►
From YouTube: App Runtime Deployments Working Group [January 26, 2023]
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
have
a
quick
look
at
last
time
or
the
action
items
from
the
last
meeting,
so
the
validate
search
job
was
constantly
read
in
the
CF
deployment.
Pipeline
and
yeah
I
simply
fixed
it.
By
allowing
to
ignore
some
paths
which
are
not
relevant
and
to
answer
questions,
question
or
I
will,
as
as
proposed
by
Carson.
It
now
also
checks
the
load.
B
A
A
This
is
for
setting
up
the
infrastructure
bubble,
environment
yeah,
so
this
is
also
done
yeah
so
for
today,
I
don't
have
too
many
topics,
because
before
we
come
to
Benjamin's
bigger
topic,
so
we
had
the
beginning
of
the
me
a
week.
A
problem
with
the
private
Docker
registry
account.
A
So
there
are
two
Cloud
Foundry
acceptance
tests
which
check
that
you
can
download
images
with
credentials
from
a
private
Docker
repository
and
they've
mentioned
something
that
you
meant
to
rotated
credentials
because
of
some
reason.
Yeah.
C
C
C
It's
just
it's
just
required
for
the.
We
just
need
some
sort
of
private
Docker
hub
for
those
cats
tests.
If
we
want
to
be
running
them
in
our
pipeline.
A
Okay,
so
it
might
make
sense
to
create
our
own
account
good,
not
sure
what
is
needed,
for
that
I
mean
we
have
enough
email,
Gmail
mail
addresses
created
by
Chris
yeah,
so
I
will
check
what
we
need
and
try
to
create.
An
own
account
sounds
good,
yeah,
yeah.
Okay,
second
topic
is
migration
of
the
not
this
one
of
the
of
our
Concourse.
A
So
it's
still
based
on
the
Bosch
CI
stem
cell
infra
project
and
that
one
is
pretty
outstated
and
we
will
follow
the
approach
from
the
app
runtime
interfaces
working
group.
They
have
set
up
their
own
Concourse,
which
means
we
will
may
have
to
temporarily
change
the
domain.
A
A
Okay,
good,
that's
everything
from
my
slide.
I.
D
We
have
an
outdated
version
of
kubernetes
1.21
and
that
one
was
deprecated
and
it
looks
like
this
old
Conquest
installation
uses
some
beta
apis
or
whatever,
and
they
notified
us
that
this
will
increase.
A
Yeah,
but
there
are
also
other
reasons
to
migrate
to
the
new
Concourse
setup.
A
Good
release
of
safe
deployment
is
out.
There
was
nothing
special
okay,
so
I
think
that's
everything
from
my
side
then
handing
over
to
you
Benjamin.
B
Yeah
I
guess
I'm
not
sure
entirely
what
we
want
to
do
with
this,
but
I
I
just
wanted
to
mention
that
we've
had
a
feature
that
allows
people
to
egress
metrics
about
the
VMS
that
are
running
on
their
system.
B
This
has
been
used
to
be
by
System
metrics
forward
or
now
is
system
metrics
agent
I,
don't
think
either
of
them
have
ever
been
mainlined
into
CF
deployment
and
I
just
wanted
to
bring
it
up
and
be
like.
Is
this
something
we
might
want
to
consider
mainlining
into
CF
deployment.
A
So
for
our
understanding,
I
think
Bosch
already
emits
a
lot
of
metrics
like
system
CPU
as
I've
seen
that
the
metrics
agent
provides
a
few
more
metrics
like
Network
utilization.
B
You
can
collect
all
of
those
metrics,
but
it
doesn't
I
wouldn't
say
it
emits
those
metrics
by
default.
So
the
goal
would
be
is
if
you
had
a
CF,
nozzle
or
or
something
like
that
or
if
you're
using
law
cache,
you
should
be
able
to
collect
it
and
using
that
method.
