►
From YouTube: k8s-infra-team's biweekly meeting 20200722
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).
A
Now,
hi
everyone,
my
name
is
bart
smikla
and
I
want
to
welcome
you
at
our
bi-weekly
community
call.
A
A
A
Okay,
so
I
think
that
we
did
the
billing
review
the
last
time.
Do
you
want
to
do
it
today,
too
tim?
What
do
you
think.
B
I
thought
it
was
interesting
to
point
out
the
that,
since
the
gtr
flip
is
happening
again
this
week
I
just
pulled
up
the
graph
just
for
giggles
and
prior
to
this.
Our
single
biggest
line
item
for
the
previous
months
was
on
the
order
of
four
thousand
dollars,
which
is
you
know
pretty
impressive.
B
So
that's
fascinating
and
once
we
finish
the
flip,
I
think
it's
really
worthwhile
for
us
to
start
thinking
about
is
this
right?
Are
we
surprised
by
this?
Are
we
serving
the
right
traffic
to
the
right
regions?
Maybe
we
can
dig
in
how
to
get
better
stats,
and
maybe
we
can
finally
revisit
the
topic
of
federating
image
serving
so
you
know
we
could
get
a
you
know
an
amazon
instance
of
vcr
also
serving
or
something
like
that,
which
requires
some
real
work
and
thinking
about
how
we're
going
to
do
that.
A
C
B
Did
you
get
that
information
from
the
report?
I
didn't
look
at
the
report.
I'm
looking
at
the
dashboard.
Is
there
how
current
is
the
report
so.
B
Yeah,
that's
about
what
I
see
in
the
in
the
console
graph
too,
but.
C
B
It
is
so
I'm
seeing
specifically.
The
line
item
is
network
internet
egress
from
america
to
america.
B
C
B
D
Yes,
the
short
answer
is
yes,
we
expected
to
finish.
I
just
sent
out
an
email
last
night,
but
based
on
previous
roll
us
that
we've
seen
that
it
takes
just
under
four
days
or
I
think,
around
90
somewhat
hours.
So
since
we
started
it
on
monday
morning,
we
started
to
finish
just
a
bit
earlier
on
on
a
friday
morning
like
I
need
to
recalculate
again,
if
I
want
to
give
out
exact
numbers,
but
you
know
basically,
friday
morning
we
should
have.
It
should
be
done.
B
E
Hey
guys,
so
just
for
a
little
bit
of
context,
and
this
might
be
familiar
to
those
of
you
who
heard
me
talk
about
what
I'm
doing
last
our
last
meeting
like
two
weeks
ago,
but
basically
I've
been
spending
my
time
making
a
are
creating
a
vulnerability,
pre-summit
check
for
the
promoter,
so
any
image
that's
about
to
get
promoted
from
staging
to
production.
E
We
just
you
know,
check
to
check
the
vulnerability
information
that
we
have
for
it
to
see
if
it's
above
a
defined
severity
threshold,
and
if
it
is,
then
we
reject
it
in
order
to
do
that,
we
also
turned
on
the
container
scanning
api
for
all
of
the
image
staging
projects
and
in
order
to
take
better
advantage
of
that,
you
know
we
realized
in
in
talks
with
me
and
linus
that
the
president
check
only
uses
that
information
like
at
the
moment
where
you're
promoting,
but
we
can
make
use
of
that
information
like
I
guess,
to
to
examine
image
health
after
promotion.
E
So
it's
just
like
to
to.
I
guess
give
people
a
better
understanding
of
what
vulnerabilities
their
images
have
consistently
and
not
just
at
the
moment
that
they're
promoting
it
currently
the
only
way
for
people
to
actually
get
information
about
what
vulnerabilities
they
have
is
by
using
the
google
cloud
platform-
and
you
know,
clicking
through
the
container
registry,
looking
at
individual
images
and
the
vulnerability
information
that
is
there.
E
E
So
the
current
thought
process
is
to
create
kind
of
a
vulnerability
dashboard
that
will
kind
of
set
up
an
open
connection,
probably
using
pub
sub
with
the
container
analysis
service,
for
you
know,
probably
starting
with
just
the
the
prod
project
so
looking
at
the
production
images,
but
eventually
also
the
staging
projects
that
can
you
know,
take
that
information
and
parse
it
and
present
it
in
a
digestible
way.
