►
From YouTube: Keptn Community Meeting - January 20th, 2020
Description
Agenda
Spec changes on https://github.com/keptn/spec/pulls
Keptn Enhancement Proposals (KEP) on https://github.com/keptn/enhancement-proposals
aBOI on K8s-managed by Keptn
Outlook
Release 0.6.0 will be available in the next days
Overview of Release 0.6.0 in our next community meeting on February, 3rd
Notes
Please note that we have updated the Zoom meeting link to https://dynatrace.zoom.us/j/451021860
Also, from now on we will stream the community meetings directly to YouTube → find us here: https://www.youtube.com/channel/UCHMn9HyAMeb81bRlaOuZyuQ
Keptn Release 0.5.x: If you are using keptn version 0.5.x, please add --keptn-version=0.5.2 to your keptn install and keptn configure domain commands. Example:
keptn install --platform=gke --keptn-version=0.5.2
keptn configure domain YOUR.CUSTOM.DOMAIN --keptn-version=0.5.2
→ Remark: The keptn.sh/docs for Keptn Release 0.5.x will be appropriately updated and released along with the Keptn 0.6.0 release.
A
Let's
go
live
on
YouTube,
maybe
the
live
on
YouTube.
We
are
awesome,
so
welcome
everyone
to
today's
captain
community
meeting.
It
is
January
20th
2028,
it's
actually
a
nice.
If
you
look
at
it's
a
nice
date,
20
20
20,
it's
the
first
Kaplan
community
meeting
in
the
new
year,
I'm
an
indie,
crap
nur.
You
have
a
couple
of
minutes
in
the
beginning
from
me.
The
maturity
of
the
presentation
will
later
be
done
by
the
team
from
Abu
doc,
who
sits
there
in
the
table.
A
I'm
Berlin
hi
guys
and
that's
also
andreas
crema,
who
I
don't
know
actually
where
he
sits
right
now,
I'm
not
familiar
with
that
meeting
room
but
I'm
sure
it's
somewhere
in
Austria.
First
floor,
no
first
floor:
okay!
Well,
okay!
So
a
couple
of
I
want
to
quickly
highlight
the
agenda
and
some
notes.
As
you
can
see,
we
always
keep
the
community
meeting
notes
on
Google
Docs,
so
feel
free
to
also
check
it
out
later
on.
There's
no
spec
changes
today
that
we
wanted
to
discuss.
Also
there's
no
captain
enhancement
proposals,
but
as
a
reminder.
A
Folks,
if
you
have
any
proposals
for
things
you
want
to
see
in
captain,
the
enhancement
proposal
is
the
way
to
go.
Click
on
that
link,
and
you
will
find
all
the
information
how
you
started
proportional
and
what
we
expect
there
to
see
now
before
I
go
into
the
presentation
before
handed
over
to
Sebastian
and
Tim
and
andreas
a
couple
of
things.
We
want
to
highlight
an
Outlook.
A
So,
first
of
all,
thanks
to
all
of
you
for
all
the
feedback
on
the
0.6
beta
releases
to
be
pushed
out
there
and
of
last
year
a
lot
of
great
feedback,
also
special
thanks
to
to
the
folks
at
avataq.
A
lot
of
these
improvements
actually
made
it
into
the
final
release,
which
will
go
out
in
the
next
couple
of
days.
A
Then
there's
also
going
to
be
more.
You
know
community
meetings
coming
up.
Obviously,
what
we
try
to
do
in
the
next
couple
of
meetings
we
always
try
to
bring
in
external
contributors
that
are
actually
you
know,
is
either
you
know
implementing
captain
on
their
projects
like
what
we
hear
from
our
talk
today
or
contributors,
whether
it
is
from
our
friends
at
notice,
whether
it's
from
our
friends
at
Citrix
that
have
contributed
to
captain
so
check
out.
What's
what's
coming
there,
also
for
those
of
you
that
alive
today,
I
think
you
found
the
link,
we
changed.
A
It
zoom
meeting
link
to
a
new
webinar
style,
which
means
you
can
still
enjoy
obviously
as
life
here,
but
you
are
not
automatically
enabled
on
audio
and
video
that
just
makes
it
easier
for
us,
because
we
can,
you
know,
keep
the
conversation
a
little
more.
Let's
say
control
if
you
want
honor.
If
you
have
questions,
there's
a
question
feature
in
in
zoom,
so
click
on
the
Q&A
button
and
please
put
into
questions
I
will
moderate
them,
which
means
also
a
heads
up
to
you
guys
in
Berlin
and
others.
A
If
they're
questions
coming
in
I
will
either
throw
them
over
to
you,
but
I'll
try
not
to
interrupt
you
only
if
I
think
it
makes
perfect
sense
right
now,
otherwise,
we'll
also
keep
it
to
the
end.
I
hope.
That's!
Okay
with
you.
Some
last
notes
the
real
life
on
YouTube
as
well.
So
there's
a
YouTube
channel
where
this
is.
B
A
So
if
you
want
to
install
or
if
you
also
want
to
do
a
configure
domain
from
a
previous
version
in
needs
to
add
this
particular
version
to
it,
okay
make
sure
they
don't
forget
that
field.
Alright,
last
remark:
you
can
I
guess
you
can
read
for
yourself
now:
I
want
to
hand
it
over
to
Sebastian
and
team
in
Berlin
who
will
give
us
an
overview
of
the
I
would
say
kind
of
the
first
public
project
that
is
using
captain.
