►
From YouTube: Weekly Mac Shared Runners Sync - 20200915
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
Okay,
so
let's
get
started
sorry:
the
current
states.
This
week
we've
been
merging,
reviewing
and
merging.
Most
of
the
mrs.
There
are
still
a
few
in
a
review,
but
the
the
xcodes
and
runner.
Sorry,
the
runner,
vm
images
are
ready,
so
we
could
theoretically
start
onboarding
users
with
those
vm
images,
even
though
the
mrs
are
not
yet
merged,
and
then.
D
D
D
Project
I'm
using
the
json
what
is
going
on?
There
was
a
couple
of
example:
projects
already
in
the
gitlab
repo
jason
was
doing
a
flappy
bird
test.
Was
this
happy
wood
app
and
there
are
a
few
others?
This
one
is,
I
don't
think
needs
a
specific
version
of
x-core
and
the
double
check,
but.
D
D
So
that's
that
one
and
then
my
number
four
did
you,
as
you
guys
saw
on
slack.
The
gdk
team
created
their
issue
as
well,
so
we
should
probably
do
it
on
dry
run
onboarding
with
them.
I
don't
know
if
you
want
to
do
it
today
or
women
might
want
to
do
it.
A
Nice
yeah,
I
still
need
to
kick
off
the
the
build
for
the
xcode
vms.
I
had
to
delete
them
due
to
quota.
We
were
reaching
quota,
so
I'll
have
to
recreate
them
once
we
merge
these.
D
Okay,
all
right,
so
I
will
drop
a
note
back
on
for
the
gdk
team
directly
in
the
issue.
I'm
going
to
use
the
issue
as
the
way
to
communicate
with
all
customers
the
status
and
starting
with
the
gdk
team.
First,
I'm
not
going
to
communicate
with
other
customers.
Yet
so
I'm
going
to
respond
back
to
you,
hey
got
it.
I
will
start
the
onboarding
process.
Maybe
tomorrow
is
that
fair?
D
D
I
have
point
five
should
I
was
thinking
in
this
kind
of
just
was
like
dragon
cross.
Should
I
create
a
schedule
on
boarding
the
customers
right
now,
not
counting
the
gdk
team?
The
last
count
I
had.
We
had
13,
mostly
customers.
I
think
slash
users
that
have
requested
access
the
course
beta.
Would
it
be
helpful
for
me
to
say:
hey
customer
a
and
b
on
friday
customer
see
in
the
monday.
Would
that
be
helpful?
I'll,
probably
want
to
approach
the
scheduling.
Do
we
have
capacity
for
different
customers?
That's
78
cpus.
D
C
B
Which
we
have
10
max
right,
yeah.
D
D
D
C
D
And
approximately
I
mean
the
way
they
kind
of
flame
us.
The
way
maximum
has
communicated
to
me
is
that,
even
though
we
have
10
you
kind
of
want
to
hit
too
bad
for
your
own
infrastructure
needs
or
whatever
okra
whatever
it
is,
theoretically
is
eight,
and
so,
if
theoretically,
and
then
yes
on
the
core
capacity,
that's
kind
of
we're
doing
it
now.
One
thing
I
didn't
do
in
this
spreadsheet
is:
if
you
notice,
I
can
row
four,
I
just
have
the
ten
total.
I
am
not.
B
Cpuis,
my
only
concern
is
point
number
six,
we're
already
jumping
over
the
threshold
of
the
one
terabyte
we
have
because
of
the
amount
of
images
that
we're
building
the
number
of
jobs
that
we're
running
at
the
moment
is
abnormally
high.
Just
because
we're
developing
on
it
a
lot
so
we're
creating
a
lot
of
merchandise
and
a
lot
of
images.
B
But
we
need
to,
like
probably
be
super
concerned,
be
super
conservative
with
the
amount
we
allocate
for
development
or
we
have
to
increase
the
amount
of
storage
we
have
in
our
poo.
D
C
So
if
we
want
five
customers
on
boarded
by
25th,
that
means
we
have
to
onboard
one
per
day
right.
D
Yeah
exactly
and
what
that
does
as
well
is,
I
can,
depending
on
kind
of
as
we
finish
up
all
the
merch
requests,
we're
doing
now
we're
building
the
vm
images
and
we
say,
hey.
You
know
we're
actually
to
your
point
on
six
we're
having
the
storage
capacity
issues.
I
can
have
my
conversations
with
the
max
leader.
Folks,
like
this
flight
is
okay.
What
do
we
need?
What
can
we
do
to
adjust?
The
capacity
of
the
storage
shares?
Are
we
bumping
into
and
those
types
of
things.
C
A
B
B
A
Yeah,
that's
why
it's!
It
would
be
cool
to
have
the
gdk
team
as
soon
as
possible,
working
on
this,
because
I'm
sure
that
we'll
find
a
lot
of
low
hanging
fruit
before
we
start
onboarding.
Others.
