►
From YouTube: Keptn Community & Developer Meeting - March 4th, 2021
Description
Meeting notes: https://docs.google.com/document/d/1y7a6uaN8fwFJ7IRnvtxSfgz-OGFq6u7bKN6F7NDxKPg/edit
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
B
A
A
A
By
the
way
we
are
following
here:
the
cncf
code
of
conduct,
so
that
basically
means
please
be
nice
to
each
other
hi
robert
thanks
for
joining,
and
yes
so,
please
put
in
your
names
and
first
part
on
the
agenda.
A
A
This
so
captain
got
accepted
into
the
lfx
mentorship
program
with
two
projects,
and
we
are
happy
to
welcome
two
mentees.
A
So
thanks
for
joining
the
captain
project,
I
think
you're
also
here
today
in
this
meeting.
So
thanks
for
joining
also
in
this
meeting,
we
are
pretty
sure
that
we
hear
more
from
your
side
and
your
development
on
captain
once
we
get
your
projects
starting
and
raj
and
heyman
will
join
us
for
the
next
couple
of
weeks
for
for
a
mentorship
project
and
it
will
work
on
on
two
two
dedicated
projects
and
I'm
sure
he
will
share
the
pro
progress.
A
Also
in
these
meetings,
so
welcome
everyone
and
welcome
to
our
mentees
really
great
to
have
you
with
that.
If
there
are
any
questions,
please
let
me
know
otherwise.
Yes,.
B
Please
I
think
we
should
mention
that,
starting
today
that
every
second
week
we
will
give
the
community
more
time
here
in
this
staff
meeting
to
show
us
their
pull
requests
to
come
up
with
questions
to
have
a
chat
about
what
they
are
working
on
and
what
they
want
to
implement
in
captain
and
then
every
other
second
week,
the
captain
core
team
that
are
building
captain
on
on
their
daily
business.
They
are
presenting
features
and
yeah
of
improvements
and
and
what
have
been
accomplished
within
two
weeks.
B
This
is
the
kind
of
new
format
that
we
want
to
follow
now
and-
and
today
is
the
time
that
actually,
the
community
gets
more
time
to
speak.
A
B
B
B
And
after
that
time,
to
discuss
okay.
A
So
we
will
actually
do
it
this
way.
Next
time
is
a
feature
presentation
march
11th,
oh
sorry,
and
the
week
after
it
will
be
march
18th
correct.
A
I
think
we
will
have
dedicated
time
for
for
community.
A
Actually,
also
a
great
part
where
we
can,
we
can
use
this
also
with
working
with
the
mentees
on
the
projects,
and
we
can
answer
questions
and
also
as
we're
seeing
more
interesting
captain
integrations.
This
will
also
be
part
of
working
on
captain
integrations.
I
know
that
there
has
been
some
work
already
going
on
in
lifting
captain
integrations
from
captain
7.3
to
the
latest
version
of
captain.
A
So
next
week
we
will
for
sure
hear
more
about
the
progress
going
on
in
from
the
captain
court
team,
and
then
we
can
use
this.
The
the
other
weeks
for
for
the
more
community
focused
content.
So
this
is
great
cool
yeah.
Thanks
for
this
input,
is
there
any
other
input
on
their
community
or
on
the
lfx
mentorship
programs
or
any
other
questions
so
far,.
A
If
not,
then,
by
the
way,
I'm
just
allowing
everyone
also
to
unmute,
then
I'm
already
handing
over
to
the
next
part
on
the
agenda,
and
I
think
that
will
be
johannes
sharing.
A
B
Really
cool
that
last
week
we
jointly
kicked
off
the
ci
pipeline
for
releasing
captain.
B
Then
we
now
took
a
week
to
finally
test
it
out
and
also
provide
the
documentation
of
the
latest
features
that
we
implemented,
and
today
I
have
the
honor
to
announce
that
captain
autodate
is
officially
released
and
you
can
now
download
it
install
it
and
yeah
work
based
on
captain
order
date,
as
you
were
used
to
0.73.
