►
From YouTube: Cloud Foundry Community Advisory Board Call [July 2020]
Description
Find the agenda here: https://docs.google.com/document/d/1SCOlAquyUmNM-AQnekCOXiwhLs6gveTxAcduvDcW_xI/edit?usp=sharing
B
A
It
okay
there.
It
is
welcome
everyone
to
the
July,
Cloud
Foundry
community
advisory
board
meeting.
It's
going
to
be
a
short
one,
I
think,
but
we'll
cover
some
some
updates.
As
we
always
do,
I
posted
a
link
to
the
cab
agenda
in
the
chat.
It's
the
normal
Google
Doc
link
that
that
rolls
up
all
the
all
the
notes
from
all
of
the
meetings
so
for
July,
15
I
want
to
kick
it
off
with
cff
highlights
and
updates
and
I
want
to
extend
a
warm
welcome
to
our
new
developer.
A
C
Yeah
sure,
thanks
Troy
hello,
everyone
I
have
been
doing
some
fun
stuff
around
DevOps
folks,
particularly
with
a
performance
monitoring
tool
and
open
shift
in
the
past
couple
of
years.
So
conversations
with
chip
led
to
this
opportunity
where
the
idea
is
to
sort
of
spread
the
cheer
among
people
who
are
not
just
enterprise
developers
and
people
who
are
not
just
in
like
companies
that
have
over
a
thousand
employees.
So
I
think
the
idea
is
to
focus
on
Cloud
Foundry
for
smaller.
D
C
C
A
C
A
When
did
you
start
with
with
the
foundation
just
a
week
ago?
Okay?
Well,
welcome.
Welcome.
Welcome
aboard
more
than
merrier
always
I'm
really
excited
to
hear
that
smaller
organizations
are
gonna,
be
a
focus
for
Cloud
Foundry
evangelism
efforts.
I've
always
long
felt
that
small
footprint
cloud
foundry
was
something
we
should
pursue
and
that
it
shouldn't
just
be
for
the
big
teams
that
anyone
can
use
this
this
platform.
So
it's
one
of
my
favorite
use
cases
to
put
it
in
a
smaller
organization
and
help
that
organization
grow.
A
So
I'm
really
glad
that's
part
of
a
focus
that
we've
got
and
hopefully
we'll
have
have
Shedrick
available
for
the
next
call
yeah.
You
don't
think
you
could
make
it
today,
but
yeah
well
welcome
his
him
as
well
and
we'll
get
a
little
bit
of
info
on
him
and
what
his
focus
areas
are
and
yeah
so
good
good
to
have
you
here?
Mr.
Clark,
do
you
have
any
CFF
highlights
you'd
like
to
share
with
us
I.
E
C
E
E
E
We
don't
have,
we
didn't
pick
any
dates
yet,
probably
in
a
couple
of
weeks,
all
majority
of
the
talks
are
now
on
YouTube,
so
you
can
find
all
those
four
for
those
that
didn't
they
couldn't
make
it
or
for
those
who
you
know,
didn't
catch
all
the
talks
you
can.
You
can
watch
them
freely
now
on
YouTube,
so
that's
cool
yeah,
that's
kind
of
all
I've
got,
though.
A
Cool
well
on
the
cff
front.
I
just
saw
that
Cedric
has
joined
so
Cedric.
Do
you
want
to
introduce
yourself
to
the
hi
I'm
Troy
I'm
from
Sousa
I'm
were
lack
of
a
better
person
to
do
it,
I'm
running
the
clouded,
the
advisory
board
meetings-
and
this
is
our
crew
today?
Why
don't
you
introduce
yourself,
tell
us
what
you're
doing
with
Cloud
Foundry
foundation.
D
Sorry,
everybody
can
hear
me
yeah,
so
my
name
is
sheetrock.
I
can
tell
and
I
recently
joined
self
on
different
dish
as
a
tip
advocates,
so
I'm
currently
involved
in
the
community.
We
our
side
of
things
and
also
getting
up
to
speed.
We
be
familiar
with
Excel
for
any
project,
so
that
kind
of
mostly
what
I
do
correctly.
So
it's
nice
to
meet
everybody
here,
hi
Troy
Wes
with
everybody
here.
