►
From YouTube: Delving into Keptn - Documentation and Exploration
Description
In this stream, the DevsOperative team starts looking at Keptn and discussing it's use-cases.
A
Today
we
are
going
to
be
talking
about
captain
and
this
stream
in
particular
we're
going
to
make
a
two-parter.
So
today
we're
just
going
to
delve
into
captain
we're
going
to
see
what
is
it?
What
does
it
allow
you
to
do
and
some
use
cases
for
it
things
like
that.
That's
today's
stream,
but
before
we
jump
into
that,
let
me
introduce
my
co-hosts,
which,
if
you
watch
our
stream,
you
all
know
we
have
sam
della
marter
right
below
me
and
below
sam.
We
have
dan
mccracken.
A
B
C
B
A
C
Anyway,
I
was
jumping
through
hoops.
I
remember
how
awesome
I
was
I
have
I
live
out
in
the
country.
I
have
a
generator
everything's
on
backup
power,
as
promised
when
we
moved
it
here.
Nikita's,
like
oh
man,
I
don't
know
the
website
website
was
up
in
20
minutes.
You
should
have
seen
me.
I
wasn't
even
talking
to
my
wife.
I
was
just
like
I
gotta,
get
it
going,
gotta
get
it
on.
A
Okay,
also,
if
you
notice
sam,
has
some
new
lighting,
so
he
looks
healthy.
He
looks
great,
very
excited.
C
A
Very
true,
okay,
so,
let's
jump
into
captain
here,
let's
start
the
conversation
and
let's
start
the
conversation
with
something
I
think
we've
talked
about
before.
Maybe
we
might
be
jumping
the
gun
with
this.
Actually
captain
works
great
when
you
already
have
a
deployment
tool
implemented.
Specifically
argo.
Argo
cd
is
a
very
popular
gitops
deployment
tool.
D
I
don't
think
we
have
and-
and
I
I
don't
know
what
captain
is
I
when
you
mentioned
it
the
other
day.
I
looked
at
this
real
quick
and
I
did
see
that
it
works
with
argo,
which
is
great
because
I
know
what
argo
does
but
yeah.
Can
you
give
the
the
10
second
pitch
of
what
is
kept
in,
because
if
it's
not
doing
the
deployments
and
automation,
what
what
does
captain
give
us
and
why
do
I
want
it.
A
I
love
it,
so,
let's
I'm
going
to
give
a
I'm
going
to
give
a
story
here,
I'm
going
to
give
a
a
white
paper
story,
so
we
are
currently
implementing
argo
cd
at
a
large
client
of
ours
and
it's
going
well.
A
Argo
cd
is
great
at
get
ops,
it's
great
at
hey.
I
want
to
roll
out
this
image
this
container,
this
helm
chart
to
this
environment
and
you
can
create
some
complex
configs
that
make
it
go
to
like
10
environments
at
the
same
time
and
you
can
keep
them
in
sync:
inversions,
a
lot
of
power,
a
lot
of
power.
Now
what
argo
does
not
have-
and
this
is
where
captain
comes
in
into
play
here-
captain
or
sorry,
argo,
sorry,
argo
does
not
have
any
workflow
mechanism.
A
It
does
not
have
any
way
to
have
approvals.
A
D
Okay,
that's
cool
so
hearing
that
the
one
thing
that
I'm
going
to
look
for
is
differences
between
what
captain
can
do
and
what
I
think
you
could
use.
The
other
argo
products
for
like
does
kept
in
really
give
me
more,
and
maybe
it's
just
ease
out
of
box,
because
I
was
watching
a
devops
toolkit
video
this
morning
on
just
automating
everything
with
argo
and
so
argo
workflows,
argo
events
and
I
mean
argo
workflow
itself.
D
C
C
Okay,
you
could
make
the
all
of
these
are
square
holes
and
square
pegs
or
whatever
I'm
bad
at
analogy
or
whatever
you
call
those
things
but
which
one
is
best
for
your
business
is
really
the
question
which
one's
best
for
your
use
case
and
how
much
user
experience?
How
much
does
user
experience
matter?
I
think
that
that
matters
a
lot
my
my
gut
on
and
the
hope
of
what
captain
will
do
or
will
show
us-
is
that
it's
really
nice
for
clicky
clicky.
It's
really
nice
for
business,
user
or
system,
analyst
non-developer.
C
Let's
say
it's
my
hope
and
it
looks
great
just
yeah.
You
know
this
one
we're
going
to
get
into
it.
I've
not
really
don't
not
really
looked
at
it
either.
That's
why
we're
doing
the
stream
go
figure
yeah.
B
A
I'm
just
going
to
take
a
look
at
the
main
page
here.