B
But
with
this
new
version
changes
come
along
and
today
I
want
to
show
you
on
a
conceptual
and
high
level
overview
or
point
of
view
what
have
been
changed
and
also
what
new
capabilities
that
you
can
now
use
using
the
new
version,
and
therefore
I
mean
oh
yeah,
maybe
if
you
are
also
interested
in
what's
coming
next,
what
what
we
are
heading
towards?
You
always
find
the
roadmap
so
to
say
on
the
captain
project
board,
which
is
on
the
captain
organization,
just
go
there.
B
B
Then
the
ui
will
get
improvements
in
regard
to
the
quality
gates
and
into
base.
First,
there
will
be
a
project
wide
overview
of
the
quality
gate
results
covered
by
this
epic
and
the
second
one
will
be
improvements
in
regard
to
the
naming
of
slice,
as
they
can
become
very
long,
and
there
are
some
limitations
with
these
long
names.
We
will
also
improve
this
part
from
time
to
time.
Please
just
take
a
look
at
this
board
and
there
you
find
all
the
information
or
towards
the
direction.
B
B
You
find
on
the
release,
notes
all
the
key
announcements
that
are
yeah,
that
we
publish
and
that
we
announce
like
the
new
way
of
specifying
so-called
task
sequences.
Then
individual
tasks
in
a
sequence
and
so
on.
Please
check
out
this
release
notes.
I
will
now
show
you
in
a
slide
deck,
how
those
things
are
actually
then
implemented,
and
you
can
make
use
of
all
right,
and
this
is
now
where
I
jump
back
to
my
presentation
and
we'll
talk
about
our
cap
nota
date.
B
The
main
part
that
has
been
implemented
is
based
on
the
new
shipyard
format,
which
has
the
version
of
oda2.
This
has
been
proposed
by
this
cap.
It's
a
captain,
huntsman
proposal.
It
incorporated
a
couple
of
specification
changes.
You
also
find
in
here,
and
this
is
what
mainly
drove
the
development
of
cap
notate,
but
before
getting
into
details,
let
me
to
all
new
participants
and
listener.
B
B
In
this
example,
the
shipyard
in
the
shipyard
you
model
the
way
you
want
to
promote
an
artifact
assuming
you
have
a
death
stage
where
you
test
out
your
your
new
artifact,
then,
when
testing
in
depth
is
fine,
you
promote
it
to
hardening
and
when
you,
when
test
and
hardening,
are
done
and
show
you
that
everything
is
going
smoothly,
then
you
finally
deploy
it
into
production.
B
This
is
the
first
thing
that
you
can
model
with
the
shipyard
and
then
within
each
stage.
You
can
then
define
the
processes
you
want
to
execute
like
a
delivery
process
or
remediation
process.
This
is
the
second
capability
and
the
second
feature
that
is
unique
or
important
for
understanding
the
shipyard,
the
shipyard
itself
you
define
on
a
project
level.
This
means
that
all
services
that
are
within
this
captain
project
they
share
the
same
shipyard,
meaning
that
they
follow
the
same
staging
process
and
they
also
follow
the
same
processes
you
define
in
this
shipyard.
B
Let
me
just
give
you
here
an
example
of
a
delivery
process.
There
we
are
default
or
it's
common
practice
or
best
practice.
You
have
a
deployment
step
where
you
just
deploy
the
artifact,
then
you
test
it
after
testing,
you
evaluate
the
quality.
When
everything
is
good,
you
ask
the
user
for
an
approval.
And
finally,
when
the
approval
is
okay,
then
you
go
on
and
you
release
it,
meaning
that
you
make
it
public
available
and
also
end
user
facing.
B
And
also,
it's
always
good
to
to
rethink
principles
that
the
captain
project
follows,
and
one
of
our
core
principles
is
that
we
focus
on
separation
of
concerns,
meaning
that,
on
the
one
hand,
in
the
shipyard,
you
define
the
processes
and
they
should
should
be
decoupled
from
the
underlaying
tooling.
