►
From YouTube: cf-for-k8s SIG (Aug 2020)
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).
C
B
D
B
B
F
F
So
I
guess,
while
we
are
waiting
so
today,
flavio
will
talk
about
our
back
profiles
for
kubernetes
that
are
cf-like,
and
then
we
have
mario
and
rohit
on
to
talk
about
quark
secret.
F
Hey
four
hour:
four
minutes
into
the
meeting,
so
I
guess
welcome
again
everybody
to
this
week's
special
interest
group
called
cloud
foundry
on
kubernetes.
F
A
A
Well,
okay,
so
we
are
from
project
quarks,
and
we
are
here
to
talk
about
the
quarks
secret
operator
and
quarks
is
the
idea
behind
this
was
to
to
build
a
set
of
operators
that
are
working
together
to
manage
qcf
and
are
useful
to
the
cloud
foundry
community.
A
Strangely,
this
hasn't
updated
the
text,
so
we
have
actually
a
few
of
them
now
and
quark
shop,
I
think,
is
one
of
the
most
interesting
ones,
which
is
about
making
kubernetes
jobs
reusable
and
storing
the
output,
also
triggering
on
certain
changes
and
some
interesting
features,
and
then
there's
quark
secret,
which
is
what
we
are
going
to
talk
about,
and
I
think
this
is
the
smallest
one
of
the
operators
and
it's
used
to
generate
secrets
in
cluster.
A
We
also
have
quark
stateful
set,
which
isn't
quite
standalone
yet
it's,
but
you
could
run
it
as
a
binary
outside
of
the
cluster
or
that
your
own
docker
image
and
we
have
the
quarks
operator,
which
is
basically
taking
the
bosch,
manifests
of
the
old
cf
deployment
and
turning
that
into
a
kubernetes-sized
cloud
foundry.
A
So
quark
secret
is
the
one
big
cid
and
you
can
choose
what
type
of
secret
form
you
want
to
generate
in
the
cluster.
There
are
passwords
certificates,
ssh,
keys,
asr,
key
rsa,
keys,
http
basic
of
configuration
right,
and
then
it
will
just.
You
will
create
the
the
customer
resource
belonging
to
the
custom
resource
definition
that
will
create
a
secret
in
the
cluster
with
that
content
and
there's
a
manual
override.
So
if
the
secret
already
exists,
it
won't
overwrite
it.
A
So
we
use
that
to
supply
default
values
in
qcf
right
when
you,
after
deployment,
manifest
with
all
those
values
you
can
create
them
manually
or
if
you
don't
create
them,
quark
secret
will
generate
them
for
you,
we
can
also
add
labels
to
the
generated
secrets
and,
I
think
annotations,
which
might
be
useful.
We
have
a
special
rotation
config
which
you
can
create,
which
will
then
look
into
the
config
and
take
the
secret,
which
is
there
and
rotate
that
we
have
also
featured
copy
secrets
into
some
into
other
namespaces,
which
needs
some
improvement.
A
A
A
So
hence
for
you,
you
can
find
us
on
hemhup
and,
for
example,
if
you
add
the
repo
you
can
search
it,
you
can
show
the
readme
for
quark
secret.
Here
you
can.
A
A
Again,
you
can
specify
the
version
or
just
leave
it
off
for
the
latest
released
one
and
one
thing
I
wanted
to
talk
about
because
I
feel
like
this
comes
up.
A
lot
is
crds.
A
What
we
currently
do
when
you
deploy
harm
is
that
we
apply
the
customer
resources
from
code.
So
that's
mainly
because
we
had
problems
back
then
with
time2
and
it
reduced
the
amount
of
scripting
we
needed.
So
we
actually
have
code
to
apply
the
custom
resource
definition,
which
is
a
difficult
resource
because
it's
cluster-wide
right.
So.
A
You
can
basically
only
have
one
of
those
the
fields
and
types
which
most
people
care
about,
like
what
can
the
city
do
are
not
actually
part
of
that
yaml.
