►
Description
Meeting notes: https://docs.google.com/document/d/1y7a6uaN8fwFJ7IRnvtxSfgz-OGFq6u7bKN6F7NDxKPg/edit
Timecodes ⏱
00:00 Intro
01:14 Gitlab + Keptn presentation by Nicas Mietz @solidnerd
30:01 FAQ
36:42 Community Contributions
Learn more: https://keptn.sh
Get started with tutorials: https://tutorials.keptn.sh
Join us in Slack: https://slack.keptn.sh
Star us on Github: https://github.com/keptn/keptn
Follow us on Twitter: https://twitter.com/keptnProject
Sign up to our newsletter: https://bit.ly/KeptnNews
A
Hi
everyone
and
welcome
to
this
edition
of
the
captain
community
meeting
today
is
october,
14th
and
yeah.
As
always,
please
indicate
your
attendance
by
adding
yourself
to
the
attendee
list
without
further
ado.
I
will
actually
already
hand
over
to
our
first
agenda
item
this
week
is
kubecon,
so
there
is
a
lot
of
people
are
also
at
cubecon
everyone
that
is,
there
have
fun.
It's
I'm
sure.
It's
it's
a
great
event.
A
Also,
this
year,
it's
it's
a
hybrid
event,
there's
also
some
captain
activities
going
on
there.
Probably
we
can
share
more
in
the
next
meeting
of
the
the
learnings
of
kubecon
and
and
what
what
we
have
our
conversations
there
and
what
we've
learned,
but
I'm
already
handing
over
to
nicholas
to
present
his
work
on
from
the
latest
weeks
or
even
months
on
on
combining
kid
club
and
captain
so
handing
over
to
you
and
thanks
for
joining
us
today,.
B
Yeah
yeah,
probably
a
short
introduction
in
the
meantime,
where
I
tried
to
share
my
screen
yeah,
so
my
name
is
nicholas:
I'm
working
for
company
who's,
doing
mostly
buzzwords,
so
I
do
blockchain
and
kubernetes,
but
that's
a
different
topic
where
not
today
for
here-
and
I
am
also
a
ditler
hero,
so
I
do
really
like
to
work
with
gitlab
and
all
the
stuff,
and
today
I
want
to
share
mostly
the
results
that
we
get
over
the
last
one.
So
it's
like
an
ongoing
activity.
B
So
looking
a
little
bit
in
the
past,
we
started.
I
think
this
one
year
ago,
mostly
so
the
most
initial
work
was
done
by
christian
hetermann.
So
it
took
a
boost
a
lot,
so
he
did
the
initial
version
and
yeah.
Then
we
figured
it
out
after
some
meantime
where
we
are
so
I've
also
presented
some.
B
I
bring
up
also
some
slides
to
get
a
little
bit
overview,
what
we
are
doing
today
or
where
we
are
currently.
So
I
made
it
through
a
change
here,
so
it
of
course,
in
the
heart
and-
and
here
we
declare
blower
so
that
we
have
everything
important
in
place
and
yeah.
So,
let's
dive
into
the
topic-
so
mostly
probably
all
of
you-
are
already
familiar
with
captain
in
this
meeting,
and
I
don't
need
to
tell
about
what
return
to
the
use
case
are.
B
So
mostly,
we
will
extend
the
use
case
in
terms
of
also
using
gitlab
for
the
part
of
delivery
and
probably
later
in
the
future,
also
more
but
it's
another
topic.
So
when
I
talked
about
the
others,
an
ongoing
project
yeah,
it's
really
a
long-range
project,
because
when
is
the
diagram
big
enough,
can
you
see
it
or
should
I
make
it
a
little
bit
bigger
because
fingers
which
I'm
interested
in
a
short
trick?
B
So
when
you're
coming
from
a
pastor,
as
I
said,
it's
like
an
android
practice
and
we
started
initially
with
version
0.7
with
gitlab
and
captain
mostly,
and
we
had
a
lot
of
different
services,
so
also
the
potato
head
was
not
there
in
the
meantime
and
we
have
also
two
different
prometheus
servers.
So
we
had
the
idea
so
or
mostly
when
christian
was
presenting
it.
He
showed
all
the
monitoring
with
dyna
trace
to
this
point,
and
I
was
really
curious
about
that.
B
Oh,
I
really
use
mostly
on
my
daily
work
and
also
my
free
time
only
prometheus
for
monitoring,
and
I
said
okay.
