►
From YouTube: Keptn Community & Developer Meeting - Feb 03, 2022
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
Sign up to our newsletter: https://bit.ly/KeptnNews
A
That
many
people
so
today
is
february
3rd
and
we
have
a
second
community
meeting
this
week.
So
yesterday
we
had
a
meeting
in
the
epoch
time
zone
where
we
discussed
quite
a
lot
of
topics.
We
had
welcome
with
matt
mark
roberts
who
joins
to
help
this
documentation.
A
Also,
we
discussed
cap
68
and
we
will
continue
today.
We
had
a
quick
update
on
cap
65
and
on
captain
incubation.
So
for
today
we
have
a
few
topics.
Basically
cap
68
continued
cap
65
continued
and
I
wanted
to
do
last
sync
up
on
the
google
summer
of
code,
because
I'm
really
related
with
this
topic.
So
a
plan
b
is
needed,
and
this
is
what
I
wanted
to
discuss.
A
So,
let's
type
starts
with
kappa
68,
then
yeah
sure.
Would
you
like
to
screen
share?
Would
you
like
me
to
screenshare
no.
B
This
captain
has
proposal
describes
a
very
handy
feature:
it's
the
future
of
retriggering,
a
sequence
yeah,
because
this
sometimes
or
almost
on
a
two
weeks
basis.
We
get
the
request
that
it
would
be
very
handy
that
I
can
re-trigger
a
sequence
that
failed,
especially
when
I'm
new
to
captain
and
I'm
exploring
and
playing
around
with
captain.
Then
I'm
confronted
with
failed
sequences
and
to
fastly
retrigger
them.
I
would
like
to
have
support
here
in
the
in
the
bridge,
as
well
as
by
the
cli,
and
there
are
two
options.
B
A
new
sequence
will
then
pop
up
here
in
the
left
and
left
part
where
I
get
new
entry
and
from
that
on
yeah.
The
sequence
will
then
get
executed
in
the
first
stage
in
the
next
one
and
so
on,
and
it's
really
like
a
new
sequence,
but
the
payload
is
identical
to
the
one
that
we
had
here
here,
I'm
also
explaining
or
just
proposing
how
this
could
look
like
from
the
cli
point
of
view.
B
And
also
what
should
not
be
part
of
that
of
that
cap
and
should
not
be
part
that
we
provide
a
dedicated
endpoints
for
those
features,
because
we
already
have
the
event
end
point
and
that
is
sufficient
to
yeah
to
implement
this
capability.
B
And
also
very
important
in
this
cap,
we
also
don't
wanna,
allow
someone
to
retrieve
a
task.
This
cap
is
really
focused
on
just
sequence
level
and
not
on
triggering
individual
tasks,
all
right,
yeah.
Some
discussion
already
happened
on
this
captain
hunter
and
proposal.
So
I
got
a
couple
of
questions.
B
I
also
explained
here,
for
example,
what
happens
if
I
what
happens
with
the
retriggering
button
when
a
sequence
was
retriggered,
then
my
that
we
disable
those
buttons
of
all
sequences
and
just
make
the
the
last
one
highlighted
and
active.
B
B
We
then
kick
off
a
release,
sequence,
just
an
example,
but
it
could
be
the
case,
and
here
we
have
a
triggered
on
property
set,
and
when
we
have
the
situation
that
the
delivery
was
successful,
then
the
release
will
be
executed,
and
in
this
case
it
would
also
be
possible
and
feasible
that
I
retrigger
both
of
them.
B
In
case
we
have
a
sequence
which
works
like
a
a
backup
sequence
or
a
sequence
which
does
a
a
cleanup
so
to
say,
then
I
would
not
allow
to
retrigger
this
one,
because
this
would
then
cause
more
confusion
or
more
problems
that
it
might
solve,
and
in
this
particular
case
the
button
should
not
be
visible
here,
and
this
is
the
situation
when
we
have
a
triggered
on
configured,
which
then
yeah
searches
or
selects
on
on
a
match
pro
on
a
result.
Fail.
B
The
left
one
is
kind
where
I'm
just
having
triggered
on.