A
A
A
D
It's
been
sweet
for
me,
I
mean
coming
from
background,
be
a
pistol.
You
want
an
engineer,
it's
it's
quite
easy
for
me,
I,
don't
know
pretty
thing
break
into
the
to
the
cloud
needs
if
technology
sector,
so
it
was
really
easy
for
me
and
so
far
the
Millennial
person.
Quite
it's
like
food
I
mean
I,
be
able
to
like
maybe
do
a
couple
of
things
and
how
many
people
watch
anymore
and
also
share
my
knowledge,
basically
who.
A
F
Thanks
Trey
yeah
thought
I'd
just
do
a
quick
rundown
of
some
of
the
project
updates
within
the
app
runtime
BMC
I.
Think
the
biggest
news
is
that
the
CLI
team
finally
released
the
first
GA
of
the
v7
CI.
So
that's
great
news.
No.
That
was
imminent
in
our
last
call,
and
we
had
a
lot
of
discussion
about
that
at
Summit,
so
again,
correct!
Congratulations
to
them
and
to
main
integration.
Projects
are
continuing
to
move
forward
relevant
released,
another
minor
version
of
CF
rockets.
F
I,
know,
Irene,
he's
still
working
on
application
tasks,
I
think
there's
a
looks
like
they're.
Nearly
at
the
end
of
that
Jules
should
be
coming
back
this
weekend.
It
looks
like
he's,
got
a
fair
amount
of
acceptance
to
clear
out
as
his
welcome
back
present
for
paternity
leave
and
they're
also
continuing
to
work
on
some
more
internal
CRD
representations
of
the
resources
that
are
in
the
Irene
API,
so
as
moving
deeper
into
the
kubernetes
api
machinery
for
their
workflows.
A
couple
other
highlights:
UA
now
has
Prometheus
available
metrics
exposed
on
its
gate
deployments.
Artifact.
F
So
they've
incorporated
the
stats.
The
exporter
see
you
translate
the
stats
D
metrics,
they
were
emitting
previously,
and
then
they
put
the
annotations
on
so
Prometheus
can
scrape
those
in
a
kubernetes
environment
and,
like
neon
metrics
I,
know,
they've
I
mentioned
they're
resolving
a
few
of
issues
they've
seen
in
CF
deployment
13
around
this
Aslaug
agent
and
then
they're
also
starting
to
explore
whether
it
would
make
sense
for
them
to
be
able
to
use
some
of
the
metadata
that
Kathy
has
been
exposing
in
the
API,
so
labels
and
annotations
to
translate
into
annotations.
F
F
A
A
A
A
Susan
might
be
able
to.
We
looks
like
we're
able
to
contribute
someone
to
that
team,
but
we'll
talk
with
with
the
members
of
that
team
about
that.
I
wanted
to
go
back
to
the
discussion
that
happened
about
the
v7
CLI,
specifically
about
the
fact
that
there
is
not
in
this
in
the
release
announcement
and
the
notes
about
it.
A
F
A
And-
and
that's
that
was
comforting
to
hear
that,
because
the
language
in
in
the
release
notes
was
a
little
alarming
to
some
of
us
who
thought
well
how
you
know
we're
gonna
have
to
plan
a
migration
where
everybody,
no
and
all
of
our
build
pipelines
and
everything
are
going
to
have
to
shift
from
v6
to
b7
when
we
upgrade
Cloud
Foundry
to
the
next
CF
deployment.
That's
not
strictly
the
case.
It's
the
wording
of
the
release,
so
so
we'll
talk
about.
Maybe
change,
maybe
expanding
that
a
little
bit
to
describe
that.
F
A
I
just
see
norm
joined
us
from
Stark
and
Wayne
welcome
norm
and
since
we're
on
the
topic
of
v3
API
and
the
V
said
the
CF
7
client,
something
that
that
we
talked
about
during
the
summit
when
it
hit
us
in
a
lab.
Was
the
higher
load
put
on
the
clock
controller,
because
a
certain
amount
of
processing
is
happening
now
server
side
rather
than
client
side
right
norm.
Could
you
just
talk
a
little
bit
about
the
concern
there
and
and
what
you've
observed?
What.