It
would
be
probably
a
good
idea
to
have
an
example
working
with
prometheus
so
and
in
the
end
we
started
to
working
on
that
so
first,
taking
up
at
least
the
old
version
with
version
seven.
B
So
then
the
time
flies
a
little
bit
and
we
had
a
problem
with
it
of
updating,
also
captain
to
the
version
to
the
latest
version
and
that
there
we
are
not
trying
to
achieve
in
the
end
so
also
in
this
meeting.
What
I
want
to
show
here,
so
what
we
are
now
the
current
state
is
yeah.
It's.
A
Updated
sorry
for
interrupting,
but
can
you
go
into
presentation
mode
because
then
it
will
be
a
little
bit
bigger.
B
B
C
B
Yeah
so
because
I
have
a
4k
monitor,
I
was
not
sure
if
it's
not
too
small
and
that's
the
reason
why
I
didn't
share
the
full
stream
yeah,
and
now
we
have
a
lot
of
updates,
mostly
also
in
the
captain
community.
So
it
means
like
we.
Switching
back
from
the
prometheus
sli
service
was
in
the
past
system,
similar
service,
and
we
had
the
prometheus
service
in
the
end,
and
this
mostly
now
converted
into
a
single
service
and
yeah.
That
means
that
okay,
we
need
to
do
also
some
changes,
mostly.
B
What
is
also
out
today
of
scope.
A
little
bit
is
the
gitlab
service.
We
will
currently
even
also
to
update
it,
to
the
latest
version
to
re
accept
all
the
new
event
files
in
this
case
so,
but
I
would
can
show
you
where
the
direction
is
heading
to
and
also
what
we
currently
will.
Change
in.
B
Video
is
mostly
that
we
also
go
in
using
instead
of
j
meter,
we're
starting
doing
the
road
test
with
with
low
cost
in
the
end
so
yeah,
that's
it
yeah.
So
mostly,
then,
it's
the
interesting
part
in
terms
of
watching
didn't
take
place
so
also
in
our
example.
So
we
use
the
concept
architecture.
So
I
added
the
club
where
I
see
currently
everywhere,
so
we
could
use
it
also
leveraging
the
little
detox
agent
for
that.
But
it's
also
another
story.
B
Mostly,
we
will
use
today
a
little
perfect
container
registry
and
also
for
the
continuous
delivery
in
terms
of
we
will
deploy
directly
from
git
lab
into
our
turbinators
cluster
and
then
will
captain
take
place
in
terms
of
doing
the
load,
testing
and
doing
this
stuff
of
getting
the
promiscuous
results,
and
so
on.
So
for
that
I
have
prepared
a
demo,
so
I
will
switch
a
little
bit
now.
B
So
let
me
check
yeah
cool
currently,
as
you
can
see
here,
so
we
have
an
own
github
instance
on
that
a
two
minutes
cluster
with
the
club
instance
where
we
can
see
at
least
the
the
hardening
stage,
so
our
project
in
ecc-
and
we
want
to
take
a
deeper
look
in
here.
So
mostly
we
have
currently
only
one
stage.
We
have
currently
seen
that
the
quality
date
is
failing,
but
that's
okay
for
the
current
state
yeah.
B
So
mostly
the
more
interesting
part
is
note
in
terms
of
no
switching
page
to
detail.
What
did
lab
is
in
the
end
doing
so
I
have
so
all
the
repositories
are
open
source,
so
everyone
can
see
it
so
they're
coming
from
where
everyone
can
contribute
groups,
so
you
can
go
under
captain
and
there
are
a
lot
of
repositories.
The
most
interesting
repository
right
to
start.
It's
mostly
the
captain
of
a
portrait
head
captain
demo
repository
while
we
really
deploying
the
application
and
starting
to
trigger
in
the
captain
pipelines.
B
So
for
that,
as
you
probably
familiar,
also
a
little
bit
with
gitlab.
So
this
is
a
pro
a
project
and
we
have
currently
our
all
the
icd
configuration
is
written
in
the
file
stored,
ci
yammer.
So
what
we're
doing
in
the
end,
the
the
configuration
so
and
for
that
we
have
now
different
stages.
So
we
will
look
now
into
the
parts
of
doing
the
build
the
testing
the
captain
onboarding.
So
it
means
like
okay,
when
you
start
up
a
new
project
in
detail.
Captain
doesn't
know
about
it.
B
For
that
case,
you
need
to
do
also
captain
onboarding
from
the
cicd
system,
so
that
everything
goes
in
the
place
of
having
the
captain
on
board
so
that
we
can
start
the
deployment.
