►
Description
A
All
right
just
guys,
okay,
so
let
me
quickly
summarize
the
in
turn
the
way
I
see
it.
We
basically
wants
to
close
a
hole
of
like
public
projects
consuming
unlimited
amount
of
the
minutes,
because,
like
our
Terms
of
Service
as
there
is
like
fifty
thousand,
we
even
gonna,
probably
like
reduce
that
amount.
Just
today,
there
is
like
no
control
to
like
to
to
define
that
or
like
to
enforce
that
within
the
system.
So
the
intent
is
basically
like
being
able
to
configure
that
within
the
system
and
kind
of
make
everyone
at
her
to
this
intent.
A
From
from
what
I
understand
as
a
subtopic,
we
want
to
consider
like
two
tiers
of
these
three
type,
penis
that
here
like
of
existing
users.
They
cannot
have
slightly
bigger
allowance
and
under
care
of
the
completely
new
users,
are
gonna,
have
smaller
allowance
of
these
criminals,
but
the
the
Internet
in
generics
like
that
everyone
gonna,
have
some
amount
of
the
minutes.
That
is
clear
and
documented
presented,
but
it's
not
unlimited
as
it
is
today.
B
Yeah
I
will
add
two
things
to
that
and
maybe
I'm
gonna
show
my
screen
use
I
think
this
has
helped
me
think
through
it.
Today,
a
in
our
pricing
page.
We
say
that
if
you're
a
free
customer
on
Ghaleb
calm,
you
are
allocated
two
thousand
five
minutes
per
group
per
month
on
our
shared
runners
and
then,
if
you're
a
public
project,
we
say
you
get
gold
status.
So
that
means
you
get
fifty
thousand
pipeline
minutes
per
month
on
our
shared
runners
and
I.
B
Think
we
want
to
reduce
both
of
these,
so
the
two
thousand
would
go
down
to
a
thousand
and
maybe
for
free
public
projects.
Instead
of
saying
you
get
everything
in
gold,
you
would
get
some
you
get
the
gold
features,
but
your
your
pipeline
allocated
minutes
is
something
less
than
50,000
and
to
your
point
that
we
would
enforce
it.
I
think
we
consciously
do
not
enforce
it
on
public
projects.
Today,
Hurley's
like
in
code.
B
A
Iii
think
it's
this
time,
probably
like
one
think
that,
like
that,
our
procedures
that
evolve,
it's
like
consideration
of
the
public
projects,
which
kind
of
like
we
kind
of
be
more
strict
like
but
considering
the
public
projects
are
separate
care,
but
rather
making
the
student
part
of
the
allowance
that
you
receive.
Oh,
but
but
overall
I
I,
think
like
we.
We
look
into
like
two
type
of
the
limits
for
the
frame
we
just
like
the
current
plus
was
the
new
users
and,
like
the
second
guy
I,
suspect
some
high-risk
and
open
source.
A
B
There
and
and
then
I
guess
what
I'm
highlighting
is.
There
is
a
that
issue
that
I've
added
the
product
issue
861.
It's
meant
to
just
be
the
R&D
efforts,
because
they're
separate,
like
legal
discussions
and
marketing
changes
that
we
would
want
to
update
the
number
of
minutes
on
the
web
page
and
can
we
legally
adjust
somebody's
contract
if
they
previously
signed
up
under
these
terms,
and
that
is
why
we
kind
of
have
to
have
that
other
tier.
A
Okay,
let
me
describe
exactly
what
type
but
different
types
of
problems
we
have,
how
how
they
actually
behave
and
how
we
also
want
get
off,
because
there
seems
to
be
some
sub
challenge
like
with
ensuring
that
we
not
break
our
workflow
today,
yeah,
just
like
very
heavy
community-focused,
sir.
My
screen.
B
B
A
A
It's
quite
connected
like
with
what
we
want
to
achieve,
and
maybe
we
figure
out
like
some
small
changes
that
gonna
make
this
easier
for
the
future
generations,
but
I
see
that
it's
already
added.
There
is
nothing
bublik,
as
there
is
nothing
private
in
this
issue.
