►
From YouTube: Agones Community Meeting July 2022
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).
B
Boom
right
also
do
recording,
I
guess
I'll
I'll
start
welcome
to
the
agonist
community
meeting
for
july
28
2022..
Let's
maybe
start
at
the
top
of
the
agenda.
I
threw
a
bunch
of
stuff
in
here,
not
a
bunch,
a
couple
things
in
here
last
night
for
us
to
to
cover
and
sort
of
put
the
word
out
about,
and
I
don't
have
the
doc
up.
But
if
I
recall
correctly,
the
first
thing
was
the
release
candidate
that
got
published
yesterday.
B
Yeah
we
had
stuff
moving
up
the
the
feature
list,
so
one
of
the
things
I
often
put
on
the
agenda
for
this
meeting
is:
where
are
we
at
on
features?
Should
we
be
promoting
features
and
with
this
release,
we've
got
two
features
that
did
get
bumped
up,
so
we've
got
one
that
moved
to
stable
and
one
that
moved
to
beta,
which
means
it's
on
by
default,
which
is
great.
A
We
also
had
fixes
for
the
alpha
feature
of
for
handling
a
graceful,
graceful
shutdown,
so
basically
so
that
there
was
a
there's,
a
feature
in
there
that
if
the
sdk
stays
alive
until
it
moves
to
a
shutdown
state
on
on
when
the
what's
going
to
call
it,
when
a
termination
request
comes
through
to
the
processes
such
that,
if
you're
using
some
kind
of
preemptable
or
spot
vm
or
something
like
that
and
the
node
gets
removed,
you
can
still
do
stuff
to
the
sdk
until
such
time
that
it
moves
to
shut
down.
B
A
Are
they
all
broken
yeah?
This
all
says
1.125.0.125.0.
A
B
B
So
the
next
thing
I
put
on
the
agenda
was:
we've
talked
about
in
the
last
couple
of
releases
about
starting
to
add,
support
for
arm
and
arm
support
that
landed
in
the
prior
release
in
the
1024
release
has
been
tested
and
works
on
gke
it.
You
know
most
certainly
works
other
places
too,
but
gke
has
announced
support
for
arm,
and
we
know
that
agonist
works
with
that.
B
So
it's,
I
think
now
in
a
pretty
good
place,
although
it's
in
from
the
agonist
perspective,
it's
in
a
similar
state
as
we
have
with
windows
where
it's
been
sort
of
tested
manually,
but
we
don't
have
any
continuous
testing
for
it
yet
so
we
know
it
works
today
we
don't
know
if
it'll
work.
You
know
next
week
or
next
year
until
we
sort
of
start
beefing
up
our
tests
a
little
bit,
but
if
arm
support
is
something
that
you
are
interested
in
or
excited
about.
B
Please
let
us
know
so
that
we
can
know
how
or
whether
to
prioritize
getting
some
of
that
sort
of
continuous
support
set
up
so
that
we
will
know
that
it
keeps
working
going
forward.
B
Let's
see
so
the
next
thing
was
we
didn't
get
the
kubernetes
version
upgraded
as
part
of
this
release
candidate.
So
it's
at
125,
but
I
think
we
are.
You
know
I
think
we
could
have
technically
via
our
policy,
which
means
that
for
126
we'll
definitely
want
to
look
at
upgrading
our
version,
so
the
release
gets
cut
next
week.
B
A
B
All
right
so
next
thing
I
suck
your
name
on
this
one
mark,
which
was,
I
noticed
yesterday
that
you'd
gone
through
and
pinged
some
some
issues
asked
about
marking
my
stale
and
I
thought
maybe
it'd
be
like.
I
started
then
like
kind
of
going
through
myself,
and
I
thought
oh,
maybe
we
should
just
kind
of
do
this
together
during
the
community
meeting
and
say
like
let's
just
start
with,
like
the
very
very
oldest
issues
and
work,
our
way
forwards,
going
through
maybe
five
to
ten
per
meeting.