So
for
that
we
will
start
them
the
onboarding.
The
onboarding
is
mostly
a
simple
script,
so
in
gitlab,
what
we're
mostly
using
here
is
a
concept
of
templates,
so
we
have
different
predefined
gamuts
in
terms
of
helping
you
to
set
up
easier,
the
icd
pipeline
stuff
and
so,
for
example,
for
config.
We
have
these
onboarding
yaml.
B
What
we
can
include
in
the
end-
and
this
will
do
various
steps
in
terms
of
getting
the
captain
project
on
board
it,
pushing
the
shipyard
yama,
pushing
all
the
resources
that
I
currently
required
in
terms
of
prometheus
and
sli
in
this
case,
and
that's
the
mostly
part
about
the
onboarding.
So
what
we
now
would
start
the
pipeline
in
the
end,
we
can
do
that,
let's
trigger
a
pipeline
in
the
project,
so
we
will.
B
Do
a
short
trigger
and
also
what
is
probably
a
little
bit
different
to
the
other
cluster.
So
it's
like
this
is
the
current
upstream
potato
head
repository,
so
I
made
only
a
small
change
in
the
I
added
a
little
folder
there,
so
you
have
to
typically
develop
the
icd
example
there.
So
it's
only
the
fake,
it's
mostly
the
same
repository
like
everything
else,
and
but
what
we
here
have
is
like
a
chart
for
doing
like
the
deployment
of
a
potato.
B
So
what
we
will
do
in
the
end
is
when
we
trigger
in
the
pipeline.
B
We
can
see
here
now
that
we
are
not
spraying
on,
but
we
can
go
in
the
previous
drop,
so
we
were
building
our
docker
image.
Then
we
starting
the
onboarding
process
in
the
end
in
terms
of
registration,
the
whole
application
to
captain
and
then
afterwards
we're
starting
the
deployment.
So
our
deployment
mechanism
is
now
not.
The
captain
will
do
the
job
for
us,
so
gitlab
will
do
it,
and
for
that
we
were
leveraging.
Mostly,
we
were
starting
with
commands.
B
We
letting
captain
knowledge
here
that
the
deployment
sequence
is
started,
we'll
see,
so
this
will
be
done
through
a
script.
I
can
will
show
you
later
how
the
script
is
working
afterwards.
We
starting
the
deployment
started
event
so
that
we
are
in
the
sequence
and
then
we're
starting
a
normal
helm,
upgrade
so
we're
doing
the
application
deployment
with
a
help
file
so
mostly
because
in
our
company,
working
mostly
with
them-
and
I
really
like
the
approach
and
stops
of
pd
clip-
is
doing
the
pushing
of
the
application
and
yeah.
B
So,
in
the
end,
when
the
whole
deployment
is
finished,
we're
starting
we're
sending
also
the
deployment
finished
event
where
we
captain
recognize
the
deployment
is
done,
so
we
can
go
into
the
next
sequence.
So
now,
when
we
are
looking
into
the
sequences,
so
you
can
now
see
here
that
the
deployment
was
triggered.
B
You
can
see
here
also
that
the
event
was
started
mostly
and
now
you
can
see.
Okay,
also,
the
department
has
finished
so
now
locross
will
be
informed
in
terms
of
server
onboarding.
Also.
What
is
you
know
interesting,
as
you
probably
saw
in
the
previous
version
of
the
the
club
captain
integration?
B
We
have
also
the
support
of
labels
here
so
that
you
can
put
meta
information
for
your
testing
or
also
for
other
services
later
so
in
terms
of
probably
when
you
are
here
now
in
cap,
you
want
to
back
which
pipeline
is
currently
running.
You
can
click
here
on
the
ci
backlink
and
then
you're
coming
back
directly
into
your
ci
cd
pipeline,
where
currently
the
deployment
was
done.
B
So
that's
really
cool
and
we
can
add
a
lot
of
more
services
in
the
end.
B
B
B
Yeah
yeah,
it's
it's
nothing
like
a
special
service.
I
will
come
yeah,
probably
it's
a
good
state
now,
because
lotus
kicks
in
regarding
the
source
service.
As
I
said
we
have,
I
have
built
it
up
scripts,
mostly
to
do
that
so,
and
we
have
here
the
captain
doctor
repository
where
I
have
like
really
simple
wrapper
scripts
in
terms
of
sending
captain
events
to
get
started
to
caption
it
now
when
a
deployment
has
started
and
so
on.
B
So
mostly
here,
we
starting
in
terms
of
our
deployment
so
we're
sending
here
the
full
event
and
also
here's
where
the
source
event
is
coming
from.
