►
From YouTube: App Runtime Deployments Working Group [July 27, 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
This
cat's,
Zipkin
test
is
still
failing.
Very
often
is
it
is
indeed
often
blocking
the
validation
pipeline.
So
Carson
started
to
work
on
that
trying
to
find
out
how
this
application
was
built,
but
there
is
no
fix
so
far.
So
if
this
test
fails,
we
have
to
keep
retrying
until
it
eventually
succeeds.
A
I
check
if
this
feature
is
officially
supported
and
documented
and
well
yes,
it
is
very
briefly,
but
it
is
documented.
So
it
is
explicitly
mentioned
that
Trace
ID
span,
ID
headers,
are
supported,
and
this
particular
test
will
just
check
that
the
mentioned
headers
are
correctly
generated
and
processed.
So
okay,
so
we
should
validate
this
feature
and
yeah.
A
Let's
hope
that
Carson
finds
out
how
to
either
reverse
engineer
the
jar
or
build
a
new,
simple
application.
I
mean
the
application
itself
doesn't
do
very
much
some
kind
of
spring
application
and
you
just
checked
out
the
headers
are
correctly
setting
go
router.
So
possibly
we
can
also
use
another
application.
A
Okay,
release
process
for
cats
should
be
documented.
Carson
did
a
very
good
job
here.
So
now
we
have
in
the
CF
acceptance
tests
Wiki
release
documentation,
it's
basically
copied
over
from
the
CF
deployment
Wiki
and
the
script
down
a
little
bit.
So
now
everyone
should
be
able
to
create
the
CF
acceptance
tests
release
if
necessary
and
as
a
reward,
get
a
cronut
with
caramel
topping
yeah.
A
A
Api
check
test
fails,
this
is
merged
and
then
this
gets
automatically
integrated
in
I
think
the
jobs
who
use
runtime
CI
that
the
jobs
use
the
the
main
branch
of
runtime
CI.
So
if,
in
the
update
releases
pipeline,
the
API
Cloud
controller
health
check
fails,
we
should
see
a
better
error
message
before
it
was
the
HTTP
code
and
the
response
body
was
not
printed
yeah
small
Improvement,
then
the
gcp.
B
A
Gateway
change
for
bubble.
It
was
just
merged
40
minutes
ago.
That
is
very
nice,
so
next
steps
we
wait
for
I
know
next
step
is
something
I
can
do
this
tomorrow.
We
can
now.
A
Set
this
ephemeral
external
IP
flag
to
false
in
the
gcp
Bosch
Cloud
config.
It's
not
needed
anymore,
as
we
have
the
nut,
Gateway
and
yeah.
When
this
is
then
released
all
together
in
a
new
hi
cousin
in
a
new
Bosch
bootloader
version,
then
we
would
not
see
these
external
IP
addresses
anymore
in
our
gcp
deployments.
A
Yeah,
okay,
new
stacks
for
cloud
native,
build
packs.
We
discussed
this
briefly
in
Tuesday's
talk
meeting
and
will
basically
we
are
a
bit
early
here.
There
is
nothing
to
do
yet,
so
we
will
just
do
busy
waiting
until
it's
clear
which
Stacks
we
have
to
support
and
validate
okay
hi
Carson.
A
We
were
just
going
over
last
times
meeting
minutes
and
well.
Basically,
everything
is
done,
except
this
cat's
Zipkin
test
where
you
try
to
find
out
how
the
spring
sleuth
jar
was
built.
Is
there
anything
new.
B
No
I
had
a
chat
with
I
did
have
a
chat
with
Dave
to
see.
If
he
might
know
somewhere,
we
could
steal
an
app
from
he
recommended
that
we
just
rewrite
the
thing
and
go
I
still
haven't
tried.
Just
extending
the
timeout
as
you
recommended.
B
However,
I
have
noticed
that
it's
been
flaking
less
than
before,
but
still
yeah
would
like
to
get
it
fixed
for
sure.
Yeah.
A
A
A
Good
okay,
so
that
were
the
two
twos
from
last
time
for
today,
I
have
only
one
point
on
the
agenda
so
far.
