►
From YouTube: Keptn Community Meeting - January 14th, 2021
Description
Meeting notes: https://docs.google.com/document/d/1y7a6uaN8fwFJ7IRnvtxSfgz-OGFq6u7bKN6F7NDxKPg/edit#heading=h.j66trjb15q1b
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
A
B
B
B
Okay,
don't
worry,
but
of
that
then
I
started
to
work
on
probably
okay
sort.
Most
people
are
mad
using
homebrew
to
install
all
her
bunnies.
So
I
started
okay,
let's
try
how
homebrew
is
working
and
how
it
can
create.
Probably
a
formula
for
that
and
I
cannot
share
my
screen
a
little
bit.
I
will
go
to
the
pull
request
and
explain
a
little
bit
about
what
how
how
room
is
working.
Probably,
I
will
also
switch
to
the
page
what
it
is
give
me
one
more
minute.
C
B
Yeah,
probably,
is
you
know
the
page
of
homebrew
literally
and
currently
we
had
the
idea
to,
as
I
mentioned,
for
pure
quest
with
christian
christian
russia.
Probably
we
could
do
it
also
to
do
it
in
our
own
tab,
but
he
found
a
better
idea
to
use
it
from
the
public,
so
everyone
can
easily
install
it
so
for
that
it's
really
easy
to
use
it.
B
B
And
literally,
the
command
to
create
the
whole
formula
was
very
straightforward
in
the
beginning,
so
you
can
easily
do
a
brood
create
and
what
homebrew
will
then
do
for?
You
is
really
to
fetch
all
the
important
changes
for
the
binaries,
so
it
will
try
to
fetch
the
information,
the
all
the
stuff.
Currently
it
says.
No,
it
already
exists
because
currently
homebrew
will
include
this
into
your
current
cross,
so
it
will
be
laying
down
in
your
current
system.
B
So
that
is
why
it's
not
working
anymore
right
now,
but
it's
not
a
problem
so
afterwards,
what
it
does
for
home,
recreate
it
will
enter
automatically
the
url.
It
gives
you
the
char
of
the
binary
and
then
I
started.
Then
I
decided
okay,
how
we
want
to
use
this.
Instead
of
consuming
the
finished
binary,
we
start
to
building
the
binary
locally
on
the
system
when
you
type
brew,
install
on
your
machine
really.
B
B
B
B
There
could
be
some
more
improvements
on
that,
so
we
could
use
github
actions
to
automatically
update
the
formula,
so
the
drawer
needs
to
handle
all
the
stuff
again
and
yeah
mostly
also
go
release
awards
for
that.
So
you
could
probably
also
use,
go
releaser
yeah,
and
this
was
the
powerful
idea
to
work
with
the
homebrew
stuff.
A
B
Yeah,
probably
when
you
want
to
support
when
you
want
to
support
more
of
that
when
you,
when
we
want
to
plan
to
start
the
door,
releaser,
there's
also
a
true
package.
It's
called
nfpm
where
you
can
automatically
build
with
fedora.
These
are
all
other
packages
so
like
that
packages
rpm
apk,
so
that
you
can
contribute
to
all
other
upstream
repositories,
and
I
can
also
probably
help
you
to
get
into
alpine
because-
and
this
is
also
the
last
time
app
and
it's
also
straightforward
to
include
us.
There.
A
Yeah,
we'll
have
a
look
at
the
go
releaser
because
I
mean
we've.
I've
actually
done
a
lot
of
work
with
github
actions
the
last
couple
of
weeks,
so
we
we
might
end
up
doing
some
more
stuff
if
it
makes
sense,
I
guess,
for
now
most
of
the
people
will
be
happy
with
the
brew,
install
command
and
maybe
with
the
quick
start
and
downloading
it,
but
yeah.
B
D
B
B
A
B
B
We
have
discussion
here,
something
like
that,
so
we
try
to
build
arm,
but
then
we
get
really
strange
errors.
So
not
really
strange,
but
it
says
cpu
info
is
not
available
and
I
didn't
know
from
which
library
was
coming
from
or
if
it's
the
core
runtime.