This
is
the
the
way
we
cue
sequences
and
and
on
the
right
side
we
have
a
selector
set,
and
whenever
there
is
a
selector,
then
I
would.
I
will
expect
that
I
don't
get
a
re
trigger
button.
B
Yeah,
I
think
we
we
have
to.
We
have
to
wait
or
we
we
need
feedback
here
and-
and
I
would
go
the
pessimistic
way
first
of
not
allowing
to
retrigger
this
particular
sequence
and
when
someone
really
has
has
a
approve
and
and
can
yeah
explain
why
he
or
she
needs
needs
it,
then
we
also
allow
it.
But
for
now
I
would
go
with
with
not
allowing
you
to
retrigger
this
particular
sequence
and
on
the
left
side
here.
B
We
because
this
is
just
the
way
we
are
q,
chaining
sequences
and
this
chain
works
whenever
there
is
a
sequence
that
passed
successfully
and
here
re-triggering
should
be
possible
and
should
not
cause
any
any
problems
on
the
right
side.
This
might
not
be
the
case.
B
And
then
it
was
also
the
question
coming
up:
can
I
retrigger
sequence
two
times
or
multiple
times?
Yes,
you
can.
I
did
this
already
by
by
sending
a
couple
of
events
to
captain
and
what
will
happen
then
yeah
you
get
a
queue
of
sequences,
and
this
cube
will
then
be
executed
from
top
to
bottom.
But
this
is
possible
and
also
not
a
big
problem
and
just
a
remark.
B
I
would
be
also
nice
to
have
a
confirmation
dialogue
when
I
click
the
retrigger
button,
because
yeah
users
should
get
informed
of
the
consequences
that
are
happening
and
all
like.
You
also
mentioned
there
whether
it
is
possible
to
do
this
via
api,
and
the
good
news
is
that
this
is
possible
as
of
today,
because
you
just
take
the
payload
of
the
of
the
initial
event
and
send
it
to
the
events
endpoint.
B
Then
good
question
to
try
it
out.
That's
the
negative
yeah!
That's
that's
a
good
one,
but
I
I
have
to
try
it
out,
but
I
would
assume
that
it
shows
up
as
an
additional
delivery
sequence
here
as
an
additional
sequence.
B
D
As
you're
showing
this
johannes
is
the
all
these
web
hooks
too,
because
we
are,
you
know
we
also
listen
for
those
those
new
tasks,
so
presumably
they
they
would
work
the
same
way.
Fine
like
if
you
retriggered
a
a
sequence
or
a
task,
it
would
just
the
web
hooks
that
subscribe
to
it
just
automatically
pick
up
with
no
problem.
A
Yeah,
it
sounds
reasonable,
yeah
yeah.
For
me,
the
main
question
right
now
is
automation,
so
basically,
a
web
interface
is
smallest
trivial,
and
this
is
what
we
need.
Well,
I
can
guess
it
will
be
more
complicated
if
we
had
parameters
etc,
but
right
now,
it's
quite
straightforward
and
for
cloud
events
and
the
company.
It's
something
important
to
specify,
because
my
understanding
that
what
we
specify
now
would
be
also
immediately
aligned
with
cd
events,
because
such
generic
actions,
like
retriggering
pipelines,
etc,
would
be
rather
there.
B
A
Cli
commands
for
that
or
how
would
the
user
interaction
look
like
beyond
the
web
interface.
B
I
define
at
this
stage
and
when
I
hand
over
already
sh
capital
context,
then
I
would
yeah
re-trigger
this
particular
actually
oops.
Sorry,
thanks
for
pointing
it
out,
the
sequence
name
is
missing
here:
captain
trigger
sequence,
delivery
in
my
dev
stage
and
with
this
sh
cap
in
context,
and
then
I
would
get
a
new
sequence
in
here.
C
So
one
opinion
on
that
as
well.
What
if
you
have
a
type
point,
sh
captain
context,
either
in
the
parameter
flag
or
copying
the
context.
You
forgot
the
character,
then
it
has
a
completely
different
meaning.
So
I'm
also
very
much
for
considering
I'm
not
saying
it
needs
to
be
like
that,
but
for
considering
a
separate,
retrigger
command.
B
Yep,
that's
true,
because
this
could
then
cause
yeah
problems
and
issues.
