►
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
Let's
start
the
kobe
builder
and
control
one
time
meeting
the
first
eating
in
the
agenda
usually
is
new
faces.
That
is
more.
If
he's
someone
that
is
new
here,
and
would
you
like
to
introduce
yourself
and
say
why
you
are
here
or
what
do
you
expected?
Please
feel
free.
B
B
A
A
No
you
you
can
go,
you
can
go
sorry,
no!
No!
It's
fine
welcome
yeah,
and
that
is
amazing
that
you'd
like
to
contribute
to,
with
the
kobe
builder
your
help
and
your
support
is
definitely
more
than
outcome.
A
We
have
assignments,
so
I
think
we
can
move
on
right.
So
let's
go
for
the
first
item
from
the
common
interest.
That
is
the
demo
tony.
Would
you
like
to
share
your
screen.
C
Awesome
so
we
just
get
started.
A
C
Thank
you
so
hi
everyone,
I'm
tony
and
I'm
a
mentee
under
the
guidance
from
camilla
for
sure
and
rashmi
for
the
jsoc
project
this
summer,
so
we've
been
working
on
creating
a
new
grafana
plugin
that
has
helped
to
simplify
visualization
the
status
of
the
operator
through
the
grafana
web
ui.
C
So
if
you're
not
very
familiar
with
profana
plugin,
it
is
it's
just
a
newly
created
one,
as
I
just
mentioned,
and
basically
it
will
help
to
simplify
the
generation
of
the
grafana
magnify
fast
based
on
the
default
metrics
generated
by
controller
runtime.
So
that's
the
user
can
directly
copy
the
content
of
the
manifest
to
their
grafana
web
ui
to
visualize
to
observe
their
operator
status.
C
So
I
recommend,
if
you
have
interest,
I
recommend
you
to
go
through
the
documentation
and
and
I'll
be
glad
if
you
can
have
a
try
on
that,
and
let
me
know
if
anything
you
have
questions
or
issues
or
any
idea.
You
want
to
bring
to
this
plugin
very
I'll,
be
very
glad
so
currently
for
the
current
release
of
the
grafana
plugin
we
have,
it
will
bring
you
two
dashboards
that
by
default
will
give
you
the
metrics
based
on
the
control
runtime.
C
So
one
dashboard
is
the
controller
resource
matrix
that
will
show
you
the
basic
cpu
and
memory
usage
of
your
operator.
So
in
the
future
we
probably
will
bring
new
resource
based
on
different
perspective
if
it
is
necessary
and
another
panel.
Another
dashboard
is
the
runtime
metrics,
so
this
basically
will
show
you
the
relative
status
of
the
reconstitution
and
the
word
queue.
C
So
it
will
give
you
the
performance
of
your
rack
consultation
and
the
status
of
the
work
queue,
so
that's
like,
or
what
previously
the
ground
plugin
can
bring
you
so
upon
that
richard
and
me
have
been
thinking.
If
ever
we
can
try
to
visualize
the
custom
metrics.
So
that's
what
I'm
gonna
bring
for
today
so
for
custom.
Metrics.
C
We're
thinking
that,
like
since
operator,
can
be
various
kind
because,
like
like
different
users
or
different
companies,
may
have
different
kinds
of
operators,
and
it
can
be
a
common
case
that
the
operator
author
will
create
or
compose
some
custom
metrics
based
on
different
perspectives.
C
And
so
it
can
be
interesting
if
we
can
help
them
to
visualize
their
custom
metrics
through
the
grafana
web
ui.
C
So
actually
it
feels
like
a
challenging
because
I
just
mentioned
like
since
it
is
customized
so
like
what
the
meaning
of
the
metrics
and
like
what
kind
of
the
panels
they
want
to
visualize
can
be
varied,
and
it
is
like
a
little
bit
hard
like
for
the
plugin
to
judge
like
what
kind
of
the
panel
and
what
kind
of
the
expressions
they
want
to
see
so
we're
starting
like
the
very
basic
things
for
like
experimental
so
in,
and
so
I
can
show
like
some
latest
some
like
specification
in
the
latest
documents
of
the
master
branch
so
again
like
the
grafana
plugin
can
be
triggered
by
the
coupe
subcommand
init
or
like
the
subcommands
added,
and
then
this
time
it
will
also
generate
a
config.yml
file.
C
That
is
an
entry
for
the
users
to
define
the
metrics
and
they
want
to
view
so
inside
the
config
gmo.
It
will
give
an
object,
called
custom
metrics.
That
is
basically
an
array
and
for
each
element
it
will
help
the
user
to
define
the
metric,
the
type
and
the
expr.
So
the
metric
is
just
the
raw
custom
metric,
the
user
defined.
The
type
is
the
type
of
the
custom
metric.
C
That
is
the
prometheus
data
type,
so
you
say
it
can
be
counter
gouge
and
histogram,
and
the
expr
is
the
composed
prompt
ql
based
on
the
metric
circuits
like
custom
magic
is
like
this.
This
is
just
the
raw
metric
and
more
in
general
cases.
We
try
to
build
some
like
complicated
problem
care
based
on
that.
C
So
this
is
the
field
for
expr,
so
for
the
so
for
the
distribute
the
metric
and
the
type
is
required
so
that
the
plugin
will
know
like
what
exactly
the
metric
it
is
and
the
expr
is
optional,
so
that
if
the
user
would
be
glad
to
bring
the
xpr,
then
the
grafana
plug-in
will
just
directly
use
that
if
the
user
not
sure
like
what
the
xpl
would
be
look
like,
then
the
plug-in
will
generate
automatically
create
one
based
on
the
metric
and
the
type.