So
then
it
seems,
but
a
lot
of
because
I
stumbled
about
this
pr,
literally
where
I
had
mentioned
that
almost
a
lot
of
other
grow
bindings
are
also
broadened.
B
A
A
One
of
the
things
I
just
mentioned.
I've
spent
a
lot
of
time
with
github
actions
and
we
can
now
finally
say
rest
in
peace
to
travis
ci.
Unfortunately,
I
have
to
say
because
they
just
they've
just
killed
their
open
source
support
literally
and
there's
no
way
to
get
in
it,
except
for
writing
like
support
tickets
every
week,
and
I
guess
it's
easier
to
just
switch
to
a
system
that
works
with
open
source
right
now.
A
With
that,
I
cannot
show
you
a
pull
request,
but
I
can
just
tell
you
a
little
bit
about.
We've
done
a
lot
of
work
to
make
sure
that
we
don't
have
to
spend
the
time
again
if
we
need
to
migrate
again,
so
we're
trying
to
make
sure
everything
is
in
shell
or
bash
scripts
and
is
reproducible
on
local
environments
in
the
worst
case.
A
So
if
we
have
to
switch
again,
it
should
just
be
the
matter
of
executing
the
scripts
on
a
different
system
and
also
we've
improved
the
reporting
by
just
creating
files
that
have
reports
in
it.
I
will
show
you
in
a
second
then,
and
also
we've
improved
with
artifacts,
so
what
I
mean
with
artifacts,
for
instance,
our
master
builds,
are
now
no
longer
stored
on
google
cloud,
but
they
are
stored
in
the
github
actions
tab
here
on
the
repository,
you
can
click
on
ci
and
select
the
branch
master
here
or
also
release
branch.
A
If
you
want,
and
then
you
see
the
last
couple
of
builds,
that
we
did
on
master
branch
and
this
one
was
four
hours
ago
and
if
you
look
at
it,
you
can
see
there's
a
prepare
step.
There
is
a
unit
test
for
the
cli
which
is
running
on
macos
and
linux.
As
far
as
I
know,
and
then
there
is
a
build
that
runs
on
mac,
os,
linux
and
windows.
So
if
you
can
click
on
it,
you
can
usually
see
a
little
bit
more,
but
the
the
column
width
is
a
little
bit
low.
A
A
Then
we
have
unit
tests
that
are
executed
for
the
microservices
unit
tests
that
are
executed
for
our
ui,
and
if
the
unit
tests
work,
we
do
the
docker
builds
and
upload
the
docker
images
to
docker
hub
and
also
what
we
do
is
we
build
the
helm
chart
and
we
upload
the
help
chart
as
well
as
the
cli
builds
here
as
a
github
artifact.
So
if
you
scroll
down,
you
can
see
the
artifacts,
and
here
you
can
get.
The
cli
builds
basically
builds
from
this
commit
on
a
master
branch.
A
So
this
is
the
latest
cli
so
to
say,
and
also
a
helm
chart
and
what's
interesting
with
the
help
chart,
I
want
to
show
you
this
one
has
a
values,
yaml
and
other
actually
value
example
is
not
the
right
one.
This
one
has
templates.
We
go
control,
plane
templates
that
I
believe,
use
the
correct
version.
A
A
Anyone
knows
out
of
their
head
where
the
version
is
stored
for
captain
it's
in
the
chat,
chat
yeah.
So
what
we
do
here
is
we
write
the
version
in
here
and
this
version
is
actually
used
for
the
docker
images
within
our
manifests,
and
I
mean
right
now.
This
is
just
the
zero
eight
one
dash
def,
but
this
might
as
well
be
zero
at
one
dash,
def
dot,
2021
january
14,
17
or
1500
whatever.
A
So
this
could
also
be
this
version,
and
that
means,
if
you
download
this
helm
chart,
you
could
basically
install
the
version
that's
coming
out
of
this
commit.
So
this
will
allow
us,
in
the
future
to
have
better
integration
tests
and
also
allow
us
for
for
end
users
to
just
download
the
artifact
here
and
apply
it
on
the
clusters.