They
are
part
of
the
type
your
operator
registers.
A
So
what
we?
What
we
have
with
hands
three,
is
the
approach
to
use
the
cds
folder
to
specify
the
cds
and
this
guarantees
that
the
ordering
is
correct,
like
cds
are
installed
first
and
custom
resources
are
installed
after
you
can
skip
this
with
the
skip
crds
flag,
but
there
is
no
templating
done
for
your
cds
because
they
want
a
stable
api.
A
In
the
cluster
before
they
apply
the
custom
resources,
there's
also
no
donation,
because
that
would
be
cascading,
so
it
would
delete
all
your
custom
resources
and
it
would
delete
all
the
objects
which
are
owned
by
those
custom
resources.
However,
this
approach
supports
updates.
So
if
you
have
a
new
cid,
which
has
a
new
version
of
that
cid
in
it,
it
would
be
added
to
the
cid
in
cluster.
A
The
other
approach
I
talked
about
is
wire
operator,
which
we
do
from
code
right.
You
can
control
it
by
an
environment,
flag
or
command
line
flag,
and
it
also
supports
updates
but
updates,
but
it
requires
special
permissions
in
cluster
to
work
like
the
operator
that
you're
running
needs
to
have
permissions
to
create
cids,
and,
of
course,
since
this
is
in
code,
you
could
make
this
more
intelligent,
if
you,
if
you
have
to,
and
what
we
currently
have
also
have
besides
the
apply
from
code,
is
the
template
folder.
A
So
this
is
like
the
old
approach,
which
I
think
is
time
to
compatible
where
you
have
the
cd
in
the
template.
Folder
and
you
have
some
templating
around
it.
For
example,
if
apply
cd
is
false,
then
that
template
will
be
rendered
so
that
results
in
arbitrary
ordering.
I
think
so.
This
might
not
be
optimal.
It
will
also
error
out
if
the
cid
is
existing,
so
it
because
ham
can't
create
an
object,
that's
already
there
and
it
also
can't
update
update
existing
objects
yeah.
A
B
G
Okay,
I
hope
my
screen
is
visible.
Okay,
so
I've
already
installed
the
operators.
The
first
one
here
is
cf
operator
and
then
the
quarks
drop
and
then
the
quark
secret
now
I'll
be
showing
all
the
features
of
quark
secret
and
how
to
use
the
quark
secret
controller.
You
can
also
install
quark
secret
controller
individually,
there's
no
need
of
other
operators.
G
So
right
now,
I'm
in
the
examples
folder
and
the
first
feature
is
generating
the
basic
username
and
password
auth
see
how
it
looks
so
here
we
have
two
things:
two
resources:
the
first
one
is
generating
password
and
username
with
a
predefined
username.
G
Here
we
specify
that
request.basicauth.username
as
myuser,
which
is
predefined,
so
it
only
generates
the
password.
In
the
second
example,
we
didn't
specify
any
predefined
values,
so
it
will
generate
both
password
and
username,
so
the
type
should
be
basic
auth
and
it
would
have
two
fields:
username
and
password.
G
G
G
G
Okay,
so
the
next
feature
would
be
rotating
a
secret.
That
is
regenerating
a
particular
secret
right.
Now
we
only
support
password
rotation.
I
think
there's
a
pr
being
put
by
mario,
which
supports
rotation
of
certificates
too.
G
So
it's
being
deployed
now,
let's
see
how
rotation
works,
so
you
need
to
define
a
config
map
and
in
which
you
need
to
give
the
list
of
all
the
quark
secrets
that
you
need
to
rotate.
In
this
case
we
are
rotating,
generate
password
that
is
generate
secret
one
and
when
I
create
it,
it
should
get
rotated.
B
G
Yep,
that's
it!
I
guess
I'll
give
back
to
mario.