C
So
I'm
gonna
show
a
very
briefcase
how
we
can
go
over.
That.
C
So
here
I
will
use
the
memcache
operator,
so
this
is
like
a
very
basic
operator
and
the
operator
sdk
test
data,
but
it's
still
like
it
can
be
driven
by
code
builder,
so
I'm
gonna
use
the
computer
edit
sub
command
with
the
plugin.
C
C
So
inside
it's
I,
I
was
going
to
create
some
configs
for
the
custom
metric.
So
before
that
I
have
added
to
custom
matrix
based
on
the
memcache
operator,
one
is
it
has
a
histogram
which
is
the
memcache
operator
reconcile
time
seconds.
One
is
the
reconcile
total.
It
is
not
necessary
to
be
very
correct,
but
just
is
a
trial
for
like
the
custom
metric.
C
And
again
I'll
define
it
as
the
counter
and
I'm
not
necessary
to
use
the
expr.
So
let's
get
started
so
now.
I
would
also
then
I
will
again
trigger
with
the
coupe
edit.
So
this
time
since
I
have
defined
the
config
gmo,
the
plugin
will
read
that
config
and
generate
the
json
file.
C
So
search
for
the
keyword
expr,
we
will
see
that
it
will
automatically
generate
a
new
problem,
qr
expressions
based
on
the
given
metric
and
expression.
So
this
is
the
histogram.
It
will
generate
like
a
histogram
quantum,
to
show
like
some
statistical
results
based
on
the
custom
metrics,
and
so
here
it
will
do
some
summary
rates
based
on
the
reconcile
based
on
the
counter
metric.
C
So
basically
like
what
the
plug-in
do
is.
It
will
try
to
compose
a
very
basic
promoqueer
expression
based
on
the
type
of
the
metric
and
it
will
insert
the
custom
metric
into
the
into
the
expressions.
C
C
So
we
just
paste
it
in
the
input
page
and
let's
say
we
chose
a
folder
playground
and
let's
select
the
relative
data
source,
which
is
prometheus
and
click
import.
So
now
we
can
see,
there's
the
so
there's
the
the
panel.
So
if
we
choose
so
let
me,
let
me
choose
different
time
range
say
like
10
hours,
so
we'll
get
skip
this
one.
So
I
probably
have
a
wrong
expressions
on
the
default
metric,
so
let
me
try
to
fix
that
so
so
that's
the
reason
this
because
I
guess
I
didn't.
C
I
didn't
correctly
spell
the
custom
metrics
as
I
defined
here.
It
should
be
buckets.
I
guess
so.
Let
me
see
if
I
can
fix
that.
C
Oh,
it
seems
it's
correct
because
there's
the
histogram
I'm
not
sure
why
it's
not
showing
outside,
but
it
looks
we
can
view
what's
happening
over
here.
C
So
that's!
Basically
what
currently
the
custom,
how
like
the
profound
plugin,
will
bring
you
the
panels
for
the
user-defined
custom
metrics.
So
it's
very
like
it
is
still
like
an
experimental
feature
and
in
a
very
early
stage
because,
like
you
can
see
that
which
is
trying
to
like
build
upon
the
very
basic
problem,
q
expressions
to
help
define
the
panel
based
on
wasted,
custom
metrics,
and
we
just
deal
with
the
the
timeline
drawing
panel
as
as
what
is
displayed
over
here
and
so
be
very
glad.
C
If,
like
you,
would
be,
have
interest
on
trying
this
and
it'd
be
very
glad.
C
If
anyone
would
have
some
ideas,
how
it
could
be
better,
so
it
can
be
like
challenging
because,
like
at
one
point,
we
don't
want
to
leave
so
many
config
fields
for
the
user,
because
it'd
be
bring
some
burden
of
learning
curve
for
the
operator
author,
but
still
want
to
make
it
functional
because,
like
since
it
is
custom
matrix,
it
seems
like
there's
like
a
couple
of
things:
that's
really
defined
by
the
user
side,
not
the
plugin
itself,
so
yeah,
that's
what
we
have
for
the
latest
feature
of
the
grafana
plugin.
Thank
you.
A
That
is
a
great
really
nice
work,
tony
that
he
has
been
amazing.
So
I
think
the
next
step
is
asking.
If
he's
someone
to
like
to
make
questions
or
to
say
something
about.
D
No
questions,
but
it
looks
really
good.
I
know
folks.
I
know
that
a
lot
of
people
like
graphs
and
one
of
the
common
things
that
comes
up
is
what
should
operators
be
able
to
like
what
type
of
metric
should
they
be
putting
out,
and
I
think
this
will
help
folks
want
to
put
metrics
and
make
sure
that
they
are
exposing
them
and
be
able
to
see
them
and
see
the
value
in
it.
So
good
job.
C
Thank
you.
Thank
you,
so
I
guess,
like
maybe
one
challenging
things
that
happens
for
the
previous
days
I
witnessed
is
like
somebody
may
have
like
some
interest
on
using
this
plugin
and
they've
met
difficulty
on
trying
to
collect
the
metrics
from
the
operator
to
their
promises.