That
is
actually
then
doing
the
execution
of
a
certain
task.
B
This
is
yeah
known
as
being
tool,
agnostic
and
here
just
visualize
this
with
different
colors,
meaning
that
the
deployment
step
is
done
by
this
tool
here
in
green
testing
by
the
red
one
and
so
on.
This
is
a
core
principle
that
we
as
a
project
follows
all
right.
Yeah,
let's
go
on
with
the
specification
of
the
new
shipyard
format.
This
now
follows
the
the
frame
and
the
pattern
of
of
a
custom
resource
known
in
kubernetes.
B
This
means
that,
on
the
top,
you
have
some
meter
data
made
information
about
the
shipyard
about
this
is
the
spec
version.
Then
the
kind
we
are
talking
about
and
then
place
for
adding
arbitrary
meet
the
info
information
like
the
name
of
the
of
the
file
you're
looking
at
then.
After
that,
you
have
the
chance
to
model
than
the
stages.
B
This
here
is
the
stage
property,
and
this
is
a
list
of
stages
that
you
and
your
shipyard
that
you
want
to
define
in
your
shipyard
here.
In
this
example,
I
have
just
one
stage:
it
has
the
name
hardening
and
then,
in
each
stage
you
can
add
sequences
sequences
are
the
processes
I
mentioned
before
and
again
you
can
add
multiple
sequences
to
a
stage,
give
it
a
name
and
there's
also
a
way
that
you
can
define
when
this
sequence
gets
triggered.
B
This
is
done
by
this
triggered
on
property,
where
you
just
reference
the
event
that
works
as
a
trigger
for
for
kicking
off
the
sequence
and
afterwards
the
standard,
the
final
piece
of
of
defining
a
sequence,
because
then
you
just
tell
which
tasks
should
be
executed
when
this
sequence
gets
x
gets
triggered,
and
here
again
this
is
also
a
list
where
you
can
define
all
the
tasks.
B
That
sequentially
will
be
then,
orchestrated
by
by
captain
back
to
my
example,
from
before
it's
a
deployment
step
after
the
deployment
a
test
gets
executed,
then
a
second
test.
They
just
differ
by
two
properties:
the
first
one
gets
the
property
functional
and
the
other
one,
the
property
performance.
B
When
this
second
test
is
finished,
we
kick
off
the
get
sli
function,
then
the
evaluation
will
be
triggered
there.
Based
on
the
evaluation,
you
can
then
ask
the
user
for
approval
and
finally,
the
releasing
step
is
executed.
B
All
right
first,
you
can
now
clearly-
or
you
can
now
define
multiple
sequences
within
one
stage.
This
is
the
first
user
benefit
that
comes
with
the
new
format.
Let
me
show
you
that
here,
I'm
just
adding
a
second
sequence
to
my
production
stage.
In
this
example,
the
first
one
is
that
that
the
first
one
I
call
deployment
it
kicks
on
when
hardening
at
the
delivery
hardening
is
finished.
B
B
B
B
When
the
tool
then
received
it
and
starts
to
work,
then
it
just
sends
up
a
deployment
started
to
inform
the
shipyard
controller
that
the
action
yeah
has
been
started.
B
And,
finally,
when
the
the
service
is
done,
then
it
just
sends
a
deployment.finished,
and
this
informs
then
the
shipyard
controller
to
go
on
and
to
start
with
the
next
test
with
the
next
task.
In
this
sequence,
and
by
following
this
mechanism,
captain
orchestrates
your
on
your
delivery.
In
this
case.
B
All
right
now,
next
to
the
user
benefit
of
separating
the
sequences,
you
can
also
add
custom
tasks
to
a
sequence.
