
we
would
just
bite
our
tongue
like.
I
have
no
idea.
H
So
but
there's
in
the
bgp
arena
you
have
the
owner
company
listed
like
it
is.
Definitely
there.
I
guess
we're
wanting
to
say
is:
are
there
ones
we
shouldn't
be
looking
at?
Would
you
like
to
bear
that
back
a
bit
yeah?
It's
not.
It
is
definitely
updated
minute
by
minute.
H
Like
the
whole
internet
knows
these
are
good,
but
what
we
would
like
is
someone
to
confirm
that
we're
actually
looking
at
the
right
arena-
and
there
may
be
some-
oh
we've
got
these
over
here-
that's
actually
not
under
the
same
company,
but
that's
actually
where
google
cloud
lives,
for
example,
would
be
a
huge
miss
on
our
part.
That
would
be
missing
the
the
reason
the
the
mapping
exists.
G
H
H
Pairing
with
those
that
do
for
a
bit
helps
us
to
understand
some
of
the
flows
and
things
that
we
don't
know
yet
tim,
and
I
did
some
of
this
really
early
on
we're
trying
to
explore
what
permissions
were
required
to
do
various
things,
and
I
wonder
if
it
wouldn't
be
useful
for
to
set
up
maybe
on
the
week
opposite
this
a
a
a
working
session.
Whereas
this
meeting
is
more
for
looking
at
the
action
items.
H
A
H
G
F
H
We'll
put
it
on
us
to
put
together
a
list
of
kind
of
like
requested
pair
sessions
that,
because
we
can
ask
the
questions
and
sometimes
they
can
get
answered,
yeah
go
ahead.
I'm
sorry!
Do
you
have
a
specific
subject
in
mind?
H
Well
in
this
in
brno's
case,
for
example,
we're
trying
to
focus
him
primarily
on
going
through
that
migrating
away
from
google's
internal
work
to
to
the
public
bucket,
and
that's
a
getting
the
the
wrapping
your
head
around
that
flow
and
trying
to
get
even
those
individual
tickets
understanding
the
the
flow
like
these.
H
That's
part
of
that,
like
the
whole
reason,
this
group
exists
is
to
migrate
out
of
that
umbrella.
Ticket
is
pretty
much
our
reason
for
being
and
and
I'm
trying
to
allocate.
You
know
basically
two
two
humans
full
time
to
it.
Okay
and
it's,
I
think
we
should
reduce
the
friction,
and
some
of
that
is
not
even
knowing
what
questions
to
ask
in
looking
at
that
body
of
work,
because
there's
tribal
knowledge
around.
Even
the
creation
of
that.
That
is
hard
for
even
me,
and
I've
been
around
peripherally
since,
since
the
beginning.
B
Okay,
okay,
I
think
we
can
try.
I
mean.
H
G
I'm
all
for
yeah
and
just
again
to
be
clear,
I'm
totally
not
against
this.
I
would
love
to
sit
and
spend
time
pairing
and
just
exploring
the
problem
space.
The
last
couple
of
weeks
in
the
next
few
weeks
is
going
to
just
be
hellish
for
me
for
totally
unrelated
reasons,
so
I'm
I'm
a
little
wary
of
signing
up
for
recurring
anything
right
now.
No,
this
isn't
definitely
so.
H
What
I'm
trying
to
do
is
if
we
can
here's
my
my
request.
Can
we
prioritize
three
of
those
that
I
can
put
burno
on
specifically
and
then,
once
we
prioritize
those
three
we'll
set
up
an
availability
for
to
pair
with
us,
and
those
are
the
three
things
we'd
love
some
time
on,
and
it's
not
a
recurring
it's
just
that
these
are
nrq
for
the
knowledge
dump.
Does
anyone
have
a
moment
and
they'll
be.
G
G
H
G
G
The
existing
project
has
google
specific
stuff
in
it
and
kubernetes
specific
stuff
in
it
and
like
I
can
get
a
sense
like
an
order
of
magnitude
of
cost
of,
say
the
you
know
release
bucket
like
we
know
that
that
is
a
large
single
line
item
and
I
can
give
you
an
order
of
magnitude,
but
that's
about
it.
But
honestly,
I
probably
couldn't
even
show
you
my
screen,
while
I'm
looking
at
that
because
there's
other
google
stuff
tangled
in
there.
B
A
H
A
Okay,
because
to
be
honest,
there
are
some
issues
that
are
extremely
tricky
to
deal
with.
I
don't
work
on
that,
because
there
are
very
specific
restriction.
A
A
H
H
Today
and
for
getting
us
sitting
at
the
meeting
invite
for
tomorrow,
much
appreciated.