First
observability
dashboards
alerting
sounds
great.
They
have
some
good
marketing
here.
A
And
I
guess
what
I
want
to
look
at
I'm
going
over
the
web
page.
I
just
want
to
see
what
they
have
here,
but
what
I
want
to
do
is
I
kind
of
want
to
play
with
their
live
demo.
I
almost
want
to
start
with
that.
What
do
you
guys
think
do
it
all
right
operations
by
closed
loop,
remediation
kept
in
offers
a
whole
bunch
of
stuff?
A
lot
of
people
use
it.
Okay
live
demo,
explore
the
live
demo.
B
A
A
Okay,
oh
potato
head
explore
the
two-stage
demo
with
the
cncf
pod
tato,
head
application
that
has
been
onboarded.
The
application
is
deployed
twice
per
day
with
one
fast
and
one
slower,
build
explore
how
the
kept
in
quality
gates
prevent
the
slower
run
to
reach
production
based
on
data
from
prometheus.
A
That's
the
one
that
I
want
to
look
at,
there's
also
sock
shop,
our
famous
shock
sock
shop,
application
with
a
three-stage
environment
and
quality
gates,
using
data
from
dynatrace
each
day.
Three,
each
day
three
builds
try
to
make
it
into
production,
but
only
two
versions
are
stable
enough
to
pass
the
quality
gates.
You
can
set
up
this
demo
yourself
by
following
the
blah
blah
blah.
A
C
A
A
D
A
C
B
A
D
I
don't
know
it
doesn't
seem
like
a
ton
that
we
can
interact
with
right
now
unless
we're
just
missing
it.
C
Well,
if
you
go
back,
let's
just
walk
through
those
little
options,
because
that
right,
there
kind
of
tells
me
a
little
bit
the
the
options
on
the
side.
So.
C
What
what
were
you
just
on
when
you
had
those
filters
on
the
left.
C
It's
really
interesting
that
could
be
a
big
list,
stages,
look
like
somewhat
environments,
but
you
could
have
a
dev
stage
that
deploys
to
a
dev
environment,
the
sequences,
delivery
sequence
and
a
remediation
sequence.
C
Health
check
or
something
else
now
you
saw
dyna
trace.
I
remember
we
were
talking
about
this
earlier.
Somebody
was.
This
is
a
spin-off
from
dynatrace
the
vendor.
It's
an
open
source
situation
from
them,
which
is
great,
so
I'm
not
surprised
to
see
that
there.
D
A
C
D
A
Still
so
you
can
define
an
slo
here.
I
I
kind
of
wanted
to
see
like.
D
C
D
A
I'm
trying
to
fi
like
because
we're
on
stream.
I
don't
want
to
just
read
the
website,
so
I'm
trying
to
find.
C
C
C
A
A
Following
this
declarative
approach,
there
is
no
need
to
write
imperative
pipeline
code.
Captain
takes
the
shipyard
file
and
creates
a
multi-stage
workflow
each
stage
having
a
deployment
strategy
either
blue-green
a
testing
strategy
and
an
optional
automated
remediation
strategy
for
triggering
self-healing
actions.
C
C
One
of
the
options
is
a
manual
right
manual,
button
click,
quality
gate
which
you
hate
to
see,
but
in
reality
there's
a
lot
of
in
reality.
You
start
somewhere.
C
Reality
isn't
greenfield
all
the
time,
and
you
know
kubecon
demos
actually
keep
going
this
year
was
really,
I
thought
quite
gritty
and
real.
I
heard
a
lot
of
real
stuff
there,
not
not
all
fluffy
demoy
stuff
to
be
to
give
them
some
credit.
A
What,
yes,
we
threw
the
stream
together
a
little
last
minute,
but
I
still
think
this
is
good.
We're
gonna
have
to
go
through
this
process.
Anyways.
A
So
I,
let's
let's
keep
going-
I
I'm
kind
of
feeling
the
whole.
You
know
we
were
talking
about
this.
Then
we
started
talking
about
quality
gates,
so
I
found
a
link
to
qualitygates.
So
why
don't
we
just
keep
following
our
train
of
thought
here
and
let's,
let's
look
at
what
qualitygates
gets
us.
A
Captain
quality
gate
process,
I'm
gonna
read:
I'm
gonna
read
a
little
bit:
captain
quality
gates
based
on
the
com
base.
Okay,
pr
we
got
a
pr.
This
grammar
captain
quality
gates
based
on
the
concepts
of
service
level
indicators
and
service
level
objectives.
A
Therefore,
it
is
possible
to
declaratively
describe
the
desired
quality
objective
for
your
applications
and
services
so
number
one.