A
We
want
to
learn
more
about
and
that's
what
I
am
interested
in
I
want
to
know
what
is
the
Foresters.
So
what
what?
What
is
actually
you
know,
what
is,
what
is
I
will
talk
to
what
is
able.
Why
did
you
select
kubernetes?
Why
do
you
select
captain
and
how
does
it
really
help
you
I
think
that's
also
going
to
be
interesting
for
our
audience,
how
to
to
understand
better
what
problems
does
kept
really
solved
for
you
this
way,
okay,.
C
Perfect,
thank
you
Andy
for
the
nice
introduction
and
thank
you
for
the
invitation
to
present
our
use
case
and
together
with
the
team
year
with
Stefan
and
with
Viktor,
we
worked
and
I
would
say,
laughs,
three
and
a
half
four
months
really
really
heavily,
of
course,
with
you
together
on
captain,
but
of
course,
as
well
on
our
use
case,
and
this
is
what
the
presentation
today
is
about.
It's
not
only
slides
as
well.
C
There's
a
technical
presentation
and
I
think,
let's
dive
directly
into
it,
so
as
you
can
see,
we
selected
one
of
our
use,
one
of
our
solutions
exactly
for
the
first
use
case
on
captain
we
want
to.
We
want
to
show
you
today
and
to
understand
what
a
boy
actually
does
for
us
and
why
it's
important
a
few
words
about
the
agenda.
Okay,
I
think
this
is
something
like
we
will.
We
will
cover.
We
were
couple
today
to
understand
what
everybody
does
for
us.
C
We
talked
a
little
bit
about
what
we
are
doing
here,
our
dock
and
then
we
dive
directly
into
the
use
case.
Look
at
the
old
existing
architecture
take
a
look
at
the
new
one
and
try
to
figure
out
if
captain
can
solve
all
of
our
challenges.
So,
as
you
can
see,
we
are
I
would
say
not
a
very
large
company.
We
are
two
hundred
people
all
over
the
globe,
heavily
focused
in
Europe
and
with
a
few
offices
in
the
APEC
region
and
what
we
are
doing.
C
It's
sitting
between
our
ticket
system,
it's
OTS,
it's
open,
source
solution
and
WebEx
teams.
This
is
the
main
collaboration
solution.
We
are
using
an
in-house
and
for
top
priority
cases,
so
we
are
talking
about
high
severity
priority,
one
priority
to
where
parts
of
the
architectures
of
our
customer
is
not
working.
We
need
to
bring
everyone
in
one
room,
and
this
is
something
we
do
with
our
a
boy
solution.
Of
course,
you
know
there
are
solutions
out
there,
where
you
can
easily
register
a
web
book.
C
The
room
gets
created,
the
message
gets
promoted,
but
what
we
actually
want
to
achieve
was
bi-directional.
It
exchanged,
so
we
want
to
have
single
source
of
truth.
Single
source
of
truth
should
be
always
the
ticket
system,
but
we
want
to
have
something
like
the
functionality
in
the
chat
room,
that
we
can
all
work
together
to
the
back
to
the
ticket
system,
and
you
see
a
high-level
overview.
What
kind
of
technologies
will
be
involved?
I
show
you
on
the
next
slide
so
and
we
built
the
application.
C
I
would
say
two
and
a
half
years
ago,
and
two
and
a
half
years
ago
we
try
to
leverage
all
the
cool
stuff
that
was
at
this
point
available
for
us
and
our
team.
As
you
can
see,
every
part
of
the
service
actually
runs
in
docker
container,
so
we
are
leveraging
a
lot
of
stuff
from
the
AWS
cloud.
We
are
using
the
RDS
database.
C
C
I
want
to
have
this
as
well,
and
if
we
have
an
escalation
going
on,
would
this
be
possible
that
we
are
invited
as
well,
and
then
we
saw
okay,
our
a
boy
solution,
quite
cool
houses
right
now
solves
all
of
just
by
what
we
see
we
need
maybe
an
a
boy
for
each
customer,
and
each
customer
has
slightly
a
different
feature
set.
So
one
customer
has
a
very,
very
large
IT
organization
with
a
lot
of
sub
tips.
C
So
and
then
we
need
something
like
additional
logic
in
the
a
boy
solution
and
of
course
we
don't
want
to
work
on
Simula,
but
nearly
the
same
project.
So
we
we
looked
out
if
we
find
something
better,
that
the
existing
architecture
we
hadn't
we
had
in
place.
As
I
mentioned,
a
boy
runs
on
an
ec2
instance.
C
Everything
is
started
in
docker
containers,
so
we
had
for
the
for
the
deployment
we
are
using
github
a
second
repo
and
leveraging
CIC
functionality
with
a
lot
of
kit
LeBron
us
we
are
using
ansible
and
you're
using
terraform,
so
I
would
say
technology
that's
in
place.
It's
really
really
cool.
But
if
we
are
talking
about
to
bring
this
to
more
customers
with
the
same
code
repo,
it
could
be
a
little
bit
difficult.
What
we
see
as
well-
it's
not
really
scalable.
So
we
need
to
make
the
ec2
instance
bigger
and
bigger
and
bigger.
C
So
there's
something
that
you
don't
want
to
do
in
the
future.
