►
From YouTube: Keptn Community & Developer Meeting - Jan 13, 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
Sign up to our newsletter: https://bit.ly/KeptnNews
A
Well,
welcome
to
the
captain
community
meeting
today
is
january
13th,
so
we
are
academy
after
a
short
break,
and
yet
today
our
goal
is
that
just
to
sync
up
on
the
projects,
I'm
going
to
talk
about
a
few
communities
so
I'll
share
my
screen.
A
B
A
A
we've
got
quite
a
lot
of
things
to
do
this
year.
So
at
the
last
meeting
we
had
a
new
year
update.
I
will
share
the
link
later
but
yeah,
I
guess,
for
the
next
year.
We
will
need
to
spend
some
time
maybe
to
review
the
roadmap
but
yeah.
There
is
a
bunch
of
topics
plus
there
were
some
updates
over
the
christmas
break,
so
you
might
have
seen
that
there
is
45
from
thomas
and
the
king,
it's
about
the
guitars
and
captain.
A
So
it's
a
great
progress
and
thanks
to
everyone
who
participated
in
prototyping
and
yeah,
I
guess
that's
it
for
that.
C
A
Okay,
so
the
next
quick
update,
so
we
will
have
user
group
next
week
and
this
meeting
will
be
about
captain
data
as
npt
data.
They
will
be
talking
about
the
kepler
in
these
cases
and,
if
you're
interested,
if
your
contributor,
if
you're
a
captain
user,
please
join
so
yeah.
The
announcement
is
on
the
same
baby
page,
take
a
look
and
we
believe
it
will
be
an
interesting
presentation
for
everyone.
A
D
There
is
a
new
service
called
resource
service
which
will
be
a
replacement
for
the
current
configuration
service,
and
this
resource
service
is
basically
a
configuration
service
on
steroids
and
it
doesn't
rely
anymore
on
an
internal
git
repository,
but
it's
by
now
disabled
by
default
and
it's
an
under
a
beta
state.
Let's
say
and
testing
state.
D
A
Yeah,
so
if
anyone
have
something
like
something
would
be
much
appreciated,
yeah
so
yeah,
then
there
was
a
topic
we
had
before
the
christmas,
so
we
have
artifact
hub,
which
is
still
within.
A
Yeah,
so
artifact
hub
is
currently
a
part
of
captain
symbols,
but
we
widely
used
this
metadata
for
the
captain
side
for
artifact
hub
and
there
was
a
proposal
to
actually
promote
it
to
production,
great
project,
so
either
captain
country
or
maybe
just
main
captain
organization.
A
E
I'm
fine
with
either
captain
contrib
or
the
captain
organization
itself,
given
that
we
don't
usually
have
third
party
integrations
living
in
the
captain
organization,
but
in
captain
corn
trip
I
think
it
better
fits
to
captain
contrib.
However,
we've
seen
in
the
past
that
different
organizations,
different
projects
might
have
different
views
on
that.
B
If
I
would
vote,
I
would
go,
but
also
with
captain
country
as
altitude
is
dealing
with
our
contributions,
and
this
is
kind
of
a
major
project
around
all
integrations,
and
so
I
think
it
would
fit
better
into
captain
country
rather
than
the
chapter
organization.
A
E
A
While
yesterday
there
should
have
been
a
cncftrc
meeting.
I
checked
the
meeting
notes.
There
was
no
discussion
for
captain
specifically,
but
definitely
all
projects
are
in
the
backlog,
so
they
prioritize
the
applications
to
the
sandbox
because
they
have
basically
insane
number
of
them,
but
they're.
Also,
looking
in
the
incubation
to
graduation
applications,
you
may
have
seen
there
was
announcement
of
elite
moscow's
just
yesterday,
so
it's
officially
incubating
project
and
what
we
have.
It
was
a
house
integration,
service
and
yeah.