B
A
Those
tags
right.
Yes,
so
I
think
it's
not
this
tag.
I
think
it's
it's
it's
a
dash
in
between.
I
need
to
check,
but
they
have
a
tag
like
this
and
just
one
thing
to
note:
we
will
clean
up
dev
images
after
30
days
or
something
so
if
you
install
this
on
a
cluster,
it's
really
for
the
purpose
of
testing
it
right
now
and
not
for
keeping
it
forever.
A
All
right,
the
second
thing-
I've
also
linked
it
here.
We've
also
spent
some
time
with
integration
tests
and,
as
you
can
see,
I've
pre-selected
integration
test
here
for
branch
master.
You
can
also
just
select
a
release
branch
here
or
any
other,
maybe
not
the
feature
branch
to
be
honest,
but
the
release
branch
works
because
we're
only
testing
or
because
you're
only
running
integration
tests
right
now.
A
If
the
build
process
has
finished
before
and
the
build
process,
essentially
a
full
build
takes
about
an
hour.
So
we
only
do
those
for
master
and
release
branches,
not
for
feature
branches,
the
integration
tests.
If
you
look
into
them
they
they
are
running
daily
right
now
on
the
master
branch
they
consist
of
five
jobs
and
these
five
jobs
then
end
up
in
a
report.
It's
stating
report
to
pull
requests
here.
This
is
wrong.
Actually,
I
need
to
ch.
I
need
to
fix
the
label.
A
It's
right
now,
just
a
report
to
artifact
and
the
end
result
of
the
tests
is
basically
stored
in
this
test
report.
I
can
show
you
that
in
a
second
for
now,
if
you
click
on
here,
you
can
see.
There's
multiple
tests
that
are
being
running
so,
for
instance,
we're
running
a
test
with
mini
shift
on
github
we're
running
k3s
on
github
in
two
different
versions.
A
A
You
see
those
different
steps
like
we're
installing
and
starting
k3s,
and
if
you
click
into
this,
you
basically
get
the
logs
of
this
step
and
I
need
to
say
this
is
much
nicer
than
travis
ci,
because
on
travis
you
just
get
like
a
bunch
of
locks
and
it's
not
grouped,
and
this
one
is
nicely
grouped
so,
for
instance,
if
I
don't
I'm
not
interested
in
the
install
process,
but
I
want
to
see
what
the
q
and
a's
notes
look
like.
I
have
a
debug
step
here.
A
A
So,
if
we're
on
a
master
branch,
it
uses
the
last
build
of
the
master
branch,
and
if
we
are
on
the
release
branch
you'd
use
the
last
build
from
the
release,
branch
and
yeah.
It
just
goes
on.
There's
a
couple
tests
involved:
you
can
click
for
it,
I'm
not
going
to
show
it
as
way
too
much
to
show
and
in
the
end
we
are
writing
a
test
report.
A
This
is
the
test
report.
Just
for
this
little
test,
and
basically
it
is
like
one
line
and
it
says
a
couple
of
successes
and
one
thing
that's
being
skipped,
and
there
is
no
header
here.
Obviously,
so
that's
that's
not
good,
but
this
all
comes
together
in
the
end
in
the
artifact
in
here
where
you
have
the
final
test
report
stored
and
this
one
is
now
a
markdown
file.
A
These
are
the
things
that
I
tested
so
captain
install
was
successful,
captain
off
was
successful,
and
then
we
have
the
tests
for
linking
stages,
quality
gates,
self-healing
delivery
assistant
and
the
continuous
delivery
use
case
in
the
end,
and
obviously
due
to
your
memory
and
cpu
restrictions
on
github,
we
cannot
run
these
on
github,
so
we
only
run
the
continuous
delivery
once
here
on
google
kubernetes
engine
and
if
we
ever
fancy
adding
any
more
target
platforms,
it's
really
quite
easy.
So
that's
also
a
great
success
of
this
change.
A
A
So
the
first
thing
you
need
to
do
you
need
to
say
you're,
not
allowing
failures
or
in
other
words,
you're
skipping
failures.