And,
of
course,
if
we
are
talking
about
the
real
cool
deployment
features
like
blue
green
and
can
array
deployment,
maybe
we
can
do
it
with
the
existing
architecture,
but
it
would
be
a
lot
of
additional
I
would
say
pipeline
works
that
we
need
to
involve,
and
now
we
are
coming
to
the
to
the
biggest
con
of
our
solution.
If
we
are
talking
about,
we
want
to
bring
able
to
a
lot
of
our
customers,
maybe
to
all
of
our
customers.
C
There
would
be
a
lot
of
pipelines
involved
and
then
needs
to
be
a
better
that
it
needs
to
be
a
better
solution
out
there
and
to
give
you
a
better
understanding
if
we
are
talking
about
one,
a
boy
deployment.
That
is
what
is
currently
in
place
or
what
we
did
before.
We
started
the
project
together
with
the
captain
guys
we
had
the
project
and
one
project
is
a
lot
of
combined
pipelines.
We
have
these
kind
of
as
a
kid
leprosy
ICT
steps.
C
We
have
the
answer
was
there
and
then
we
have,
of
course,
to
terraform
steps,
and
this
is
something
like
the
combined
pipeline
and
if
we
are
talking
about
that,
the
pipeline
steps
I
mentioned,
and
we
now
want
to
bring
it
to
more
and
more
customers.
We
would
have
something
like
this
formula
and
in
the
end,
we
need
a
sounds
like
a
dedicated
person
was
just
working
on
the
pipeline's.
C
We
are
here
at
our
track
and
software
development
team,
15
people,
and
it
would
not
work
this
way
because
we
have,
of
course,
other
solutions
and
services
we
are
working
on
and
we
can't
hire
a
new
additional
pipeline
engineer.
Every
time
we
want
to
bring
out
a
new
service,
so
there
needs
to
be
a
better
solution
out
there
and
maybe,
as
you
all
know,
the
kubernetes
branch
wagon
is
I.
Think
it's
I
think
it's
a
really
hot
topic,
not
only
for
us
for
all
of
our
customers.
C
Everyone
is
talking
about
and,
of
course,
we
took
a
look
at
the
solution
as
well
and
said.
Maybe
this
could
be
something
for
us.
I
see
something
like
purlins
back
in
the
days
when
VMV
and
virtualization
finally
took
off.
It
was
something
like
the
platform
was,
was
over,
everyone
decided.
Ok,
this
will
be
the
next
platform
and
all
of
our
application
should
be
developed
exactly
on
this
platform,
and
then
we
took
a
look
and
found
some
really
really
really
cool
things
about
it.
Every
public
cloud
vendor
is
offering
a
kubernetes
solution.
C
We
already
have
the
state
management
integrated
in
the
solution.
So
if
something
goes
wrong
and
the
hardware
on
the
underlying
hardware
communities
is
taking
care
of
this
as
well,
we
have
the
integrated
scaling
of
the
workload
is
going
up
and
what
we
saw
as
well
is
great
platform
Aaron's
like
him
and
is
you
are
available
and
the
reason
why
I
think
we
are
thinking
that
this
kind
of
cool
platform
add-ons
and
now
develop
at
such
speed.
C
It's
because
kubernetes
seems
to
be
typically
every
now
and
for
us
I
think
this
was
something
like
the
main
decision.
Ok,
we
need
to
go
into
kubernetes,
because
the
future
are
the
platform
of
the
future
quite
cool,
but
what
we
saw
as
well.
He
itself
not
all
challenges
if
we
are
looking
at
this
kind
of
blue-green
deployments,
if
you're
looking
at
the
pipeline
complexity
or
the
outer
remediation
things
that
are
going
on
outside
right
now,
it's
not
really
possible
with
equipment.
C
It
just
alone,
and
this
is
the
reason
why
we
looked
further
in
further
and
one
perform
event
in
Singapore.
I
was
in
contact
with
Andy
Gardner
and
he
talked
to
me
and
said:
maybe
we
have
a
solution
for
this,
and
this
was
the
something
like
the
informal
kickoff
exactly
for
the
April
project,
to
run
it
on
kubernetes
managed
by
captain
and
that
what
what
we
want
to
show
you
today
is
that
we
actually
implemented
this.
So
we
have
right
now
and
a
boy
running
on
communities
managed
by
captain
in
two
stages.
C
We
are
using
production
and
staging.
Obviously
we
will
deploy
a
new
feature
and
we
will
see
that
the
new
feature
is
directly
deployed
to
the
staging
environment
and
in
the
background,
before
we
promoted
through
production,
we
will
see
some
internal
and
integrated
tests,
and
hopefully
everything
in
the
picture
has
developed
very
cool
and
we
can
promote
it
automatically
to
to
the
live
environment.
And
this
is
something
we
will
see
and
we
were
covered
today.
C
So
and
let's
take
a
look,
as
I
mentioned,
we
are
using
WebEx
team.
So
it's
a
collaboration
solution,
that's
very
similar
to
slag
on
Microsoft
teams,
whatever
you,
whatever
you
use
in
your
organization
and
as
you
can
see,
we
created
two
rooms
for
preparation.
One
room
is
with
staging
environment
and
oneness
with
the
production
environment,
and
now
I
will
show
you
our
triggered
bot.
The
bot,
of
course,
we'll
send
out
automatically
messages
and
create
rooms.