D
G
F
B
G
Field
so
in
the
spec
you
have
a
particular
field
called
secret
annotations
and
you
specify
the
list
of
annotations
that
the
generated
secret
should
have,
and
then
the
generated
secret
would
have.
This
have
those
particular
annotations,
and
if
these
generated
secrets
are
used
anywhere,
then
they
get
updated.
A
I'm
not
sure
if
this
should
be
the
default
for
everything.
I
think
kubernetes
will
notice
if
you
mount
a
secret
as
a
volume
and
then
eventually
restart
the
pot
or
something
but
there's
no
guarantee
when
this
is
going
to
happen.
There's
something
about
this
in
the
docs.
So
in
some
use
cases
you
might
want
to
control
this
in
others,
you
might
not
care.
A
Right
other
certificate,
rotation
yeah.
I
I
think
that
was
mainly
a
bug
but
yeah
it's
it's
multi-step,
because
we
create
the
key
first
and
then
we
wait
for
it
to
be
approved
right.
There's
this
auto
approval,
then.
Finally,
all
this
is
collected
and
put
into
the
secret
we
have,
which
is
like
a
combination
of
the
different
secrets
which
were
there
before
right.
We
combine
those
outputs
into
one
secret
and
that's
the
final
step.
A
H
Right
I
mean,
usually,
you
would
generate
a
new
certificate
roll
that
out
so
that
it's
also
valid
so
that
you
have
multiple
certificates,
valid
right
and
then
and
then
you
would
have
another
step
to
remove
the
old
one.
A
Yeah,
that's
not
supported
by
the
quark
secret
control
at
the
moment.
I
wonder
if
that.
A
A
A
Yep
definitely.
H
Yui
sure,
like
here
or
like.
A
Yeah
docker
config
json,
I
mean
this
third
rotation
is
still.
Nobody
has
been
using
it
apparently
so
great
to
see
some
ideas
there,
docker
config
secrets
are
also
pull
requests.
We
are
working
on
because
you
have
to
have
a
private
docker
registry
in
kubernetes.
You
need
a
special
kind
of
secret
and
we
want
to
generate
that
too.
There
we
do
have
the
secret
copying
feature
where
you
can
generate
a
secret
and
copy
it
into
another
namespace,
so
something
in
another
namespace
can
use
it.
A
I
think
we
have
a
bit
of
a
design
back
there.
That's
also
an
open
issue
where
we
want
to
improve
the
robustness
of
the
deployment
where
you
you're
sure
to
get
the
secret
with
the
right
values
and
you
don't
run
into
invalid
secrets
and
on
the
other
end,
you
don't
want
to
overwrite
arbitrary
secrets
in
the
target.
Namespace
right,
that's
the
conflict,
and
here
the
the
last
item
is
the
samware
releases.
We
do
have
releases
now,
but
they
are
very
much
focused
around
ci.
So
it's
it's
not
really
like
a
real
release.
A
So
if
anybody
would
need
that
we
yeah
that
would
be
an
open
issue.
A
Yeah
thanks
here
are
some
urls,
so
m
hub
and
dark
side.
I'm
done
with
that.
You
know
if
you
have
any
questions,
otherwise
I
would
hand
off
the
vlog.
F
F
How
is
it
generally
used
there
like
it's
the
plan
to
kind
of
use
it
for
pretty
much
any
secrets
and
then
probably
to
to
ue's
point?
Also,
according
to
my
recollection,
there's
often
this
two
or
three
step
process
like
how
is
that
actually
done
in
in
cfo.
F
F
C
Okay,
sorry
got
confused
by
zoom
ui.
Okay,
sorry
about
that,
all
right,
so
hi
everybody,
I'm
flabby
castelli,
I'm
from
souza,
where
I
work
on
kubernetes,
and
I'm
here
today
to
talk
about
this
organization
controller.
C
The
long
story
short
is
that
you
can
actually
partition
a
kubernetes
cluster
so
that
multiple
tenants
get
their
own
dedicated
space.