So
we
can
specify
here
the
source
and
then
we
can
check
it's
the
typical
json
out
of
a
captain
spec,
mostly
with
all
provided
values
in
terms
of
to
triggering
view
events,
so
that
the
event
will
be
treated
in
the
correct
stage.
Afterwards
we
centered
this
file
with
the
context
id.
B
We
need
to
store
the
context
because
it's
important
to
be
in
the
same
secrets
in
the
end,
so
for
that
we
now
can
start
with
the
deployment
started.
So
for
that
we
have
the
department
started
deployment,
and
this
is
really
doing
the
same.
It
does
a
little
bit
more
magic
or
not,
I'm
not
saying
magic
in
the
end,
but
what
it
does
in
the
line
27
to
line
30
it
waits
until
because
captain
I
have
doing
with
development.
B
I
found
autocad
currently
re-triggering
the
delivery,
but
the
problem
is
that
the
deployment
event
will
be
generated.
So
I
need
to
wait
for
that
and
that's
the
reason
why
I
loop
here
mostly
in
the
captain
group,
so
I'll
wait
here
until
this
events,
starting
so
I
check
with
I
get.
I
fetch
all
the
events
from
the
captain
cli
and
then
doing
with
a
simple
jq
check,
checking
when
the
deployment
trigger
id
is
there,
because
I
need
to
get
this
id
in
the
end
to
handle
the
deployment
started
event
in
the
correct
order.
B
Otherwise
the
flow,
the
sequence
will
never
finish.
That
was
a
problem
that
I
found
out
after
a
little
bit
debugging
and
so,
and
here's
where
also
the
source
here
is
the
source.
So
we
can
also
put
here
in
eclipse
iso
for
better
name.
Okay,
that's
a
good
mark
and
also
then
you
can
see
here
all
the
labels
and
that
we
saw
in
the
chapter
ui
are
also
here
mostly,
we
can
define
them
mostly
in
the
way.
That's
used
interesting
that
these
sheets
are
mostly
ignored
in
some
way.
B
We
should
remove
them
afterwards
and
then,
when
we
are
finished
with
the
deployment
started
event
we
can
determine,
we
will
do
it
in
vci
cd,
mostly
when
we
starting
with
the
gitlab.
B
So
we
have
here
twitter
now
the
deployment
starting
executing
the
hem
upgrade
process
after
the
hem
upgrade
process
was
done.
We
are
sending
the
deployment
finished
to
cabins
for
captain
now
cp.
Now
we
need
to
go
into
the
next
sequence,
to
ask
mostly
so
and
then
automatically
the
test
stage
will
be
triggered
to
configure
in
our
stages
and
then
we
starting
the
evaluation,
because
the
evaluation
comes
in
terms
of
from
the
configuration
mostly.
B
So
we
have
also
probably
forget
that
in
the
beginning,
when
we
going
back
into
delivery,
gitlab
going
into
the
shipyard
diamond,
so
we
passed
now
mostly
the
deployment
stage
now
we're
on
the
testing
stage,
and
now
we
are
in
the
evaluation
phase.
B
And
so,
as
I
said,
for
evaluation,
I
really
like
want
to
use
prometheus,
and
mostly
these
days,
a
lot
of
people
set
up,
not
a
similar,
simple
prometheus
service
when
you
arrange
in
cubanitas
you're,
probably
also
using
the
prometheus
operator
in
the
end,
to
do
the
full
deployment
so
that
what
was
also
outdoor
okay.
So
we
want
to
have
a
really
self-burned
service
in
terms
of
so
that
we
can
also
leverage
the
prometheus
operator
for
that.
B
So
I
will
switch
now
yeah,
because
now
we
see
also
here
the
results,
so
we
can
also
check
okay.
The
evaluation
is
not
currently
failing,
but
this
is
mostly
tight,
so
we're
getting
here
the
values
correctly
back,
but
I
assume
that
something
with
the
percentage
is
not
correct
the
evaluation
so
that
something
is
still
affecting
in
the
prometheus
source,
because
I
got
a
value
about
98
percent
or
it
depends
on
how
the
past
youtubers
implemented.
B
So
they
have
currently
not
for
clue
why
this
is
not
evaluated
correctly,
but
I
will
we
will
figure
it
out
afterwards.
So
yeah.
C
Nico
yeah,
can
it
do
me
a
favor?