C
Ups
interfaces
I
just
asked
for
for
40
help
and
I
do
the
same
in
the
life
system,
and
we
will
see
that
we
get
back
exactly
the
same
amount
of
features.
So
and
now
what
we
want
to
do
is
exactly
what
I
described
earlier
on
the
slide.
We
want
to
deploy
a
new
version
that
is
directly
deployed
to
the
staging
environment,
and
then
we
will
see
that
some
tests
will
take
will
take
place
in
the
background
and
in
the
time
we
wait,
I
will
hand
over
to
Andrea's
kuma.
C
You
will
tell
us
a
little
bit
more
about
captain
and
how
you
can
use
captain
and
nice
other
features
that
you
actually
can't
pull
off.
If
you
leverage
captain
as
management
platform
for
preventative
and
to
actually
deploy
a
new
feature,
I
prepared
everything
here,
you
work
with
the
with
the
captain
bridge
through
this
CLI
interface
and,
of
course,
we
generated
all
the
images
in
our
container
registry.
C
So,
as
you
can
see,
maybe
in
the
background
we
connected
captain
as
well
with
our
WebEx
tips
platform
and
now
you
can
see
a
configuration
change
was
detected
and
we
will
see
if
the
command
is
finished
and
then
we
will
quickly
check
out
if
the
if
the
command
is
now
working
in
the
staging
environment
and
we
will
check
again
in
the
live
environment
if
the
command
is
still
not
available
and
then
I
will
handle
that
one.
Yes,.
A
And
the
busting
as
well
we're
waiting
here,
just
a
quick
reminder
for
everyone
that
is
in
attendance.
If
you
have
any
questions,
please
put
them
into
the
Q&A
field
of
the
zoom
meeting.
Well,
a
couple
of
questions
already
came
in
also.
You
can
also
use
the
chat
feature
you
can
send
it
directly
to
me.
I
got
a
couple
of
questions,
but
I
will
keep
them
to
the
a
and
probably
a
second
one
interrupt
your
flow,
but
I
thought
now
it's
a
perfect
time
to
remind
folks.
C
Thank
you.
We
will
have
a
Q&A
session
in
the
end,
but
if
something
is
coming
up,
sets
makes
sense
to
answer
to
and
please
feel
free
to
show
the
message
to
us
and
what
we
see
we
got
an
answer
can
be
Greek
god,
another
event
from
Captain
the
deployment
is
finished
and
now
it
should
be
available
in
the
staging
environment.
As
you
can
see,
I
triggered
a
bot
once
again
waiting
of
the
reported
reply,
and
then
you
see
we
will
maybe
showcase
this
use
case
as
well.
C
C
Yeah,
you
can
see
this,
and
now
we
will
wait
for
the
tests
that
run
in
the
background
against
our
SL
Ison,
that
defined
s
ellos.
And
if
we
match
the
test
see,
the
feature
will
be
promoted
through
the
live
environment
and
with
this
I
will
jump
back
to
the
presentation
and
actually
hand
over
to
you
under
yes-
and
you
just
give
me
please
assigned
if
I
should
take
forward
a
slide
and
yeah
all.
B
Right,
thank
you
really
cool,
because
we
already
have
seen
now
captain
using
for
making
a
new
deployment,
so
you
pushed
a
new
version
on
which
had
a
tag
called
feature,
and
now
I
would
like
to
explain
you
what's
really
happening
in
the
underground.
What's
captain
is
doing
for
you?
So
let's
get
started
at
a
really
really
high
level
and
therefore
I
will
explain
you.
What
is
captain
in
one
sentence.
B
So
captain
is
an
event-based
control
plane
for
continuous
delivery
and
automated
operations
for
cloud
native
applications,
so
Canton
provides
you.
Therefore,
a
set
of
events
which
allow
you
to
model
the
processes
in
continuous
delivery
and
also
for
automated
operations
and
Kenton
is
unique
compared
to
pipelines
compared
to
terraform
scripts
and
so
on,
and
now
I
would
like
in
the
next
slide.
Please
so
again,
please,
because
captain
gives
you
a
scribblers
delivery
and
operations.
What
does
that
mean
without
to
write
a
single
line
of
pipeline
code?
B
You
get
out
of
the
box,
Bluegreen
deployments,
also
automated
operations
and
so
on,
and
of
course,
this.
The
descriptions
can
be
shared
across
a
number
of
micro
services,
so
you
can
reuse
them,
and
here
we
are
separating
the
concerns
between
certain
site,
reliability,
engineers,
DevOps
and
tests
so
side,
reliability.
Engineers
are
mostly
responsible
for
the
process.
For
example,
they
have
to
need
to
add
a
further
stage
which
can
be
used
for
security
checks.
B
Its
actions
next
and
finally
having
the
set
of
events
also
allows
you
to
trace
the
events
which
are
necessary
for
doing
the
deployment
for
doing
the
operation
stuff
and
this
set
of
immense
and
can
be
visualized
in
a
bridge.
And
finally,
you
get
a
full
observability,
what
has
been
deployed,
what
has
been
tested
and
so
on.
Okay,
and
now
before
we
have
seen
how
we
can
trigger
captain
having
a
new
artifact
and
now
I
show
you
on
a
real
high
level,
how
this
continuous
delivery
process
really
works
on
captain.
B
So
here
we
have,
we
got
triggered
or
Kevin
got
triggered
by
having
a
new
artifact
available.