You
can
do
that
by
using
features
that
are
built
into
kubernetes
or
you
can
bring
in
some
extra
features
from
the
kubernetes
ecosystem
and
by
leveraging
built-in
features
or
external
projects.
You
can
get
to
a
pretty
good
level
of
isolation.
C
This
is
not
going
to
be
comparable
to
running
dedicated
clusters
like
one
per
tenant,
but
you
can
get
to
to
a
pretty
good
isolation
level,
so
the
features
that
are
built
into
kubernetes-
I
guess
everybody
here
knows
about
them,
so
the
kubernetes
namespace
is,
is
what
we're
going
to
use
to
to
create
a
dedicated
space
for
for
the
tenants
to
to
run
their
workloads.
C
C
We
can
use
resource
quota
to
put
limits
on
the
resources
and
finally,
we
can
use
bot
security
policies
to
just
tie
it
make
everything
more
secure
on
on
the
low
level
side
of
things,
but
still
with
these
with
this
features
provided
by
kubernetes,
there
are
still
some
challenges,
which
is
you
have
to
find
a
way
to
map
the
organization
of
your
company
into
into
the
kubernetes
cluster.
C
You
also
want
to
define
all
these
policies
through
you're,
defining
the
policies
through
these
built-in
features,
but
you
want
to
have
them
defined
in
a
central
place
and
then
from
this
central
place,
to
have
them
applied
to
the
users
of
the
different
name
species.
C
So
the
organization
operator
is
part
of
a
solution
to
to
these
challenges.
So
it's
basically
copying
what
cloud
foundry
does
so.
It
introduces
two
custom
resources.
One
is
called
organization
which
is
used
to
identify
the
different
tenants
and
the
other
one
is
called
space,
which
is
what
identifies
the
areas
that
the
teams
of
organization
can
can
interact
with.
These
are
their
their
own
private
sandboxes,
which
are
baked
inside
of
kubernetes
by
an
actual
kubernetes
namespace.
C
The
goal
is
to
define
the
policies
into
a
central
location,
as
I
was
saying
before,
and
then
find
a
way
to
map
these
policies
to
all
these
kubernetes
namespaces.
The
way
to
do
the
mapping
is
done
by
leveraging
a
label
selector,
so
the
namespace
will
have
a
series
of
labels
that
are
used
to
figure
out
which
policies
have
to
be
applied
to
this
namespace.
C
So
this
is
a
short
overview
of
the
relation
between
the
different
custom
resources.
So,
as
you
can
see,
the
organization
has
a
name.
It
has
a
list
of
administrators,
editors
and
viewers,
and
these
can
be
either
groups
provided
by
the
external
identity
provider
or
they
could
be
users
and
depending
on
the
role
they
have.
They
of
course,
have
different
kind
of
privileges.
C
So
this
is
where
we
set
the
baseline
of
the
policies
that
we
want
to
enforce
at
the
namespace
level,
and
then
each
organization
has
one
or
more
spaces
which
can
have
their
own
extra
set
of
administrators
editors
and
viewers
to
give
some
flexibility
to
the
end
user,
and
then
each
space
has
its
own
namespace,
which
has
a
name
that
is
calculated
starting
from
from
the
space
itself,
and
it
has
this
list
of
inherited
labels.
C
These
labels
are
are
calculated
starting
from
what
is
defined
inside
of
organization
and
what
is
also
defined
inside
the
space
extra
config
custom
resource
resource.
Where,
through
the
extra
labels
attribute,
we
can
define
some
extra
policies
that
we
want
to
apply
to
a
certain
namespace
or
we
want
to
override
to
a
certain
namespace
look
a
bit
deeper.
C
So
the
organization
object
is
created
only
by
the
admins
of
the
cluster.
It's
a
cluster-wide
resource.
Every
time
you
create
a
new
organization.