C
So
that's
definitely
another
story,
but
I
guess
like
if
someone
have
met
similar
issues,
so
it's
not
maybe
very
specific
to
graffana
can
also
reach
to
me,
because
I
can
probably
help
with
to
see
how
we
can
make
the
metrics
exposed
to
certain
promises,
so
that
can
help
the
users
to
see
the
metrics
on
their
profana
to
test
the
plugin.
A
That
was
about
the
permissions
right,
yes,
yeah,
because
when
we
install,
I
think
operator,
promotions
operator
or
it
could
be
prometheus
or
predictor.
I
think
that
is
the
number
of
the
project.
The
project
will
be
suggesting
the
document
for
someone
that
is
liking.
Stunning
is
creation.
Project
changes
trying
to
check
how
the
things
works.
A
If
you
install
by
default
only
the
the
full
name,
space
prometheus,
has
permission
to
check
the
metrics.
So
has
many
ways
I
think
we
now
we
have
in
the
documentation.
If
I'm
not
wrong
a
warning
to
let
people
know
as
well.
So
has
many
ways
for
you.
A
Let's
promise
you
change
the
configuration
from
the
promotions
and
say:
okay,
you
can
get
the
data
from
the
holder
cluster
or
you
can
get
the
data
from
these
specific
namespace
and
when
we
create
a
project
with
the
kobe
builder,
we
will
create
a
namespace
and
deploy
his
namespace,
that
is
system
and
the
name
of
the
project
with
the
name
of
the
project.
So
by
default
that
name
space.
A
If
you
just
installed
the
prometheus
via
ts
operator,
you
will
not
have
right,
but
this
information
is
also
in
the
getting
started
of
the
promoters
of
predictor.
How
you
can
allow
prometheus
operator
check
all
namespaces
or
a
few
of
the
and,
however,
just
to
bring
a
little
bit
of
context.
B
So
I
I
have
one
question
like
regarding
this.
I,
though
I
have
not
explored
the
grafana
even
alpha,
but
I'm
just
asking
like.
Is
it
all
the
like?
If
you
want
to
give
any
points,
I
mean
the
in
the
config.ml.
We
can
provide
all
the
pointers
like
which
dashboard
we
want
to
show
or
something
like
that
or
is
we
need
to
do
some
recording
as
well
other
than
writing
in
the
complete
moment.
B
B
A
That,
okay,
when
you
create
the
project
with
a
copy
builder,
you
have
some
metrics
that
are
exported
by
the
foo.
These
metrics
are
provided
by
control
runtime
and
actually,
if
you
create
a
manager
using
control,
runtime
you'll
have
these
metrics
right
in
the
discount
folder
of
the
kobe
builder,
in
discomfort
that
is
done
by
the
food
for
you.
A
You
see
that
she
we
exposed
the
metrics
in
a
entry
point
college
metrics,
the
graphene
apple,
the
graphene
plugin
will
only
scale
fold
the
manifest
for
you'll
be
able
to
create
the
graphs
in
graphene,
but
that
if
you
check
the
documentation
of
the
graphene
plugin,
you
see
these
steps
as
well.
You
need
to
ensure
that
you
have
parameters
configured
accordingly
right
for
be
able
to
to
to
to
use
the
metrics.
Have
you
all
permissions?
A
You
also
need
to
enable
thatching
the
diffuse
cathode.
The
metrics,
so
has
sunny
steps
that
you
are
in
the
document
like
how
you
enable
the
metrics
right,
because
this
top
fold
is
done
by
you,
but
you
need
your
neighbor.
You
need
to
say.
Okay,
I
want
using
this
option
from
the
project
right
here.
B
Yes,
yes,
thank
you
hi,
so
I
have
question
regarding
the
metrics.
So
when
you
say
custom
metrics,
what
does
that
exactly
include
like
other
than
the
native
metric
that
comes
with.
A
Let's
see
if
I
understood
your
question
when
using
custom
metrics,
what's
more,
you
can
see
okay,
the
custom
metrics,
the
you
have
all
these
metrics
that
you
are
exported
by
for
you
by
default
right.
You
don't
need
to
do
anything
if
you
are
using
controller
on
time.
You
have
some
metrics,
which
I
think
is
the
common
casey
scenarios.
A
Then
you
have
one
option
for
your
go
that
and
say:
okay,
I
want
to
create
my
own
metric,
which
is
the
custom
metric
which
you
call
like
a
custom
metric.
If
you
checking
the
ssdk
repository
now,
you
see
that
it
has
inside
under
the
test
data
main
cache
between
simple
you.
You
see
that
in
the
controller
and
the
outside,
you
see
that
it
has
a
new
directory
metrics.
I,
if
I'm
not
wrong
the
name
you're
seeing
example
of
the
custom
metrics.
A
So
in
that
case
and
just
to
show
how
that
can
be
done,
the
metric
is
done
to
count
how
many
times
the
reconciliation
was
made,
and
the
number
of
the
instances
that
you
say
I
wanted
three
instances
of
them
in
cash
right
here.
Put
that
in
the
custom
resource.
In
the
example-
and
today,
if
you
put
you
change
this
value,
you
go
inside
of
a
conditional
in
in
the
reconciliation,
so
the
metrically,
your
country,
how
many
times
they
value
the
desired.
A
The
number
of
instances
on
the
cluster
was
not
equals.
The
reality
right.
The
designing
state
is
not
equal
to
what
it
can
be
found
on
the
cluster.
C
A
A
And
to
watch
it
on
today,
you'll
be
like
if
you
create
a
custom
attic.
That
is
something
that
is
valid.
Only
for
your
specific
case,
then
you
can
go.