It
doesn't
matter
so
it's
green,
that's
a
little
bit
bad,
but
at
the
same
time
it
prints
out
all
the
errors.
If
there
are
any
in
the
end
report
here
so
here
you
can
see
on
mini
shift.
There
is
an
error
where
it
says:
checking
source
expected,
lighter
service
got
empty
string
and
we
could
now
dig
deeper
by
clicking
here
and
see
where
it
appears.
A
So
this
is
basically
these
are
all
the
errors
that
happened,
and
this
would
help
us
nailing
down
the
errors
actually
and
yeah.
At
the
same
time,
we
also
say
continue
on
error,
because
we
want
this
report
to
be
generated
in
the
end,
so,
for
instance,
this
error
here
on
mini
shift,
means.
If
we
look
at
the
final
test
report,
it
should
say
that
something
failed
on
mini
shift,
and
indeed
it
actually
did
so
here
we
have
a
failure
on
a
mini
shift.
B
A
A
Okay,
then
last
actually
yeah.
Let's,
let's
skip
this
next
thing.
The
captain
service
template
I'll
talk
about
it
next
week
anyway.
Last
thing,
the
quick
start
that
I
mentioned
before
got
updated.
So
previously
you
had
to
write
this
command
and
then
sudo
minus
e
bash.
If
you
wanted
to
install
captain
cli,
why
it
is
quick
start
and
you
didn't
fancy
using
brew
or
any
other
tools.
A
This
is
now
working
without
sudo.
That's
the
first
update
it's
calling
pseudo
in
the
script.
So,
if
you
look
in
the
script,
you
will
see
that
sudo
is
being
called.
It's
actually
being
checked
that
it
is
available
and
if
not,
we
will
exit
with
an
error
message
and
yeah.
The
second
thing
this
has
been
working
before,
but
I
just
want
to
mention
it
again.
If
you
want
to
install
a
captain
version,
that's
not
the
latest
stable
version.
You
can
influence
the
script
by
just
saying
which
version
you
want
and
yeah.
This
works
now.
A
If
you
want
to,
if
you
don't
specify
it,
it's
automatically
determined
using
a
system
command,
but
this
also
means
by
the
fact
that
it's
automatically
determined,
if
you
don't
specify
it
and
you
run
it
on
an
arm
mac
right
now,
I'm
not
sure
if
it
actually
works.
If
I
have
a
fault
back
in,
let
me
check.
A
No,
I
don't,
I
don't
think
it
actually
works
so
when
it
tries
to
download
a
file
that
doesn't
exist
because
we
don't
have
the
architecture,
it
basically
tells
you
an
error
occurred.
So,
but
this
is,
in
my
opinion,
this
is
fine,
because
once
we
release
the
next
version,
I'm
pretty
sure
we
will
have
to
target
architecture
set
just
one
thing
to
keep
in
mind.
A
E
Let
me
just
share
my
screen
all
right
today.
I
want
to
show
you
two
per
requests,
so
the
first
one
is
about
the
helm
service,
so
the
home
service,
when
every
time
it
deploys
a
new
artifact
or
applies
a
configuration
change
for
a
certain
service,
it
will
add
the
deployment
uri
for
that
service
and
include
it
in
the
deployment.
E
Dot
finished
event
this
and
this
information
is
needed,
for
example,
by
the
jmeter
service,
which
will
need
how
to
access
the
service
and
then
execute
the
performance
test
about
that
and
up
until
now,
this
deployment
url
was
composed
using
a
fixed
scheme.
So
within
the
service
we
do
have
the
availability
to
specify
the
ingress
protocol
so,
for
example,
http
or
https,
then
the
service
project
and
stage
are
figured
out
by
the
home
service
when
it
retrieves
the
deployment.
E
Then
it
deploys
the
service
and
that's
how
the
deployment
url
looked
and
then
christian
came
up
with
the
idea
of
an
enhancement
to
make
that
url
configurable
for
this,
for
that
url
by
defining
a
hostname
template
environment
variable
where
you
can
use
those
those
parameters
or
environment
variables
and
basically
use
that
variable
to
specify
the
format
of
the
deployment
url
yourself
and
basically
that's
what
has
been
done
in
this
pro
request.