Oh,
can
you
do
me
a
favor
and
jump
to
the
column
that
is,
you
know,
highlighted
on
the
left
or
supposed
to
know
no
in
the
heat
map.
There
is
this
screen
to
the
left.
Please
there
is
this
cream
color
or
the
the
last
cream.
A
Date,
2021
and
may
16th,
so
this
will.
A
C
In
the
heat
map,
as
jung
said
and
pointed
out,
just
select
the
column
with
the
date
16th
of
may
12,
25.
A
If
you
go
left
here,
yeah
left
it's
it's
in
the
first.
Yes.
C
A
B
Yeah
this
there
was
doing
the
testing,
so
I
just
started
independently
evaluation
yeah,
and
here
we
have.
C
A
value
of
zero
and
at
the
current
value
that
we
measure
is
above
it's
0.9,
and
this
is
an
increase
of
yeah
more
than
100
percent.
That's
the
reason
why
this
sli,
currently
failing.
B
Why?
Because
for
rt,
because
it's
a
smaller
liquids
and
we
are
increasing
greater
than
100
percent-
okay
yeah,
then
I'd
rather
here
and
it's
yeah.
I
need
to
fix
in
our
ship
here.
So
our
potato
head
service
is
really
good
in
terms
of
improving
and
for
testing,
but
thanks
and
I
have-
and
why
yeah
and
mostly
now,
coming
to
interest
point
in
terms
of
getting
all
the
metrics
data
here.
So
I
set
up
the
prometheus
operator.
B
We
can
see
here
also
that
it's
running
currently
the
prometheus
operator
is
deployed.
So
here
we
have
a
different
concept.
So
mostly
you
would
not
write
directly
in
your
promiscuous
theorem
of
the
changes.
B
So
the
previous
operator
has
the
concept
of
a
service
monitor
and
the
service
monitor,
can
scrape
multiple
kubernetes
services
for
that
in
the
metrics
and
the
hard
finding
or
the
most
interesting
point
out
of
that
was
okay,
the
drop
name
mostly
because
the
permission
service
has
currently
resumption
so
that
the
the
job
needs
to
be
suffixed
with
dash
canary.
B
We
also
need
to
rename
the
standard
pattern
in
the
service
monitor
a
little
bit,
but
I
will
show
you
now:
how?
Technically,
it's
working
so
because,
mostly
when
we
are
looking
going
back
into
the
so,
for
example,
now
fetching
the
request
of
the
of
a
potato
head.
So
we
can
see
it
currently,
240
requests
were
done
and
also
it's
coming
from
veterinary
service
and
the
potato.
It
was
mostly
tested
so
and
this
the
reason
where
the
value
is
coming
from.
B
So
for
that
I
was
switching
on
my
screen,
but
I
need
to
find
my
there
live
in
one
second.
B
Cool
so
yeah,
you
see
a
little
bit
more
here.
I
configured
also
the
community
service.
Mostly
now
we
can
look
into
that.
So
we
have
here
a
normal
checkpoint
set
up.
Mostly.
We
have
here
also
the
prometheus
service
configuration
so
for
that
and
when
you
look
into
the
where
currently,
where
our
previous
operator
is
living,
the
prometheus
operator
is
living
in
another
namespace,
so
in
the
tube
prometheus
state
namespace.
B
So
currently,
here
prometheus
is
living
with
operator
and
all
the
stuff
for
monitoring
incentive,
operator
namespace
and
now
we
need
to
instruct
two
parts.
So
first
off
would
be
to
let
the
application
now
so
the
previous
know,
which
application
needs
to
be
scraped
so
and
for
that
we're
using
a
custom
resource
export
service
monitor-
and
this
is
mostly
placed
in
this
namespace,
so
the
potato
head
name
space
and
we,
when
we
are
looking
on
that,
I
mostly
make
a
small
change
on
that.
B
Cool,
I'm
sorry,
that's
true,
and
currently
I
define
here
like
the
service,
so
this
is
like
the
instruction
which
22
should
be
straight
and
you
can
also
dictate
which
label
should
be
used,
and
I
tricked
it
a
little
bit
out
in
terms
of
that.
I
used
here
that
by
default
it
would
use
the
service
name,
but,
as
I
said,
we
need
to
have
this
dash
prefix
canary,
I
said,
okay
use
a
label
from
the
service
and
the
label
is
called
captain
drop.
So
when
I
now
look
into
the
service.
B
You
can
see
yeah
here's
the
name
of
the
jobs,
that's
the
reason
why
it's
so
hardcoded
in
the
end-
and
this
is
also
in
the
handshake-
to
put
veterinary
field
on
that
for
the
premium
so
so
that
it
can
scrape
the
events.
