►
From YouTube: Kubernetes, monitoreo con Prometheus y Grafana - Cinthya Palomo y Jackeline Benitez, USAC
Description
Esta charla trata sobre una forma de monitoreo con Prometheus y Grafana además de la utilización de estos últimos.
A
B
Everyone
I
hope
you
are
enjoying
the
lectures
and
are
learning
a
lot
on
this
occasion.
I
have
the
honor
of
Introduce
the
presenters
to
Jacqueline
Benítez
and
Andrea
Paloma,.
They
are
going
to
introduce
us
to
the
subject
of
committed
monitoring,
turbinates
and
Graf
Anna.
I
hope
they
pay
attention
and
he
will
give
them
the
microphone.
A
To
present
a
conference
about
monitoring,
commit
and
graf
anna
is
mainly
intended
for
students
who
are
starting
out
or
well
people
who
are
starting
to
venture
into
the
world
of
the
government
container
box
&.
It
is
because
we
know
that
later
someone
will
probably
look
for
some
resource
on
how
start,
then
this
could
be
a
good
start.
So,
let's
start
first
introduce
you.
My
name
is
jacqueline
benítez,
my
partner's
is
andrea,
paloma
andrea.
Would
you
like
to
introduce
please?
Yes,.
C
Good
afternoon
I'm
andrea
palo
I'm,
not
a
student
at
San,
Carlos,
University
I'm
in
the
eighth
and
ninth
semester
of
the
engineering
degree
in
sciences
and
systems,.
It
is
a
pleasure
to
be
here
and
present
this
conference
and
to
be
able
to
give
you
a
little
special
about
what
this
topic
is
and
well,
leave
it
with
my
partner,.
There
is
here
so
that
she
can
present
herself
Thank
you
very
much,.
A
Andrea.
Yes,,
like
Andrea,
I,
am
also
a
student
at
the
Faculty
of
Engineering,,
a
student
of
engineering
in
sciences
and
systems
at
the
University
of
San,
Carlos,,
Guatemala,
and
well,.
We
had
the
pleasure
of
sharing
many
projects
with
Andrea
throughout
the
degree.
So,
well,,
more
than
grateful
to
be
in
a
new
project,,
also
with
her
now
as
speakers.
A
A
Let's
start
well,
let's
start
with
what
is
monitoring
when
someone
talks
about
monitoring
the
first
thing
they
probably
think
is
to
control
and
monitor
something,
for
example,
most
parents
when
they
are
new
or
have
many
children
or
a
baby
son,
regularly
buy
some
type
of
monitors
to
be
able
to
monitor
these
babies
so
that
nothing
happens
to
them
to
prevent
an
accident.
So
the
same
thing
happens
with
the
monitoring
in
the
computer
environment
in
a
term
deeper
into
the
environment.
A
Ok
now,,
because
it
should
be
monitored
first
or
that
the
use
of
resources
should
be
monitored.
First,
for
example,
the
CPU.
memory
disk
space.
So
it
is
quite
important
to
monitor
also
the
request
rate
if
there
is
increasing
or
decreasing
traffic.
The
error
rate.
This
is
quite
important
to
see
if
system
errors
were
introduced
and
see
how
they
can
be
handled
and
also
the
number
of
instances
of
any
app.
If
we
have
any
service
verify
how
many
there
are,
if
there
aren't
any
it's,
because
the
service
doesn't
work.
A
A
Now
we
see
that
there
are
two
types
of
monitoring:
black
box
or
black
box
monitoring
and
white
white
box
or
white
monitoring:
black
box
monitoring
is
done
at
the
application
or
host
level
since
they
are
observed
from
the
outside,
but
this
can
sometimes
result
quite
limiting,
because
we
only
refrain
from
seeing
only
what
the
application
is,.
Instead,
the
white
box
monitoring
can
use
this
more
to
verify
and
all
the
performance
of
the
internal
components
where
the
application
or
service
is
running
very
well.
We
are
going
to
continue
and
now,
well,.
What
is
it?
A
We
promise
that
it
is
a
monitoring
tool
that
we
are
going
to
use.
It
is
a
time
series
database
and
a
monitoring
and
alert
system..
This
type
of
database
is
especially
for
time
to
save
times
to
order
them.
chronologically,
measuring
variables
over
time
and
well,.
The
databases
are
also
focused
on
time
series
that
are
called
times
and
gray,
and
we
will
see
it
later
because
it
has
one,
that
is,
it
has
some
meaning,.
It
has
meaning
that
is
quite
important
to
understand.
What
is
it
that
prometheus
throws
at
us?
A
Then?
It
is
also
important
to
mention
that
this
project
has
a
fairly
active
community
of
users
and
developers
that,
since
its
creation
in
2012,
has
evolved
and
has
become
a
fairly
independent
and
open
source
project,
and
that
It
is
also
part
of
the
Cloud
Computing
Foundation
since
2016,
and
it
is
the
second
hosted
project
after
Cúber
Net.
It's
good.
Now,
we
are
going
to
mention
some
of
its
main
features.
The.
A
First
one
we
have
is
multidimensional
data
modeling.