There,
use
the
plugin
configure
the
dutch
and
you
have
a
dashboard
to
visualize
the
data
export
franchise
custom
metric
using
graphene.
A
Also,
you
can
customize.
It
is
because
it
has
everything
that
is
done
by
the
tool
right.
So
the
tool
is
called
fold.
The
files
show
for
you.
How
to
do
is
a
gucci
start
point,
but
you
also
can
go
there
and
you
try
to
improve
and
to
change
things
as
pleasio.
C
I
I
think
most
of
the
thing
you
say
is
pretty
correct
and
just
imagine
like
if
so
it
may
be,
not
that
much
so
complicated
you
just
like
like
go
for
like
like
try
to
import
as
like
any
kind
of
metrics
you
would
like,
and
so,
if
you
have
already
configured
your
promises
and
then
you're
running
your
operator
correctly,
so
it
will
automatically
collect
your
metrics,
and
so
you
don't
need
you.
So
the
only
thing
you
need
to
do
is
you
just
define
that
and
then
you
can.
C
C
A
A
The
other
thing
that
I
open
here
to
show
is
the
example
that
she
was
adding
the
main
cache
operator
in
the
operator
sdk
app.
So
if
you
came
here,
cheese
director
is
not
scam,
followed
by
the
full
from
for
sgk
or
kobe.
Be
there.
However,
this
is
a
customization
done
on
top
to
show
how
the
custom
metrics
works.
So
if
you
go
here,
you
see
the
implementation
of
the
metric,
so
that
is
exactly
what
I
was
trying
to
share
to
share
the
total
number
of
times
the
deployment
size
is
not
the
same.
A
Is
here
so
here
we
check
it
right
is
the
number
the
fine
energy
in
the
spec
in
the
custom
resource
equals.
The
number
of
that
I
can
find
on
the
cluster.
When
I
I'm
checking
the
object,
if
not
right,
that
is
different,
then
we
are
calling
the
metric,
and
that
is
counter
plus
one
in
the
total.
It's
just
a
very
simple
example
just
to
show
how
the
chicken
works
right.
A
A
A
If
you
came
in
the
master
branch
easier,
okay,
if
you
came
in
the
master
branch,
plugins
graphene,
where
are
you
here?
Okay,
here
check
all
documentation
in
gg
with
the
steps
that
was
added
by
tone
about
his
new
featuring
to
work
with
the
custom
metrics
right
so
has
the
gif
is
as
well
and
aston
describe.
If
you
have
any
problem
face
any
issue,
or
would
you
like
to
make
suggestions
as
usual?
Please
feel
free
to
raise
an
initial
to
contribute
to
people
requires,
as
you
see
fit,.
B
Sometime
that
just
mentioned
that
that
matrix,
like
a
graph
of
a
reconciliation
group
loop,
what
other
things
that
graph
shows
like
in
reconciliation,
like
one
thing
like
bruce,
also
mentioned
that
failed
deployments
and
all
so
what
all
things
is
possible.
B
B
Can
we
do
this
sometime
back?
You
mentioned
that
the
graph
like
which
reconciliation
matrix
graph
obvious
it
was
for
the
complete
stuff,
or
only
for
the
single
reconciliation.
B
B
A
B
Hear
us,
okay,
I
was
just
trying
to
get
that
if
like,
if
is
there
somewhere
like
in
matrix,
we
can
analyze
our
reconciliation,
single,
recall,
reconciliation
thing,
but
all
stuff
has
been
fair,
like
the
frequency.
What
all
resources
has
failed
or
passed
like
that
way.
I
was
trying
to
get
so.
I
thought,
like
diani,
just
mentioned,
that
a
single
reconciliation
graph,
so
I
was
just
thinking
in
that
way.
A
I
I
think
he,
the
metrics,
has
more
than
only
the
reconciliation.
If
you
came
here
tony
help,
you
with
dutch
as
well
in
the
reference
you'll,
find
all
the
full
metrics.
A
I
don't
have,
in
my
mind,
all
user
cases
so
far,
but
these
ones,
for
example,
are
related
to
the
reconciliation
she's
one's
is
webby
hooks.
A
A
So
you
also
can
have
a
metrics
energy,
monitor
the
manager
or
the
operator
how
you
want
to
save
like
the
deposit
which
he
is
running,
the
binary
with
your
code,
the
ceo
manager,
all
controllers
as
well
right,
so
that
it's
another
point
that
you
can
use
in
the
metrics.
A
B
C
Yeah,
I
guess
some
additional
thing
I'd
like
to
say
is
so
I
think
that
for
the
default
metrics
given
by
runtime
they're,
like
very
relative
to
like
reconciliation
and
the
work
queue
and
the
relative
performance
on
like
the
web
book,
or
something
like
that,
but
like
as
you
can
see
that
we
so
the
graphing
plugin
also
brings
you
the
resource
dashboard.
C
That
is,
for
the
cpm
memory,
and
for
these
things
some
of
them
are
not
necessary
to
be
used
default
metrics,
and
so,
if
you're,
using
like,
like
other
like
metrics
systems
like,
for
example,
coupe
state
metrics,
you
can
definitely
like
observe
your
controller
in
a
different
perspective
that
is
outside
the
operator
and
so
like
for
the
custom
metrics.
C
If,
if
I'm
understanding
correct
it's
like
so
for
the
most
cases,
we're
trying
to
like
compose
how
the
reconcile
will
be
looks
like
so,
I
imagine
like
in
most
cases
when
you
define
a
custom
metrics,
you
will
be
mostly
working
around
the
rat
constellation,
but
that
reconciliation
is
mostly
customized
by
yourself.