Are
you
aware
of
this
issue
and
how
yeah.
B
A
Copy
that
into
the
it's
easier
for
me
to
write
that
in
vs
code
for
now.
Thank
you
because,
because
of
the
formatting
and
monotype
font,
it
breaks
every
demon.
You
try
to
do
some
things
like
that
on
good
luck,
but
also
like
cert
runners
and
things
like
that
they
are
the
documentation
so,
but
maybe
quickly,
some
are
of
all
types
we
have.
Let's
continue
that
group
runners
so
can
be
used
by
anyone.
A
Group
runners
are
assigned
to
single
group,
so
group
runners
internally
kemi's
can
be
assigned
to
multiple
groups,
but
the
current
design
is
like
that.
The
group
runner
is
assigned
to
the
single
group
today
it
doesn't
matter
whether
this
is
like
subgroup
or
top-level
group.
Basically,
every
project
that
has
the
group
runner
assigned
can
use
that
runner
and.
B
A
Yes,
but
wait
but
rank
you
cannot
assign
the
group
runner
to
multiple
concurrent
groups,
let's
say
like
two
different
namespaces:
this
is
what
you
ask:
project
runners
are
assigned
to
many
projects.
So
basically,
as
long
as
you
have
permissions,
you
can
attach
the
given
project
runner
to
multiple
projects.
It
doesn't
matter
the
affinity
of
the
groups,
it's
it's
fully
free,
like
freely
allocatable
system.
You
want
in
the
project
a
that
is
in
group
a
it's,
no
problem,
if
you
want
in
the
project
being,
would
be
it's
also
not
a
problem.
C
A
So,
like
I,
think
about
of
that
as
different
personas
managing
these
runners,
so
runners
are
being
managed
by
the
system,
administrators
that
are
managing
it,
love
and
maybe
managing
the
whole.
The
whole
installation
group
runners
are
being
managed
by
regressor,
a
DevOps
of
like
the
specific
group
team
or
whatever
private
runners.
You
may
be
used
registered
a
specific
one
or
like
performing
some
complicated
deployment
to
some
of
your
production
environments,
so,
like
these
different
levers
are
basically
to
be.
A
They
are
used
to
like
to
represent
different
personas
managing
these
groups
also,
like
you,
cannot
transform
project
runner
into
group
runner,
and
so
it
wander
into
group
their
story.
There
is
no
a
way
to
like
to
migrate,
one
runner
to
another
type.
There
is
also
the
runner
cannot
be
of
multiple
types.
Their
owner,
II
always
have.
The
single
type
assigned
can.
B
A
It's
it's
so
everyone
aware
it's
created,
it
has
assigned
type
and
it
allows
the
mic.
Usage
of
this
type
are
lost
under
like
the
management
of
this
runner
and
like
what
type
of
the
permissions
are
being
executed,
and
what
is
executing
this
is
information
that
is
being
used
as
an
ID,
optimization
part.
Oh,
there
is
like
one
reason.
The
second
reason
is
it's
not
that
you
can.
You
cannot
do
that.
Probably
we
could
implement
a
control
that
would
allow
it
migrate
that,
but
it
has
a
few
security
challenges
connected
with
that.
A
It's
basically
based
on
the
assumption
that
it's
easy
to
register
a
new
runner
that
solves
a
very
specific
intent,
but
it's
very
hard
like
to
provide
you
a
controls
and
ensure
that,
like
our
security
aspects
of
managing
the
runner
and
migrating
that
between
different
types
and
ensuring
that
this
representation
is
like
reflected,
it's
saving
significantly
more
harder
to
do.
Okay.
B
C
So
so
what
is
the
primary
motivation,
though,
to
have
these
different
types?
Is
that,
like
permissions
or
like
access
control,
like
you
mentioned,
security
concerns?
I,
don't
quite
understand
like
why?
Why
do
we
have
this
sort
of
breakdown
at
all,
rather
than
just
saying,
like
a
runner
as
a
runner,
and
like
you
assign
you
know
whatever
jobs,
you
define
it
to
run.
C
A
Consider
ourselves,
we
have
the
third
one.
Our
state
is
used
by
everyone
that
is
using
github
dot-com.