A
D
A
A
A
E
If
it
allows
me
to
share
just
one
screen,
I
can
share
firefox.
I
guess
there
we
go
so
yeah
we've
been
working
on
several
things
in
the
background,
automation,
etc,
and
also
did
some
bug
fixes
for
the
job
executor
service
and
the
end
result
is
now
that
we
can
easily
create
a
new
release
and
the
pre-release
for
shop
executor
service,
and
we
basically
also
just
did
that
for
the
job
executor
service.
E
If
you
want,
please
check
out
the
release,
notes
for
version
0.1.5,
just
keep
in
mind.
With
this
release.
We
are
switching
from
kubernetes,
manifest
based
installation
to
help
chart
based
installation
which
makes
a
lot
of
things
so
much
easier.
Yeah.
Essentially,
you
can
install
it
now
using
helm,
install
and
you
just
refer
to
the
helm,
chart
that
we've
attached
to
the
release.
E
Key
features
that
we've
added
is
official
support
for
captain
0.10.
It
also
works
with
0.11.
It
just
means
this
is
the
target
version
that
we've
worked
on.
We
added
an
image,
pull
policy
feature
for
the
tasks.
E
Let
me
quickly
show
you
what
this
looks
like.
Essentially,
it
means
when
you
specify
an
image
that
you
want
to
run
with
job
executor.
You
can
now
also
specify
the
image
pull
policy
for
this
image
same
as
you
do
with
kubernetes.
So
it's
a
great
way
of
saying
use
the
image.
If
you
already
have
it
or
always
pull
it
in
case
you
use
the
latest
image.
Yeah
just
gives
you
more
flexibility
and
beside
that
some
minor
improvements,
documentation,
tasks
etc.
So
yeah,
that's
it
from
my
side.
Any
questions.
B
E
Yes,
there
will
be
an
automated
pull
request,
an
artifact
hub
around
midnight.
Regarding
this
new
release.
However,
I
think
I
need
to
tweak
some
things
of
this
automation
and
manually
because
we're
switching
from
the
cube
anita
space
deployment
to
a
help
chart
based
deployment
and
I
believe
in
artifact
hub.
You
can
directly
then
link
this
helm
chart
which
will
make
it
1000
times
better
right.
E
D
D
Yes,
you
should
be
able
to
see
it
now
right,
okay,
so
the
first
task
we
work
on
is
on
creating
a
new
repository
for
for
our
helm,
charts
right
now.
The
helm,
charts
deployed
or
created
by
captain
for
the
captain,
installation,
plus
the
jmeter
service
and
the
home
service,
are
currently
under
a
google
cloud
storage
and
we
would
like
to
get
rid
of
this
and
instead
use
some
special
repository
over
github
to
store
them
and
to
store
our
hemp,
charts
and
yeah.
D
E
Is
there
like
a
grace
period
for
which
both
repositories
will
be
filled
with
latest
versions?
Will.
D
No,
the
old
one
will
be
cl
available
for
another
extra
release,
so
basically
with
captain
14
or
15,
we
will
delete
the
odd
repository,
but
it
will
only
contain
the
hem
charts
until
captain
12..
D
D
If
not,
then
we
can
jump
to
the
next
one,
so
it
also
provides
some
more
automation
towards
renovate
our
new
dependency
manager
that
we
have.
If
tests
are
green,
when
a
new
pr
is
provided
by
renovate,
it
will
be
also
automatically
merged,
which
is
a
really
cool
feature.
So
we
don't
need
to
look
at
the
prs
anymore
if
the
bumping
of
our
libre
library
works
and
also
improved
the
integration
tests.
Now
we
also
check
for
out
of
memory
errors
and
that's
everything.
B
F
I'm
going
to
present
two
pure
requests
today,
so
the
first
one
is
about
showing
a
dialogue
to
prevent
data
loss
on
unsafe
forms,
and
in
this
specific
case
we
have
the
project
settings
page,
which
will
in
future
also
allow
to
change
other
project
settings
and
therefore,
when
you
change
data
there,
yeah.