G
I
observed
is:
is
that
the
when
you're
doing
the
individual
calls,
because
I
do
so
many
extra
calls
back
to
the
back
end
I
see
that
the
operation
of
basic
calls
like
CF
apps
running
about
two
to
three
times
slower
than
the
CF
CLI
for
version
6,
what's
interesting
to
me
is,
is
when
you
use
the
CF
curl
command
that
doesn't
exist
at
all.
I,
don't
see
the
slowdown,
I
mean
it
makes
a
major
difference.
G
One
of
the
things
I
saw
that
was
in
the
channel
of
this
CF
CLI
people
as
they
were
reworking
how
they're
doing
some
of
the
logging
in
and
it
looks
like
they're
going
to
in
version
1.8
change,
how
they
do
things.
So
more
things
will
be
cached
on
the
client
side,
which
should
resolve
that
issue
of
that
slowness,
but
they're
not
planning
to
put
that
out
into
1.8
fine
Frank.
G
It
back
six
to
eight
months
ago,
but
because
they
were
so
underwater,
they
didn't
want
to
look
at
that
issue
at
all
and
say
they
actually
took
my
issue
and
closed.
It
says
we
can't
we're
not
going
to
resolve
this,
and
so
just
leaving
it
open
and
say
this
is
going
to
be
resolved
sometime
in
the
future.
When
you
look
at
it,
I
gave
them
timing,
estimates
and
everything
else,
along
with
it
show.
Why
is
this
true
thanks
for
doing
that,.
H
Speaking
from
sa
P,
we've
done
same
observations
and
also
reported
that
the
copy
team,
so
in
one
particular
setup,
was
easy.
F7,
CLI
and
some
beta
versions,
we
observed
some
performance
degradations
from
about
five
minutes
was
C
of
six
in
a
space
was
lots
of
applications,
I
think
it
was
70
applications
or
something
like
that,
which
was
rather
big,
okay
and
took
two
hours,
and
ten
minutes
was
yeah.
H
That's
why
we
said
okay.
This
is
something
we
cannot.
We
don't
have
to
talk
about
that
and
with
the
part
of
the
pack
and
their
improvement.
So
we
don't
have
yet
see
you
have
certain
dot
five
and
our
big
landscapes
due
to
some
other
reasons.
So
we
cannot
do
this
comparison
with
the
GA
version
and
we
got
some
feedbacks
that
there
are
improvements,
but
I
think
it's
a
long
way
to
go
until.
G
They
actually
build
a
full
set
up
up
front
or
v6
was
actually
doing
aging
pulling
in
the
data
as
needed,
so
that
so
that's
so,
even
though
it
took
you
five
minutes
in
the
old
one,
because
you
were
doing
page
by
page
by
page
and
restricted
your
request
on
the
CFC
lie,
it
would
be
the
full
request
and
get
all
the
data
and
then
aged
out
so
the
time
to
get
all
that
data
was
because
it
was
doing
a
full
table
scans.
Instead
of
you
know,
index
scans,
it
was
a
great
deal
slower.
G
H
Pretty
sure
that
would
be
more
so
once
we
have
our
landscape
updated,
we
will
also
do
some
performance
investigations
and
there
will
be
lots
of
feedback
I'm
afraid,
but
anyway,
yes,
it
has
to
be
addressed
in
sand
right.
If
we
look
into
the
specific
cases,
I'm
sure
that
this
can
be
improved.
Quite
a
lot,
I
mean
there
were
some
ideas
was
in
the
two
API.
D
H
It
was
in
in
the
v2
but
yeah
I
guess
it
can
be
improved.
That's
not
something
I
would
have
seen
and
as
a
point
regarding
the
versions
that
I
want
to
bring
so
I
tested,
actually
is
a
ga
version
against
our
landscapes,
which
are
not
yet
on
working
dot.
Five.
So
it's
supposed
not
to
work.
Unfortunately,
well,
there's
an
error
message
missing.
That
explains
you
why
it
doesn't
work,
so
that's
also
something
that
would
put
to
worse
just
to
print
out.
Okay
is
this.
Cf
version
is
obviously
oh
it
and
what
you
get
is
here.
