►
From YouTube: wg-k8s-infra biweekly meeting 20200219
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
A
B
C
Can
see
him
John,
okay,
so
it's
John's
story,
I
totally
forgot
about
your
part.
I
can
only
stay
for
a
few
minutes.
I've
got
to
go
to
a
doctor's
appointment
with
my
kids.
So
if
possible,
can
I
front-load
my
updates
and
then
I
can
drop
off
yep
okay
I
spent
time
yesterday,
turning
up
Kate's
that
IO
in
Tripoli,
it
is
up
it's
running,
it
is
I,
did
not
flip
the
DNS,
so
I
sent
a
pull
request
last
night,
which
needs
to
be
reviewed.
C
Specifically.
The
certificate
I
asked
runners
to
look
at
and
the
group
names
were
wrong,
so
I
have
to
that's
on
the
same
PR
and
then
the
last
part
is
that
it
flips
DNS
I
chose
not
to
flip
it
at
11:30.
Last
night,
I
thought.
Maybe
that
was
a
bad
idea
and
I'm
disappearing.
Now,
so
I
haven't
flipped
it
yet
this
morning,
but
practically
we
could
flip
it
in.
You
know
six
hours
or
something
and
it
should
just
work
all
the
tests
pass
against
it
in
the
new
cluster.
So.
C
Self-Tuning,
okay,
there
were
like
there's
not
actually
any
changes
from
the
old
config.
Just
take
a
look
at
it
make
sure
that
it
still
represents
sort
of
the
best
practices
for
this
particular
version
of
cert
manager,
and
if
we
need
to
upgrade
cert
manager,
we
can
do
that
in
a
separate
track.
Yeah
sounds
good.
Okay,.
A
D
C
C
Yeah,
the
vast
majority
of
it
is
build
time
high,
CPU
32.
We
can
dig
into
that,
but
there's
10,000
minutes
of
build
time,
I'm
guessing
that's
release,
team
doing
stuff
and
then
the
next
is
the
clusters
and
then,
after
that,
it's
DNS
which
is
down
to
a
mere
366
million
queries
this
month,
as
opposed
to
the
billion
and
a
half
last
month
yeah.
So
we
can
say
that
it
was
fixed.
D
C
Okay,
anyway,
I,
don't
think
anything
seems
particularly
out
of
line.
Maybe
we
should
poke
them
just
to
see
if
they
are
spending
it
the
way
that
they
expect
to
be
spending
it.
You
know
it's.
It's
always
useful
to
put
a
dollar
amount
on
there,
you're
testing,
but
I,
don't
think
they're
doing
anything
wrong.
I,
don't.
D
Know
if
you
can
see
with
a
screen
head
up
now
Tim,
but
we
do
have
a
a
month
to
date
by
service
and
project,
and
it
does
show
indeed
Kate
staging
release.
T-Test
is
cloudbuilt
$400,
and
that
is
the
the
big
one
and
we
can
see
like
the
next
one
is
the
next
classic.
Our
bonus
is
cluster
API
and
then
some
project
called
cops
and
then
staging
artifacts
P,
something
so
but
yeah.
The
lion's
share
is
about
build.
So
we
have
this
one
already.
E
Yes,
so
last
time
I
think
I
touched
on
addressing
the
essentially
the
separation
of
I
started.
Actually
I,
don't
know
how
much
I
can
go
into
this,
but
there
are
some
security
related
things
that
I
have
to
figure
out
or
iron
out
from
the
Google
side,
so
I'm
in
the
middle
of
doing
that.
The
long
and
short
of
it
is
that
I
think
end
of
this
quarter
is
the
so-called
tea
date
or
the
date.
When
we
say
that
Google
containers
will
enter
a
freeze.
E
E
Well,
as
a
read
me
a
few
days
ago,
so
that
will
outline
the
process.
The
hope
is
that
that
will
happen
end
of
this
quarter.
So
that's
the
date.
So
we
need
to
like
send
out
I
suppose,
maybe
a
more
official
announcement
on
probably
the
other
mailing
lists
out
there
I'm
detailing
this,
so
it
actually
doesn't
affect
open
source
site
that
much
because
people
are
already
using
the
promoter
to
promote
to
the
new
Kate's
artifacts
prod.
E
Kids
on
FX
product
instead
of
Google
containers,
so
that's
my
update,
I,
don't
really
have
anything
else.
Are
there
any
questions
on
this
I
know
a
claudio
that
you
are
asking
on
the
slack
channel
about
why
there
are
so
many
different
registries
and
how
could
they
don't
redirect
its
the
correct
region?
And
that's
because
this
domain
flip
hasn't
happened
yet
so
after
March
31st,
you
should
be
able
to
just
push
to
Kate's.