This
can
be
done
using
our
CLI
as
Sebastian
baited,
but
it
can
also
be
done
using
the
REST.
Api
Kevin
provides
you
as
first
step.
We
update
the
respective
configuration
in
your
case.
You
have
a
deployment.
Equipment
is
deployment
in
the
background,
and
here
we
are
setting
the
new
image
and
then
this
new
deployment
will
be
stored
in
the
get
repo.
B
B
Bluegreen
deployment,
so
we
still
have
two
versions
and
as
next
step,
we
can
now
trigger
our
tests
with
muchacha
meter
tests,
and
this
G
meter
tests
run
performance
tests
and
will
produce
metrics,
and
the
metrics
then
are
of
course
monitored
either
by
a
notorious
Prometheus
on
mutants
and
captain
allows
you
to
validate
this
matrix
using
a
so
called
lighthouse
service,
and
we
will
see
details
how
this
will
work
in
the
next
slide.
But,
finally,
the
lighter
service
gets
back
an
emulation
result.
B
This
can
be
successful,
warning
or
failed,
and
if
it's
failed,
captain
triggers
a
rollback
to
the
gold
old
version,
or
in
case
it's
good.
We
can
repeat
those
steps
for
the
other
stages
perfect
and,
as
you
see,
the
core
of
this
continuous
delivery
process
is
really
the
built-in
quality
gate
and
this
disk
quality
gates.
We
can
automate
on
quality
checks
next
slide.
Please-
and
this
is
really
essential,
because
in
the
cloud
native
world
you
have
multiple
deployments
per
day
and
it's
simply
impossible
to
do
this
quality
takes
by
hand.
B
So
you
need
a
way
to
automate
the
quality
checks,
and
here
captain
uses
well-known
concepts.
First,
it
allows
you
to
define
service
level
indicators
and
this
service
level
indicators
can
be
metrics
like
to
respond.
Some
error
rate
superior
usage,
but
also
infrastructure
and
business
metrics
and
using
this
SL
ice
then
allows
you
to
define
s
loz
perfect,
and
this
as
a
noise
can
be
compared
against
fixed
threshold,
but
also
against
previous
runs
off
of
your
deployment
and
deployment
process
and
technically.
B
B
We
are
using
as
it
is
common
in
the
cloud
native
El
Gamal
for
formulating
the
s
ellos,
and
here
our
first
version
is
1.0
of
these
Asano's.
Then,
as
next
configuration,
you
can
specify
the
comparison
strategy
here
we
are
using
a
single
result,
which
means
we
compare
against
the
result
of
the
previous
run,
and
then
we
are
already
ready
to
define
the
s
lies
as
Eliza
our
service
level
indicators,
and
here
our
first
SLI
is
the
response
time
and
Mort
Italy.
B
We
are
using
the
95th
percentile
and,
as
this
is
like
and
then
be
used
to
define
criteriums
criteria,
and
here
you
can
always
differentiate
between
a
pause
and
a
warning.
So
everything
is
fine.
We
get
the
pass
if
the
response
time
is
below
200
millisecond,
we
get
a
warning
if
the
response
time
is
below
500
millisecond.
So
this
is
our
first
SLI
next.
The
next
SLI
shows
us
and
the
SQL
statements,
and
now
we
are
comparing
whether
this
SLI
and
is
SQL
statements
change.
B
So
we
get
a
pass
if
the
SQL
statements
did
not
change
at
all,
comparing
the
previous
run
with
the
current
deployment,
and
we
accepting
it
as
a
warning
if
we
only
have
a
relative
change
of
plus
minus
5
percent
cool.
Now,
maybe
this
looks
a
little
bit
abstract.
Let's
see
how
how
Kevin
does
the
emulation-
and
here
we
go
through
an
example
on
using
the
grading
here
on
the
right
side
if
the
value
is
between
200,
we
get
a
pass
and
for
each
pass.
We
get
one
point
for
warning.
B
B
So
we
get
the
pass
if
the
overall
we
get
a
pass.
If,
if
we
have
90%
of
the
overall
points
which
are
possible
and
we
get
the
warning
if
75
of
the
overall
impossible
points
are
reached
so
now,
let's
have
look
at
the
example,
which
makes
it
re
really
easy
to
understand.
Let's
assume
the
response.
Time
is
400
milliseconds
and
you
can
see.
400
millisecond
doesn't
fulfill
the
past
criteria,
but
it
fulfills
the
warning
criteria.
Therefore,
we
get
0.5
points.
B
C
As
you
can
see,
we
got
a,
we
got
another
captain
event
and,
as
you
see,
the
evaluation
did
not
pass
thanks.
Thanks
Richter
fire
muting
me.
Thank
you,
as
you
can
see
what
andreas
described
the
test
took
back,
took
place
in
the
background
and
it
didn't
pass.
So
what
should
actually
happen
to
the
interface?
If
we
are
looking
in
the
staging
environment,
the
command
should
be
still
available
because
was
a
direct
deployment.
C
As
you
can
see
it
still
the
old
world
and
the
new
commanders,
the
new
commanders
missing
what
we
want
to
do
now.
We
want
to
try
with
a
different
version
we
of
course
prepared
beforehand,
and
we
will
trigger
to
CLI
once
again,
but
this
time
with
another
text,
that
we
are
pulling
another
image
from
our
up
from
our
container
registry.