B
We
could
label
them
as
sort
of
triage,
so
we
know
how
far
we
got
last
time
if
there
are
things
that
we
leave
open,
but
I
think
if
we
start
with
the
oldest
ones,
we'll
find
a
lot
of
things
that
are,
you
know,
obsolete
or
haven't
been
touched
in
ages
or
that
maybe
can
be
sort
of
closed
out
to
reduce
the
backlog.
B
A
A
This
is
one
that
has
come
up
before,
so
this
is
game
server
is
an
allocated
state
can
respond
to
fleet
updating-
probably
maybe
not
maybe
edit,
that
title
here
a
couple
of
times
in
slack.
Basically,
the
scenario
is
say:
you
have
a
long
running
game
server
process
like
say
a
persistent
world
or
like
a
lobby
just
even
just
a
lobby's
fine
and
you
run
a
rolling
update
and
you
want
to.
A
Actually,
I
think,
there's
actually
two
scenarios
here,
because
I'm
thinking
about
this
tonight,
there's
there's,
I
think
two
actual
things
I
think
we
can
do
one
is
you
may
want
to
stop
out
allocating
or
reallocating
to
those
persistent
lobby
servers,
because
they're,
an
old
version
or
and
or
you
want
to
tell
those
ones-
hey
like
tell
all
your
players
that
they
need
to
get
out
in
like
10
minutes
or
or
something
like
that.
They
need
to
be
like
aware
somehow
that
a
rolling
update
has
occurred.
B
So
basically,
basically,
we
want
to
allow
people
to
drain
right.
So
I
think
yeah,
especially
in
in
some
of
our
scenarios,
with
high
density
and
with
reusing
game
servers-
and
this
is
the
game
server
by
default-
will
just
kind
of
keep
keep
going
right.
It
won't
know
that
you
definitely
won't
know
exit
right,
and
this
would
basically
be
one
way
to
signal
the
game.
Server,
like
hey,
there's,
there's
a
new
version.
You
should.
B
You
need
to
do
so
that
you
can
exit.
So
if
it
typically
gets
reused,
you
know
five
or
ten
times.
Maybe
it
stops
going
back
into
the
ready
state
and
just
exit
at
the
end
of
the
next
one
or
or
it
applies
some
like
a
little
bit
of
jitter
if
you
have
a
lot
of
them
so
that
you
don't
yeah
everybody
exit
at
the
same
time
and
then
like
for,
like
the
persistent
one,
you
can
do
something
similar
where
you
like,
you
said,
stop
allocating.
B
So
you
stop
reusing
in
like
the
high
density
scenario,
because
if
you
keep
reallocating
it,
it
might
last
forever
and
that
sort
of
will
effectively
start
draining
it.
And
then,
if
the
game
server
wants
to
watch,
you
know
the
label
annotation
or
wherever
we
put
on
it.
It
can
then
also
start
draining
itself
right.
It
can
move
existing
players
off
if
it
wants
to
so
that
it
can
cleanly
shut
down.
So
I
think,
having
a
signal
that
the
game
server
knows
that
it
is
now
out
of
date
makes
a
lot
of
sense.
A
Yeah,
so
the
idea
that
I
actually
had
so
the
other
thought
I
had
as
well
is
like
if,
if
somebody
doesn't
want
this,
there's
probably
no
reason
to
go
and
apply
labels
or
annotations
or
something
to
all
the
game
servers,
because
it's
just
extra
load
on
on
ncd
and
like
that's,
not
necessarily
a
good
thing.
So
my
actual
thought
here
was
to
add
a
section
to
the
fleets
back
where
you
could
have
like.
I'm
not
sure
what
to
call
it.
A
We'll
call
it
action
for
lack
of
a
better
term
where
you
could
specify
what
label
and
or
annotation
you
would
want
to
apply
to
your
allocated
game
server.
When
a
rolling
update
occurs-
and
I
think
I
was
like
because
by
default
does
nothing
like
all
the
allocated
ones
are
left
alone,
nothing
happens,
but
that
gives
you
the
flexibility
to
be
able
to
be
like.