G
F
User
should
be
notified
before
leaving
the
page,
and
this
is
done
or
a
little
bit
changed
in
regards
to
how
it
was
in
the
past.
In
the
past,
you
would
have
immediately
seen
a
data
changed
pop-up
here
now
we
changed
that
so
that
as
soon
as
you
try
to
leave,
for
example,
I
tried
not
to
enter
the
environment
screen
yeah.
F
You
get
a
notification
here
that
you
have
pending
changes,
and
I
have
the
options
to
stay
on
this
page
or
to
discard
changes
and
leave
the
page
and
if
the
form
would
have
been
yeah
filled
out
correctly,
so
that
I
can
save
the
changes.
I
can
also
save
them
from
this
pop-up,
as,
of
course
as
if
I
choose
to
stay,
my
changes
will
not
be
lost
and
I
stay
on
this
page.
If
I
choose
to
discard
changes
yeah,
they
will
be
discarded
and
I
will
be
redirected
to
the
next
page.
F
And
if
I
choose
to
save
my
data,
then
it
will
save
the
data
and
then
redirect
or
for
example,
in
this
case,
because
the
credentials
are
not
correct,
show
the
warning
or
whatever
yeah.
One
known
issue,
which
I'm
still
working
on
is
that
if
I
have
changed
so
my
form.
G
F
Unsaved
data
here
and
I
changed
the
project
here.
On
top
you
see
it's
changed
now
to
the
test
project,
but
I'm
still
on
the
stock
shop
page,
so
yeah
not
sure
how
to
figure
that
out,
but
looking
to
fix
that
as
well,
so
that
the
value
here
changes
only
if
the
redirect
really
works.
F
If
page,
oh
yeah
sure,
so,
if
I
try
to
refresh
the
page,
then
first
I
will
see
the
browser
native
popup
here,
because
there
is
no
workaround
to
this.
Only
the
browser
can
enter
or
cancel
the
refresh
or
reload
to
another
page.
But
if
I
click
on
a
board,
so
I
choose
to
stay
on
the
caption
page,
then
I
will
also
see
the
notification
here
on
top
on
the
bottom.
F
Yeah
and
the
second
pro
request
I
would
like
to
present
is
introducing
captain
color
scheme,
so
we
used
in
captain,
as
you
can
see
on
my
page,
so
my
local
captain
installation
here
comparison
component
as
library
and
they
come
with
a
default
theme
having
these
turkeys
buttons
and
yeah.
This
color
scheme
there,
and
now
we
also
introduced
this
pro
request-
is
not
yet
merged.
F
So
that's
why
I'm
just
showing
the
request
introduced
now:
yeah
the
captain
blue,
as
you
know
it
from
the
title
here
from
the
header
bar
across
the
page
and
also
other
interaction
items.
So
basically,
when
we
have
a
look
here,
yeah
just
some
places
where
we
used
colors,
we
have
the
default
hover,
active
color
and
those
are
used
for
interaction
items
like
buttons
check,
boxes,
radio,
buttons
and
so
on.
Yeah,
so
that
we
have
a
consistent
color
scheme
across
the
whole
captain's
bridge.
G
F
Branding
directory,
so
if
you
use
custom
branding,
then
you
would
need
to
define
all
these
colors,
of
course,
in
your
custom,
css
and
yeah.
G
F
The
values,
as
you
can
see
here
or
if
you
don't
provide
any
values
as
here
in
this
captain
theme
css,
then
it
will
use
the
default
barista
components
colors.
So
this
turkey's
buttons
again.
H
You
should
now
see
my
screen.
If
not
please
yell
alright.
Today,
I've
won
four
requests
to
show
you
or
actually
it's
the
result
of
a
couple
of
requests,
which
is
the
service
that
germany
has
already
told
you
about.