We
have
group
runners
that
we
assign
for
every
project
in
the
keep
and
grow
it
could
be.
Our
runners
consider
perfect
runners
when
you
assign
a
very
specific
runner
that
can
perform
a
deployment
to
only
specific
environment,
and
you
don't
want
this
runner
that
is
assigned
to
the
project
to
be
shared
between
any
other
projects
like
meeting
access
controls
or
like
like
usage
patterns.
Okay
is
primarily.
A
B
Yeah
I
think
I,
don't
know
if
this
is
a
true
statement.
Camille,
but
I
would
imagine,
there's
also
like
a
cost
control
like
if
I'm,
the
administrator
of
an
instance
there's
a
certain
amount
of
runners
that
I
want
to
make
available.
But
if
the
project
wants
to
have
more
capacity
or
needs
some
unique
capacity,
they
managed
because
it's
it's
gonna
compute,
so
they
have
to
bring
their
own
compute.
That
is
then
isolated
and
only
available
to
that
project
or
group
and
they're
bearing
the
cost
of
that
computer.
B
A
Like
maybe
not
really
about
the
cost
control
but
more
about
the
specific
requirements
control
it's
much
easier.
Basically
and
like
we
know
how
of
organization
the
choice
they
bring
their
own
runners
because
they
need
some
fancy
workflows
to
run
and
it's
a
simpler
than
trying
to
provide
that
for
everyone,
but
also
like
search
because
it's
served
by
everyone.
You
need
to
be
additionally
more
careful
exactly
how
you
run
that
work
like
this
untrusted
workload
in
the
search
space
on
the
private
runner.
A
You
always
know
that
you're
running
on
in
the
given
project,
so
there
is
like
less
likely,
be
T
of
sharing
some
confidential
data
from
your
private
secrets
or
whatever
else,
because
you
are
not
showing
the
same
like
that,
the
same
space
before
the
praise
it
would
be
running,
conduct
I
mean
let's
move
on
it's.
It's
not
really
that
important,
but
it's
good
to
like
know
exactly
what
type
of
levers
we
have
and
basically
each
project
has
an
ability
to
enable/disable
showrunners,
a
name
or
disabled
group
runners.
A
Basically,
there
is
a
button
in
the
project
settings
when
you
can
say
I
want
to
use
short
runners.
I
want
to
use
kilometers
and
this
kind
of
like
a
switch
for
the
whole
project.
So
you
you
don't
want
to
use
through
partners
aware
you
don't
want
to
ever
use
search
runners.
You
always
want
to
use
only
private
ownership.
You
can
really
limit
that
and
your
baby
not
using
it,
but
but
let's
just
kind
of
go
into
more
a
third
runners
territory.
A
Okay,
so
it
wanders
for
github.com
has
a
control,
speed
that
allows
you
to
assign
or
or
more
like
sorry,
and
that
allows
you
to
validate
amount
of
the
minerals
that
you
are
consuming
and
the
CI
minute
is
only
a
property
of
the
third
runner.
It's
not
a
property
of
the
group
and
it's
not
a
property
of
deprave
runner.
It's
it's
based
on
the
assumption
that
like
if
you
are
running
the
big
instance,
you
are
sharing
the
resources
group
project.
It's
like
much
smaller
entity,
so
we
don't
need
to
have
additional
controls
of
the
usage.
A
Only
smaller
entities
so
like
we
are
only
tracking
amount
of
the
minerals
that
you
consumed
on.
The
third
runners
then
set
runner
minutes
is
not
a
property
of
the
runner.
It's
not
that
like
runner
device,
every
group,
namespace
or
project
can
use
that
amount
of
the
minutes.
It's
rather
like
every
product
receives
some
allowance
on
the
of
the
project
authority
of
the
minutes
that
can
use
across
all
available
shell
runners,
for
example,
search
runners.
A
A
There
is
there.
Is
this
discussion
about
like
some
kind
of
metric
to
measure
or
different
type
of
the
sizes
of
the
machines?
We
don't
really
do
that
today,
so
we
considered
each
runner.