C
I
didn't
quite
dive
deep
into
like
how
like
that
would
be,
looks
like,
but
just
basically
I
think
that
would
be
the
case.
So
camilla
you
can
fix
me
if
I'm
sims
no.
A
No
specifically,
I
I
think
you
spent
a
long
time
on
this,
probably
more
than
I
all
my
life
in
the
metrics,
so
yeah,
you
are
totally
fine
and
you
might
understand
his
touch
as
well.
You
have
a
data
metrics
that
you
can
get
from
kubernetes
api
directly
you'll
have
data
that
you
can
get
from
control,
runtime
that
are
more
specific
to
the
controllers,
and
in
this
way
the
reconciliation
right
ngo
also
can
create
your
own
things.
A
A
For
than
next,
I
think,
can
we
move
forward
or
please
feel
free
to
raise
any
other
questions.
A
The
silence
you'll
be
a
yes
right,
so
I
proposed
some
tops
here
as
well.
We
have
like
20
minutes,
so
I
don't
usually
speak
about
everything,
because
you
we
also
have
topics
from
another
person.
I
just
see.
Would
you
like
to
speak
with
me
about
it,
two
of
them
enjoying
my
movie
for
the
another
person
right
before
they
don't
take
over
of
all
meeting,
so
the
first
one,
I
think,
will
be
very
very
fast
if
he
we
check
here.
B
A
We
do
tests
against,
we
have
a
integration
test.
Engineering
stresses
actually
angel,
we
do
test
is
again
see
many
versions
of
the
kubernetes
by
default
the
project
she
supports
the
latest
three
ones
right,
because
it's
the
same
support
ability
that
e-commerce
provides
and
I
think,
to
reduce
the
the
force.
Do
you
know
they
run
in
parallel,
but
I
think
we
could.
I
think
I
don't
see
so
many
benefits
when
you
test
so
many
versions.
A
So
I
would
like
to
propose:
we
try
to
keep
only
the
list
three
and
also
it
will
help
us
clean
some
code,
because
today
we
have
some
tests
that
we
need
to
do
things
different,
because
in
one
day
2022
we
have
a
duplication
of
apis.
A
We
have
a
removal
of
apis
actually,
so
we
have
some
some
code
that
is
like
before
this
version.
We
do
something
after
this
person
do
other
thing
right
to
ensure
things
that
she
was
provided
before.
A
B
E
That
matches
with
what
upstream
wants
to
do
I'll,
say
in
case
it's
helpful
downstream
in
my
own
projects.
We
do
something
similar
where
it's
like.
We
run
the
latest
versions
for
pr's,
but
we
do
a
nightly
of
like
older
versions
just
to
keep
track
of
whether
they
continue
to
work.
So
maybe
you
want
to
do
that
just
in
case.
You
still
do
want
to
know
if
some
of
those
older
versions
are
working.
A
A
Yeah,
we
usually
don't
break
the
things
we
try
our
best
for
don't
worry
the
things.
Unless
is
something
out
of
your
control,
because
we
have
these
stable,
plugins
right,
so
the
plugins.
Actually
we
try.
We
have
the
policy
of
distancing
to
those
break
chains,
but
she,
if
he
we
have
like
a
deafening
that
she
breaks
that
king
forces
trying
to
move
forward,
and
you
mean
it
should
not
fight,
but
we
avoiding
in
the
max
break,
something
like,
for
example,
before
we
have
a
flag
that
is
clg
verizon
yokojin
forum.
A
If
you
want
better
one
right,
but
it
is
like
you
cannot
use
it.
If
the
list
is
very
soft,
it
could
be
builder
because
the
list
is
very
soon.
You
have
it
as
adapts
controller
on
time
and
controller
tools,
versions
that
it
doesn't
support
v1
beta
1
because
they
are
also
bumping.
Kubernetes,
lisa's
release
and
cheese.
Api
no
longer
exists
right.
So
in
this
case
the
flag
is
deprecated
and
you
ain't.
Actually,
you
can
work
with
this.
A
B
Because
I
was
gonna
say
I
noticed:
122
was
a
little
over
a
year
ago.
123
was
a
little
bit
less
than
a
year
ago.
I
think
it
might
be
nice
to
support
everything
that
was
like
around
a
year
so
like.
If
it
was
going
to
be
like
four
releases
a
year,
then
I'd
say
we
would
support
four.
But
if
it's
gonna
be
three,
we.
A
D
A
D
Either
either
either
you
keep
n
minus
three
so
that
even
though
kubernetes
supports
n
minus
two,
you
keep
one
extra
or
you
go
n
minus
two
and
you
always
clean
up
every
time.
You
add
one
well
either
way
the
cleanup,
I
think
you
will
still
have
to
do.
Even
if
you
went
with
n
minus
three,
you
still
to
get
rid
of
your
oldest
ones,
the
oldest
one.
So
I
think,
as
part
of
each
update,
one
of
the
things
we
should
include
is
getting
rid
of
the
oldest
test
in
the
process.
A
A
B
Yeah,
I
agree,
I
think
n1s2
is
good
and
actually
that
would
be
23
now
right,
because
25
is
out
so
one.
I
think.
Whenever
we
do
the
125
bump
for
in
front
stuff,
we
can.
We
can
remove
everything
before
23.