So
actually,
this
new
service
was
a
collaboration
between
andre,
ana
and
myself.
H
So,
first
of
all,
thank
you
to
you.
Both.
You
did
a
very
great
job
in
helping
to
bring
the
service
to
life
and
yeah.
Basically,
this
is
the
swagger
documentation
for
the
new
service,
which
will
then
be
supposed
to
replace
our
current
configuration
service
and
because
of
that,
it
pretty
much
provides
exactly
the
same
functionalities,
with
the
additional
prerequisite
that
it
needs
a
mandatory
git
remote
upstream
repository,
but
that
in
turn
gives
gives
us
a
number
of
advantages.
H
The
first
important
one
is
that
we
are
now
able
to
run
multiple
replicas
of
the
service
and
because
now
the
remote
upstream
repository
is
a
single,
the
single
source
of
truth,
and
we
also
actually
don't
need
a
persistent
volume
attached
to
the
individual
parts
anymore.
But
instead
they
can
use
empty
deer
for
storing
the
data
which
is
created
at
each
start
of
the
pot.
H
So
in
order
to
reduce
the
probability
of
running
into
any
rate,
api
rate
limit
restrictions,
for
example
with
the
github
api,
and
that's
why
we
added
the
option
of
when
a
specific
resource
is
fetched.
We
now
provide
the
option
of
providing
a
git
commit
id,
which
means
that
the
resource
for
that
given
commit
id
should
be
returned
and
that
allows
the
resource
service
to
first
check
in
its
local
git
history.
H
H
So
if
and
we
also
plan
to
allow
the
captain,
execution,
plane,
services
and
the
shipyard
controller
to
kind
of
or
use
the
same,
git
commit
id
and
refer
to
that
one
over
the
course
of
a
complete
sequence.
H
And
that
means
that,
for
example,
if
evaluation
sequence
for
that
particular
git
commit
that
is
executed,
then
we
should
actually
have
at
most
n
calls
to
the
actual
upstream.
Where
n
is
the
number
of
replicas
of
the
resource
service
that
is
currently
running,
so
that
should
reduce
the
number
of
costs
to
the
remote
upstream,
quite
a
bit
and
yeah.
What
else
is
there
to
say?
Yeah,
maybe
one
nice
thing,
because
we
all
like
things
that
are
nice.
H
If
you
look
at
the
coverage
statistics
here,
you
see
that
the
current
configuration
service
has
a
very
sad
value
of
8,
which
is
the
lowest
in
the
list.
But
now,
if
we
direct
our
attention
to
the
resource
service,
which
is
the
its
successor,
it
is
now
our
coverage
champion.
It
has
83.44
so
yeah
that
was
a
nice
by-product
of
this
new
implementation.
Having
this
additional
coverage.
I
Yep,
hopefully
you're
seeing
my
screen
so
some
also
some
good
news
regarding
what
fullerene
mentioned,
creating
a
new
resource
service
which
will
be
a
successor
of
the
configuration
service.
There
is
a
need
when
converting
from
configuration
service
to
resource
service,
to
have
ups
to
have
an
upstream
repository
for
each
project.
Therefore,
we
added
a
warning
for
the
missing
upstream
in
specified
projects
during
the
captain
update,
upgrade
comment.
I
This
should
act
to
actually
warn
the
user.
That
yeah
please
create
an
upstream
repository
for
this
project.
We
have
kept
an
update
and
therefore
it
should
be
working
properly.
Also
the
update
yeah.
The
second
thing
was
actually
removing
the
cli
command,
which
was
deprecated.
This
command
was
captain
on
board
service.
I
This
command
was
actually
until
now
available
in
captain,
but
from
now
it
won't
be
available,
so
our
users
will
need
to
switch
to
from
captain
onboard
service
to
captain
create
service,
that's
kind
of
a
new
thing
too,
and
also
we
have
a
new
integration
test,
developer
documentation
which
should
be
nice
for
developers,
the
the
which
one
to
which
one
to
contribute
to
the
project.