They
were
not
two
nudes.
The
first
one
was
from
rob.
A
A
Okay,
but
if
these
three
times
happen
in
parallel,
then
so
we
won't
merge
these
pipelines,
so
we
will
trigger
three
competing
pipelines
after
what
sequences
at
rating
against
the
same
environment,.
B
Here
we
should
or
like
you
should
note
that
they
are
not
running
in
parallel,
then
they
are
queued
and
I
executed
one
by
the
other
meaning
when
two
are
coming
in.
At
almost
the
same
time,
then
yeah
the
first
one
will
kick
off
and
then
we'll
get
executed
by
captain
the
other
one
is
in
acute
state
and
just
when
the
first
one
is
is
finished,
then
the
next
one.
B
But
I
agree:
this
can
happen
that
that
there
are
coming
in
that
multiple
sequences
are
coming
in,
but
I
will
leave
it
then
up
to
the
user,
to
a
port,
the
other
ones
or
yeah
to
let
them
run.
A
So
well
later,
we
could
implement
some
kind
of
q
merge
or
we
could
just
prohibit
triggering
if
it's
already
triggered
by
returning
special
status.
A
I'm
just
concerned
that
if
we
talk
about
delivery
sequences
that
multiple
reprovisioning
will
actually
wreak
the
havoc
in
user
in
target
environment,
especially
if
the
service
gets
brought
down
because
you
have
a
delivery
it
passes
and
then
you
shut
down
the
service
again
because
you
reprovision
the
instance
and
my
understanding
that
currently
it's
not
dependent
on
the
status.
So
basically
you
can
retrigger
even
the
successful
sequence
right
right,
yep
is
it
it's
designed.
B
That
is
as
designed
here.
That's
that
you
cannot
read
that
you
can
or
that
the
status
is
not
considered
when
hitting
a
retrigger.
A
Yes,
well,
there
are
restarts
from
milestone,
etc,
which
behave
a
bit
differently,
but
in
principle
the
status
doesn't
matter
and
the
jenkins
actually
has
multiple
statuses,
not
just
past
failed.
There
are
special
statuses
for
aborted,
for
interrupted
and
for
unstable.
A
D
Is
there
a
way
to
put
a
check?
Oh
there's
a
chat.
Sorry,
I
think
this
is
okay
to
put
here,
but
just
you
know,
as
you
guys
are
thinking
about
it,
I
mean
other
tools.
You
just
mentioned
jenkins.
I
mean
there's
a
link
to
another
product
that
will
that
you
know
is
trying
to
solve
that
same
problem.
You
know,
do
you
rerun
the
whole
workflow?
Do
you
rerun
individual
steps?
D
So
you
know
just
as
you
survey
options,
there's
there's
one
reference
of
how
someone
else
did
it,
but
maybe
to
that
point
maybe
we
need.
I
know
we
have
an
abort,
the
sequence
button,
but
maybe
we
need
also
an
abort
the
the
task
button.
So
in
that
scenario
you
fired
off
three
deliveries
by
mistake:
you
could
you
can
nuke,
you
know,
or
you
know,
cancel
out
two
of
the
three.
You
know
that
are
still
cued
without
having
to
disrupt
the
whole
sequence,
basically,
research,
how
other
projects
behave
in
such
situations
yeah.
D
So
I
just
had
two
comments
there.
One
was
right,
take
a
look
at
how
others
do
it
and
then,
secondarily,
coming
back
to
that
use
case
of
I
accidentally
did
two
more
deliveries
and
they
haven't
run
yet.
Maybe
the
user
can
have
a
little
like
a
little
x
like
we
have
on
the
the
sequence
level,
but
just
have
another
x
there.
So
have
it
at
that
level
as
well,
so
you
can
abort
cued
tasks
that
haven't
run
yet
or
or
support
tasks
that
are
hung.
Maybe
I
don't
know,
what's
the
thought.
D
And
I
don't
know
if
other
tools
distinguish
between
a
rerun
versus
a
regular
run.
You
know
what
I
mean,
so
I
don't
know
that's.