H
G
G
Bumps
in
the
road
and
I've
seen
is,
if
you
attempt
to
use
of
these
seven
CLI
against
the
v6,
you
lie
and
Hades
involve
because
the
ageing
parameter
names
have
changed,
they
will
error
out,
saying
you
can't
find
it,
but
these
seven
stuff
will
air
it
out.
Saying
can't
find
its
parameters.
Page
side,
look,
let's
say
it's
pages.
Instead
of
page
size,
you
know
get
through
words,
you
could
write.
At
least
it
tells
you
what's
going
on
the
recognized
port.
G
G
My
only
concern
is
is
but
trying
to
I
don't
I
wish
it
would
continue
updating
the
CLI
from
version
6,
because
until
everything
moves
over
the
version
7,
you
got
to
support
both
and
are
not
doing
that
at
this
point,
they're
just
they're
trying
to
avoid
doing
any
updates
to
the
version,
6,
stuff
and
I.
Think
that's
I
think
that's
also
an
issue
because
at
least
for
the
next
six
months,
I
think
they're
in
a
frame
transition
period.
Then
he
didn't
support
both
not
just
one.
This.
A
Is
why
it's
it's
good
to
have
this
discussion
here
with
Eric
here,
but
also
we
did
get
a
little
bit
of
clarity
from
the
CLI
team
that
they,
you
know
for
an
urgent
security
issue
with
v6
they
would.
They
would
go
ahead
and
patch
that
so
I
think
with
this
feedback
that
we're
getting
that
I
mean
very
excited
about
having
cf7
I,
like
that,
we
got
the
the
v3
API,
but
it's
got
some
teething
problems
and
we're
gonna
need
a
bit
of
overlap.
A
There
I
think,
is
the
consensus
from
from
people
who
are
testing
this
at
scale
that
we're
gonna
need
that
v2
API
endpoint
to
stick
around
for
a
while
I
know,
for
instance,
cube
CF
has
is
just
gonna,
be
because
cat
cats,
the
the
tests,
are
going
to
move
over
to
C
f7
soon
and
so
we're
we're
bumping.
A
cube.
Cf
is
not
yet
on
the
version
which
works
with
CF
7,
so
we're
gonna
have
to
bump
that
and
change
the
tests
over
and
so
we'll
be
able
to
contribute.
H
F
There's
been
some
previous
announcements
about
duplication
schedules,
but
as
far
as
I
understand
that
the
v3
acceleration
team
is
still
working
through
a
lot
of
the
transitions
and
replacements
of
v2
endpoints
to
v3,
and
so
we're
not
even
at
a
point
right
now,
where
we
could
say:
ok,
we're
starting
that
Deptford
window
for
the
entirety
of
the
v2
API,
so
I
think
that's
that's
they've
still
got
that
to
complete,
and
then
we
definitely
won't
have
long
enough
transition
periods
away
from
those
old
end
points,
even
once.
They're
done
so.
A
Okay,
the
seven
is
actually
calling
a
number
of
v2
end
points
for
those
okay,
I'm
glad
we
had
this
little
talk.
It's
almost
good
that
we
we
have
time
when
this
in
this
call,
because
we
didn't
get
the
presentations,
so
we
had
that
chat
and
because
they
don't
have
some
extensions
updates
really
e
to
share.
We
do
have
a
discussion
point
unless
somebody
wants
to
I
want.
A
A
A
It's
been
announced,
and
that
was
a
surprise
to
me
so
yeah,
it's
fun.
We're
gonna
have
some
potentially
have
some
people
from
that
that
team
or
that
community,
that
would
would
come
and
participate
here.
Let's
I
hope
some,
some
really
smart
people
that
would
have
a
lot
to
contribute
to
this
community
as
well.
A
We
do
have
a
proposal,
that's
out
there
on
CF
dev
that
Yui
wanted
to
talk
about
and
I
want
to
support.
This
is
the
proposal
to
change
the
default
branch
names
of
cloud
foundry
projects
from
master
to
main,
as
I
think
the
front-runner.
At
this
point,
you
want
to
talk
a
little
bit
about
about
what
you
posted,
yeah.
B
I
posted
that
I'm,
hoping
we
can
get
to
the
point
where
all
the
the
PM
CDs
have
have
bought
in
and
agreed
to
to
make
this
particular
change.