B
So
now
we
have
this,
then
the
next
interesting
point
to
see
how
now
the
prometheus
source
gets
the
data.
So
the
previous
service
lives
in
the
tap
namespace.
It
has
a
little
different.
B
Configuration
so
here
we
can
see
so
mostly
the
biggest
impact
is
so
that
I
pointed
out
to
the
existing
previous
instance
and
for
premier
uses
namespace,
but
everything
by
that
is
anything
else.
It's
like
the
default,
so
there's
nothing
changed
in
terms
of
that
the
slis
are
worked
on.
It
doesn't
work
in
the
end,
so
these
the
most
change
that
you
need
to
do
to
have
it
up
and
running
so
the
premium
endpoint
is
also
important
here.
This
is
the
default
one.
B
You
can
also
overwrite
it
with
a
secret
so
that
you
can
say
also
for
your
project
to
mention
a
secret
with
another
prometheus
endpoint.
You
can
also
configure
that,
but
for
easier
bootstrapping.
I
said
the
prometheus
operator
is
like
the
default
instance
for
every
service
that
we
use
in
the
end,
and
that's
it
mostly
from
the
technical
part
to
get
everything
up
and
running
with
the
previous
operator
in
gitlab
and
captain.
But
let's
look
back
into
the
stage
of.
B
In
my
turn
now,
but
give
me
one
second
now
close
the
transition
either:
okay,
yeah
so
and
that's
all
the
data
that
we
have
now
in
terms
of
now,
we
could
assume
until
we
could
probably
go
on
into
the
next
stage
with
it
up
and
running
mostly
here.
We
to
see
it
now
from
the
with
all
the
concepts.
It's
mostly
right.
Okay,
gitlab
is
doing
the
job
of
deploying
our
applications
so
doing
the
deployment,
and
then
we
could
coming
back
into
a
later
point.
B
Okay,
we
could
say
that
also
gitlab
gets
notified
and
run,
and
you
probably
a
new
pipeline,
for
example,
doing
the
next
release
promotion
all
the
stuff.
For
this
we
have
faded
lab
service,
but
the
gitlab
service
needs
to
be
a
little
bit
polished
again
in
terms
of
that,
it
need
to
listen
to
the
correct
events,
but
that's
another
topic.
B
So
when
we
are
looking
back
going
into
our
slides,
so
we
saw
literally
the
whole
flow
in
terms
of
what
gitlab
is
currently
doing
it
doing
the
deployment
of
a
potato
head.
We
saw
the
methods
of
the
sli
metrics.
The
prometheus
server
is
connected.
B
The
promises
operator
controls
the
prometheus
and
the
log
manager
and
mostly
having
a
full
flow
of
deploying
all
the
stuff
yeah,
so
the
more
interesting
parts
what's
coming
now
up
next
is
really,
of
course.
I
added
here,
of
course,
for
delivery.
Vid
lab
icons,
so
mostly
in
the
slides,
so
everyone
is
aware
of
that
yeah,
but
for
more
interesting
parts.
B
B
So
the
next
steps
are
mostly
bringing
up
the
quality
baked
quality
dates,
some
information
from
the
quality
dates,
also
in
gitlab
again,
and
the
bigger
path
is
mostly
to
bring
the
captain
gitlab
servers
back
to
life,
so,
for
example,
that
we
can
fulfill
the
use
case
of
trading
when
the
deployment
from
the
deep
lab
servers
to
gitlab.
B
So
it
says:
okay,
please,
deploy
an
overproduction
stage
so
that
we
can
also
have
multi-stage
delivery
from
gitlab
instead
of
using
this
listen
captain
so
also
re
reimagination
is
the
same
concept
in
terms
of
deployments
only
being
back
and
also
would
be
really
cool.
What
I
see
mostly
what's
important
so
from
my
products
currently
also
is
like,
so
that
some
automated
responses
in
merch
requests
open.
B
So
when
you
would
do
like
a
testing
phase,
the
test
results
will
be
put
into
directly
in
the
mercury
tests
and
so
that
you
can
easily
see
everything
in
the
club
ecosystem
and,
of
course,
I
think
the
next
bigger
step
for
the
community
is
here
so
to
have
like
more
guided
walks
through
and
how
to
set
up
the
current
setup
that
we
have
so
that
everyone
can
try
it
out,
because
mostly
it's
right
in
a
state
where
it's
working,
but
it's
not
like
no
in
a
repeatable
way
working.