Oh,
apply
this
label,
which
will
drop
it
out
of
say,
like
the
allocation
pool
or
override
a
label
that
you
would
normally
use
and
like
turns
it
off.
A
B
Because
if
we
ask
people
to
put
stuff
in
their
gamble
and
that's
like
extra
config,
you
have
to
put
in
every
single
time
you're
creating
a
new
fleet
and
you're
applying
a
fleet
update
versus
if
it's
something
the
system
just
does,
then
it's
something
that
maybe
you
can
rely
on
so
yeah?
I
guess
the
question
is
like:
is
it
something
everybody's
going
to
want
and
the
system
should
just
do
it
by
default
or
is
it
something
that
applies
only
to
a
subset?
A
Yeah,
I
think
if
you
have
a
session
based
game
that
is
like
one
session
per
container.
This
is
something
you
probably
don't
care
about
or
like.
If
your
timeout
is
like,
if
you're
running
like
you
run
five
sessions,
then
you
delete
or
something
like
that.
You,
you
probably
don't
care,
I
mean,
there's
a
minor.
There
might
be
a
minor
optimization
there,
but
you
probably
don't
care.
B
Short
right
like
if
you're
three
minutes
each
and
you
run
ten
of
them
before
restarting
you
know,
maybe
you
do
care,
because
that's
a
long
enough
period
of
time
that
you
want
to
recycle
that
container.
If
the
sessions
are,
you
know
a
minute
each
and
you
run
five
of
them
like
five
minutes.
This
may
be
fine
to
just
let
it
go
right.
A
One
is
one
it
solves
the
it
makes
it
really
easy
if
you
want
to
like
in
a
reallocation
type
scenario,
if
you
want
to
be
able
to
just
turn
off
that
allocation
without
having
to
conform
to
whatever
we
provide
so
having
that
configuration
option,
and
then
it
also
allows
for
if
at
any
point
in
time,
we
want
to
add
extra
actions
to
some
other
feature
that
we've
added
in
the
future
to
be
like.
A
Oh
because
there's
the
second
one
that
I
want
to
talk
about,
which
is
like
waiting
as
well
like
if
you
wanted
to.
If,
if
we
had
a
waiting
situation,
we
could
add
to
that
set
of
being
like,
oh,
you
could
set
a
label,
you
could
set
an
annotation,
you
could
set
your
weight
to
zero.
You
could
do
whatever
makes
sense
for
your
game,
and
so
I'm
sort
of
leaning
more
towards
the
flexibility
side
than
necessarily
the
like.
Let's
do
it
for
you
side.
B
Right
I
mean
it
also
allows
you
to
say
like
on
this
update.
I
want
to
set
a
label
to
take
it
out
of
allocations,
but
on
the
next
update,
maybe
I
don't
like,
maybe
you
don't?
Sometimes
you
care,
and
sometimes
you
don't
the
system.
Does
it
every
time,
then
all
of
a
sudden
you're
having
to
change
like
your
games
for
allocation
code
and
your
matchmaker
at
the
same
time,
you're
doing
a
rollout
like
just
to
kind
of
you
know,
make
the
behavior
change
one
time
right.
B
So
I
think
yeah
sort
of
decoupling
those
where,
if
you
always
put
something
in
your
game,
server
allocation,
you
don't
always
have
to
apply
that
label
on
update,
but
you
can,
if
you
don't
want
to
reallocate.
A
Okay,
so
it
sounds
like
that
vaguely
makes
like
it
does
sort
of
resonate
I'll
write
up
I'll,
write
up
some
sort
of
design
at
some
point
and
then
we'll
we'll
argue
over
names
for
a
while,
exactly
that
is
that
is
the
and
as
such
as
it
should
be.
A
Yeah,
so
this
came
up.
This
came
up
a
while
ago
and
I
kind
of
went
nah.
This
was
before
we
had
before.