That
is
mark
assured
to
be
of
the
same
of
the
same
cost
and
basically
the
artery.
Today,
it's
very
simple
like
there
are.
There
are
like
two
level
of
the
limitations,
so
so
trainers.
A
A
A
A
So
when
the
Dove
finishes,
we
then
accumulate
see
I
mean
oats
being
consumed
for
for
a
given
namespace.
So
there
is
like
one
assumptions
built
into
that
system
which
in
some
cases
is
wrong,
but
it
actually
it's
much
easier
to
model
its
assumption
is
that,
like
the
system
allows
you
to
escape
the
image
because
we
validate
amount
of
the
consumed
minutes
before
actually
executing
the
workload,
so,
for
example,
we
will
consume
59
minutes
out
of
your
60
a
lot
minutes.
We
don't
really
know
ahead
of
the
time
how
long
you're
gonna
take.
A
C
A
So,
if
you
look
at
the
2,000
minutes,
we
may
escape
is
2,000
me
notes
by
amount
of
the
jobs
that
are
being
picked
at
the
given
time
multiplied
by
this
maximum
duration.
That
is
allowed
for
your
job
to
execute,
and
this
was
this
was
like
the
aspect
of
the
system
that
was
being
was
being
abused
by
the
abusers
because
they
would
fire,
for
example,
100
jobs
in
the
private
project,
and
each
of
these
jobs
would
take
3
hours
and
basically
they
would
escape
these
limit
quite
significantly.
A
A
There
was
like
one
extension
being
thought
of
being
Fausto
to
this
process
like
if
we
discover
the
gate,
beaded
fire
or
whatever,
as
so,
you
did
go
over
that
he
need
it
would
proactively
go
and
cancer
of
your
existing
jobs
ensure
that
you
are
not
consuming
any
more
minutes
as
of
now.
If
this
was
like
one,
one
extension
that
was
considered
and
as
for
the
as
for
the
accumulation
accumulation
currently
deliver,
any
attorney
ignores
public
projects.
This
is
by
design,
and
this
is
also
like
the
original
idea
behind
that
that
we
only
want
to
that.
A
We
wanted
to
meet
the
private
projects
and
internal
projects,
but
the
public
projects,
since
they
are
openly
aloud
basically
to
consume
pretty
much
unlimited
resources,
which
was
in
some
cases,
we've
been
abused
so
now
like
there
is.
How
does
he
have
minutes
or
like
CI
minutes
limit
is
being
said,
because
there
are
like
two
ways
how
how
this
limit
can
be
set,
and
this
is
exactly
how
it
currently
functions.
We
have
the
customers
github.com
portal,
which
basically,
we
link
name
space
with.
A
According
to
a
plow,
but
this
part
I
think
it
only
looks
at
the
paid
users,
so
it
has
where
it
is:
zero
knowledge
about
the
three
users
and
for
the
three
users
what
we
do
is
like
for
three
users:
we
have
application,
setting
where
there
is
like
CI,
CD,
being
woods
limit
or
something
of
sorts.
So
basically
this
CI
CD
minutes
we
need
takes
precedence.
A
Sorry,
it
does
not
take
precedence.
This
CI
c
demeanors
limit
is
a
fallback
if
there
is
no
more
granular
or
image
search
for
a
given
name
space.
So,
for
example,
if
the
name
space
is
being
said
by
this
customers,
gift-wrap
comm
portal
to
let's
say
4,000,
this
limit
gonna
be
used,
and
this
is
gonna
be
enforced.
If
this
thing
is
like
not
such
we're,
gonna
use
the
image
that
he
said
like
it,
lab
comm
system
wide
instance,
which
is
application
on
settings
and
CI
CD
minutes
limit.
A
A
A
We
run
CI
for
git
repositories.
Basically,
we
use
a
mix
of
all
possible
runners,
like
we
have
project
runners
which
are
named
private,
eggy
top
runners
manager,
it's
basically
how
to
scale
architecture.
It
is
assigned
to
some
of
our
projects,
which
is
like
the
price
that
are
being
used
by
my
arethey
of
the
get
lot,
which
is
get
up,
get
up
false
security,
key
fob
and
omnibus
key
flood,
and
whatever
else
the
price
are
there.