I
guess
that's
part
of
that
research
like
when
I
think
of
other
tools
like
you
just
when
you
run
the
pipeline
again,
you
get
a
new
job
id
and
it
just
kind
of
does
its
thing
right.
You
know.
B
D
This
subtask,
within
the
existing
sequence,
yep
get
that
right
right
totally
get
that
yeah
totally
makes
sense.
I
wasn't
thinking
of
that.
Second,
one,
like
that's,
that's
cool,
I
was
just
thinking.
We
were
gonna,
be
allowed
to
run
the
whole
sequence
again.
So
to
me
that's
more
than
I
was
thinking
this
would
do
so.
This
is
great.
B
Cool
yeah
thanks
for
the
feedback
already
and
we
will
not
start
implementing
it
immediately.
This
also
gives
all
of
us
some
more
time
to
think
about
it,
but
when
you
have
some
thoughts
ideas,
please
just
yeah
watch
watch
out
this
captain
hudson
proposal
and
add
a
comment
there
happy
to
answer
all
the
questions.
B
I
briefly
want
to
jump
to
the
next
enhancement
proposal
that
came
in
in
the
last
couple
of
weeks,
and
it's
all
about
the
cap
directly
declarative.
B
The
author
of
that
is
actually
thomas
is
thomas
on
the
call
today
nope,
no
he's
not.
He
did
here
a
great
job
of
describing
a
new
way
of
declaring
configuration
for
captain
and
then
also
aligning
this
with
another,
more
advanced,
git
ops
approach.
B
I
don't
want
to
recall
the
entire
captain
enhancement
portal.
I
want
to
directly
jump
to
the
last
action
item
that
we
derived
and.
B
B
B
Why
we
are
splitting
it
because
we
think
that
here
we
have
two
concerns
where
also
two
teams
can
work
on
without
dependency,
even
though
there
is
a
little
bit
of
a
dependency,
because
the
githubs
operator
is
a
requirement
for
a
gift
operator,
but
to
have
the
topic
separated,
we
would
like
to
yeah
also
split
this
cap
and
the
status
is
the
following:
I'm
waiting
for
thomas
to
update
this
captain
hunter
proposal
with
his
latest
updates,
and
once
this
is
done,
I
will
yeah
take
out
the
part
of
the
kept
operator.
D
All
right
so
can
I
just
ask,
I
mean
sorry,
I'm
just
reading
this
up
on
someone
can
at
a
high
level.
What
would
this
do?
What
would
this
do
for
from
a
an
end
user
point
of
view,
okay
structure,
my
files
in
a
different
way,
like
kind
of
like
I
know,
there's
a
lot
of
internal
stuff,
but
what
what
does
it
do
to
me?
As
the
user
of
this.
B
For
an
end
user,
the
change
will
be
that
you,
as
an
end
user
user.
You
don't
type
in
captain,
create
project.
Captain
create
service,
captain
trigger
sequence.
You
don't
do
this
anymore.
You
just
define
in
your
captain
configure
in
your
git
repository,
which
stage
you
want
to
have
which
service
you
want
to
have
on-boarded,
which
sequence
you
want
to
get
triggered.
You
just
save
it,
commit
it
to
your
repository
and
the
github
operator
is
watching
it
or
yeah.
It's
watching
your
repo,
and
whenever
a
change
is
detected,
then
yeah.
B
A
D
D
Well
I
mean
it's
a
you
can
confirm
where
we
I
have
to
have
a
blank
repo
when
I
create
my
project
so
that
currently
captain
will
build
all
the
directories
it
needs.
Then
I
can.
I
I
don't
even
know
if
I'm
allowed
to
add
additional
project
files
into
that
same
repo
and
kind
of
manage
it
with
my
code,
because
we
always
talk
about.
D
I
have
the
captain
files
and
my
code
all
in
the
same
place,
but
you
have
to
start
with
the
blank
repo
to
even
begin
and
and
it's
like
a
chicken
and
the
egg
problem
it's
like.
Well,
I
have
this
repo
with
my
get
with
some
code
that
I
want
to
use
as
my
starting
point
and
then
make
it
enabled
with
captain,
but
then
I
can't
do
it
that
way.
Like
now,
I
have
two
repos
right.
It's
captain
repo.