The
process
of
evaluating
a
quality
gate
can
be
triggered
either
via
the
captain
cli
or
the
captain
api
once
triggered.
Captain
fetches
the
slis
from
a
data
provider
like
prometheus
or
dynatrace
captain
evaluates
the
sli
against
the
slo
that
are
defined
for
the
application
or
service
after
evaluation
and
scoring
captain
returns.
The
result
that
can
be
either
processed
in
an
automated
way
by
an
existing
cd
pipeline
or
by
the
user
to
manually
decide
the
next.
A
Wow
wow,
okay,
so
let
me
I
know
I
just
read
that
talk
for
a
little
bit,
but
I
want
to
give
my
quick
thoughts
and
I
want
to
hear
your
guys's
things.
Thoughts
on
it.
A
You
know
we're
trying
to
we're
our
goal.
Right
now
is
to
implement
this
product
with
a
customer
and
obviously
we're
learning
as
we
go.
I
think
that's
very
clear
from
the
stream,
but
our
customer
writes
some
integration
tests
and
writes
some
smoke
tests
and
they
run
those
right
now
we
run
them
using
jenkins
or
we
run
them
manually
this
reading.
This
has
me
thinking
that
that
is
going
to
have
to
change
dramatically,
especially
the
captain
fetches,
the
slis
from
a
data
provider
and
then
evaluates
those
against
the
slos.
C
Let's,
what
are
the
data
provider
options,
because
my
my
gut
is
that
jenkins
is
a
data
provider?
Maybe
okay.
A
D
Analyzing
the
the
metrics,
but
is
there
something
that
is
driving
traffic
or
is
this
just
you
know
we
deployed
it?
It
should
be
getting
whatever
traffic
dev
test
or
you
know
staging
gets,
and
you
know
based
on
that,
whereas
like
what
we
currently
have
right
is
somebody
or
as
a
part
of
the
build
or
deploy
process.
D
A
So
there
is
in
the
captain,
sandbox
github
repo.
There
is
a
tutorial
for
integrating
jenkins
with
your
slis
and
slos.
C
A
So
let's
go
back
to
what
sam
was
talking
about,
because
I
I
definitely
agree
so
in
this
picture
that
we
were
looking
at
with
the
arrow
moving
through
the
the
multi-stage
delivery.
Here
we
have
dev
equality,
gate
and
then
hardening
and
dev
is
surrounded
by
a
blue
box
and
has
a
arrow
pointing
to
it
that
says:
functional
tests.
D
D
A
I
see
it,
I
see
okay,
so
back
in
the
yaml,
there's
a
test
strategy
and
it
says
functional
and
then
the
next
one
the
test
strategy
says
performance,
so
these
are
defined
somewhere.
D
Cool
yeah
can
we
get
like
the
docs
on
that
yaml.
A
A
Let
me,
let's
do
a
search.
C
C
C
He's
just
being
a
small
brain
about
it,
don't
I'm
just
trolling
but
like
how.
A
B
B
A
C
D
I
mean
I
think
by
next
week.
Maybe
we
can
have
some
of
the
pre-racks
set
up
and
you
know
we
can
create
a
gke
cluster
to
do
a
full
tour
but
yeah
doing
that
live
in
the
next.
B
A
See
this
is
this
page
right
here
I
have
come
to
captain
many
many
times
over
the
last
year,
literally
after
over
the
last
whole
year,
maybe
even
more,
and
I
always
stop
because
their
tutorials
are
so
heavy-handed.
To
be
honest,
I
mean
look
at
these.
Every
single
one
requires
you
to
set
up
a
kubernetes
cluster
and
I
just
want
to
you
know,
get
a
feel
for
it.
That's
all
I
want
right
now
I
don't.
I
don't
want
to
set
up
a
kubernetes
cluster.
A
C
A
C
I
have
something:
okay,
a
a
brilliant
gentleman
in
the
chat
did
give
us
a
really
a
nice
little
link.
I
just
followed
it,
it's
not
a
troll,
although
he
is
known
to
troll.
Are
you?
Are
you
at
a
place
where
you
can
click
the
chat?
Yes,
I
am.
We've
got
it's
kind
of
good
because
I
think
these
tests,
these
testing
thingies,
are
sequences.
C
Got
a
couple
cloud
native
journeymen
in
the
chat:
maybe
they
can
help.
C
C
Right
and
that
that
might
be
great
as
a
quality
gate,
because
I
mean
that's
literally:
look
the
quality
of
your
service
hasn't
degraded
like
you're
at
I'm.
Still
at
this
level
of
thinking
where
I
I
just
need
it
to
work,
is
it
working?