I
There
is
a
detailed
description
of
the
structure
of
the
integration
tests,
how
the
new
integration
tests,
how
to
run
the
integration
tests
remotely
on
github
or
also
locally,
when
running
the
test
locally.
There
are
steps
to
how
to
set
up
your
environment
on
k3d
and
also
mini
shift,
and
also
how
to
run
the
integration
tests
with
captain
yeah.
So
that's
actually,
all
from
my
side
are
there
any
questions.
C
Yeah
one
question:
did
you
add
the
commentation
updates
or
remove
missing
eye
command
so.
A
Yeah,
so
thanks
everyone
for
the
demos
and
yeah.
I
just
wanted
to
quickly
discuss
google
summer
of
code.
A
We
discussed
it
a
bit,
maybe
one
or
two
months
ago,
so
the
applications
will
open
next
week
and
yep.
The
question
is
whether
we
want
to
apply
as
a
captain
community.
I
think
that
technically
it
would
be
possible
if
we
could
have
a
few
mentors.
Usually
first-year
projects
just
get
one
or
two
student
slots
and
yeah.
It
would
mean
three
months
of
part-time
coding
in
sabaan
by
a
student
if
a
project
gets
accepted.
A
So
I'm
not
sure
what's
commonly
a
practice
for
the
teams,
so
do
you
usually
do
summer
mentorships.
B
I
mean
in
the
past
we
had
mentorships,
one
was
for
lfx
and
the
other
one
was,
I
think,
what
was
it
called
community
bridge
or.
B
Right
and
yet,
but
this
was
not
over
summer,
it
was
more
in
an
autumn
or
yeah
or
spring
time
frame
and
I'm
not
sure
or
I
think
we
we
should
think
about
that.
B
A
A
I
want
to
just
create
a
repository
use,
whatever
project
ideas
with
some
documentation
and
then
should
we
decide
to
apply
to
google
summer.
Of
course
we
can
just
use
and
apply
quickly,
so
we
will
still
have
a
few
weeks
to
actually
whether
you
want
to
do
it
or
not,
maybe
find
potential
mentors
like
we
do
in
other
projects
and
yeah.
After
that,
we
can
see
whether
we
have
enough
interest
and
enough
commitment
to
apply,
because
google
summer,
of
course
is.
Rather
it
requires
several
hours
per
week
from
each
mentor.
A
So
it's
still
a
kind
of
commitment,
especially
during
the
summer
when
people
usually
take
vacations
and
from
what
we've
seen
in
many
projects,
you
need
at
least
three
mentors
to
be
sure
that
your
student
doesn't
get
left
without
mentors
that
particular
period
of
time
in
the
summer.
So,
let's
see
yeah,
I
will
try
to
prepare
a
more
concrete
proposal,
maybe
for
the
next
meeting
committing
in
two
weeks
and
then
we
can
see
what
we
will
actually
apply.
A
A
There
were
several
threads
about
backstage
integration
being
raised
by
different
people.
So
basically
we
have
two
parts
of
the
question.
One
is
backstage
service,
which
is
probably
not
the
most
critical
part,
because
many
things
could
be
done
by
cli.
Hence
job
executor
service,
if
packaged
properly,
another
site
is
actually
having
a
plugin
for
backstage,
which
would
display
some
key
metrics
from
captain,
and
this
part
was
brought
up
multiple
times
over
the
past
months.
So
I
think
that
we
definitely
need
to
look
in
this
story
right
now.
A
But
yeah
immediately,
I
didn't
forget
about
this
story,
because
yeah
also
I'm
also
a
big
fan
of
backstage.
There
was
also
a
proposal
just
yesterday
about
having
a
backstage
instance
for
captain.
So
who
knows.
G
B
I
mean
with
the
help
of
istio.