This
time
we
are
not
taking
the
the
git
command
with
the
registry
image
with
tech
feature.
Now
we
are
taking
the
image
with
your
tag
fixed
and
we
would
start
and
trigger
to
deployment
once
again.
C
B
B
B
So-
and
this
is
the
use
case
when
there
is
already
the
obligation,
for
example,
in
production
and
for
example,
you
see
that
a
problem
or
code
so
next
animation,
please
so
for
those
who
are
using
the
energize
you're,
definitely
family
with
problem
events
which
occur
when
dynaTrace
automatically
detects
a
problem
for
those
who
are
promised
users-
and
you
also
are
probably
aware
of
the
alert
concept
and
this
either
in
case
of
a
problem
or
not.
This
can
trigger
captain
and
based
on
this.
B
Based
on
this
problem
event,
we
can
look
up
better
to
use
a
defined
remediation
action
which
should
be
executed
in
production
in
order,
for
example,
to
heel
the
autumn,
the
application.
So
the
first
step
is
looking
up
indicator,
a
PO
whether
there
is
a
defined
remediation
action
and
if
so,
we
will
execute
this
remediation
action.
But
you
can
see
he
visualized
is
a
simple
rollback
of
one
service
from
version
to
version
one,
but
here
you
can
be
creative.
You
can
toggle
feature
flag.
B
It
is
again
meant
or
not,
and
we
can
also
wait
here
where
the
diner
is
closed,
with
the
problem
or
prometheus
closed
alert,
and
this
rehabilitation
is
done
using
the
lighter
service,
and
if
this
is
the
case,
we
can
either
inform
the
the
users
by
a
slack
or
the
WebEx
channel,
and
that
the
problem
was
solved
or
we
can
escalate
in
case
it
is
not
solved
yet
so
this
is
now
or
the
automatic
remediation
use
case
most
now.
The
second
part
captain
addresses
you
can
again
see.
This
is
really
similar
to
the
first
views
case.
B
C
Perfect
then,
ok,
perfect
thanks.
Thanks
for
the
explanation,
what
captain
can
do
for
us
and
now
a
few?
What
what
captain
actually
has
done
for
us
if
you're
looking
at
the
challenges
I
talked
about
earlier,
so
we
are
actually
achieving
this
kind
of
script.
Let's
delivery
for
operations,
so,
as
you
can
see,
we
just
triggered
with
one
CLI
command
and,
as
andreas
correctly
stated
in
the
end,
how
we
use
it
here.
C
We
are
not
triggering
which
CLI
we
connected
our
gitlab
code,
repository
with
with
a
weblog,
and
if
we
are
committing
new
code
to
the
right
branch,
then
automatically
and
you
build
well,
it
gets
triggered
and
the
new
build
will
be
automatically
deployed
by
captain
and
we
have
this
kind
of
blue
blue
green
deployments
and
canary
deployments.
We
already
saw
because
the
woman
version
that
was
obviously
where
the
back
didn't
get
promoted
to
to
production
and
we
have
only
as
we
have
less
pipeline
work
involved.
C
This
is
what
we
actually
want
to
add
short,
but
we
were
actually
want
to
achieve
what
we
saw
as
well,
and
this
is
something
we
discovered
during
the
process
working
on
the
project.
The
separation
of
concerns,
of
course
we
ran
our
team,
it's
quite
a
small
team.
We
have
these
kind
of
roles
already
in
place,
but
before,
if
you
ran
in
some
errors
and
some
challenges
in
our
staging
or
not
production
environment,
we
need
to
have
something
like
an
all-hands
situation.
C
Everyone
need
to
come
to
one
virtual
or
physical
table
and
fix
the
pipeline
fix
all
these
steps,
and
now
we
have
exactly
this
kind
of
separation
of
concerns.
David
developer
can
clearly
focus
on
the
features
on
the
artifact
delivering
great
new
additional
features
for
our
product
and
it's
not
longer
involved
in
the
and
the
overall
pipelining
work,
and
what
we
saw
as
well
is
the
event
based
extensibility
of
captain.
As
you
can
see
these
kind
of
messages,
what
is
kept
doing
to
our
to
our
environment.
We
provided
an
add-on.
C
We
were
contributed
to
the
captain
project
with
our
team
during
the
during
the
work
on
this,
and
it
was
really
really
easy
and
I
encourage
everyone
in
the
audience
and
everyone
who
likes
to
work
and
with
Captain
side
to
contribute
to
the
project
as
well
and
last
but
not
least,
we
have
this
kind
of
observability.
For
me,
it's
transparency,
I
and
my
role
as
a
manager
want
to
know
what's
going
on
in
the
development
Eve,
what
kind
of
environments
they're
working
on
and
I'm
most
interested
in
do
I'll
be
running
in
some
kind
of
errors.
C
Is
there
something
where
we
need
to
take
action,
and
this
is
clearly
achievable
with
whis
captain
and
all
these
kind
of
automated
irritations
that
automated
messages?
What
we
will
see
now
we
will
go
back
to
the
demo
and
see
if
the
deployment
went
through
successfully
and
then,
if
it's
not
the
case,
we
will
go
to
the
Q&A
session.
I
can
imagine
that,
even
if
you
have
no
questions
from
the
audience
that
Andy
Gardner
has
some
questions
for
us
and
then
we
will.
We
will
see
how
we
move
forward
from
this.