You
see
our
I/o.
A
E
Sorry,
one
more
thing
because
you
mentioned
terraform,
so
I
actually
also
emerged
or
submitted
a
PR
for
deploying
the
cloud
run
service
on
GCP.
This
is
the
auditing
service
that
will
check.
You
know
any
changes
to
GC.
Are
they
okay?
Are
they
not?
Okay,
that's
currently
written
in
bash,
but
some
you
know
g-cloud
sprinkled
in
there.
So
that's
like
a
great
candidate
to
terraform
eyes.
It's
already
been
merged.
People
can
look
at
it.
It's
a
small
unit
of
work
that
I
think
is
right
for
terraforming.
Maybe
that's
a
good
place
to
start.
You.
A
Okay,
so
there
is
a
topic
which
I
have
added
on
behalf
of
one
of
the
users
about
the
emigration
of
publishing
bot.
But
what
I
see
is
in
the
comments
that
there
are
already
some
action
items
which
should
cover
this
topic,
because
there
was,
there
was
some
discussion
about
moving
copy
pasting.
Some
part
of
code
deploy
the
publishing
bots,
which
some
people
didn't
like,
and
the
action
item
right
now
is
to
push
some
changes
to
the
board
itself
and
I.
Think
that
it's
at
this
point
it's
it
covers
this
discussion.
A
D
Justin,
yes
sure
I,
just
added
this
certain
last
minute,
we
wanted
to
rotate
the
AWS
they're,
a
bunch
of
eight
of
us
test
accounts
that
are
used
for,
for
example,
cluster
API
provider,
AWS
kappa
testing,
and
we
wanted
to
rotate
those
that
have
been
a
while.
So
we
created
new
ones,
those
are
not
currently
managed
by
this
group.
D
I
guess
I
did
manage
to
figure
how
to
script
it,
which
was
a
big
improvement
over
how
I
did
it
the
first
time,
but
I
guess
the
question
is
whether
we
want
like
those
scripts
to
live
in
well.
First
of
all
consider
just
manage
outside
of
this
world,
like
in
shadow
IT,
whether
we
put
those
back
scripts
into
the
Cates,
IO,
repo
or
and
or
whether
we
try
to
move
that
to
terraform
and
manage
it.
A
Personally,
think
that
is
the
good
place
to
put
them
in
the
case
in
from
our
repository
and
also
I,
think
that
I'm
in
the
position
where
it
would
be
good
to
start
already
writing
it
in
terraform.
But
if
it's
easier
faster,
we
can
start
with
bash
and
then
I
write
it
to
terraform.
But
that's
my
own
opinion
I.
D
G
A
B
Quite
a
long
time,
I
haven't
got
like
a
major
discussion
point,
but
mostly
given
we've
got
45
minutes,
I
thought
I'd
quickly,
say:
Tim
mentioned
about
upgrading
cert
manager.
At
some
point,
I
need
to
go
and
test
out
my
credentials,
and
things
is
a
it's
a
little
bit
more
involved.
This
particular
upgrade
that
we're
doing
because
we're
basically
we're
still
using
the
old
CRTs
that
use
the
kata
IO
suffix.
B
F
B
I
think
there's
not
too
much
to
add.
I
need
to
go
test
out
my
credentials.
They
give
me
access
to
the
server
manager.
Namespace
I
think
I'll
need
to
modify
things
like
cluster
roles
and
cluster
role
bindings,
because
obviously
it's
a
cluster
that
we're
service
and
then
also
double
check
all
of
the
different
resources
across
the
clusters
to
make
sure
everything
that
we
previously
had.
B
We
now
have
in
the
new
format
and
in
the
repo
and
everything
so
I
can
go
through
put
together
all
the
PRS
for
that
and
get
it
all
prepared
ahead
of
time
in
terms
of
actually
rolling
it
out.
It
might
I
mean
I
can
just
hand
off
to
someone
but
I'm
also
quite
happy,
and
maybe
it's
safer.
If
I
do
that,
maybe
on
my
comparing
session
with
someone,
so
we
can
actually
make
sure
it
was
multiple
people
looking
at
I,
just
don't
know
about
credentials.
Wise
I
might
need
certain
permissions
beyond
who
I
help.
A
My
suggestion
is
I'm
very
open
to
do
some
per
session.
If
you
want
I
think
that
I
have
credentials
and
if
you
will
need
something
we
can
sit
together
and
work
on
it
and
just
check
what
you
need
and
I'm
going
to
discuss
with
other
if
it
really
just
easy
to
just
run
through
these
permissions,
or
we
should
do
it.
Some
other
way,
I'm
very
open
to
do
perception.