B
So
most
of
terms,
it's
a
lot
of
scripts,
but
more
like
that.
It
works
over
like
what
I
currently
can
see
in
terms
of
waypoints
yeah
and
that's
it's
mostly
from
videoclip
integration
of
the
current
state
of
integration.
C
B
Used
the
latest
version,
but
I
need
to
do
a
change
on
that.
I
will
send
up
also
pull
requests,
because
currently
the
previous
server
is
probably
not
working
because
it
started
up
to
service
on
port
8080
because
there
will
be
an
http
server
started
and
then
the
cloud
event
servers-
and
this
is
currently
a
blocking
part.
But
I
will
send
up
a
pull
request
to
fix
it.
Cool
awesome,
yeah.
B
A
Pretty
cool
to
see
this,
this
working,
and
also
now
we
have
a
it's
not
yet
that
here
on
the
slides,
we
see
that
the
guy
did
walk
through,
but
we
can
already
see
it
in
action.
We
already
have
now
everything
if
someone
wants
to
to
rebuild
it.
D
B
Yeah,
I
would
say,
probably
in
terms
of
cli,
would
be
interesting
to
have
like
something
in
the
direction
with
me,
pointing
back
to
my
slides
and
going
back.
So
we
could
think
about
community
would
be
probably
cool
to
have
like
the
events
so
that
we
could
template
the
events
in
some
way,
but
I
think
it
should
be
also
complex
in
terms
of
advertise,
because
I
really
leveraging
the
captain's
cri
everywhere,
where
I
can
so
mostly
that
I
don't
need
to
do
anything
of
using
other
parsing.
B
But
what
be
a
crude
other
cri
feature
is
mostly
to
have
also
more
light,
or
probably
are
not
aware
of
that.
So
when
I
checking
for
events
so
that
I
can
select
to
a
different
type
of
events
instead
of
that,
I
need
to
do
like
via
jq.
We
are
parsing
the
correct
event
type
to
get
the
event
yeah.
That's
the
current
cli
stuff
that
I
currently
can
see
iphone.
D
You
can
pass
the
event
type
in
here,
okay,
but
there
there
is
different
options.
So
maybe,
depending
on
what
you
choose,
you
might
not
be
able
to
provide
some
other
options,
but
I
think
it's
possible.
D
Nevertheless,
I
think
the
event
sending
itself
is
something
we
should
consider
like.
I
mean
in
this
case,
like
the
data
has
project
stage,
service
etc.
Like
you
will
repeat
this,
this
is
a
repeated
pattern
right
same
for
labels.
Maybe
an
option
to
just
provide,
let's
say
minus
minus
values,
similar
to
what
you
can
do
in
helm,
when
you're
doing
an
upgrade
to
helm
install
could
help,
because
then
you
would
avoid
having
to
write
spec
versions,
source
type,
data
content,
type
etc.
D
So
like
just
really
sending
data
attributes
in
form
of
a
values
file
or
whatever.
That
would
probably
reduce
your
a
lot
of
code
from
your
pipeline
yeah.
B
Yeah
for
sure
yeah
yeah.
That
was
also
the
most
challenge,
because
when
I
were
in
the
version,
0.7,
okay,
the
events
was
a
little
bit
different
and
then
I
need
to
understand
the
concept.
Okay,
you
need
to
trigger
this
and
this
when
I
would
do
it
like
in
a
higher
programming
range
in
dolan,
or
something
like
that.
It
would
be
a
little
bit
easier
so
but.
D
B
Okay,
yeah,
but
anything
else
runs
mostly
smoothly,
so
they're
currently
counts.
The
improvement.
B
You
it
will
be,
it
will
be
really
interesting,
so
I
will
probably
pick
up.
I
currently
can
don't
give
an
ata
on
that,
but
I
will
probably
pick
up
the
word
now
again
on
widget
club
servers
and
tunes
is
also
a
little
bit
up
so
that
we
get
it
also
like
for
the
usage
that
we
have
already
to
have
it
also
up
and
running
again,
because
I
think
this
resembles
more
value
in
terms
of
that
you're
3d.
B
It
could
be
then
two
systems,
but
if
you
have
all
the
information
in
one
system,
mostly,
for
example,
also
for
developer
focus
stuff,
it's
really
true,
so
that
you
can
add
this
yeah.
I
think
this
is
a
big
change
in
cabinet
all
right,
so
because
we
have
too
many
services
or
we
have
a
lot
of
services
and
getting
all
the
integration
mostly.