C
I
would
like
to
check
out
the
environment
and,
let's
check
out
what
kind
of
messages
we
did
get
from
captain
in
the
background,
so
this
looks
actually
quite
promising.
We
we
we
saw
an
evaluation.
That's
done,
let's
check
out
the
staging
environment
once
again,
if
you
have
the
new
comment
and
then
let's
check
out
prod,
if
we
got
a
new
comment
there
as
well,
so.
C
B
C
So
maybe
I
just
hit
the
send
button
and
perfectly
the
right
nanosecond,
so
I
would
say
that
the
deployment
when
said
successfully
through
and
now
we
can
move
to
the
Q&A
session.
If
there
are
questions
from
the
audience,
otherwise
Andy,
of
course
I
guess
you
have
some
some
ideas.
Some
questions
for
us
yeah.
A
There's
actually
well,
first
of
all,
thank
you
for
a
great
combination
of
kind
of
explaining
us
who
you
are.
What
Abel
is
why
you
decided
on
kubernetes?
Why
captain
and
then
also
andreas
crema,
for
giving
an
a
background
of
the
mains
we
use
cases.
So
that's
the
first
thing.
It
actually
has
a
couple
of
questions
that
came
in
and
came
to
me.
Can
you
go
back
to
your
WebEx
team's
client,
because
I
think.
A
A
C
A
So
because,
basically,
if
I
can
rephrase
you
had
a
one
version
deployed
in
the
beginning,
then
you
try
to
deploy
a
new
version
that
was
called
feature
and
that
feature
version
was
no
good
right.
It
was
basically
stopped
in
staging
because
it
had
bad
performance
or
problems.
There
was
never
updated
and
then
he
brought
in
another
version
which
was
called
fixed
this.
When
did
this
made
it
through
staging
and
then
automatically
promoted
into
production,
which
is
why
we
are
now
seeing
the
new
features
capabilities,
also
in
production
environment,
so
that
was
great,
so
Boston.
A
C
Actually,
that
was
our
first
project
on
kubernetes,
so
we
we
talked,
and
the
team
I
would
say
for
one
that
half
years
about
it,
so
they
had.
This
could
be
future
solution
because
we
saw
our
costs
on
AWS
was,
was
rising
more
YC
two
instances.
Is
there
a
better
way?
We
are
already
containerized,
so
it
was
something
like
the
obvious
choice.
Then
we
saw
all
of
our
customers
are
jumping
onto
anita's
as
well,
and
we
know
we
need
to
go
there,
but
we
saw
we
will.
C
A
And
I
think
that's
also
a
good
point.
You
said
once
everything
is
set
up,
because
obviously
you
didn't
show
how
to
create
a
project
and
how
to
onboard
a
service.
I
mean
you
mentioned
that
there
is
there's
no
pipeline
anymore,
because
everything
is
declarative
is
andreas
said
it
because
there's
a
couple
of
questions
that
came
in.
So
how
do
you
define
the
stages?
It's
it's.
A
If,
if
folks
are
interested
in
more
on
how
to
skip
work,
just
go
to
captain
daughter's
age
or
go
to
the
github
repo,
we
have
tutorials
that
walk
you
through
the
whole
process.
It
is
super
easy
within
five
minutes.
You
are
up
and
running
and
you
will
really
see
there's
no
pipelines
per
se
anymore,
where
you
have
to
define
what
should
happen
in
each
individual
stage.
Sebastian
has
another
question
on
again
on
kubernetes:
have
you
decided
on
a
managed
kubernetes
service
already?
Have
you
decided
on
a
on
the
cloud
vendor
and
if
so,
why.
C
So
I
would
say
we
are
still
open
right
now
we
on
Google,
it
is
engine,
but,
as
you
know,
from
from
my
introduction,
we
are
still
infrastructure,
heavy
and
most
of
our
customers,
so
are
still
infrastructure
heavy
they
using
when
leaders
in
their
own
data
centers,
with
some
support
from
Cisco
and
other
vendors
and
all
these
kind
of
stuff.
So
we
as
a
development
team
in
this
space
needs
needs
them
to
be
open.
So
there
are
some
customers
who
opted
for
Ella.
There
are
some
customer
who
opted
for
AWS.
C
A
B
A
Yes
and
as
you
just
go
to
a
website
and
everything
is
there
next
question
goes
back
to
the
Ithaca,
isn't
Berlin,
it
seems
it's
from
somebody
that
has
experience
with
captain.
So
captain
is
using
helm.
Have
you
used
helm
before?
Did
you
have
to
move
your
deployment
scripts
to
helm
and
if
you
had
to
move
it,
what
it
used
before
so.
D
So
when
we
get
started
with
kubernetes,
there
was
also
held
involved,
and
so
we
had
some
basic
knowledge
about
helm
and
how
you
have
to
structure
your
folder,
and
why
do
you
have
to
put
all
your
charts
for
kubernetes
to
be
ready
and
at
the
end,
if
you
prepared
your
charts
for
the
resources
you
need
to
deploy
in
kubernetes,
then
it's
not
how
to
just
switch
to
twelve.
It's
just
another.
You
have
just
talk
to
your
father
at
the
end,
so
let's
not
the
real
problem:
yeah,
okay,.
A
B
A
list
of
tool
integrations
can
be
found
in
our
captain
contrib
organization.
Here
you
can
find
different
services.