C
The
interesting
thing
to
point
out
is
that
these
edit
admin
and
view
roles,
they
are
basically
built
around
some
cluster
roles
that
are
defined
by
kubernetes,
which
are
pretty
flexible,
and
they
are
basically
allowing
us
to
to
have
some
good
security
basic
in
there
without
having
to
define
new
roles.
C
Last
but
not
least,
the
space
extra
conflict.
This
is,
as
I
was
saying,
used
to
define
extra
labels
that
can
be
added
to
an
organization,
and
this
is
what
allows
an
admin
of
a
of
a
cluster
to
say
instead
of
these
space
that
belongs
to
to
this
organization,
I
want,
for
example,
to
have
less
isolation
in
terms
of
network,
or
I
want
to
put
more
constraints
in
terms
of
quotas
and
and
so
on
and
so
forth.
C
So,
at
this
point,
as
a
cluster
admin,
you
can
go
ahead,
define
different
organizations
once
the
organizations
are
defined.
The
tenants
can
create
the
spaces
in
this
self-service
way
and
every
time
a
space
is
created,
then
we
create
a
namespace
for
that
there
is
still
something
that
has
to
be
done,
which
is
as
a
cluster
administrator.
C
So
all
the
namespace,
with
with
the
label
with
key
size
and
value
small,
are
going
to
have
the
namespace
config
operator
create
this
resource
quota.
That
puts
limits
on
the
amount
of
cpus
and
memories,
and
you
can
do
that
for
for
any
kind
of
stuff
you
want.
So
in
this
case
it's
a
resource
quarter,
but
you
can
create
any
kind
of
namespace
resource.
C
So
that's
basically
it
it's
it's
pretty
a
lot.
So
there
is
a
poc
on
github
in
this
url
and
inside
of
the
github
repo.
There
is
also
a
link
to
this
google
document
that
provides
a
more
in-depth
view
about
the
wall
operator
and
that's
it.
E
That's
sort
of
an
open
question
should
should
irene
be
using
this
to
segregate
user
applications
into
cf,
orgs
and
spaces
or
or
orgs
into
name
spaces
that
are
organized
like
the
cf,
orgs
and
spaces
on
the
cf
cluster,
because
right
now
it
doesn't.
I
think
it's
important
that
the
iranian
team
know
this
is
here
if
they
want
to
start
segregating
the
applications
which
are
which
is
in
the
plan,
so
I'm
gonna
make
sure
demetrius
and
that
and
that
team
knows
about
this.
H
Yeah,
I
don't
think
anyone
from
the
irony
group
is
on
the
call
today
so
probably
worth
sharing
sharing
the
recording
from
here.
B
H
You
can
also
reach
out
separately,
but
I
think
I
recall
when
the
cafe
team
was
investigating,
like
what
are
what
are
some
possible
mappings,
that
that
people
desire
like
different
isolation,
things
they
they
they
put
out
a
a
dock,
trying
to
describe
different
scenarios
and
some
of
them
are
like.
Okay,
are
you
trying
to
stripe
across
multiple
clusters,
a
single
app
across
multiple
clusters,
and
then
what
does
that
mean?
H
Are
you
trying
to
map
and
and
then
therefore,
like
okay,
this
name
space
in
this
cluster
and
this
name
space
in
this
other
cluster,
but
you're,
trying
to
like
spread
your
app
across
both
of
those
those
clusters?
And
how
might
you
represent
that
in
in
cloud
foundry,
as
as
one
of
the
scenarios
they
described
a
few
other
scenarios
of
like
okay?
H
Is
this
what
you're
trying
to
do
in
terms
of
isolation,
and
it
it's
more
like
you
know
the
the
thought
exercise
of
trying
to
understand
what
are
the
different
isolation
desires
people
have,
rather
than
the
implementation
side
of
it.