B
A
I
know
after
I
I
go
okay,
so
we
have
13
minutes.
I
don't
really
speak
about
the
cheese
ones
to
let
our
colleague
have
the
chance
enjoy
your
move,
which
is
ones
for
the
next
one
right
just
say,
for
a
happy
other
person
be
able
to
to
share
the
thoughts.
A
E
Yeah,
it's
sorry.
It's
been
a
while,
since
I've
been
it's
probably
been
a
couple
months
since
I've
actually
been
on
this,
I
used
to
come
to
it
more
frequently
so
we've
talked
before,
but
I
definitely
haven't
been.
I
haven't
come
back
in
a
while.
I
just
wanted
to
throw
a
question
out
there,
so
I
have
noticed
a
pattern
in
the
operators
and
controllers
and
stuff
that
I've
been
working
on
across
a
multiple
projects,
and
I've
noticed
this
pattern
out
in
the
wild
of
with
map
functions,
silently
throwing
away
errors.
E
So
if
you
click
on
the
example
that
I
put
in
there,
this
isn't
my
code
by
the
way
this
is
from
contour.
It
was
just
something
publicly
available
to
kind
of
illustrate
the
fact
that
this
happens
else
like
pretty
this
happens,
commonly
the
basic
idea
of
this
map
function.
Is
you
know
you
put
this
map
function
on
your
watch
and
it's
saying
I'm
watching
gateway,
class
custom
resources,
and
what
I
want
is
the
gateways
that
that
relates
to.
So
it's
for,
like
object,
relational
things
you
might
map
right.
E
That's
what
the
map
functions
are
there
for
the
problem,
is
you
see
it
on
line?
126,
basically,
is
that
it's
pretty
common
to
put
things
that
can
error
out
in
here,
and
so,
if
I'm
understanding
it
correctly
it's
fairly
minute,
but
in
bad
conditions
on
like
the
api,
server
and
stuff
like
that,
you
can
straight
up
just
lose
an
enqueued
object
like
you
would
never
in
queue
in
some
rare
conditions,
you
would
end
up
never
in
queuing,
so
I
guess
my
question
is:
there's
some
ways
like
to
work
around
this.
E
So
on
my
own
projects,
we
try
to
use
ownership
as
much
as
possible.
It's
not
always
like
this
example.
Ownership
would
not
be.
You
wouldn't
use
ownership
for
this,
so
you
can't
have
that
direct
link.
That
gets
you
back
to
the
thing
and
also
it's
you
know
the
the
client
is
cached.
So
in
theory,
this
is
a
really
rare
condition
when
you're
actually
going
to
have
the
error,
but
are
there
better
ways
to
do
this?
Would
we
be
open
to
having
map
functions?
That
part
of
the
signature
could
be
an
error?
E
D
B
D
E
B
B
D
Done
yeah,
but
that's
that's
not
his
his
question
that
part
he's
not
questioning
that
part.
He's
questioning
that.
Let's
say
there
are
five
gateways
out
there
when
I
go
to
list
them.
That
list
call
can't
find
them
because
there's
an
error,
so
it
will
log
the
error
and
then
it
returns
new.
So
then,
the
rest
of
the
machinery
says:
oh,
there's
no
gateways,
but
there's
five,
but
we
never
saw
them
because
for
some
reason
something
happened.
That
list
call
errored
for
a
variety
of
reasons.
E
E
So
I'm
looking
for
better
ways
to
do
this
because
I've
got
some.
I
have
some
use
cases
where
I
could
I'd
like
something
better.
It's
not
like
burning
me
right
now
with
like
actual
production
failures,
but
I
have
seen
problems
with
this
happen
and
we
have
some
mechanisms
right
now,
like
we
actually
have.
One
of
our
controllers
will
just
eventually
just
kind
of
re-reconcile
these
things
on
its
own,
like
just
double
check
them
and
that's
one
way
to
work
around
it,
but
it's
not
a
great
way.
E
So
if
there's,
if
there's
like
a
known
better
way
to
do
this,
ownership
is
the
obvious
one.
It
gets
a
little
bit
easier
to
do
it
with
like,
because
if
you
have
the
object,
then
you
can
just
look
at
its
ownership
and
put
in
a
reconcile
request
for
the
name.
You
have
to
do.
No
lookups,
you
won't
have
an
error,
but
that's
not
always
possible,
like
in
this
case.
E
A
Is
because
this
one
is
a
reconcile
request
right,
let's
see,
I
need
to
think
about
that.
That
is
the
most
appropriate.
But
yes,
I
would
suggest
you
anyway
create
an
issue
in
the
context.
Runtime
put
the
scenario
and
you
bring
that
for
discussion,
and
then
we
also
have
time
to
think
to
analyze
what
you're,
saying
and
you're
thinking
the
options
and
to
try
to
understand
why
you
are
doing,
which
is
this
person
is
doing
it
this
way
and
then
so
insert
that's
more
appropriate
unless
someone
would.
E
E
E
Kind
of
fills
in
all
the
blanks,
then
that
sounds
good.
I
just
want
to
double
check
on
the
call.
First
in
case
I
was
just
missing
something
obvious
like
there's
some
obvious
somebody
knew
something
so
yeah
I
could
be
done
with
this,
then
thank
you
for.
E
E
A
B
E
Yep
yep
sounds
good
I'll.
Do
that.
The
next
item
on
the
agenda
is
also
from
me
and
it
is
related
to
the
last
time
I
was
on
the
call
so
I've
been
trying
to.