B
Let
me
just
show
you
the
example
here
I
have
the
delivery
process,
but
in
for
a
certain
case,
I
want
to
add
a
migration
step
before
kicking
off
the
deployment,
because
I
have
to
migrate,
for
example,
a
database
or
I
have
to
do
any
kind
of
migration
step,
then
I
just
add
this
task
in
front
of
the
deployment
task
in
my
sequence,
and
this
is
enough
that
captain
will
then
yeah
take
care
of
sending
out
the
migration.triggered
event,
and
you
have
then
to
implement
the
component.
B
That
does
the
migration,
but
when
this
is
plugged
in
then
your
process
can
be
executed
with
this
new
task
in
there.
B
With
this
feature
in
mind,
yeah,
it's
quite
powerful,
because
now
we
can
actually
model
any
process
that
you
want
to
implement
as
we
are
flexible
and
can
accept
any
tasks
in
there.
B
B
The
trigger
is
yeah
an
element
or
a
property
of
a
sequence.
It
also
takes
a
list,
and
this
list
is
just
a
key
value.
Pair
of
the
key
is
event
and
the
value
is
then
the
name
of
the
event
that
this
sequence
will
then
the
name,
the
event
that
this
sequence
will
be
kicked
off
by
which
the
sequence
will
be
kicked
off.
Let
me
just
give
you
here
an
example.
B
Here
I
have
an
environment
that
has
a
death
a
hardening
and
a
pre-production
stage.
B
I
left
out
the
sequences
from
the
depth
and
hardening
to
yeah
due
to
limitations
of
space,
but
debris
broad
has
a
delivery
sequence,
and
then
I
have
two
additional
environments:
a
production
environment
for
customer,
a
and
a
production
environment
for
customer
b.
B
You
can
also
see
it
here
in
this
yeah
visualization
and
I
modeled
it
in
a
way
that
production
customer
did
the
stage
for
production
customer
a
and
b
they
just
listen
and
get
kicked
off
when
the
delivery
in
the
pre-prod
environment
has
been
finished,
and
this
is
just
done
by
adding
here
the
triggered
on
property
and
specifying
the
event
pre-prod
dot
delivery
finished,
and
this
is
enough
that
we
can
react
on
what
happened
before
to
kick
off
the
sequence
in
this
particular
stage
and
exactly
the
same
is
true
for
the
production
customer
b
stage.
B
Multiple
stages
for
multiple
customers,
there
is
also
the
need
that
these
environments
may
run
on
different
in
clusters
and
just
by
giving
this
example
here
we
have
the
stage
or
the
environment
for
the
custom
a.
We
assume
that
this
one
is
running
on
on
gte,
and
then
we
have
environment
for
customer
b.
This
one
is
on
an
eks
cluster
on
aws.
B
This
is
now
also
possible
to
implement
using
the
new
captain,
and
we
call
this
the
multi-cluster
support
and
the
way
that
you
can
set
up
this
is
explained
here.
B
B
You
want
to
expose
your
captain
to
the
outside
world
and
when
it
comes
now
to
the
execution
plane
that
is
on
on
the
edge
and
not
linked
to
your
control
plane,
then
you
have
to
take
care
of
deploying
the
execution
services
there
by
default.
We
provide
support
for
the
helm
service
and
the
gmeter
service,
and
we
also
got
a
way
of
deploying
these
services
using
helm,
charts
you
find
the
helm
charts
on
the
github
release.
0.8
yeah
just
feel
free
to
check
them
out
and
to
download
them.
B
You
can
also
take
a
look
into
the
way
this
helmet
chart
is
specified,
and
then
you
have
to
you
have
to
configure
this
service
by
just
defining
that
this
service
runs
on
remotely-
and
here
you
just
say
in
the
values.yaml
of
the
helm,
chart
enabled
a
remote
control
plane
enabled
to
you
define
the
protocol.