E
E
Wouldn't
it
be
great
if
kubernetes
had
these
simple
structures
like
orgs
and
spaces
with
which
admins
could
readily
get
their
head
around,
how
to
partition
a
kubernetes
cluster
for
users,
so
this
is
totally
borrowed
from
cloud
foundry
and
then
just
developed
independently,
but
at
the
same
time
as
we
try
and
figure
out,
I
demoed
irene
for
somebody
a
while
ago
and
they
just
assumed
like
okay
and
obviously
these
are
all
going
into
separate
name
spaces
based
on
and
I'm
like.
E
So
they
wanted
to
be
using
the
kubernetes
structures
as
for
segregation,
because
that's
that's
what
they
understood
so
actually,
cloud
foundry
was
the
odd
concept
and
they
just
naturally
assumed
that,
because
of
our
organs
oregon
space
structure,
that
we
were
using
name
spaces.
E
So
I
really
think
it
would
be
good
if
they
they
could
adopt
this
or
or
at
least
explore
it
and
see
if
it
it
could
be
used
rather
than
implement
it
again
independently,
like
like
the
same
way
secrets,
it
became
a
very
handy
thing
for
for
cf4k8s
to
use
quark
secrets,
so
I
think.
D
This
might
be
more
interesting
to
the
cappy
team
right.
I
think,
if
you
think
about
it,
when
you
create
orgs
and
spaces,
you'd
want
the
cloud
controller
to
create
those
objects
in
cube
and
then
with
irene.
You,
you
you'd,
hope
to
just
be
able
to
tell
irene
just
put
my
app
in
this
space.
H
H
The
the
campy
and
irony
team-
they
have
some
sync
up
meetings,
kind
of
on
a
cadence
where
they
they
talk
about
concerns
that
that
crossover
and
definitely
the
topic
of
further
segmentation
is,
is,
as
as
you
say,
troy
like
a
lot
of
people
want
it,
and
it's
more
a
case
of
how
how
what
what
is
that
model,
and
when
are
people
gonna
work
on
it.
E
Yeah
there's
another
use
case,
though,
that
this
was
this
was
mainly
driven
by
before
I
even
thought
about
irene
was
that
you
could
give
a
cloud
foundry,
uaa
user,
similar
permissions
on
the
kubernetes
cluster,
to,
for
example,
deploy
helm
charts
for
services
that
aren't
in
a
broker
so
or
or
something
you
know,
basically
to
give
regular
developers
or
a
certain
class
of
developers
access
to
the
kubernetes
api
that
is
similar
to
their
access
or
the
same
as
their
access
to
the
cf
api.
With
with
similarly
scoped.
I
I
You
know
is
probably
like
a
broader
concern
right
like
it.
I
think
it
probably
does
go
a
little
closer
to
cappy
like
it
just
depends
on
what
component
of
cloud
foundry
is
responsible
for
doing
those
transforms
or
if
we
want
them
to
be,
that
to
be
distributed
across
across
the
system.
As
somebody
who.
F
I
F
So
maybe
one
more
question
to
you
to
flower
video,
as
I
also
saw
some
reference
to
like
other
crds
like
from
their
type,
they
seem
to
have
been
created
by
red
hat.
Is
there
like
a
broader
conversation
in
the
broader
kubernetes
community
around
that,
like
the
multi-tenant
special
interest
group
there,
or
did
you
like
build
that
specifically
for
a
cloud
foundry
use
case.
C
So
the
reddit
operator
is
interesting
because
it's
not
part
of
ansible
sorry
of
open
shift
is
something
that
comes
from
a
community
of
best
practices
around
reddit.
So
it's
it's
done
by
some
reddit
employee,
but
it's
not
part
events
of
openshift
it's
so.
The
the
main
goal
of
organization
organization
operator
is
to
create
this
comp
set
of
having
organization
spaces,
which
is
something
that
is
not
done
by
any
project
that
is
inside
of
a
multi-tenancy
special
interest
group
of
kubernetes.