I
want
to
do
this
one,
but
I
want
some
help,
so
I
I've
been
like
I've
been
after
this
one
for
like
more
than
a
year.
I
think,
and
it's
just
been
kind
of
slow
going,
but
I
have
so
many
coo
builder
and
now
operator,
sdk
code
repos.
E
That
kind
of
have
problems
related
to
this,
and
it
would
be
really
nice
to
get
it
fixed
and
I'm
happy
to
help
fix
it,
but
I
don't
really
know
where
to
start
so
like
three
months
ago
I
I
was
on
this
call
and
I
asked
I
want
to
say
it
was
jesus
jesus,
I'm
not
sure
exactly
the
pronunciation
that
you
want.
Okay,
I
think
we
talked
and
somebody
else
had
had
like
said
ping
us
on
this
thing
and
we'll
get
it
going
again,
for
you
ask
for
some
help
on
here.
E
I
did.
I
didn't
get
a
response,
so
I'm
back
again
to
say:
can
somebody
push
me
in
the
right
direction
here
I'm
willing
to
even
like
jump
on
a
zoom
and
have
somebody
just
kind
of
do
a
little
bit
of
walking
through
the
code
with
me
to
get
me
so
I
can.
I
would
be
happy
to
try
to
fix
this
thing.
I
just
don't
know
where
to
start
right
now.
D
A
A
A
E
One
is
getting
controllers
in
internal
because
we
also
are
used
were
also
used
as
a
go
library
in
some
of
these.
So
we
literally
don't
want
those
to
be
part
of
the
public
api.
We
want
them
to
be
under
internal,
that's
that
was
kind
of
one
of
the
bigger
ones
and
then
having
api
not
be
the
only
thing
outside
of
package.
That's
basically
part
of
the
package.
Pkg
would
also
be
nice
like
being
able
to
put
api
under
pkg.
D
So
camilla
open
that
look
in
the
pull
request
and
let's
see
which
one
is
yeah
that
one
so
in
this
particular
one
I
mean
it
doesn't
solve
it
the
same
way.
You
asked
shane
in
terms
of
putting
controller
into
internal,
but
what
this
does
is.
Apparently
it
allows
you
to
create
a
go
workspace
so
that
the
api
directory
can
have
its
own
go
mod
and
you
can
import
it
from
another
project,
but
not
import.
Your
entire
operator.
B
Well,
what
what
this
vr
does
is
it?
It
creates
a
multi-module,
go
project
kind
of
has
like
a
mono
repo,
so
essentially
how
this
pr
works
is
it
creates
the
go
mod
in
the
api
folder
and
that
will
in
turn
essentially
make
it
a
separate,
go
module,
but
still
within
the
same
repository.
B
So
this
doesn't
do.
I
think
exactly
what
shane's
looking
for.
I
think
shane's
wanting
me
for
like
following
good
go
practices
by
putting
that
in
the
package
holder,
and
then
I
guess
potentially
having
it
end
up
doing
like
an
internal
for
controllers
and
moving
the
controllers
into
an
internal
folder,
but
it
does
that
sound
right.
Shane.
E
Yeah,
what
we
actually
do
right
now,
at
least
for
one
poo
builder
project-
is
we
have
it
in
internal?
We
literally
put
it
in
there
and
we
move
it
out
for
when
we're
using
the
coupe
builder
cli,
there's
like
a
script
to
like
dance
it
back
out
to
where
it's
expecting
it
and
then
dance
it
back
in
it's
kind
of
janky.
A
He's
his
new
plugin,
I
think
he
solved
the
problem.
I
want
export.
Only
the
apis
right
here
is
one
option
to
solve
that
problem,
that
your
issue
for
change,
the
language,
the
discussion
in
the
past
we
thought
he
was
was
like
that
he
what
he
is
a
good
practicing
where
portuguese
things
is
more
like
something
a
bit
opinionated
right,
everybody.
You
have
some
preference
right.
So
we
have
ice
standards,
we
have
a
cheese
layout
and
everybody
who
wants
and
some
change
some
customization.
A
A
A
Also,
I
don't
know
we
can
think
about
your
other
options
as
well
like
it
has
people
that
say,
I
would
like
to
to
change
like
customize
the
path
of
the
directors.
Have
you
any
idea
how
we
could
achieve
that
like?
If
someone
has
some
idea
like,
would
you
like
to
propose
things
on
other
ways
to
solve
as
well?
All
that
option
is
like
a
thinking,
apple
game,
that
you
sculpt.
What
do
the
things
different
and
you
a
third
option
you'd
be
like?
Would
you
like
to
really
change
the
default
scaffold
for.
A
E
A
It's
a
flag,
you
are
because
now
we
have
uploading
ecosystem
right.
So
the
code
is
the
plugin
you
everything
is
inside
of
the
plugin,
so
the
flag
that
the
api
the
clip
provides
for.
You
is
the
plugin
flag,
so
you
use
it
by
default.
We
have
the
golem.
Plugin
ngo
is
called
fortuitous
way.
You
could
you
have
like
a
camilla's
golden
plugin
and
to
do
this
composing.
The
way
that
camilla
sings
is
better
right
for
foreign.
E
I
get
you
I
it
I
I
think
I
get
yeah,
it
just
seems.
A
E
I'm
not
really
sure
I've
never
built
a
plug-in
for
coup
builder
so
like
to
me.
It
seems
like
it
should
just
be
like
on
coup
builder
init.