B
You
have
basic
functionality
to
configure
filters
on
the
on
the
edge
and
these
filters
work
on
project
level,
on
stage
level
and
on
service
level
and
based
on
my
example
I
had
before
or
a
similar
one,
where
I
just
want
to
split
production
and
hardening
on
different
clusters,
then
I
would
say
that
my
service
that
is
running
on
this
cluster
has
to
listen
on
events
for
the
hardening
stage
and
the
execution
services
that
are
running
on
the
cluster
b.
They
have
to
listen
to
events
from
the
production
stage.
B
And
next
to
defining
the
communication
way
or
the
way
to
communicate
with
the
control,
plane
and
adding
the
one
or
other
filter,
you
then,
are
good
to
go
to
deploy
your
execution
plane
service
on
the
on
the
cluster,
and
this
is
done,
as
I
said,
by
using
helm,
install
and
that's
basically
it
for
yeah
deploying
the
service.
B
Finally-
and
this
is
very
handy,
you
can
test
that
the
communication
to
the
control
brain
works,
but
just
running
the
command
helm
test.
Then
the
name
of
your
service
and
the
namespace
you
installed
it
in
and
then
the
test
report
should
tell
you
that
the
communication
worked
or
did
not
work
in
case
it
did
not.
Then
we
provide
here
a
little
bit
further
down
a
troubleshooting
section
where
you
can
again
get
more
information.
What
may
be
wrong
and
how
you
can
fix
this
communication
problem.
B
Okay,
this
brings
me
now
yeah
to
the
end,
the
very
central
component
that
also
has
to
be
implemented
that
has
been
implemented
with
the
captain.
0.8
release
is
the
shipyard
controller,
as
it
is
the
brain
for
the
captain
control
plane.
The
shipyard
controller
understands
what
you
are
modeling
in
your
shipyard
specification.
B
It
takes
over
responsibility
for
sending
out
the
events,
and
it
also
is
the
com.
It
is
also
the
component
that
acts
upon
events
that
you
are
sending
back
to
the
controller
in
order
to
successfully
orchestrate
the
processes
that
you
came
up
with,
and
you
want
to
get
executed.
B
All
right,
oh
yeah,
here
again
I
have
a
slide.
I
think
I
mistakenly
moved
it
moved
it
to
the
wrong
location,
but
just
for
your
understanding
about
the
multi-cluster
setup.
B
As
I
said
when
you
have
here
two
different
clusters,
one
for
custom,
a
one
for
customer
b,
one
on
g
key,
the
other
one
on
on
eks.
Then
you
have
to
install
small
components
there.
The
components
that
take
over
just
the
task
of
deploying,
for
example,
as
I
said,
and
that's
basically
it
for
for
setting
up
the
master
at
the
multi-cluster
environment.
A
It's
it's
really
a
big
release
and
a
big
improvement,
how
how
captain
works
and
which
new
and
about
all
the
new
use
cases
that
are
now
possible.
We
have
this.
B
Decision
this
opens
now
a
wide
variety
of
use
cases.
That's
true,
and
now
captain
really
jumps
into
the
role
of
becoming
a
control
plane.
A
It
wasn't
so
I
shared
I
in
the
middle
of
your
presentation.
I
lost
the
audio.
My
internet
connection
did
not
allow
me
to
hear
you
other
other
questions
coming
in.
I
don't
think
so.
I
I
know
that
just
robert
and
I
think
christian
they
did
in
the
in
the
slack.
They
discussed
the
pr,
but
maybe
that's
one
pr
that
we
can
have
in
one
of
the
next
developer,
focused
or
community
focused
community
meetings
here
of
captain
once
it's
it's
merged
yeah.
A
If
there
are
no
more
questions
or
feedback
to
this
johannes,
can
you
do
me
a
favor
and
just
briefly
go
back
to
the
catherine
community
meeting.
B
A
Because
I
forgot
something
earlier-
and
I
just
want
to
do
this
now-
I
want
to
have
you
already
switched.
Maybe
it's
my
internet
connection
that
I
cannot
see.
Oh
now,
you
can
see
the
documentary
I
already
switched
there.