So
people
have
a
you
know
a
destination
where
they
can
always
look
up
vulnerability,
information
for
their
images
very
easily.
E
My
question
for
the
community
is
one:
is
this
something
you
know
that
we
would
see
as
useful?
I
you
know
it
doesn't
add,
necessarily
an
extra
charge,
because
we
already
have
the
container
scanning
api
enabled
and
from
from
what
I've
seen
container
analysis
doesn't
charge
based
on
reads,
so
we
wouldn't
be
incurring
any
extra
charges.
So
it's
really
just
making
more
use
of
the
information
we're
already
paying
for.
E
The
second
question
is:
is
there
does
kubernetes
kind
of
have
a
standard
tool
for
creating
the
infra
or
the
ui?
That
would
be
necessary
for
a
dashboard,
yeah.
A
So
those
are
the
two
things
I
have
for
you
guys
so
about
the
second
question.
I
don't
think
we
do
and
as
I
was
playing
with
the
graffana
in
the
last
month
or
two
and
there
is
the
working
instance
in
our
infrastructure
right
now,
but
I'm
using
it
for
tests
of
the
monitoring.
A
E
B
So
yeah,
I
was
just
gonna
agree
with
that.
I
have
no
objections
to
using
something
like
grafana.
It
seems
great.
I
don't
know
it
very
well
if
you
think
you
can
like
try
it
and
and
if
you
can
express
things
we
want
to
express
in
it
then
cool.
I
also
don't
know
whether
sort
of
the
the
normal
thing
is
to
have
a
shared
grafana
instance
that
we
would
use
for
multiple
purposes
or
whether
we
want
to
run
different
ones.
E
Okay,
linus
linux
and
I've
been
talking
for
the
past
couple
days
and
we
you
know
we
also
came
up
with
the
the
idea
of
using
grafana.
If,
if
that
seems
like
not
necessarily
an
idea,
you
guys
are
sure
about
but
are
open
to.
I
can
look
more
into
that
and
see
kind
of
what
steps
I
would
take.
E
No
that's
so.
This
is
kind
of
like
an
area
that
I
wasn't.
I
guess
super
knowledgeable
about.
That's
why
that's
why
I
was
coming
to
you
guys
just
to
see
if
there
was
something
or
any
any
ideas
that
you
guys
had.
D
Just
to
add
as
another
alternative,
so
there's
prior
art,
you
know
we
have
a
basic
dashboard
that
we
use
internally
and
that
one
is
just
just
an
html
page
with
just
a
table,
and
that
is,
I
think,
sufficient.
As
a
very
you
know,
simple
baseline,
so
short
of
griffon.
I
think
that's
also
an
option.
B
I'm
like
it
doesn't
need
to
be
fancy.
I
I'm
very
fine
with.
D
A
Feel
free,
as
I
said,
there
is
working
one
grafana
instance
I'm
open
to
to
help,
and
we
can
do
some
call
around
this
topic
too.
E
Okay,
I
can
yeah
I'll
look
into
the
grafana
thing,
I'll,
also
look
into
kind
of
yeah,
as
lionel
said,
the
the
kind
of
vulnerability
dashboard
that
google
has
internally-
or
I
guess
just
to
just
to
replicate
the
kind
of
like
html
table
that
we
have
there.
E
Okay,
if,
if
everybody
thinks
that
this
is
kind
of
like
a
it
would
be
a
useful
tool,
then
I
can
just
continue
working
on
the
idea.
D
I
think
it's
useful.
I
think
I
speak
for
everyone
when,
when
I
say
that
it
is
very
useful,
it's
a
very
useful
thing
to
have
just
more
visibility.
If
there
are
any
objections,
please
raise
them.
A
Consensus,
okay,
so
was
was
your
first
question
also
answered
or
I
because
I
just
missed.
I
think
that
first
question.
A
B
So
I've
been
I've
been
a
little
bit
absent
in
the
last
couple
of
weeks
with
code
freeze
and
everything
else.