I.
Think
since
then,
I've
noticed
a
number
of
other
foundations
are
so
adopting
policies
around
other
language
and
preferring
terms
like
allow
and
deny,
and
things
like
that
of
our
blacklist.
Whitelist
so
I
think
that
that's
worth
considering
some
of
those
terms,
I
believe
in
the
networking
area,
for
example,
shannon
has
long
where
you
would
expect
the
term
like
blacklist
or
whitelist.
A
I've
shared
it
with
the
extensions
pmc
recommending
that
at
their
ease,
easiest
convenience,
I
don't
want
to
introduce
a
lot
of
mandatory
churn,
but
at
their
earliest
convenience
to
make
that
change.
Certainly
Stratus
is
on
board.
I
know
from
from
my
cube
CF
contacts
that
that
they
are
on
board
with
this.
It's
just
you
know,
pipelines
have
to
change
stuff
has
to
change,
so
it
might
be
a
while.
There's
no
I
haven't
seen
any
objections
in
principle
to
this.
A
It's
mostly
just
you
know
it's
going
to
take
time
to
figure
out
when
when
to
do
it
and
how
to
do
it,
I
would
say
to
the
community
if
anyone
has
a
particular
issue
with
this
or
if
you
don't
feel
comfortable
talking
about
it
in
the
wider
group,
and
it's
like
you
have
an
objection
or
a
concern,
or
you
have
a
particular
favorite
branch
name
that
you'd
like
to
change
to,
and
you
don't
feel
comfortable
you
can.
You
can
approach
myself
and
you
a
maybe
you
as
well
directly
on
on
slack
to
raise
those
concerns.
A
F
That's
that's
on
my
plate.
We
had
discussed
it
with
the
attendees
at
the
time
PMC
meeting
last
week,
and
people
were
generally
supportive
of
both
branch
renaming.
We
may
settle
on
some
different
patterns
as
as
is
appropriate
for
the
development
lifecycle.
I
know
somebody
posts
are
already
doing
a
develop
and
then
release
promotion
branch
workflow
anyway,
and
so
it
may
just
be
that
that
more
stable
release
branch
gets
renamed
to
something
other
than
master
yeah.
F
A
F
F
A
A
G
A
I,
don't
know
to
outside
of
databases
which
are
consumed
by
Cloud
Foundry.
If
that's
something
we
control
I
mean
remember,
we
we
are
going
to
be
consuming
a
lot
of
other
projects
which
will
have
different
role,
naming
conventions.
So,
to
the
extent
I
mean
we
wouldn't
go,
I,
don't
think
to
the
extent
of
like.
A
G
A
G
A
A
Want
to
if,
unless
there's
any
further
comments
on
that,
I
wanted
to
open
it
up
for
potential
free
for
all
items
of
concern
things
we
want
to
talk
about.
You
know
what
you
had
for
breakfast
today.
Do
we
want
to
do
social
time?
We
have
25
luxurious
minutes
to
share
our
thoughts
on
Cloud
Foundry,
so
I'm
gonna,
open
the
floor
for
questions,
comments,
etc.
I
So
you
can
see
when
think
a
big
drain,
but
think
that,
being
you
know,
processed,
we
now
show
things
like
the
hypervisor
and
the
inspections
so,
and
we
also
just
to
chime
in
on
the
last
topic.
We
also
have
the
conversation
and
we're
making
the
changes
in
the
Bosch
team,
where
we
have
to
work
on
the
pipelines
to
change
the
naming.
Anyone
else
well,
I
decided
chime
in,
and
so
you
know,
the
box
team
is
still
attending
and
still
happy.
Ok,.
A
A
E
E
Basically,
we
have
a
team
looking
into
CF
for
Kate,
so
we
have
been
looking
also
into
it
YouTube
CF
recently
or
for
quite
some
time,
and
recently
we've
decided
to
kind
of
fit
the
focus
a
bit
to
to
look
in
to
see
if
okayed,
so
we
have
a
TV
looking
into
it
and
also
starting
to
I
would
say,
contribute
some
first
enhancements
there.
So
things
like
being
able
to
specify
an
external
database
was
something
that
that
the
team
was
was
actually
working
on.