Unless
you
have
something
for
the
cats
pipeline,
we
deploy
Cloud
Foundry,
run
the
smoke
tests,
the
cats
unit
tests
and
then
the
Cloud
Foundry
acceptance
tests.
A
I've
noticed
that
there
is
no
destroy
CF
job,
so
the
cloud
Foundry
deployment
keeps
running
and
I
just
wanted
to
save
some
codes.
Would
it
make
sense
to
do
a
destroy
CF
here
at
the
end,
as
we
do
for
most
other
validations
too?.
B
We
could
my
understanding
is
that
the
cats
doesn't
get
torn
down,
because
both
the
smoke
tests
and
the
cats
pipelines
use
it,
and
neither
of
them
neither
of
them
actually
care
about
deploying
cfd.
So
to
make
this
that
both
those
pipelines
quicker.
They
just
have
this
beginning
part
of
cats
that
anytime,
a
new
cfd
version
is
created.
A
B
If
we
want
to
change
this
to
clean
it
up,
we
can
we're
sacrificing
a
little
speed
in
the
pipelines.
That's
the
trade-off,
but
honestly
I,
don't
think
we
get
enough
changes
to
where
that's
going,
to
make
a
difference
so
I'm
personally
trying
to
to
spit
it
down
we're
just
going
to
have
to
rework
both
the
cats
and
the
CF
smoke
tests
pipeline.
Not
just
the
cats
pipeline.
B
Test
and
release
that,
if
you
click
on
that
group
yeah,
it
uses
the
cat's
environment,
they
share
an
environment.
Okay,.
A
A
A
No
one
has
complained.
It's
just.
Maybe
that
I'm
transferring
Sap's
cost
saving
Behavior
now
to
our
accounts.
I
think
it
makes
sense
right.
A
Okay,
it
would
make
sense,
I
mean
we're
doing
it
for
for
all
other
environments
as
well,
so
everything
every
deployment
for
update
releases.
We
do
even
Bubble
Up
and
Down
each
time.
But,
okay,
that's
not
a
problem,
as
there
is
Maybe
One
release
update
per
day
and
for
the
CF
deployment
environments.
We
do
always
abortionally
deployment,
your
okay
good,
then
we
should
also
try
to
do
it
for
cats.
I
will
great
tasks
and
check
if
there
is
anything
else
to
do.
If
this
we
don't
get,
how
many?
A
Well,
every
two
or
three
days,
maybe
yeah
and
pretty
much
it
gets.
Dependable
appears
also
about
one
or
two
per
day
or.
B
If,
if
we
do
this
refactor,
it
might
be
worth
considering
giving
smoke
tests
its
own
environment
so
that
those
two
pipelines
aren't
always
blocked
on
each
other.
If
we
are
introducing
a
complete
spin
down,
then
it
it
doesn't.
Increase
costs
that
much
more
to
add
one
more
environment,
I
think
that's
usually
in
the
director
State
and
we're
talking
about
potentially
holding
up
one
of
the
two
pipelines
by
I.
Don't
know
like
45,
you
know
like
two
hours
while
we
run.
A
Environment
in
that
world
a
new
bubble
end
for
smoke
tests,
we
just
need
a
domain
and
then
we
have
the
jump
box
in
the
director.
So
we
should
do
this
first
of
course,
and
then.
B
A
B
A
Nuts
has
a
major
upgrade,
but
as
I
understand,
it
should
not
affect
us
too
much.
They
are
adding
some
yeah,
some
checks
that
in
in
post
start
and
here's
the
smaller
PR.
So
when
this
comes
in,
we
should
check
if
there
are
any
problems
while
upgrading,
not
yeah,
and
then
there
was
this
smaller
issue
with
the
here
is
a
bit
chaotic.
A
There
was
a
client
of
extended
key
usage
missing
for
one
of
the
certificates,
and
yes,
this
is
a
bit
messy,
so
we
have
certificates
with
both
client
server,
some
only
server,
some
only
client
and
yeah.
It
was
not
so
apparent
that
this
was
missing,
because
it
also
depends
on
the
TLs
client
you're
using,
but
okay,
this
will
be
easily
fixed
yeah
thanks
to
Johan.