D
B
It
does
yeah,
I
mean
internally,
the
the
captain.
Repo
will
be
there,
but
not
not
visible
to
the
end
user,
because
the
only
interaction
point
will
then
be
via
the
code
repository
the
the
dev
or
the
dev
is
working
on,
and
this
is
just
the
only
way
to
to
work
with.
Captain
then
and
thomas
did
also
a
great
job
in
summarizing
all
the
issues
that
are
currently
open
and
related
to
this
topic
and.
D
D
D
Yeah,
that
would
be
great
yeah,
because
it's
yeah,
because
I
did
a
demo
yesterday
to
a
meetup
group
and
that's
where
they
didn't
ask
that
question.
But
you
know
I
have
you
know
two
different
repos.
I
was
flip
flip
flipping
back
and
forth.
You
know
here's
my
here's,
my
one
repo,
where
I
had
github
actions
calling
and
invoking
pipelines
that
called
my
existing
project
and
had
another
repo
that
captain
was
managing
for
the
captain
project
files
and
then
it
just
sort
of
looked
you
know
just
like.
D
B
Cool
yeah.
We
also
think
this
would
be
a
really
awesome
enhancement
to
captain
and
also
the
way
you
as
a
user
would
interact
with.
D
Question
maybe
like
do
we,
I
mean
if
people
start
to
review
this
is
sort
of
like
just
my
first
time
really
coming
to
these,
but
as
these
proposals
are
out
there,
how
can
someone
say
I
mean
without
just
like
you
know,
I
you
know
want
to
plus
one
this
idea
like
or
just
say
like.
I
really
want
to
you
know
give
feedback
just
in
general,
generically
like
yeah.
This
is
a
value
to
me
or
yeah.
Please
make
this
a
priority
like.
How
would
you
recommend
we,
you
know
give
that
feedback?
D
B
Yeah,
and
also
speaking
about
this
topic
will
help
a
lot
I
mean
you
now
are
aware
of
it,
and
also.
D
The
importance
yeah
because
there's
a
lot
of
these
enhancements
that
right
and
that's
what's
helpful
to
know
like
right
these
I
mean
right
just
where
to
find
the
these.
You
know
the
conversations
and
the
discussions
and
proposals.
That's
just
an
education
too
that
where
to
find
these
links,
you
know
maybe
that's
feedback
for
you
all
like
like
where,
like
you
know
in
the
community
or
getting
started
kind
of
guides
like
you
know,
know
that
we
have
enhancements
proposals
here.
D
A
So
we
try
to
put
the
key
items
on
the
roadmap,
so
here
you
can
see
roadmap
labels,
yeah,
and
actually
I
suggested
that
we
add
them
to
67,
68
and
64.,
because
all
of
them
seem
to
be
reasonable,
so
64
and
67
would
sorry
67
and
65.
Where
is
it
well?
A
Yeah?
I
just
lost.
I
still
haven't
been
split
to
two
caps
right.
No,
what
will
happen?
Okay,
so
once
it
happens,
we
I
suggest
created
the
roadmap,
is
work
in
progress
and
trigger
sequence
somewhere
under
triage.
We
have
this
state
so
that
users
can
discover
that
so
at
least
my
vision
is
that
the
key
items
that
are
actually
being
worked
on
should
go
to
the
roadmap
and
at
some
point
we
need
to
think
what
we
actually
do
with
caps
that
actually
that
are
stale,
because
it
looks
like
we
have
a
bunch
of
them.
B
Yeah
right
captain
open
telemetry
is
currently
on
ice
and
zone.
C
If
I
might
get
back
to
recap
68,
I
have
two
things
I
wanted
to
say.
One
is.
I
know
that
thomas
is
writing
this
from
a
very
technical
standpoint.
I
think
there
would
be
a
benefit
if
we
can
get
like
the
end
user
perspective
in
a
little
bit
more
like
what
does
it
feel
like
for
the
end
user?
C
I
know
why
but
like
from
an
end
user
perspective,
it's
not
clear
and
I
I
could
think
that
it
is
possible
to
go
into
a
single
repo
like
also
putting
jenkins
up
as
an
example
again
in
chanki's,
you
just
configure
one
repository
that
you're
pulling
your
jenkins
file
and
your
source
code,
and
whatever
else
from
so
this
should
also
be
possible.