E
A
You
mentioned
it
was
really
nice
to
see
see
for
Kate's,
adopt
quark
secrets
as
the
secret
management
mechanism,
showing
that
there's
good
communication
happening
between
those
teams.
I
know
a
lot
of
us
were
concerned.
I
was
very
concerned
that
we
would
have
two
parallel
and
non
interacting
branches
of
cloud,
foundry
or
kubernetes
development
I.
Think
we've,
we've
normalized
that
in
a
way
that
makes
sense.
Of
course,
cube.
Cf
s--
plan
is
to
adopt
the
components
as
they
as
they
as
they
mature
in
CF
for
Cades
I.
A
Think
UAA
is
next
on
their
list
of
things
that
they're
gonna
pull
in
from
the
the
new
kubernetes
idiomatic
releases
from
upstream
and
then
there's
some
more
major
stuff.
They
will
have
to
come
in
all
at
once.
I
think
routing
is
something
that
has
to
come
in
all
at
once
and
the
move
of
course.
The
kpac
eventually
will
be
a
similar
sort
of
big
big
shift,
but
otherwise
it's
going
to
be
an
incremental
moving
of
things.
E
A
E
Not
like
I
believe
they're
still,
they
don't
forget.
Motor
affairs
I
think
there
was
one
of
the
standard
language
is
missing
if
I'm
not
mistaken,
but
but
like
an
hour,
and
it's
more
I
would
say
things
inside
the
belt
back
that
have
slightly
changed
for
for
some
of
the
stuff.
So
one
thing
that
I
recall
is
that
I
believe
there
has
been
a
change
with
bringing
in
JDBC
drivers
and
how
that
works
compared
to
the
existing
thanks.
A
Thought
about
something
that
we
could
talk
about.
I,
don't
know
if
we've
got
any
one
from
OSP
API
on
the
call,
but
I've
noticed
that
the
cloud
providers
are
deprecating
our
end
of
life
in
OSP,
API
brokers
for
their
platforms,
most
recently
I
sure
as
stopped
supporting
their
oh
s,
be
API
broker,
and
it's
something
that
is
a
bit
of
a
concern.
I
understand
the
move
for
cloud
I
understand
the
motivation
for
cloud
providers
to
expose
their
own
operator
force
controlling
service
provisioning.
That
seems
to
be
a
model
it's
popular
in
kubernetes,
to
get
that.
A
Unfortunately,
this
means
that
every
cloud
provider
or
every
service
bihter
will
expose
its
own
operator
in
its
own
CR
DS,
and
there
will
be
no
common
functions
for
platforms
to
plug
into
this
is
gonna.
If
this
trend
continues,
which,
once
you
have
the
three
cloud
providers
off
of
us
Bobby,
that's
that's
a
significant
move.
A
It
introduces
more
complexity
because
we
previously
had
a
single
interface
that
we
could
agree
upon
when
we
want
to
ask
for
a
service
in
Cloud
Foundry.
This
is
what
we
call.
We
call
AHS
baffi
ozb
appease
broker.
Implementation
will
call
whatever
it
needs
to
call
to
provision
of
service
and
return
connection
credentials.
A
We
now
have
to
have
a
broker
between
the
operator
and
an
Cloud
Foundry
Sue's
is
doing
some
work
on
this
related
sort
of
speculative
work
based
on
our
mini
broker.
That's
in
the
community
SIG's,
but
I
think
it's
something
we
could
keep
an
eye
on,
maybe
I'll
get
to
liova
cease
or
somebody
from
Suzhou,
or
maybe
convince
somebody
from
the
house
back
B
team
to
talk
about
it
in
the
next
community
advisory
board,
because
I
think
it's
an
important
thing.
It's
an
underrated
method
for
connecting
platforms
to
services.
A
It
works
very
well
it
just
it
sort
of
got
overshadowed
by
people
talking
about
service
mesh
and
talking
about
operators
and
they
sort
of
I
think
people
are
forgetting
the
the
core
utility
of
AWS
bappy
I.
As
a
community.
We
got
to
remind
remind
people
how
useful
it
is
any
thoughts
on
that
Wayne's,
awfully
quiet.