We
had
high
density
and
it
was
kind
of
a
bit
of
a
if
I,
if
I
like,
it,
was
a
bit
of
a
fudge
to
be
like
how
do
I
get
dynamic
resource
allocation
without
actually
having
dynamic
resource
allocation,
yet
inside
kubernetes
was
kind
of
where
it
was
going
and
eventually,
basically,
what
I
suggested
down.
A
The
bottom
is
like
you
could
set
like
three
levels
of
labels
and
then
use
the
preferred
selectors
or
now
we
have
just
selectors
to
be
able
to
be
like
okay,
high
importance,
medium
points,
low
importance
and
like
you
could
set
it
appropriately,
and
you
could
still
do
that.
But
basically
I
was
like
okay.
Let
me
think
of
this.
A
Like
the
big
scenario,
I
think
of
is,
if
you
have
high
density
game
servers
and
maybe
you're
doing
it
such
that,
which
I
wouldn't
be
surprised
if
you're
like
one
one
game
server
per
node,
which
would
actually
make
a
lot
of
sense
in
many
ways,
how
do
you,
with
the
current
system,
ensure
that
you're
packed
because
the
allocator
is
gonna
just
be
like?
Oh
there's,
one
game
server
on
each
node.
I
have
no
idea
unless
you're
using
player
tracking,
which
you
might
be,
but
you
might
not
be.
I
have
no
idea.
A
What's
got
the
most
amount
of
load
on
it
now
you
could
use
some
labels,
but
that
doesn't
tie
into
like
fleet
down
scaling
or
anything
like
that,
and
so
I
was
like.
Maybe
we
actually
should
revisit
this
now
that
we
have
high
density
stuff
because
it
basically,
my
other
thought
here
is-
is
there
are
requests
we
have
for
like
rooms
like
count?
A
How
many
rooms
I
have
and
like
how
many
game
sessions
I'd
like
to
be
as
a
little
bit
more
flexible,
I'm
trying
to
think
of
good
ways
to
do
that,
so
that.
B
A
I
mean
we
could
also
do
something
like
weighted
could
be
similar,
as
you
might
want
to
like
change
your
weight
as,
like
you
add
a
room
or
something
like
that,
but
I
just
I'm
like,
even
if
we're
doing
high
density,
even
within
a
like,
if
you
had
a
node
that
had
three
allocated
game
servers
and
another
node
that
has
three
allocated
game
servers
on
it,
agonize
has
no
idea
which
of
those
game
servers,
has
more
people
on
it
or
doing
more
important
things,
whatever
that
important
quote-unquote
thing
is,
and
I've
also
had
conversations
with
people
who
are
like
we
want
to
reallocate,
but
we
want
to
reallocate
to
like
the
game
server
that
has
the
most
interesting
thing
happening
to
it.
A
Right,
like
I
don't
know,
there's
like
this
big
battle
or
like
we're
tracking
events
inside
it,
and
I
was
like
yeah
well
then,
you
maybe
just
add,
like
an
interesting
label
to
it
and
then
like
you,
can
send
it
to
that
way.
But
that
gives
you
no
there's
no
granularity
into
it.
It's
either
interesting
or
it's
not,
and
so
I
was
like.
Basically,
I
was
like
I
wanted
to
bring
it
up
and
be
like.
Does
this
make
sense,
I'm
happy
to
write
this
up.
A
I
think
it
affects
a
lot
of
interesting
things
as
well
like
it's
not
just
allocation
but
like
we
should
also
be
getting
rid
of
the
stuff
that
is
the
lower
weight
like
when
we
scale
down
like
we
should
like
go
that
way,
so
it
hits
a
lot
of
marks,
but
I
was
going
back
and
forth
in
my
mind
of
like
do
we
need
to
do
this
or
does
it?
A
B
Right,
I
think
I
think,
with
both
with
re-allocation
and
with
high
density,
because
before,
if
like
everything
was
equal
and
we
assumed
that
everything
sort
of
ran
once
and
exited,
you
know
this
very
homogenous
set
of
things,
and
so
like
adds
a
person
puts
in
the
solution
like
weighting
them
all
equally
and
saying
they're
all
homogeneous.