Somehow
we
can
have
this
as
the
default
that's
there,
but
I'm
really
wondering
if
we
can
make
it.
D
A
A
Yeah
in
jenkins,
we
basically
have
infrastructures
code,
so
it's
configuration
for
jenkins
itself.
It's
usually
one
repository
and
another
repository
are
project
files
like
jenkins
file,
etc.
A
They
usually
go
to
the
project
repository
because
it's
how
pipeline
is
code
is
documented,
and
we
also
have
separate
repositories
for
libraries
if
relevant,
so
libraries
are
still
being
discussed
for
captain,
but
it
was
brought
up
many
times,
but
for
infrastructure
it's
called
supporting
one
repository
would
be
really
good,
and
this
infrastructure
is
code
can
also
contain
sequence
definitions,
but
at
the
same
time
I
think
we
need
to
support
both
options,
but
we,
you
have
sequence,
definitions
right
inside
the
repository
or
you
have
them
on
the
project
site.
A
D
D
D
All
right
so
yeah,
I
guess
that's
the
suggestion
is
yeah
like
a
a
maybe
some
additional
information
in
the
pr
around
what
the
motor
the
motivations
to
do
this.
From
an
end
user
point
review,
yeah,
okay,.
B
All
right
yep
for
today
I
will
stop
here
for
today
and
happy
to
continue
the
discussion
in
two
weeks.
Yeah.
A
So
for
the
next
item
for
gsoc,
I
would
rather
take
it
offline
because
yeah
I
mean
I
have
time
constraints
crestron,
but
what
I
am
going
to
do.
I
will
send
a
message
to
the
developer
channel
today
so
that
we
can
see
what
we
can
commit
to
and
yeah
the
rest
of,
like
work
actually
wanted
to
do
it
a
few
weeks
ago,
but
force
them
and
other
stuff
completely
deleted.
A
A
If
you
consider
participating,
please
let
me
know
or
if
you
have
any
project
id
in
mind
because
currently
for
us
it
would
be
nice
to
have
this
project
list
accumulated
in
this
github
mission,
and
then
we
can
see
what
actually
flies.
What
not.
But
for
starters,
we
need
to
this
framework
right.
A
A
Yes,
so
basically
what
we
have
better
github
integration,
something
can
to
expand
and
we
have
k6
id,
which
is
documented
check.
This
executor
circus
for
captain
captain
plugin
for
backstage,
which
needs
to
be
revised
because
we
have
pending
configuration-
and
we
briefly
discussed
this
chris
with
thomas
so
that
github's
for
captain.
We
might
find
something
there
too,
but
all
these
ideas
they're
currently
really
just
ideas
without
much
description.
D
Meg,
maybe
helping
out
with
our
workshops
in
tutorials,
not
workshops,
but
the
tutorials.
I
know
it's
not
code,
but
there
is
you
know
again
for
students.
You
know
they
would
kind
of
learn
and
test
and
make
sure
that
yeah.
A
A
D
You're,
I
don't
know
if
people
get
to
see
the
the
the
chat
in
the
recording
but
I'll
put
that
link
there
for
people.
A
Yeah,
so
what
I'll
do
next
developer
painting
on
next
thursday?
I
will
present
what
we
have
basically
from
end
to
end,
and
we
will
see
if
everyone
is
fine
in
the
application.
A
A
We
will
be
moving
developer
meeting
to
epoch
time
zone
because
the
most
of
the
team
voted
for
that
and
we
will
still
keep
doing
this
community
meetings,
but
every
two
weeks
so
not
every
week,
but
every
two
weeks.
A
A
So
I
think
that
yeah,
basically
whoever
wants
to
present
just
join
whether
it's
community
meeting
on
developer
meeting
and
we
figure
it
out.
A
Okay,
so
yeah,
then
thanks
everyone
and
see
you
next
week
I
will
publish
the
recording
tonight.
The
recording
from
yesterday
is
already
uploaded.
I
just
didn't,
publish
it
because
I
need
open
graph,
but
everything
will
be
done
today.