You
can
just
say
where
you
want
things
to
be,
and
then
they
are
that
way
forever
or
after,
like
maybe
configured
then
after
in
your
project
file.
But
it
sounds
like
you're
saying:
that's,
not
easy,
but.
D
E
D
D
It
is,
but
you
want
to
obey.
A
E
A
A
Can
you
please
feel
free
to
bring
in
these,
like?
I
can
share
with
you
the
documents,
at
least
for
you
having
some
idea,
why
the
plugins
and
how
that
works
is
angel.
What
was
the
motivations
behind
autism
stuff
and,
if
you're
checking
here,
for
example,
where
we
generated
the
samples
you
see
that
it
has
plugins
and
so,
for
example,
it
has
gold
f3
that
she
usually
did
the
full
and
then
we
pass
the
other
plugin
that
is
his
new
one
like
that
default
is
trying
to
propose
right.
So
by
you.
A
You
also
call
the
plugins
in
a
chain.
So
you
call
you
can't
do
things
like
uploading
that
you
just
do
something
a
helper
or
just
do
a
customization
on
top,
so
you
can,
for
example,
create
a
that
he
change
the
the
layout
as
you
please,
after
generating
the
things
that
you
know
and
fix
the
things
for
you
and
you
use
that
in
all
your
projects.
B
D
D
Oh,
I
was
just
gonna
say:
come
on
just
to
kind
of
reiterate,
you're,
basically
suggesting
having
a
plug-in
that
can
be
chained
and
that
way,
if
you
created
a
go
project,
then
it
it
knows
that
the
controllers
and
the
apis
are
going
to
be
where
the
go
wants
to
put
it,
and
then
they
can
move
it
to
where
shane
would
want
it.
And
then,
as
long
as
that
plug-in
is
configured
in
your
project,
it
would
know
and
work.
A
Yeah,
the
the
first,
the
default
one.
Never
you
know
about
the
the
external
one
right
that
we
don't
know
about
like,
but
he
the
they
can
make
any
person
can
do
a
plugin
that
he
runs
after
the
default
discovery.
Do
the
change
energy?
I
don't
know
that
is
an
option
right.
Maybe
it's
that
it's
possible.
It's
one
of
the
ideas.
One
of
your
options,
like
I
don't
know,
are
you
having
a
camilla?
Go
v3
angel,
never
use
it.
Here's
one
more!
A
It's
like
this
one
and
you
try
to
use
the
kobe
builder
as
a
limit
to
import
as
much
as
possible,
and
you
don't
need
to.
You
only
do
the
implementation
about
what
you
need
to
change,
because
you
can
import
everything.
You
can
call
the
method
that
you
call
the
image
from
the
goal:
v3
plugin
right
and
you
don't
need
to
hang
clement
out
and
play
change.
Obviously,
and
you
only
do
things
on
top
also.
We
have
now
in
the
libby
many
options
like
if
our
you
replace
or
insert
your
code
as
well.
A
You
have
some
many
many
notes,
we're
having
our
autos
option,
plugging
notions
that
you,
you
will
find
some
methods
that
help
you
try
to
customize
things
on
top.
So
that
is
an
option
as
well.
So
you
you
not
have
so
much
code
to
customize,
but
when
you
call
the
create
api,
also,
your
plugin
needs
to
knows
how
to
work
with
the
create
api.
A
A
Wait
we
crossed
six
more
six
minutes.
We
crossed
six
minutes,
really
sorry
that
she
was
my
mistake.
I
I
was
not
with
a
lot
of
attention
in
the
time.
I
would
like
to
thank
everybody
and
please
bring
it
to
stop
again
the
next
one
we
can
discuss
fingers.
I
I'm
more
than
happy
to
share
with
you
how
all
documentation,
everything
that
we
have
about
today,
plugins
and
show
for
you,
some
ideas
and
options,
and
you
also
can
came
with
other
ideas
and
options
right
and
the
last
sisters
and
see
how
how
you
can
help.
B
Hi
camilla,
I
just
had
one
question
like
if
I
were
to
contribute
to
a
new
feature
in
this
queue
builder
project.
How
would
that
be
or
how
to
explore
that.
A
You
are
more
than
welcome
to
contribute
with
everything
in
the
project
right
everybody
please
contribute
to.
If
you
could
be
builder,
you
only
make
us
happy
with
that.
If
you
you
can,
is
that
new
feature
is
something
that
you
have
in
mind
or.
B
A
A
Some
labels,
which
is
for
she
gujashu
if
I'm
not
wrong,
I
don't
know
if
you
have
too
much
if
it
is
labeled,
just
one
good
first
issue
right.
So
if
you
want
to
look
on
that
and
start
for
that
or
starts
for
documentation,
usually
is
the
easy
ones.
I,
when
I
start
to
contribute
to
a
new
project.
The
first
thing
that
I
do
is
try
to
take
something
very
simple:
just
you
to
see
how
the
things
works,
and
today
I
started
to
get
other
things
here-
is
the
contribution
guys.
A
A
B
Yeah
yeah
thanks
and
the
recordings
of
these
of
meetings
are
placed
in
which
talking
we.
A
Have
I
started
to
put
today:
we
don't
have
it
from
all
meetings,
we
usually
record,
but
we
we
have
not
been
doing
a
good
job
in
publish,
but
yes
has
a
link,
and
please
ping
me
after
choose
one
that
you
share
with.
You
has
at
least
in
your
in
your
youtube-
and
I
also
you
put
that
here
for
everybody
knows
right.