We
probably
also
have
the
group
when.
B
B
B
A
B
A
A
Basically,
like
the
product
runners
grouponers
allows
you
to
only
run
the
CI
within
the
given
project
or
like
the
given
group,
but
Forks
is
worried.
They
are
the
products
that
are
leaving
in
the
separate
namespaces
so
currently,
today,
there
is
no
function
that
would
allow
you
to
inherit
the
runner
from
like
from
your
parent
project.
For
example,
you
have
a
six
runner
assigned
to
write
to
your
first
name
project
under
under
github.com,
and
you
want
every
one
of
your
contributors
to
use
this
ICS
runner.
There
is
no
fun.
A
There
is
no
ability
to
do
that
today,
so
it's
kind
of
like
any
meeting
factor
because
we
have
to
them
register
a
cert
runners,
because
we
need
to
support
Forks.
Basically,
all
Forks
run
on
the
shelf
runners
and
the
runners
that
are
being
named
basically
like
that.
It
kind
of
indicates
that
this
are
the
shirt
runners
that
are
meant
to
be
used
for
for
geet
lab
internal
purposes.
They
just
have
different
tags,
which
is.
A
Why
it
is
important
that
that
I
mention
it's
important
from
for
two
reasons
like
first?
This
is
the
only
way
to
round
Forks
today,
because,
like
these
get
larger,
burners
are
more
optimized
to
run
this
workflow,
they
don't
run
in
the
privileged
mode,
so
we
don't
have
to
recycle
machines
every
time,
also
just
because
they
run
only
our
workload.
A
So
there
is
no
data
security
issues,
but
the
second
reason
is
that
we
can
make
them
short
runners
because
they
run
public
projects,
so
they
do
not
differ
anywhere
as
from
the
existing
cert
runners,
because,
basically,
like
you
have
shirt
runners
that
are
for
each
up
and
for
everyone
else,
and
basically
accounting
of
the
minutes,
if
someone
would
use
the
club's
hurt,
runners
manager
would
work
basically
the
same.
There
is
no
accounting
for
the
public,
but
there
is
accounting
for
the
for
the
private
projects.
A
A
B
A
A
C
C
C
A
Fourth,
fifty
percent
is
something
different,
so
not
really
today,
but
if
we
change
I'm
a
girl
be
a
small
security
issue
that
have
to
be
addressed
in
the
future,
because
because
now
like,
let's
consider
the
third
runners
case,
we
need
to
continue
using
that
for
Europe
for
time
game
being
whatever
we
do.
But
the
problem
is
like.
A
We
don't
like
even
our
community
contributors
to
have
free
me
to
contribute,
get
lock.
Basically,
so,
like
you
basically
need
to
say
on
the
github
site
runners
which
can
steal
everyone
use,
there
is
no
limit
on
the
amount
of
the
minutes
being
used.
So
when
someone
discovers
that
aspect
of
the
system,
it
could
basically
use
github
sultanas
for
unlimited
purposes
as
it
is
happening
today.
It
basically
comes
from
the
problem
of
the
lack
of
easy
way
for
sharing
or
like
sharing,
but
solitaire
is
so
much
different
from
like
someone.
A
A
You
designate
some
runners
to
be
shared
with
the
forks
and
run
in
the
context
of
the
fork
or
whether
you
mark
that
you
want
the
merge
request
from
the
fork
to
be
run
in
the
context
of
the
your
parent
project,
which
is
like
we
don't
have
neither
of
these
feature
yet,
but,
of
course,
being
prompt
in
the
future.
So
this
is
something
that's
going
to
be
improved
in
the
file
in
the
in
the
next
milestones
and.
B
Just
so
I'm,
maybe
understanding
a
connection
here.
That
would
mean
if
we
have
the
ability
to
like
a
sine
row
like
I,
think
if
it
has
a
checkbox
in
the
CI
settings
for
a
group
that
said
anyone
who
Forks
from
these
projects
can
use.
My
group
runners,
then
we
actually
wouldn't
need
the
for
get
lab.
Shared
runners
is
that
right.