D
So,
on
that
point,
how
about
this
there,
I'm
gonna
add
a
sec
so
on
the
scheduling.
So
how
about
this?
This
whole
scheduling
topic,
I'm
missing
a
spot
of
0.5,
sorry
how
about
gdk.
So,
let's
just
say
september,
what
are
we
on
today's
15
15.,
yeah,
okay,
so
september
16th,
vdk,
team,
right
and
then
some
september?
D
I
need
like
three
monitors.
Otherwise
I
need
a
way
to
have
a
calendar
open
because
I
think
it's
it's
definitely
old
age
kicking
in
like
non-remember
dates
off
the
top
of
my
head.
Is
that
bad?
Oh,
this
has
been
recorded
everyone.
I
do
one
short
day
so
september,
16
on
what
adjudicate
team
so
september.
17Th,
I'm
just
doing
business
days,
18th,
19th,
gjk,
18.,.
D
No
additional
customers,
that's
what
I'm
thinking
and
then
that
gives
us
17,
18
and
steve
you're
out
and
you're
also
on
monday,
as
well
you're
out
right.
Yes,.
B
I'm
off
18th,
21st
and
22nd
so
I'll
be
back
online
on
the
twenty.
D
So
maybe
if
it's
I
don't
know
if
our
schedules
permit,
maybe
we
move
this
call
to
the
23rd,
so
we
can
all
be
on
that
call,
and
maybe
so
then
what
I
would
suggest
is
september
and
the
third
is
first.
D
Not
updated
the
onboarding
issue
says:
hey
is
where
we're
at
the
first
customer
will
be
on
august
september,
23rd
pinging,
the
exact
customer
degree
and
then
defined
in
your
issue.
C
B
C
D
B
Is
that
good
for
you,
or
is
that
too
late?
For
you
me
yeah.
D
No,
no,
I
know
I
what
the
way
I've
communicated
this
back
in
the
meta.
Epic.
Is
these
big
belts
of
time
close
beta?
I
forget
cue,
something
open
beta
whatever
it
is,
and
then
ga
that's
fine.
The
only
thing
the
only
risk
that
I
have
to
work
through
is
that
it's
back
to
the
capacity
of
the
compute
layer.
So
we
have
the
10
macros
right
but
mac
in
terms
of
the
pricing
mac
stadium,
gifted
us
four
mac
pros
for
three
months
right:
we've
already,
you
blew
one
through
those
three
months
right.
D
D
D
A
B
That's
a
really
tricky
question.
I
would
say
it
would
still
be
closed
beta
at
that
point
just
because
we
still
have
a
specific
set
of
capacity.
I
I
like
for
the
open
beta.
Do
we
want
the
auto
scaler
to
do
everything
or
are
we
gonna
still
be
handling
vms
ourselves?.
D
Okay,
so
for
the
open
beta,
I
will
light
the
order
scaling
in
place.
Okay
before
we
open
beta.
However,
I'm
not
sure,
even
if
the
auto
scaler
was
in
place,
I'm
not
sure
I
want
every
single
gitlab
user
to
be
able
to
spin
up
a
mac
os
vm.
Is
it
the
data
I'm
having
internally?
In
my
mind,
right
now,
just
because
of
the
capacity.
C
C
B
If
we're
just
worried
about
capacity,
it's
we
can
limit
it
to
run
like
six
jobs
and
that's
it
and
then
depending
queue
is
gonna
be
forever.
For
example,.
B
That
is
something
that
we
have
to
build
in
the
auto
scaler,
because
right
now,
with
the
windows
auto
scaler,
we
just
are
going
under
the
assumption
that
we
have
unlimited
resources,
which
kind
of
we
do
with
windows
with
gcp.
B
So
we
just
keep
spawning
up
vms,
but
we'll
have
to
add
a
new
setting
in
the
auto
scaler.
Hey
only
run
six
jobs
at
the
same.
No.
Actually
we
can
do
that
already
nevermind.
We
can
because
the
runner
configuration
settings
we
can
say
how
many
concurrent
jobs
we
can
run
at
the
same
time,
and
we
just
put
it
to
three
or
four
right,
so
our
mac
os
shared
runner,
can
only
run
four
jobs
at
the
same
time,
for
example.
So
then
that's
our
capacity.
B
We
only
have
four
mac
os
shared
runners
for
the
public
right
and
then
they
can
just
wait
for
the
pending
queue
to
cheer
through.
D
B
D
Me
ask
you
this
question,
which
I
know
I
could
probably
only
answer
too,
but
I'm
going
to
ask
it
anyway,
because
I
probably
don't
know
the
answer.
There's
this
other
issue,
that's
sitting
on
our
running
back
for
some
time
called
runner
priority.
This
whole
concept
of
being
able
to
do
our
prioritization
setting.
Is
it
possible
for
us
to
know
today-
and
this
makes
no
sense,
but
maybe
it
does
the
mac
os
runners
are
only
going
to
be
available
to
the
github.com
instance.