They
all
have
equal
weights.
B
So
then
you
can
sort
based
on
you
know,
which
node
has
the
most
allocated
ones
to
figure
out
how
to
pack,
but
all
of
a
sudden,
I
think,
with
with
things
being
reused,
it
becomes
easy
to
end
up
with
one
thing:
that's
recycling
itself
on
a
node
holding
that
node
from
being
auto
scaled
down
or
high
density.
Like
you
said,
it's
easy
to
put,
you
know
five
players
on
every
game
server
when
they
can
hold
100,
because
you
don't
realize
necessarily
like
I
should
be.
I
should
be
packing.
B
You
know
over
here
right,
so
it
seems
like
it
fits
in
with
the
the
sort
of
meta
sort
of
project
goal
of
packing
and
density
and
cost
reduction,
to
have
a
better
way
to
do
that.
Now
that
we
have
more
sophisticated
features
right
and
that's
basically
sort
of
what
was
described
in
this
issue
was
like
the
defaults,
if
you're
using
it
sort
of
in
a
basic
way,
is
everything
has
equal
weights
yeah,
but
when
everything
isn't
actually
equal
anymore
like
how
do
we
express
that?
B
So
I
think
weight
was
sort
of
the
initial
proposal
of
how
to
express
these
things.
Aren't
equal
anymore
and,
and
maybe
wait-
is
sort
of
the
right
way
to
express
that
or
maybe
there's
something
else
that
we
can
use
to
express
that.
But
basically
I
think
that
the
point
is
valid
about
re
reordering,
this
sort
order
to
to
get
more
optimal
packing.
Like
that's
really
the
goal
here
right.
A
There's,
I
think,
there's
yeah
no
absolutely
100
degree
and
what
was
I
going
to
say.
A
B
Also
sort
of
degrades
gracefully,
if
you
will,
in
the
sense
that
if
you
don't
set
this
everything
works
as
it
does
today
right
exactly.
This
is
a
an
extra
level
of
influence
you
can
have
over
the
the
packing.
If
it's
something
that
applies
to
you,
but
if
you
don't
care
about
it,
everything
works
as
it
used
to
yeah.
There's
a
really
nice
way
to
have
a
feature
where
it.
It
doesn't
impose
any
cognitive
load
on
people
who
don't
care
about
it.
But
if
you
do
care
about
it,
it
gives
you
more
control
right.
A
Yeah,
because
I
feel
like
I'm
also
in
slack
and
outside
I've,
had
some
conversations
where
people
are
doing
multiple,
like
multiple
sessions
per
game,
server
instance
as
well,
and
a
lot
of
the
time
they're
like.
A
Could
you
just
count
rooms,
but
I
feel
like
much
like
we
do
with
player
tracking
essentially,
but
I
feel
like
that
is
a
very
specific
answer
to
a
problem
that
I
think
we
could
generalize
better
through
some
of
these
tools
so
that
we
can
support
a
wider
variety
of
more
flexible
types
of
orchestration,
basically,
rather
than
just
being
like
okay,
we
have
like
a
thing
because
we
yeah
the
other
and
we
can
discuss
this
in
design.
The
other
thing
that
makes
this
fun
is
like
in
a
distributed
environment.
What
does
weight
do?
A
B
A
A
Maybe
the
answer
there
in
terms
of
like
that
scenario,
where,
like
I
want
to
allocate
to
like
the
most
interesting
game
server
or
I
want
to
allocate
so
that,
like
you,
might
want
to
keep
players
in
a
group
up
into
a
point
like
you
might
be
like.
I
want
to
make
sure
there's
at
least
10
people
on
this
server
or
like
we
can
do
some
of
that
with
player
tracking
to
a
degree
but
like
weights
makes
that
a
little
bit
easier.
A
B
B
A
Do
that,
how
do
we
want
to
sort
I
like
to
sort
by
recently
updated?
Because
if
you've
put
something
on
it,
then
it
shows
up
at
the
top,
so
you
can
see
that's
the
way
I
like
to
do
it.