A
B
Because
it
sounds
like
you
are
saying,
the
only
reason
why
we
need
does
like
unique
or
for
gitlab
shared
runners
is
because
we
want
to
support
Forks
and
if
we
have
some
way
to
do
that
through
group
runners
that
are
available
to
other
projects
that
aren't
in
our
group
but
are
because
their
Forks.
We
could
just
use
gitlab
group
runners
for.
A
Yes,
so,
basically,
if
there
would
be
some
way
for
you
to
reuse
the
same
runners
that
you
have
today
assigned
to
your
plate
and
designate
some
of
these
runners
either
to
be
shirt
or
mark
some
of
these
runners,
as
long
as
that
can
run
untrusted
workflow,
which
may
be
hub,
which
could
be
a
case
that,
like
you,
make
your
marriage
Equus
pipeline
from
the
fork
to
run
in
the
context
of
the
parents,
then
you
don't
really
need.
The
short
run
is
for
the
github
repositories,
because
this.
A
C
A
I
mean
it's
done
today,
because
this
is
the
only
way
to
supporting
Forks.
It's
the
only
way
to
be
more
cost
effective
and
running
github.
Otherwise
you
would
be
using
everyone
else,
sure
runners
which
are
significantly
more
expensive
because
they
are
show
away
machines
technically
and
guitar
runners
they
by
default,
run
multiple
subsequent
jobs
on
the
same
machine
which
increases
the
cost
effectiveness
because
you
are
not
removing
and
recreating
the
machine
over
and
over,
which
is
which
is
which
has
some
implications
on
additional
cost
catches
in
the
cloud.
B
Can
I
go
back
to
something
you
said
about
where
free
users
plans
are
so
the
customer
team
pointed
me
to
some
lines
of
code
that
they
say,
outlines
the
plan
and
has
a
free
plan,
but
I
wonder
if
that
is
just
some
code.
That's
not
actually
used
I,
just
post
pasted
it
in
here,
but
it's
it
has
one
line
that
says:
free
plan
is
2,000
but
I'm
wondering
if
you're
saying
actually
that
line
is
never
used.
It's
actually
only
that
instance
wide
setting
that
is,
is
setting
the
free
I.
A
B
A
A
B
A
B
B
A
How
we
support
third
runners
give
up
repository
sport
flow
for
folks,
because,
like
whatever
we
do,
we
like
we
cannot
really
start
counting
the
case
of
usage
of
our
runners
for
our
internal
purposes,
which
may,
unless
there
is
better
future
for
migrating,
the
third
runners
of
github
to
be
like
the
group
search
runners.
Sorry,
the
project
search
with
the
forks
runners
or
whatever
is
like
this
one.
With
these
two
cases,
it
potentially
open
us
to
some
potential.
B
A
I
I
mean
it's
not
really
like
about
the
importance
of
that
issue,
but
it
really
depends
on
how
we
implement
the
modeling
of
these
aspects
and
if
we,
basically,
this
issue
could
be
pretty
much
like
the
the
first
iteration
be
done
for
free.
If
we,
if
we
chose
a
different
way
of
modeling
this
aspect,
so
I
I,
don't
think
like
that.
My
intent
is
to
solve
this
other
issue,
but
I
want
to
like
to
keep
this
issue
from
the
back
of
the
head
to
think
exactly.
A
If
you
would
have
to
implement
this
issue,
how
we
would
do
that,
because,
maybe
by
doing
very
small
change
through
like
solving
these
aspects,
we
can
basically
get
this
issue
in
the
initial
form
for
free,
but
it
really
depends
on
how
we
want
to
model
and
how
you
want
to
address
these
aspects,
which
are
like
the
main,
the
main
like
the
main
goal,
not
here
because
like
I
I
had
my
perception,
how
it
could
be
done,
but
I
don't
want
to
spoil
it.
Yet,
let's
talk
about
like
the
ideas.
B
Cool
I,
the
only
idea
I
have
is
that
when
I
suggested
of
using
group
I
do
think
of
them
as
separate
I
was
thinking
a
path
forward,
be
immediately
adjust
in
our
marketing
plans,
the
amount
that
we
say
we're
gonna
give
to
a
thousand
and
ten
thousand
four
thousand
for
private
route
projects
and
ten
thousand
for
public
projects.