Thank
you
yeah.
It's
just
my
internet
connections.
It's
a
bit
slow
today.
I
just
want
to
announce
again
that
we
kicked
off
a
working
group
on
auto
remediation.
A
That
was
already
announced
in
in
the
previous
captain
community
meetings
and
also
on
slack,
so
you
might
already
already
be
familiar
with
it.
But
there
is
now
a
meeting
document.
There
is
a
dedicated
slack
channel
so
if
you're
interested
on
shaping
how
auto
remediation
should
work
and
if
you're
interested
on
on
the
future
of
auto
remediation,
also
in
general,
not
only
in
captain,
of
course,
it
will
be
then
implemented
in
captain.
A
But
this
meeting
group
or
this
working
group
is
really
about
the
concepts
and
on
on
the
requirements
of
auto
remediation,
how
it
should
work
and
if
you
have
any
feedback,
if
you
have
any
insights
on
tools
or
if
you're,
already
building
some
kind
of
auto
revelation
inside
your
organizations.
A
Let
us
know
or
just
join
our
next
working
group
we
meet
bi-weekly
on.
I
think
we
agreed
on
wednesdays.
A
A
Otherwise,
there
are
public
meetings
anyway,
you
will
also
find
the
soon
invite
in
the
meeting
talks,
but
yeah
we're
looking
forward
also
to
seeing
you
in
this
in
this
working
groups,
if
you're
interested
in
working
on
this
yeah
thanks,
I'm
handing
back
to
johannes,
but
this
was
just
something
I
I
missed
earlier,
as
I
was
a
little
bit
blocked
with
with
meetings
today.
B
Perfectly
fine,
thanks
for
adding
this
item
to
the
bullet
list
very
important
to
to
mention
I'm
actually
also
through
my
content
that
I
prepared
today
yeah
giving
you
a
quick
overview
about
the
capabilities
of
the
new
shipyard
and
the
captain
features,
and
I
think
the
only
the
only
thing
that
has
to
say
is
download
it
test
it
out
and
let
us
know
what
can
be
improved.
What
is
your
use
case?
What
you
want
to
have
also
implemented
so
that
we
then
can
jointly
drive
yeah
the
next.
B
The
next
big
release
for
captain
all
right.
Okay,
are
there
questions
from
from
someone.
A
I
think
no
there's
just
feedback
from
robert
that
the
eight
features
look
great
and
you
did
a
great
job
on
the
demo
excellent
demo,
excellent
presentation.
I
just
want
to
to
mention
that
there
is
this
migration
guide
for
contrib
services
and
also
captain's
handbook
services.
I
know
that
christian
has
a
put
a
lot
of
effort
in
this.
How
to
upgrade
captain
services
and
captain
integrations.
There
is
a
youtube
video
out
there.
A
A
A
Well,
cool:
are
there
any
more
items?
Otherwise
I
think
we
can
close
it
a
little
bit
earlier.
I
will
just
found
the
link
from
the
video
from
christian
and
I
will
put
it
in
here
there
we
go.
A
Okay,
then
thanks
everyone
for
joining.
We
meet
again
next
week
same
time
and
I
think
we'll
then
have
some
presentations
feature
presentations
of
the
captain,
14
and
then
every
other
week
we'll
have
feature
related
presentations
and
community
focused
meetings
where
we
can
talk
about,
integrations
and
and
also
the
work
from
our
mentees.
A
So
thanks.
Everyone
for
today
have
a
great
evening
and
see
you
all,
maybe
in
the
captain,
slack
or
latest
next
week,
and
please
everyone
download
captain
0.8.
A
I
know
the
team
has
put
a
lot
of
effort
in
it
and
it's,
I
think
it's
basically
you're
the
first
ones
that
that
know
of
the
release,
because
we
have
not
even
yet
put
it
out
on
slack
or
on
twitter.
So
it's
really.
The
captain
community
meeting
is
the
best
place
to
to
get
this
news
first
hand.