C
C
The
main
reason
why
I
started
with
with
the
namespace
config
operator
is
because
I
personally
like
the
approach
of
of
using
labels
to
associate
policies
to
the
namespaces.
I
I
I
like
this
this
concept.
While
with
hierarchical,
namespace
controller,
you
would
have
to
create
the
policies
inside
the
top
level
namespace
and
then
create
children
of
that
namespace.
In
order
to
have
this
policies
propagated.
C
Also,
I
thought
that
the
article
namespace
has
too
much
freedom,
because
you
can
do
this
infinite
nesting
of
namespaces,
which
can
be
somehow
confusing,
while
here
with
this
design
right,
it's
just
a
one
space,
one
namespace,
nothing
more,
nothing
less!
So
more
straightforward!.
F
And
probably
another
one,
so
so
like
you,
you
have
been
listing
kubernetes
means
to
to
kind
of
isolate
workloads
from
each
other.
The
thing
that
that
I
always
feel
is
kind
of
most
of
the
time
missing
is
is
something
like
network
bandwidth
like
it's
not
totally
related
to
to
your
presentation,
but
I
I
keep
wondering
if,
like
somebody
is
working
on
on
that
topic
or
if
or
to
put
it
the
other
way
around.
C
I'm
not
aware
of
of
any
work
in
this
area.
It
would
be
something
to
experiment.
We
could.
We
could
leverage
c
groups,
maybe
for
that,
but
then
we
would
have
to
build
some
abstractions
into
kubernetes
to
to
gather
the
metrics
and
build
them
up
at
the
name.
Space
level.
B
J
Guess
if
you
should
be
able
to
do
this
to
some
extent,
at
least
I
don't
know-
I
haven't
looked
in
the
details,
but
at
least
if
we
use
it
anyway
in
inside
cf4k8,
probably
it
should
be
able
to
to
at
least
solve
the
noisy
neighbor
problem
for
cf
apps.
It
won't
like
solve
every
problem
for
for
every
app,
but
for
the
cf
case.
I
think
that
should
be
doable.
F
Okay,
any
other
questions
to
flavio
doesn't
seem
to
be
the
case.
So
I
guess
last
topic
for
today
is
vacation
time
great
at
least
I'll
be
out
of
office
in
like
two
weeks
from
now-
and
I
guess
first
question
will
be
how
how
this
is
the
general
attendance
in
two
weeks,
because
we
we
would
actually
have
a
topic.
So
benjamin
fuller,
I
believe,
from
from
vmware,
suggested
to
to
talk
about
logging
on
the
platform
and
also
how
how
it
ties
into
logging
in
in
the
cloud
foundry
on
vms
based
world.
F
So
I
think
we
we
definitely
have
oh
and
we
have
ben
on
the
call
as
well.
I
just
realized
hey
so
so
I
think
I
I
would
say
if
there's
enough
people
in
in
this
call
that
we
could
just
go
ahead
in
two
weeks,
then
I
guess
I
would
go
ahead
separately
and
find
like
at
least
a
moderator,
or
you
can
kind
of
arrange
that.
But
I
just
want
to
make
sure
that
we
don't
like
invite
ben
in
two
weeks
and
then
pretty
much.
Everybody
is
out
of
office.
F
F
Oh,
I
think
we
definitely
have
a
couple
of
people
that
are
interested,
but
I
would
need
to
to
actually
ping
them
separately
on
on
the
topic,
but
also
philip
is
in
the
call.
So
I
I
think
we'll
kind
of
arrange
something
and
find
somebody
who's
who's
actually
in
the
office.
J
F
J
Be
there
in
next
in
two
weeks,
awesome
done
yeah.
F
F
Awesome
good!
So
then,
unless
there's
any
last
minute
topics,
that's
it
for
for
this
time,
thanks
everybody
for
attending
and
then
see
you
in
four
weeks
from
now,
but
like
the
others,
probably
meet
even
earlier
thanks,
everybody
take.