A
B
B
I
believe
this
is
a
example,
or
this
is
one
of
the
PRS
there's
there's
two
of
them
actually,
the
other
one
is
the
same
thing,
but
if
you're
using
Windows
I
believe
at
the
very
least,
it
might
be
interesting
to
move
them
from
experimental
to
to
a
regular
Ops
file.
A
B
A
A
I
can't
really
tell
if
it's
if
it
would
make
sense,
I
mean
we
would
also
have
a
few
new
credentials
by
default
right,
yeah,
yeah,
so.
D
Feedback
on
and
then.
A
D
Out
the
let's
say,
pros
and
cons
for
for
both
approaches.
If
we
default
it,
maybe
then
we
would
need
at
least
an
Ops
file
to
to
switch
it
off
easily.
D
Yeah,
okay
yeah.
We
need
to
ask,
we
should
also
ask-
and
let's
say
Stephanie
was
Johan
from
from
the
lock
regulator
team
from
on
the
one
side,
what
does
it
mean
on
on
additional
load,
Etc
and
big
Landscapes?
That's
usually
the
first
question
that
comes
to
our
minds.
A
Yeah
I
mean
there
would
be
a
lot
to
consider
if
we
start
adding
new
components
on
each
of
our
VMS.
Yes,
it
could
have
huge
impact
for
us
on
system
load
and
and
whatever
so
yeah,
okay
good,
but
so
then
I
would
open
an
issue
and
describe
what
this
is
about.
I
do
not
have
all
the
technical
details,
so
I
would
then
ask
you
Benjamin
to
fill
in
the
gaps
around
and
ask
for
feedback
and
then
decide
how
to
proceed.
A
A
Let
me
share
again,
so
the
the
experimental
fs4
job
is
still
red
with
exactly
the
same
result,
so
these
four
are
missing,
but
nevertheless
each
one
of
those
updated
their
GitHub
issues.
So
here
is
the
parent
issue
and
it's
it's
all
the
same,
I
think
and
here
for
example,
two
days
ago,
someone
create
a
task
built
dependency.
A
D
A
It's
in
progress,
yeah.
C
C
Nope
all
right
I
can't
find
it.
It's
probably
fine.
Oh
there
was
a
request
from
Yovan
and
some
other
sap
folks,
because
about
to
inoculate.
C
Exactly
we
should
maybe
have
a
chat
about
that
real,
quick
that
someone,
some
I,
think
Stephen
lay
Stefan,
lay
called.
A
C
That
the
removal
of
the
traffic
controller
broke
broke
a
bunch
of
stuff,
so
there
was
a
suggestion
that
logger
Gator
could
be
pinned
back
to
106.7.5
for
the
time
being
to
get
around
this.
C
This
issue
for
some
period
of
time
until
it's
resolved
I
guess
the
only
alternative
I
can
think
of
or
only
Alternatives
I
can
think
of,
is
to
move
forward
and
just
try
and
fix
it
fast
I,
don't
know
on
it
sounds
like
Steve
Stefan
would
prefer
not
that
the
or
like
what
reverting
the
change
in
log
reader
itself
for
some
period
of
time.
A
C
The
weird
part
would
be
cutting
a
release
of
cfd
and
saying
we're
we're
going
back
to
this
old
way.
I,
don't
think
we
would
lose
any
features,
but
it
would
be
like
a
downgrade
in
CF
deployment,
not
on
a
component
that
maintains
a
database
so
potentially
not
a
failure
mode,
but
probably
worth
testing
out.
If
that
upgrade
isn't
a
problem,
what
do
you
all
think
about
pinning
versus
other
options.
A
They
are
pinning
for
how
long
I
mean
this.
The
this
social
context,
the
the
one
of
the
last
CFT
blood
releases,
contained
an
update
to
lock
regulator,
agent,
seven
or
something,
and
this
removed
the
recent
locks
endpoint,
and
it
turned
out
that
many
components
here
at
sap
still
make
use
of
this
endpoint.