Each
of
these
services
has
written
MD,
and
this
written
is
will
explain
you
what
the
service
is
doing
and
which
events
it
is
listening
to.
So
you
can
find
all
those
tools
on
captain
Kahn
trip,
perfect.
A
B
A
perfect
question
because
we
would
like
to
motivate
you
to
write
your
own
service
and,
yes,
we
do
have
templates
available
and
explanations
available
and
therefore
again
I
would
like
to
refer
you
on
to
captain
at
this
age.
Then
you
can
find
point
docs
to
torian's
and
then
write
your
own
cabin
service.
But
technically
I
can
also
under
this
question,
because
it's
simple
as
listening
to
20,
TP,
reek,
post
to
requests
and
then
you
get
the
information
inside
the
event
process
it
and
executing
action,
and
that's
it
perfect.
A
One
more
for
you,
andreas
different
topic
now,
but
in
your
quality
gate
example
that
he
showed
you
had
a
sample
SLO
file.
So
there
was,
it
was
good,
but
the
question
was:
where
are
these
SLI
is
defined,
so
you
had
two
examples.
Two
database
example
and
think
the
reason
once
time,
but
where
the
SL
is
defined,
because
you
just
had
a
logical
name,
are
these
hard-coded
or
are
we
flexible
so.
B
Canton
provides
out-of-the-box,
some
default,
metrics
like
to
response
time
medium
or
the
90
or
95
percentile
or
the
error
rate,
and
but
you
can
also
define
your
own
SL
ice,
and
this
is
lies
then,
live
in
the
github
repository
because
we're
following
such
an
approach
that
everything
lives
in
get,
and
here
you
can
use
your
Prometheus
queries
on
your
time.
Interest
queries.
So
you
basically
in
the
net
rate,
you
can
query
the
complete
API
and
in
from
easiest
you
can
simply
define
your
own
queries.
A
Folks,
if
you're
interested
on
the
banditry
side,
I
have
a
performance
clinic
where
I
talked
about
the
calculated
service
metrics,
which
was
a
performance
clinic
where
I
also
talked
about
the
new
metrics
v2
API
and
I
actually
had
examples
in
there
how
you
can
enrich
the
cap
and
quality
gates
with
metrics
from
the
trees.
Next
question
goes
back
to
Berlin.
A
C
C
A
C
D
Okay,
so
actually
this
is
the
captain
to
captain
repo.
You
need
to
provide
this
when
you
set
up
a
project
of
capital
and
then
captain
will
go
to
the
repository
and
will
automatically
create.
If
you
you
need
to
give
a
shipyard
file,
so
maybe
I
can
just
show
this
once
I
have
maybe
or
do
it?
Okay,
it's
perfect,
okay,
cool.
D
So
this
is
the
shipyard
file,
so
you
just
define
your
your
stages
you
want
to
have,
and
if
you
need
to
define
your
deployment
strategy
and
your
test
strategy,
as
you
can
see
for
our
project,
we
have
just
two
stages:
staging
and
prod
and
deployment
strategy
for
the
staging
stage.
The
station
stage
is
actually
direct
and
the
other
one
is
blue
green.
D
So
we
have
a
canary
deployment
running
there
and
if
you
create
a
project
with
the
circular
file,
captain
will
automatically
give
you
two
additional
branches
which
are
then
brought
in
staging
and
then
you
can
jump
into
staging
and
you
will
have
your
service
your
services
here.
So
in
our
case
we
have
free
services
for
the
project
and
it
will
pick
up
the
ham
charts.
It
will
pick
up
the
jmeter
tests
and
it
will
also
pick
up
the
ESL
ice
Nessarose,
so
in
helm
its
to
the
color.
D
D
You
have
some
custom
queries
which
you
have
to
define
beforehand
or
the
default
ones
will
be
the
first
five
lines
here.
So
you
have
the
throughput
error
rate
response
time
with
the
different
percentiles
and
then,
of
course,
you
can
specify
your
own
metrics.
If
you
leverage
from
the
dynaTrace
calculated
metrics,
then
you
can
define
them
in
dynaTrace
and
you
can
pick
them
up
via
the
API
and
then
you
can
use
MSDS
advice.
So
what
was
this
detailed
enough
for
the
one
who
asked
the
question
or
any
more
needed.
B
A
B
A
Perfect
awesome
and
maybe
a
follow-up
that
I
forgot
to
mention
that
we
talked
about
the
next
community
meeting
the
Kaplan
community
meeting
on
February
3rd
On
February
3rd.
We
will
actually
you
know,
dedicate
that
community
meeting
to
the
0.6
GA
so
will
give
you
a
detailed
overview
of
of
everything
that
is
in
that
release.
So
that's
going
to
be
a
good
overview
for
people
that
are
new
or
for
those
that
one
or
upgrade-
and
it
seems
you
have
some
nice
dashpot
here
as
you
are
as
you're
on
it's
right,
I
mean
if
you're
sure.
D
A
Mean
I
think
this
is
also
now
with
the
o6g
a
we
actually
took
these
dashboards
and
are
now
automatically
creating
a
dashboard
in
the
ANA
trace
for
every
project.
We
automatically
get
a
management
zone
in
dynaTrace.
You
automatically
get
the
alerting
setup
correctly,
and
this
is
a
lot
of
great
stuff
that
went
in
from
the
engineering
team
to
automate
a
lot
of
stuff
that
we
initially
had
to
do
manually.