So
now
it
will
not
use
this
fixed
scheme,
but
whatever
you
have
defined
in
this
environment
variable
if
you
have
defined
something.
E
A
So
so
the
idea
behind
this
actually
was
not
to
just
make
this
customizable,
because
why
the
hell
do
you
need
this
customizable
right
just
to
mention
this?
If
you
want
a
wildcard
ssl
certificate,
you
cannot
have
a
double
wild
card,
so
you
cannot
have
asterisk.asterisk.sumhost.com,
it's
just
not
possible,
or
at
least
not
easily
possible.
So
if
you
replace
this
little
dot
by
a
dash,
you've
solved
this
problem,
and
this
now
means
you
can
have
a
wildcard
certificate
for
your
captain
installation.
If
you
want.
E
E
E
We
introduced
a
new
property
to
this
get
project
payload,
which
is
called
the
shipyard
version
and,
as
the
name
suggests,
it
will
tell
tell
you
the
the
shipyard
version
of
that
particular
project.
So,
for
example,
if
you
haven't
updated
the
project
yet
that
version
will
be
set
to
0.1.7
and
after
you
have
upgraded
it,
you
will
get
the
version
0.2.0,
and
this
will
allow
you
to
keep
track
of
what
project
you
will
need
to
upgrade
and
what's
not
yeah,
it's
a
very
small
extension.
E
Yeah
yeah
all
right,
those
are
the
two
requests
I
want
to
show
you.
Also.
Let's
have
a
quick
look
at
the
breaking
items.
I
think
there
are
one
or
two
bug
fixes
that
I've
worked
on
so
okay,
this
one
is
still
on
the
review,
so
there
was
a
memory
leak
in
the
shipyard
controller,
but
that's
solved,
hopefully
with
the
pr
I
made
and
yeah
here
that
use
the
two
blocks
I
fixed.
E
E
To
and
use
that
to
to
switch
between
the
captain
installation
if
that
new
keyboard
at
this
context
happens
to
be
on
a
different
cluster
than
you're
currently
connected
to.
This
is
a
really
cool
feature,
but
there
was
an
error
when
that
environment
variable
wasn't
set
correctly.
So,
for
example,
when
it
was
pointing
to
a
file
that
didn't
exist.
E
E
D
A
D
D
What
it
looks
like
when
you
open
the
sequences
screen,
you
see
all
the
sequences
that
were
executed
here
are
some
information
still
missing,
because
we
have
separate
issues
to
yeah,
show
more
details
here
on
this
tile
about
this
sequence,
for
example,
which
services
is
affected
yeah
when
it
was
started,
how
long
it
took
and
so
on
yeah.
But
one
thing
that
you
already
see
now
here:
is
you
see
a
ongoing
task,
so
the
deployment
and
as
long
as
it's
ongoing?
D
So
it
started
it's
showing
that
yeah
it's
ongoing
and
when
it's
finished,
then
you
will
see
here
the
status
if
it
was
finished
successfully
or
with
an
error.
For
example.
Here
a
few
minutes
ago
I
sent
in
an
artifact
delivery
and
yeah
you
see
the
deployment,
then
the
test,
evaluation
and
release
step
was
executed
here
and
yeah.
You
see
also
ongoing
steps
so,
for
example,
yeah
when
the
deployment
as
it
was
now
takes
longer.
You
see
it's
still
ongoing.
D
A
D
A
D
It
that's
also
the
thing
so
I
group
it
based
on
this
starting
key
and
if
there
is
just
the
triggered
event
yeah,
it
would
show
that
it's
triggered
and
like
waiting
as
soon
as
I
get
the
started
event.
It
shows
this
yeah
loading
indicator,
which
will
be
also
animated
in
the
final
version
on
the
when
it's
finished
then
yeah
we
show
it
either
a
green
or
red
icon
with
the
status
of
that
event.