And
then
we
have
to
figure
out
whether
or
not
we're
to
figure
out
how
we
can
grandfather.
But
we
can
start
publishing
that,
so
anyone
who
signs
up
afterwards,
we
can
just
we
didn't.
B
We
inviolate
any
terms
of
their
contract,
but
we
can
say
we're
gonna.
Do
it
first
figure
out
how
we
can
enable
current
people
to
have
their
extra
ones
or
have
their
current
quota
and
then
adjust
our
quota
and
then
plug
the
holes
of
the
existing
system
when
it
comes
to
public
projects
and
when
doing
that,
make
sure
that
we
don't
mess
up
Forks
and
maybe
by
that
time,
we'll
have
this
group
runners
used
in
fork,
contexts
that
would
enable
us
individually
to
solve
for
it.
B
A
B
We
have
the
that
issue
that
you
mentioned.
That's
a
schedule
for
1210.
If
that
issue
ended
up
being
enabling
a
project
to
allow
the
project's
runners,
the
group
size.
If
that
enabled
the
group
to
allow
the
group's
runners
to
be
able
to
use
in
fork
contexts,
then
we
would
kind
of
solve
the
question
mark
of
apply
like
I,
get
the
sense
that
we
can't
apply
the
public
project
quota
today,
because
it
would
ruin
our
workflow.
B
So
we
can't
just
say:
oh
tomorrow,
we're
gonna
start
enforcing
10,000,
minute
quotas
or
50,000
limit
quotas
on
public
projects,
because
it
would
ruin
Siemens
and
those
other
customers
we
talked
about,
who
have
who
are
consuming
shared
one
hour
minutes
in
an
effort
to
contribute
together.
But
if
we
can
enable
get
loud
specifically
to
provide
access
to
those
minutes
in
that
by
having
it
be
the
group
runners
in
winning
four
contacts,
then
we
can
plug
the
whole
of
public
projects
having
kind
of
unlimited
quota
today.
So.
A
Iii
think,
like
my
sentiment
behind
that
topic,
is
that
it's
calm,
it's
very
hard
like
to
do
it
right.
So
I
would
not
based
our
approach
on
this
issue
to
be
sorting
that,
given
my
snagger
and
also
like
being
roll
out
because
I'm,
if,
like
I,
think
we
could
be
disappointed
that
it's
not
going
to
be
done
all
right.
It's
not
gonna
be
working
properly
as
the
way
as
we
expect
so
I
was
rather
like
make
us
think
about
the
concept
that
allow
us
to
model
these
aspects
without
dependence
on
of
debt.
A
By
like
acknowledging,
that
could
be
like
this
more
like
a
wiki,
how
that's
gonna
be
solved
by
by
people
shaping
that
feature
to
which
we
migrate
over
time.
But
it's
this
significantly
better.
Like
then,
it's
today
but
like
we
are
not
dependent
on
this
being
shipped
or
not
suit,
because,
like
now,
if
we
are
dependent
on
this
being
shipped
or
not
shipped,
if
this
is
not
shipped,
basically,
we
cannot
start
enforcing
anything
which
is
basically
quite
terrible
and
I.
A
I'm
actually
started
thinking
like
how
I
could
solve,
like
the
distinction
between
existing
and
the
new
users
and
I
think
that
this
is
pretty
simple
on
to
solve
because
like
if
you
would
prevail,
all
the
existing
type
ports
where
I
quit
early
meet.
As
of
now.
This
would
be
like
that
we
meet
that
system
would
start
using
for
the
private
projects.
Now,
if
we
change
customers
and
application
settings
to
a
new
image.
B
A
B
Thought
you
were
saying
that
application
settings,
so
if
a
new
person
signs
up
I
get
there,
they
never
touch
customers.
They
never
get
an
explicit
limit.
If
I
send
up
today
and
then
tomorrow
we
change
it.
They're
gonna
get
the
new
application
settings
on
it,
not
one
that
was
said
today
right.
This
is.