►
From YouTube: Lightning Talk: Thank you Keptn Obvious! Making SLOs observable with Prometheus a... Andreas Grabner
Description
Don’t miss out! Join us at our upcoming hybrid event: KubeCon + CloudNativeCon North America 2022 from October 24-28 in Detroit (and online!). Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Lightning Talk: Thank you Keptn Obvious! Making SLOs observable with Prometheus and Keptn - Andreas Grabner, Dynatrace
This talk shows how to not only collect metrics and provide application monitoring using Prometheus, but to also make them visible as SLOs and then act on them in a fully automated and cloud native way.
A
All
right
well
thanks,
first
of
all
for
being
so
dedicated
to
kubecon,
it's
5
30
on
a
beautiful
day
and
you're
here
and
not
at
the
bars.
So
thank
you
so
much.
I
also
want
to
say
thank
you
to
olek.
He
was
supposed
to
give
this
talk,
but
unfortunately
he
had
to
travel
and
leave
early.
So
I'll
do
my
best
to
give
him
justice,
but
the
great
slides
that
you
see
are
actually
from
him.
So
if
you
ever
see
all
the
ghana
chefs
then
give
him
some
kudos
talk
today.
A
Thank
you,
captain
obvious
making
observability
actionable
with
prometheus
and
captain.
Obviously
I
try
to
make
it
obvious.
I
really
like
captain.
My
name
is
andy
gregner.
I
work
for
dynatrace
during
the
day,
but
I'm
also
a
captain
maintainer
and
a
contributor,
and
I
really
like
to
show
you
what
we
are
doing
with
captain.
A
If
you
think
captain,
if
you
heard
about
captain,
you
think
captain
is
yet
another
cicd
tool
and
it's
a
if
this
standpoint
is
told
that
hopefully
I
prove
you
wrong,
because
I
want
to
give
you
a
little
different
perspective
on
captain.
Captain
has
also
changed
over
the
last
couple
of
years,
but
if
you
want
to
get
shirts
or
heads,
we
also
have
a
booth
at
the
pavilion
so
check
us
out
tomorrow.
But
let
me
first
get
into
my
talk.
A
I
talk
about
observability,
making
observability
actionable
and
when
I
talk
up
and
think
about
observability,
I
think
about
essays.
Sovs
are
using
observability
data
to
keep
systems
reliable
and
resilient,
but
you
may
say:
well,
one
does
not
simply
do
a
sre.
Well,
I
counter,
I
think,
actually
captain
makes
it
pretty
simple.
This
is
a
post
from
tarash.
A
So,
first
of
all,
let
me
get
started
on
what
problem
we
really
want
to
solve.
We
all
know
that
modern
cloud
native
systems
are
no
longer
simple.
You
have
a
lot
of
different
tools
and
you
need
to
connect
them
in
one
way
or
the
other,
which
means
you
have
to
build
a
lot
of
point-to-point
integrations
and
trying
to
put
in
also
your
ability
to
make
better
decisions.
I'm
pretty
sure
some
of
these
tools
are
familiar.
The
question
is:
how
can
you
make
better
informed
decisions
if
you
don't
have
visibility?
A
Well,
fortunately,
we
obviously
talk
a
lot
about
observability
in
this
conference.
I
also
want
to
give
a
shout
out
to
henrik
in
case
you
have
not
seen
his
phenomenal
youtube
channel.
It's
called.
Is
it
observable
check
it
out?
He
is
looking
at
new
frameworks
in
the
space
all
the
time
and
he
gives
his
perspectives.
He
also
creates
tutorials
so
check
it
out.
A
The
point
is
observability
is
a
must,
because
if
we
don't
have
observability,
we
cannot
make
good
decisions.
We
don't
even
know
if
the
system
is
broken.
The
problem
is
right.
Observability
is
not
easy.
If
you
know
that
you
have
a
big
landscape
of
tools-
and
you
don't
really
know
how
you
actually
get
data
out
of
this
well,
fortunately
again
we
have
movements
where
we
are
trying
to
figure
out
how
to
standardize
mobs
ability.
A
I
don't
need
to
remind
you
about
some
of
these
great
products
or
frameworks
and
projects
here
that
are
really
trying
to
make
it
easier
for
us
to
get
insights.
Prometheus
is
one
of
them
right.
I
think
with
hunt
and
again
this
is,
if
you
like,
these
slides
ovec
is
the
one
to
put
them
together.
I
really
like
it
right,
sauron's,
either
prometheus.
I
will
give
you
all
the
visibility
and
what
happens
in
your
environment
what
we
want
to
bring
to
the
table.
If
you
have
data,
then
we
as
captain
make
it
actionable.
A
If
you
have
a
desired
state
based
on
the
data,
the
system
should
be
helping
healthy,
it
should
be
up
and
running.
It
should
be
performing
in
a
certain
way.
We
can
act
on
this
data
to
always
keep
the
system
in
that
interstate
that
you
expect
from
your
system
based
on
your
stability
data.
So,
in
other
words,
captain
is
an
observability
driven
orchestration
engine
for
your
cloud
native
apps.
What
this
means
in
again
another
cool
slide
from
oleg
one
platform
to
orchestrate
them
all.
A
A
What's
really
important
is
that
orchestration
and
observability
drifting
orchestration
is
at
the
core,
and
what
this
means
is.
If
I
come
back
to
the
slide
that
I
had
earlier,
the
clicker
works.
As
expected.
Somebody
needs
to
observe
the
signal
here
there
we
go
all
right.
One
option
that
you
have
you
can
build
orchestration
yourself.
You
can
connect
all
of
these
tools,
but
what
we
try
to
do
is
I
just
click
here.
A
Instead
of
you
having
to
connect
all
of
these
tools
and
making
decisions
on
what
to
do
next,
captain's
control
plane
reaches
out
to
your
observability
platform
and
based
on
how
the
state
of
your
system
is,
then
all
the
right
tools
to
actually
bring
the
system
back
to
its
healthy
state.
So,
at
the
core
of
captain
we
have
slo
service
level
objectives.
You
define
the
desired
state
of
what
your
system
you
expect
the
system
to
be,
and
then
we
orchestrate
the
tool
to
bring
the
system
always
back
to
the
desired
state.
A
But
captain
has
a
really
cool
rest
api.
We
are
standardizing
on
cloud
events
and
we
are
in
the
cdf
events
spec,
and
I
see
them
out
of
time.
What
I
would
like
you
to
know
is:
if
you
have
an
app
it's
instrumented
with
open
telemetry,
you
get
some
data,
we
can
act
on
it
to
bring
the
system
always
back
to
its
desired.
Based
on
your
observability
captain
is
extendable.
There's
a
lot
of
stuff
that
we
have
out
there
check
out.
Nana's
video.