A
As
you
can
see,
then
you
can
be
like
these
are
the
ones
that
were
most
recently
touched.
A
Dynamic
yeah
or
actually
is
there
anything
here
that
jumps
out
to
you
as
something
that
could
potentially
be
stale?
It's
probably
the
easier
way
of
doing
it.
B
A
B
It's
at
the
end,
now,
okay,
so
just
let's
walk
through
from
the
bottom
up,
so
support
for
dynamic
port
ranges.
A
B
Okay,
so,
let's
march
this
as
help
wanted,
then
so
I
think
yeah
this,
it's
basically
saying
like.
Instead
of
giving
me
a
port,
I
need
10
ports
and
I.
B
B
Let
me
I
mean
it's
because
it's
a
little
bit,
I
mean
I
guess
I
don't
know
why
you
need
a
continuous
range
of
ports.
I
mean
if
you
needed,
10
ports.
You
could
always
put
10
ports
in
your
spec.
They
might
not
be
contiguous
ports,
but
you
would
get
10
at
that
point.
A
And
plus
three
yeah:
we
have
a
couple
of
thumbs
up
on
it
I'll
check
it
helpful.
B
A
B
A
B
Right
so
the
next
one
I
filed,
which
is
telling
me
the
process
to
push
container
images.
I
think
this
was.
I
can't
remember
if
I
filed
this
for
security
concerns
like
around
how
many
people
have
access
to
blob
or
things
or
if
this
was
more
about
like
we
should
define
when
we
actually
update
our
examples,
because
those
are
two
separate
issues
right.
B
But
I
I
think
you
have
the
the
newer
issue
about
artifact
registry,
which
allows
us
to
separate
things
into
different
buckets
and
have
a
larger
number
of
people
be
able
to
push
examples
than
maybe
you
can
push
releases
and
so
forth.
So
I
think
maybe
we
should
just
go
ahead
and
close
this
and
link
to
the
other
one
and
make
sure
that
as
we're
migrating
images
over
that
we
I
mean,
there's
not
really
even
much
here
to
take
into
account.
B
I
I
don't
think
I
think
this
probably
can
be
marked
as
obsolete
based
on
our
future.
You
know,
thought
slash
plan
to
move
more
over,
because
I
know
you
moved
the
first
step
over
already.
A
Yeah,
the
other,
the
other
thought
I
have
on
this
is
it's
probably
duplicate
to
the
work
that
I'm
not
sure
if
we
have
a
ticket
for
of
moving
a
lot
of
the
release
stuff
into
cloud
build
once
we
have
that
as
well.
All
right,
let
me
find
the.
A
Oh
here
we
go
all
right.
So
let's
say
this
is
duplicated
by.
A
A
B
That's
true
all
right:
this
is
the
reason
I
am
out
here
with
tracing.
A
B
No,
there
are
no
updates
to
this
one.
I
mean
the
other
thing.
It's
this
one,
this
one's
also
particularly
old.
I
think
a
lot
of
stuff
has
has
changed
since
then.
You
know,
I
think
open
tracing
and
open
census
have
now
actually
sort
of
come
together
and
open
telemetry.
That
was
like
a
future
thing
when
this
was
was
created,
so
I
think
now,
I
think
we
could
sort
of
go
with
open,
telemetry
and
and
feel
a
lot
more
future-proof.
A
A
A
A
A
Yeah,
there's
a
there's,
some
interesting
questions
about
open,
telemetry
versus
open
census,
but
we'll
worry
about
that
when
somebody
actually
makes
the
decisions,
because
they're
they're
still
it's
still
in
this
like
the
tracing
stuff,
is
stable,
but
the
metric
stuff
doesn't
seem
to
be
yet
and
it's
taking
a
long
time,
and
it
may
just
make
sense
to
just
keep
open
senses
open
census
for
now,
because
it
works
but
either
way
having
something
is
good,
metrics
players.
We
should
absolutely
do
this
like
at
some
point.