We
have
the
use
case
of
doing
blue
green
deployments
where
we
deployed
a
new
version
next
to
the
old
one,
but
I
agree
with
you
that
we
don't
collect,
then
the
metrics
from
both
deployments
and
then
visualize
them
or
compare
them.
We
just
take
a
look
at
the
new
one
and
keep
the
old
one
as
a
backup
in
case
we
need
to
do
a
rollback,
but
right
now
there
are
no
efforts
towards
ap
testing
or
doing
comparison
between
the
old
and
new
version.
G
F
A
G
Exactly
so,
basically,
in
a
b
testing,
you
would
like
to
have
a
sophisticated
comparison
logic.
So,
as
captain
calculates
the
aggregated
score,
the
score
aggregation
between
a
two
sets
of
execution
may
be
a
bit
more
challenging.
G
B
G
B
Very
good
hint,
dimitri
yeah,
that
was
a
very
good
hint,
because
what
we
do
have
is
a
an
integration
that
works
with
argo
cd,
argo
rollouts,
and
this
allows
you
to
then
perform
a
kind
of
blue
ap
testing
using
argo,
rollout.
A
So
basically,
what
we
will
need
is
visualization
part,
maybe
maybe
some
opportunity
to
trigger
a
sequences
based
on
events
happening
in
this
tab
so
yeah.
Maybe
we
would
need
to
root
comparisons
back
to
cloud
events,
I'm
not
sure
so
that
they
go
through
the
loop
of
basically
creating
well,
basically,
exposing
these
differences
as
a
part
of
slas.
G
Correct
and
when
we
will
go
to
the
detailed
design,
I
can
suggest
a
few
statistical
approaches
to
have
a
comparison
between
a
two
sets
of
metrics,
because
a
small
cellos
being
added
is
more
difficult
to
perform
the
comparison,
because
in
opposite
to
blue-green
deployment,
when
you
need
to
get
to
the
suggested
thresholds
or
not
get
under
the
suggested
thresholds.
G
Here,
you
need
to
have
an
a
kind
of
a
comparison
between
two
distributions,
so
several
statistical
methods
could
be
used
in
order
to
have
to
have
a
verdict
which
one
is
better.
B
An
engineer
correct,
correct
yeah,
as
as
steven
drew
now
brought
up
the
the
idea
or
hint
regarding
argo.
I
then
figured
out
that
we
discussed
this
topic
a
year
ago
and
therefore
we
filed
the
captain
enhancement
proposal,
which
proposes
to
use
argo
ct
for
a
b,
blue,
green
and
canary
deployments
and
dimitri
before
you
file
a
new
issue.
May
I
ask
you
to
read
over
this
enhancement
proposal
and
just
directly
provide
feedback
there
or
collect
information
from
that
enhancement
proposal.
Absolutely.
A
B
Is,
or
was
the
first
discussion
that
we
had
in
in
this
direction?.
C
A
One
reminder
to
everyone:
we
have
an
open
poll
for
meeting
in
a
park,
slash
friendly
time
zone
if
you
haven't
voted
yet,
please
do
if
you
want
to
participate.
So
the
idea
is
that
this
meeting
is
not
really
suitable
for
asian
and
pacific
region.
A
Maybe
just
moving
one
of
the
captain
community
meetings
maybe
doing
some
other
kind
of
arrangement,
but
I
think
it
would
be
nice
to
at
least
try
and
see
whether
we
could
have
critical
mass
of
users
and
contributors
to
do
something
efficient
there,
because,
if
so
yeah
we
could
really
move
some
of
the
meetings
so
not
creating
new
meetings,
but
just
doing
sometimes
on
balancing.
A
So,
if
it
happens,
I
would
expect
it
to
be
something
like
eight
or
nine
a.m.
Utc,
it's
basically
a
friend
of
europe
friendly
for
epac
region
and
also
friendly
for
those
on
the
west
coast
who
go
to
the
bet
lately,
wait
so
yeah.