Although
it
has
been
deprecated
I,
don't
know
quite
some
time
ago,
so
fixing
this
on
our
site
for
all
components
is
likely
to
take
months,
at
least
and.
B
A
Don't
know
if
anyone
has
made
efforts
to
to
assemble
a
timeline
for
all
the
components,
and
so
this
could
take
long
yeah,
and
so
we
would
have
to
pin
to
a
lower
version
for
quite
some
time,
and
that
means
no
updates
anymore
for
Locker
data,
aging
or
lock
regulator
release
one
of
those
two
I
guess.
The
best
way
for
us
would
be
Seven.
Something
introduces
recent
locks
again,
and
we
continue
with
that
for
a
while.
A
B
So
I
believe
there's
cats
tests
that
directly
test
the
the
streaming
endpoints
for
V1
and
V2
I.
Think
recent
is
only
a
thing
that
exists
on
traffic
controller,
but
the
only
way
that
recent
was
being
tested
was
through
the
CLI
and
the
CLI
switched
specifically.
Does
that
sound
right,
Carson
yep.
D
So
I
guess
from
CF
deployment
point
of
view,
leaving
sap
out
at
the
moment.
I
think
the
best
approach
would
be
that
we,
let's
say
re-enable
this
endpoint
on
a
recent
version,
seven
and
so
that
the
normal
fixed
flow
Etc
continues
and
because,
if
we
would
pin
in
CF
deployment,
this
is
a
a
shitty
situation.
Right
yeah.
D
If
we
put
soon
in
two
or
four
weeks
this
year
and
talk
about
the
bugs
that
cannot
be
fixed,
and
this
is
this
is
not
nice.
And
then
maybe
we
could,
let's
say
in
the
release.
Notes
of
CF
deployment,
explicitly
say
that
this
removal
of
this
endpoint,
if
that's
technically
possible,
I,
would
assume
that
now
that
we
say
it
has
been
re-established,
but
it
is
definitely
deprecated
and
we
remove
it
in
and
then
maybe
talk
first
with
Stefan
like
about
it
and
then
say
a
deadline
in
I.
D
Don't
know
what
it's
realistic,
I
can't
say
it,
because
I
do
not
I'm
not
involved
in
this
discussion
about
the
other
components,
but
something
about
three
months
or
six
months.
So
not
too
far.
In
the
future,
I
mean
the
traffic
controller
is
deprecated
since
long
time
ago,
and
at
the
end
of
the
day,
sometimes
you
have
to
kick
people.
C
D
C
A
C
Clarity
is
Clarity,
is
also
good
and
and
announcing
things
and
I
think
the
fall
of
personally
I,
just
I
think
this
fall
of
like
the
CF
Dev
mailing
list
has
meant
that
there's
less
communication
about
breaking
changes
like
this,
so
it's
a
good
good
call
out
from
from
everyone
involved,
sounds
like
the
cfd.
Our
working
group
opinion,
which
maybe
we
should
throw
into
a
thread,
is
hey.
We
don't
want
a
pin.
A
D
C
A
D
D
Maybe
what
we
from
time
to
time,
doing
and
let's
say
Cloud
controller,
for
instance-
is
set
then
somehow
in
the
local
list
we
could
add
a
feature
flag
to
enable
or
disable
this
endpoint
so
that
we
give
some
tools
to
the
hands
of
consumers
to
control
the
usage.
D
C
Cool
yeah
I
dropped
the
thread
where
it
seems
like
the
major
discussion
is
going
down
so
I.
Oh
I
can
plan
to
drop
our
opinion
as
it
were,
but
otherwise
you
know
feel
free
to
chime
in
or
follow
along
as
well
foreign.
A
A
D
We
need
to
run
and
yeah
I
think
our
first
step
is
to
set
up
the
new
thing
and
before
we
start
with
migrating
some
pipelines
Etc,
then
we
of
course
reach
out
again
and
so
that
everybody
knows
what
is
going
on,
because
we
don't
want
to
lose
the
ability
to
cut
releases.
Of
course,.