A
A
When
installing
it's
done
is
using
the
yellow
file,
helm,
tls
certs
that
are
used
by
service
as
a
default
to
detect
inserts.
That
is
true.
Ladies
may
inadvertently.
Use
the
solution
without
replacing
the
certificate
and
expose
that's
bad
place
before
inserting
play.
Certificates
generally
should
not
offer.
B
That's
definitely
a
good
point.
I
think
we
definitely
encourage
people
to
use
helm
in
production.
For
other
reasons,
I
think
a
lot
of
people
end
up
using
helm
for
configurability.
You
know
yaml,
I
think,
is
sort
of
the
quick
getting
started
path
where
you
can
just
apply
something
and
be
up
and
running
for
you
writing
and
testing
stuff.
A
Out
I,
I
would
be
curious
to
know
the
opinions
of
people
who
have
more
experience
on
security
than
I
do,
which
is
unlike
it's
an
intern
like
the
certs
are
used
inside
the
cluster
and
how
much
that
matters?
Is
it
more
important
that
the
certs
are
different
or
is
it
just
like?
Okay,
that
there
are
certs?
But
yes,
that
is,
that
is
not
unreasonable,
but
I
don't
know
how
you
can
do
a
yaml
installation
without
having
a
static
set
of
certs.
You
don't
really
have
a
choice.
B
A
You
know
we
could
do
is
actually
have
helm
generate
the
starts
rather
than
the
file.
So
every
time
you
run
a
release,
the
cert
will
get
rotated,
because
every
time
you
generate
an
index
it'll
just
create
a
whole
new
cert,
which
is
basically
how
home
works.
Now,
because
when
it
generates
right
with
it
does
gent.
A
Yeah
as
soon
as
you
run,
gen
like
if
yeah,
if
you,
if
you
make
a
change
to
any
of
the
yaml
files
like,
then
you
will
get
a
new,
a
new
thing
right.
To
make
a
note
about
that
yeah,
I
thought
that.
A
Or
whatever
it
is
from
static
files.
B
Our
a
lot
of
10
minutes
of
triage-
I
think
this
is
this-
has
been
useful.
I
think
if
we
just
spend
a
few
minutes
every
meeting,
we
can
make
it
longer
or
shorter,
but
I
think
if
we
keep
going
sort
of
grooming
the
backlog
this
way,
I
think
that
will
both
remind
us
of
things
like
we
saw
a
couple
things
here,
we're
like.
Oh,
yes,
like
there's
duplicate
issues,
or
this
is
something
that
came
up
recently
in
different
contexts
like
maybe
that's
something
we
should
work
on
or
oh.
B
B
So
we
can
go
back
to
the
agenda
and
we
had
one
new
person
join.
Is
there
anything
you
wanted
to
add
to
the
agenda
any
questions
you
have
for
us.
A
A
A
I
don't
think
I
have
anything
else,
I'm
just
testing
for
that.
Typo.
A
I
really
want
to
get
that
fleet
specific,
prometheus
metric
stuff
done.
I've
been
talking
about
for
like
two
months.
A
There
was
actually
there
was
one
other
issue
that
I
really
wanted
to
get
done.
That,
I
think,
is
a
good
optimization
where,
if
I
can
find
it
now
that
we're
here.
A
Yeah,
which
is
this
this
is
the
thing
I
want
to
fix,
and
you
saw
this
the
other
day
robbie,
that
if
you
scale
down
a
fleet-
and
there
are
game
servers
that
are
not
ready
yet
they
get
left
behind
rather
than
get
it
deleted.
First,
I
don't
know
why,
but
that
seems
that
seems
silly.
That's
been
there
for
a
while.
B
Yeah
that'd
be
great
and
there
were
a
couple
other
things
we
mentioned
during
the
call
of
health
wanted
things.
That
would
be
awesome
if
anybody
wants
to
pick
up-
and
I
think
the
other
thing
that
people
would
love
for
you
to
work
on
mark
is
the
player
tracking
stuff
yeah
yeah,
which
I
know
it's
a
big
big
bucket
of
work.