B
The
event
handling
is
quite
different
in
terms
of
when
you
probably
work
in
a
zoomed.
B
D
Also
regarding
the
gitlab
service,
as
you
mentioned,
this,
please
check
the
current
development
we
have
with
captain.
We
have
a
web
hook
service
that
we're
working
on
and
it
will
allow
just
sending
post
get
whatever
kind
of
http
request
to
anywhere.
So
maybe
the
gitlab
service
might
not
be
the
primary
point
of
implementation
anymore,
and
it
could
be
used
like
via
the
web
hook.
Service
directly
would
make
it
easier
for
the
end
user.
Obviously,
because
you
don't
have
to
install
a
separate
service
yeah
I
can.
D
I
can
provide
you
an
example
for
that
with
jenkins.
If
you
want.
B
A
Yeah
cool,
thank
you
so
much
any
other
questions.
A
Okay,
then,
I'm
thanks
again
nicholas
for
presenting
this
really
cool,
really
cool
stuff.
Then
I'm
heading
over
to
I
don't
know
who
is
actually
working
on
the
next
agenda
item.
The
next
review
will
be
presented
using
captain
on
captain.
Anyone
who
wants
to
comment
on
this
or
put
this
on
the
agenda.
A
Okay,
then
I
will
put
it
actually
at
the.
I
just
say
here
reminder:
oh
sorry,
I
was
not
sharing.
Unfortunately,
I
need
to
jump
because
yeah
today,
I'm
in
in
another
dynatris
lab,
I
will
hand
over
the
host
and
I
think
I
will
also
hand
over
to
suraj
for
the
next
agenda
item.
So
if
you
want
to
discuss
the
prs,
please
go
ahead.
E
E
Okay,
so
I
think
the
first
one
florian.
E
E
Now
you
can
read
about
more
about
the
issue
and
and
more
about
the
pr
industry
and
yeah.
I
just
picked
up
this
issue
yesterday,
where
we're
basically
switching
from
cli
config
manager,
which
is
a
struct
that
we
use
to
read
in
captain
config
and
using
switching
it
with
wiper.
Instead,
we
were
using
wiper
in
the
root
cli
command,
but
I
removed
it
and
then
this
is
an
issue
to
add.
B
E
So
the
earlier
bru
is
an
implementation
of
cli
manager
in
the
root
command
as
well.
It's
a
different
year
altogether,
so
I
wouldn't
talk
much
about
it,
but
it's
basically
replacing
the
cli
config
managers
implementation
using
wiper,
so
the
pr
is
a
work
in
progress.
So
I
think
I
started
working
on
this
yesterday
and
one
more
thing
was
new:
maintenance
maintenance
do
not
appear
in
the
captain
organization,
so
I'm
not
sure
if
that
is
how
it
should
be,
but
I
think
jurgen
said
something
along
the
line.
E
C
Okay;
okay,
thanks
for
for
pointing
out
the
situation
that
new
members
or
maintainers
don't
show
up,
can
you
please
share
again
your
screen?
Sorry,
sorry,.
C
Yeah,
and
can
you
show
us
who
shows
up
in
this
list.
E
C
E
So
I
think
this
is
when
it
probably
started
so
the
last
started
member
was
anna,
I'm
not
sure
if
I'm
now
seeing
that
indeed
does
not
appear
in
the
members
list.
C
E
C
E
C
Cleaned
it
up
thanks
for
pointing
out-
and
we
will
see
you
get
this
fixed.
C
All
right,
then,
let
me
just
take
it
over,
but
actually
finish
this
session,
because
we
don't
have
any
additional
item
on
the
agenda.
So
I'm
asking
if
there
are
questions
from
someone.
C
Okay,
silence
not
the
case,
then
again,
thanks
nico
for
presenting
yeah
the
power
demo
of
of
how
you're,
using
captain
it's
always
cool,
to
see
the
use
cases
and
the
way
that
yeah
users
are
stretching
captain
for
their
for
implementing
their
use.
Cases
really
awesome
and
if
you
are
kind
and
could
file
the
pr
for
the
producer
service
yeah,
we
would
really
really
appreciate
that.
C
Okay,
thanks
for
joining
thanks
for
listening
in
this
week
stay
tuned
next
week
we
will
go
over
new
features
and
what
have
been
accomplished
this
week
and
the
next
one
yeah,
so
that
we
will
see
the
new
enhancements
that
have
been
implemented.
C
That's
then,
all
to
say
have
a
nice
day
stay
tuned
and
see
you,
then
next
thursday,
bye.