I
imagine
other
people
have
been
under
pressure
in
the
same
way,
so
I'm
hoping
that
later
either
later
today
or
this
week,
I
can
start
thinking
again
about
what's
next,
we
we
ran
some
projections
on
the
current
numbers.
Once
we
flip
all
the
gcr
stuff
over.
B
Next,
I
have
no
strong
feelings
about
what
particular,
but
we
have,
I
feel
like
we've
done
a
lot
of
the
easy,
easy
quote
quote
stuff,
so
the
next
ones
will
probably
be
more
challenging.
B
I
don't
know
if
aaron
is
here
today,
but
I
felt
like
he
had
some
thoughts
about
what
he
might
want
to
tackle
next,
so
yeah,
just
as
a
group,
we
should
figure
out.
What's
the
next
sort
of
most
painful
pain,
point
for
people
who
want
to
be
excuse
me
co-administering,
stuff
and
start
the
next
one.
B
Yeah,
I
mean
it's
not
urgent,
but
this
is
what
we're
here
for
we've
we've
reached
a
good
plateau,
I
think
with
getting
all
the
gcr
stuff
done
and
honestly
I've
been
sort
of
holding
back
on
trying
to
think
about
anything
big
in
the
meantime
once
that
bridge
is
crossed
and
when
and
we're
sure
we're
not
going
back
across
it,
then
let's
burn
it
and
move
forward
onto
the
next
topic.
B
A
I
don't,
but
I
was
not
very
actively
looking
into
this,
so
maybe
it
is
a
good
time
to
ask
the
questions.
I
will
definitely
do
it
internally
and
maybe
I
will
think
how
to
get
this
knowledge
to
okay.
A
Actually,
that's
a
good
idea.
I
think
that
I
will
ask
somebody
from
the
release
team
to
come
to
our
next
meeting
and
give
us
some
update,
because
I
saw
they,
they
created
some
apps,
actually
some
cli
tools
which
are
doing
some
jobs
which
were
done
before
by
bus
scripts,
and
I
actually
am
not
very
familiar
with
it.
So
it
would
be,
I
think,
good
for
us
to
have
better
knowledge.
B
About
this
that'd
be
great,
I
also
I
saw
as
we're
moving
the
the
gcr
stuff.
I
noticed
that
the
release,
storage
and
serving
was
in
the
same
project,
so
it's
been
hard
to
sort
of
disentangle
the
costs
of
the
two.
So
honestly,
we
we
may
be
surprised
at
the
end
of
this
gcr
flip.
B
The
so
it'd
be
fun
to
get
the
release
stuff
moved
over,
it's
going
to
be
more
challenging
because
it
doesn't
have
a
vanity
domain
in
front
of
it,
and
so
the
the
real,
the
literal
bucket
name
has
been
hard
coded.
In
some
places
we
might
be
able
to
do
some
back-end
magic
to
move
the
bucket
between
projects
and
orgs,
I'm
not
sure
if
that's
possible
and
if
not,
then
we'll
have
to
think
about
how
you
strategize
how
we
want
to
move
stuff
over
yeah.
That's
a
good
idea.
A
And
I
will
ask
team,
pepper
and
the
guys
that
will
be,
I
think,
good
for
us
to
have
better
understanding
of
the
process
current
process
great.
This
is
my
extra
item
to
do
it.
Okay
and
also
justin
wrote
in
the
chat
about
this.
That
was
my
I
wanted
to
ask
because
I
saw
justin.
You
started
some
work
about
the
artifact
promotion.
C
Look
yes,
so
I'm
trying
not
to
bother
people
while
we
are
focusing
on
the
image
promoter,
but
once
the
imagery
motor
is
is
good.
Then
I
imagine-
or
I
hope
we
will
do,
the
same
thing
for
binary
artifacts
and
I'm
basically
working
on
getting
those
processes
going.
We
did
the
last
most
recent
cops
release
from
sort
of
following
the
patterns.
C
I
did
have
to
run
the
binary
promoter
manually,
but
I
created
like
the
pr's
and
did
all
that
sort
of
stuff
and
we
can
either
merge
the
pr's
or
just
put
them
in
on
hold
until
we're
ready,
but
they
are.