Obviously
right
you
have
to
be
on
github.com
to
use
this.
B
B
Not
there
is
a
license
attached
to
the
namespace,
but
that
is
only
exposed
through
periscope
and
things
like
that.
Yeah.
B
B
Because
the
license
itself
is
hosted
in
gitlab,
so
a
namespace
can
have
a
license
and
we
know
if
the
user
has
a
license
or
not.
So
we
can
make
it
just
enterprise
edition
thing.
But
that
is
something
that
we
have
to
build
custom
and
I'm
not.
That
will
be
all
in
ruby
and
I'm
not
sure
how
it
would
work
to
be
honest
because
we
don't
have
any
gating
logic
between
enterprise
or
standard
edition
for
scheduling.
D
D
I
know
I'm
kind
of
off
about
the
agenda,
but
theoretically
we
can
have
the
auto
stealer
right
and
we
have
that
prioritization
letter,
but
only
four
jobs
at
a
time,
and
then
we
can
say
in
our
marketing
material
and
of
course
we
have
to
get
approval,
because
this
is
all
pricing
and
it
has
to
be
involved
in
this
right.
But
the
mac
os
closed,
beta,
open
bait
is
only
going
to
be
available
for
say,
enterprise
right
and
so
in
our
docs.
D
B
B
Yeah,
like
looking
at
our
usage
for
windows
spot,
we
get
six
jobs.
Seven
jobs
per
customer
like
overall,
outside
outside
of
in
gitlab
runner,
we're
gonna
handle
that
very
easily.
If
it's
the
same
for
mac
os.
D
D
And
so
they're,
basically
saying
that
all
those
costs
are
blended
right
and
so
basically
so,
instead
of
saying
I'm
trying
to
do
like
my
I'm
trying
to,
I
buy
a
mac
os
minute
at
making
up
a
number
a
dollar
a
minute
right,
and
then
I'm
selling
it
to
you
at
two
hours
a
minute
and
I
buy
my
gcp
linux
runner
at
25
cents,
a
minute.
That's
up
to
you
like
50,
submit
you're,
saying:
okay,
no,
I
buy
a
thousand
minutes
and
it
happens
to
be
some
of
it.
D
I'm
paying
a
vendor
my
stadium,
x
dollars
and
some
of
them
obtain
google
right.
So
my
total
cost
is,
let's
say
three
dollars
for
how
many
minutes
I
have
and
then
you're,
just
basically
giving
a
blended
price,
knowing
that
that
the
the
actual,
true
margin
that
I
make
for
the
mac
is
less,
but
because
I'm
blending
and
my
price
point
kind
of
works
out
so
that
why
I'm
blending
but
I'm,
but
I
still
have
to
double
check
the
math,
makes
sense
over
time.
B
D
Going
to
be
the
same
price,
potentially
that's
one
suggestion
that
was
made
and
that's
why
I'm
double
checking
with
the
pricing
team
over
the
weekend
because
they
closed
really
fast.
They
closed
the
pricing
issue
for
windows
and,
I
said:
hey
we
haven't
done
the
work
on
this
yet
and
the
response
was
right.
Now
the
usage
of
windows
is
so
low.
It
doesn't
make
sense
for
us
to
have
different
pricing
for
windows,
because
yeah,
but
to
your
point,
mac
is
different.
D
That's
why
you
double
check
the
the
math
with
math
this
week
to
see
if
that
still
makes
sense.
Otherwise
you
still
have
to
go
back
to
the
work
that
camille
pointed
out
with
the
credit
system
right
exactly
exactly
and
then
figure
out
how
we
actually
measure
the
usage
of
the
win
of
the
mac
minutes,
because,
right
now
in
good
lab
is
just
one
minutes
block
right,
like
hey,
you're,
a
thousand
minutes
and
you're
not
on
500.
Anyway,
I
digress.
B
D
My
last
point
was
on
this
on
the
on
the
worksheet
that
I
sent
you
guys
there
a
second
ago.
There's
a
tab
called
closed
beta
capacity
planning.
You
have
to
look
at
it
right
now.
I
was
just
going
to
capture
in
here,
like
maybe
customer
name,
I
don't
know,
there's
some
kind
of
way
to
kind
of
like
really
quickly
see
if
we
auto
capacity
or
act
capacity.
D
So
if
you,
when
guys
want
to
take
a
look
at
this,
I
hack
it
up
to
see
if
we
have
the
right
parameters
in
here
from
tracking
this.
This
was
just
my
first
cut
and
and
so
on.
So
this
is
wide
open
for
all
us
all
three
of
us
in
the
entire
team
to
like
hack
on
and
kind
of
figure,
the
right
way
to
manage
yeah.
How
do
we
do
this?
That's
not.
C
D
Awesome
all
right,
hey
guys,
I'm
gonna
drop
off
and
start
doing
a
bunch
of
stuff
I'm
supposed
to
get
done.
This
was
super
cool,
yep,
okay,.