Can
I
get
good
at
knowing
if
it's
working
and
what
captain's
already
solving
for
is
what
I
want
to
get
to
with
my
customers
in
in
my
business?
C
A
It
is
a
okay,
so
it
is
a
task
and
a
sequence
is
made
up
of
a
list
of
tasks
and
a
stage
is
made
up
of
a
list
of
sequences.
So
it
goes
stage
sequence
task.
A
A
B
A
C
Yeah,
you
know
there
there
really
might
be.
We
should
we'll
keep
going
on
this.
There
really
might
be
a
place
in
the
or
a
gap
here,
just
in
the
the
language
where
what
we're
trying
to
do
just
hasn't
been
articulated
very,
very
well
yet,
or
for
this
type
of
use
case
for
where
we're
coming
from,
in
theory,
we're
a
target
market.
A
A
D
A
A
A
With
this
signature.
A
A
A
A
C
I
feel
like
it
might
just
be
the
the
the
file
path,
the
relative
file
path
of
that
command,
but
I'm
with
you
it's
just
a
jmeter
file,
because
this
is
like
a
cute
cuddle.
This
is
a
cli
commander.
C
It's
really
interesting:
we've
got
slab
in
the
chat,
I'm
not
familiar
with
shipyard,
but
it
looks
like
you
define
a
shipyard
file
at
the
project
level
from
there
you
define
the
sequences
and
tasks
for
those
sequences,
and
captain
has
a
reserved
set
of
tasks.
You
can
use
test
being
one
of
them
or
you
can
define
your
own,
not
sure
how
you
define
your
own.
Yet.
B
A
B
A
A
C
Probably
in
the
tutorials
primary
repo,
their
examples,
repo.
C
C
Actually
I
think
we
were
on
one
eight
and
then
they
went
to
two
dot
and
it
was
just
magically
better
and
then
2.1
was
just
like
get
at
me
better.
So
yeah,
I'm
I'm
expecting
quite
a
bit
of
a
a
nice
journey
here,
assuming
there's
good
community
involvement,
which
is
what
we're
trying
to
promote
community
involvement.
That's
where
we
are.
A
A
D
Yeah,
I
think
some
of
that
is
like
the
branding
and
verbiage
they
chose
that
just
doesn't
feel
as
natural,
as
maybe
some
other
products
that
we've
used,
but
yeah-
and
I
mean
tim
mccracken's
point
like
this-
is
an
evolution
piece
like
this.
Isn't
you
know
a
day
one
type
thing
like
evaluating
metrics
of
your
new
release
and
making
decisions
off
that
that's
pretty
advanced
along
your
cloud
name
journey
right,
like
you,
don't
need
that
day
one.
D
C
Yeah
and
I
think,
if
you're,
if
you're
green
field
and
you're
designing,
you
know
your
app
from
the
start,
you
you
could,
you
could
definitely
do
this,
but
even
then
you
know,
if
I'm,
if
I'm
writing
an
app
for
the
first
time,
and
I
really
don't
care
right
now.
You
know
I
just
want
to
get
the
thing
to
work.
C
I
want
it
out
there
and
then
ever
you
know
part
of
the
evolution,
maybe
before
production
production
workload,
if
you're
doing
if
you're,
if
you've
got
money
on
the
line,
then
you
invest
and-
and
you
really
hone
it
in
and
you
tune
some
of
these
metrics
and
you
spend
energy
there.
But
that's
asking
for
a
lot.
Let's
ask
for
a
lot
right
off
the
bat,
but
I
think
I
think
there's
a
lot
here.
A
lot
to
build
on.
I
mean
I'm
excited
to
see
more.
B
A
So
we're
gonna
continue
this
we're
not
gonna.
Let
this
one
fall
off.
We're
gonna
do
part
two
next
wednesday.
That
being
said,
I
think
as
homework
for
the
three
of
us.
We
should
watch
a
couple,
youtube
vids,
get
up
to
speed
and
then
maybe
actually
trial
it
a
little
bit
on
the
side.
So
we
have
something
a
little
bit
more
cohesive
for
the
stream
yeah.
A
I
will
do
the
same
and
for
all
our
viewers
out
there
tune
in
next
week
for
a
hands-on
captain
implementation,
where
we're
going
to
take
their
tool
and
we're
going
to
provide
you
with
clear
and
concise
instructions
on
how
to
actually
use
it.
A
So
any
closing
thoughts
or
remarks
before
we
turn
off
the
stream
for
the
day.
C
A
A
A
So,
okay,
thank
you
guys
for
showing
up
to
the
stream,
and
we
will
see
you
next
time
on
the
cloud
native
coffee
hour,
wednesday
afternoons
at
4pm.
Thank
you
for
tuning
in.