Those
artifacts
were
built
by
google
cloud
build
and
were
promoted
by
google
cloud
built
triggered
from
prowl
and
were
promoted
by
the
binary
artifact
promoter
following
the
manifest.
So.
B
Very
cool
one
of
the
things
we
need
to
figure
out
with
the
image
promoter
is
the
bulk
imports.
Maybe
maybe
we
never
have
to
do
it
again,
but
we
know
that
there's
a
sort
of
opening
for
running
out
of
api
quota,
and
if
anybody
remembers
about
a
month
ago,
it
triggered
the
alert
that
just
never
cleared
itself.
We
had
to
go
manually
flush,
the
queue
of
alerts,
so
I
know
line,
has
to
spend
a
little
time.
Thinking
about
that,
we
should
think
about
how
important
that's
going
to
be
yes
and
honestly.
D
Sure
yeah
the
the
basically
scaling
issues.
I
just
call
it
aps
gaming
issues.
That's
still
on
my
radar.
D
I
have
not
forgotten
forgotten
about
it,
just
I
guess
for
the
record
yeah
for
the
times
that
we
did
have
to
do
bulk
imports
like
the
past
couple
weeks,
when,
for
example,
last
week
when
I
prepared
the
new
project
to
have
the
backfill
of
all
google
containers
images,
I
think
we
promoted
like
200
or
something
well
times
three
regions,
so
like
600
plus
images
and
just
for
those
bulk
imports
for
those
promotions
that
I
did,
I
just
basically
disabled
the
auditor
from
working
and
then
I
flipped
it
back
to
switch
traffic
to
the
regular
one
after
that
was
completed.
D
B
D
Oh
right
and
actually
for
what
it's
worth,
there
have
been
just
regular
traffic
going.
You
know
promoting
new
images
and
for
those
things
or
promotions,
it
has
been
running
their
regular
code
and
we
haven't
seen
any
alerts
like.
I
just
want
to
stress
that
it's
not
like.
That
was
the
only
thing
that
happened
in
the
past.
B
What
might
be
cool
linus?
What
might
be
really
neat?
It
would
be
to
show
off
what
we've
got
like
where,
like?
How
would
somebody
here
go?
Look
at
the
logs
and
see
what's
been
promoted
and
when,
just
as
a
as
a
walk
through
of
like
look,
this
is
the
system
that's
running
and
somebody
besides
linus
should
be
able
to
figure
it
out
yeah
they
might
even
make
full
recording,
actually
or
or
something
to
show
off
to
a
larger
audience.
D
Sure,
okay,
I'll
take
a
note
of
that
as
an
action
item.
I
guess
we
could
probably
do
it.
Maybe
in
the
next
meeting.
B
No
pressure,
no
rush,
you
know,
but
I
think
it
would
be
worthwhile
to
do
because
we,
the
goal
of
all
of
this
work,
was
to
get
people
who
are
not
the
normal.
You
know
four
or
five
googlers
to
be
able
to
administer
these
things
so
having
it.
Knowing
that
it's
a
well
written
down
and
be
well
explained
would
be
useful.
D
I
guess
the
in
preparation
for
that
it
would
be
nice
to
just
have
other
people
in
the
same
like
groups,
for
example
the
artifact
admins
groups,
that
it's
kind
of
sparsely
populated
at
the
moment,
but
yeah.
B
Yeah
it'd
be
great
to
get
a
non-google
volunteer,
somebody
who
doesn't
have
a
ton
of
other
privileges
in
our
systems.
So
not
you
bart
and
probably
not
you
justin,
but
get
a
volunteer
to
sign
up,
for
that
would
be
great
yeah.
B
A
So
are
there
any
questions,
suggestions
or
other
topics,
because
I'm
kind
of
close
to
idea
of
giving
back
30
minutes.
A
Okay,
so
I
think
30
minutes
for
us
back.
Thank
you,
everyone
for
being
here
and
just
small
information.
I
probably
won't
be
able
to
to
host
our
next
meeting.
So
please
be
aware
of
it
and
thank
you
all.