This
is
data..
It
is
time
series
data
identified
by
metric
name
and
paired
value
keys,.
We
also
have,
among
other
features,,
a
flexible
language
of
query
sets,
that
is,
we
are
going
to,
well,,
we
promise,.
It
allows
us
to
make
queries
intended
merely
for
the
multidimensionality
of
the
time
it
throws,,
since
it
can
throw
us
a
lot
of
significant
data
from
many
things.
We
are
monitoring
always
time
and
performance
oriented.
A
We
also
have
stand
alone,
single
server
service
nodes,
good
due
to
low
reliance
on
distributed
storage
time
series
collections
happen
via
a
pull
model
via
http,
and
by
last
we
have
a
static
configuration,
and
this
allows
the
objectives
to
be
determined
by
static
configuration
or
service
discovery.
Well
moving
forward
a
little
bit
and
to
understand
what.
A
Prometheus
returns,
we
have
a
language
that
handles
that
it
is
necessary
to
understand.
First,
we
have
sampol,
which
are
samples
is
a
set
of
times
time
plus
value.
The
value
always
in
promised
is
going
to
be
float,
and
here
I
give
you
an
example
with
an
analogy
to
real
life.
Obviously
this
is
not
going
to
be
thrown
directly
by
prometheus,
but
to
understand
it
better.
A
We
also
have
the
time
series,,
which
is
a
list
of
samples
ordered
in
the
time,
and
for
this
we
could
exemplify
the
temperature
in
Guatemala
of
the
last
8
days
and,
what's
more,
if
we
imagine
a
little
more
and
how
this
could
work
already
completed
with
graf
anna,
we
could
imagine
a
bar
graph
where
the
temperature
of
all
is
shown
the
last
8
days
and
we
see
how
it
varies,
and
we
also
have
time
series
data.
You
see
that
it
is
because
what
I
was
using.
A
Ok,
now
we
are
going
to
see
the
types
of
metrics
that
They
can
handle
Prometheus
and,
to
begin
with,
we
have
the
counter
that
this
is
a
type
of
metric
that
only
grows,,
that
is
to
say
that
there
will
be
no
chance
that
it
will
decrease,.
It
will
only
grow
and
grow
and
grow,
for
example,
an
example
of
something
the
same
with
a
little
bit
of
analogy:
real
life,
two
people,
for
example-
we
can
say
the
number
of
times
the
first
person
gets
angry
with
the
second
person.
A
You
can
only
increase
an
example
to
understand
it
better.
Now,
for
a
more
technical
example,
we
could
mention
the
total
number
of
requests,
a
controller
received
or
the
total
number
of
errors.
Ok,
the
second
of
scouts
or
indicator
is
a
type
of
metric
that
can
increase
and
decrease.
This
is
more
variable,
so
the
same
to
do
a
type
of
oenology.
A
person
during
the
day
can
feel
hungry
and
then
can
be
full
if
You
can
feel
hungry
and
then
it
doesn't
grow
and
it
decreases,
it
will
always
be
variable,.
A
A
The
purpose
of
a
program
is
to
estimate
how
useful
it
is
that,
well
there
we
go
a
little
statistically,
but
because
it
is
time
it
is
used
mainly
with
request
latency,
and
we
also
have
sommer,
and
what
is
the
summary
and
is
this
The
type
of
metric
is
the
summary
of
the
histogram,
and
this
was
a
little
before
the
histogram,.
Therefore,
it
is
recommended
that
to
calculate
which
one
and
cantiles,
histograms
be
used
and
not
summaries,,
so
this
is
based
on
Prometheus
and
now,.
Before
going
to
graph,.
A
A
Ok,
so
this
is
the
graphical
interface
of
fiancés
of
fiancées
and
well.
What
I
did
to
install
it?
I
did
it
locally
and
I
did
it
in
windows?
It's
quite
simple:
we
just
looked
for
the
official
page
of
averages.
We
download,
extract
and
execute,
and
then
I
think
that
it
is
not
visible
here,
because
I
am
not
sharing
the
entire
screen,
but
when
we
want
to
enter
the
interface,
we
promise
and
enter
through
the
local
address
host
90
90,
and
it
will
throw
us
this
page.
A
A
That
it
takes
out
this
golden
http
request
averages
and
we
execute
it.
It
will
throw
them
here
because
what
it
obtained,
what
it
stored
in
its
database
and
also
what
it
we
can
see
in
graphs,
even
the
graph
here
we
can
change
it
to
align
it.
We
can
put
the
base
color
to
be
able
to
visualize
it
better.
We
can
increase
or
decrease
the
time
to
see
its
growth
or
progress
over
time,
and
obviously
it
is
already
complemented
with
graphene
and
with
a
some
service
or
something
that
we
can
evaluate.
A
A
So
graph
anna
is
a
very
good
option,.
In
fact,
it
is
one
of
the
best
to
visualize
this
type
of
monitoring
reports
and
well,.
The
compatibility
represented
by
graph
anna
is
highly
recommended,
since
it
works
very
well
with
cloud
solutions.
Now.
We
know
that
many
things
are
already
migrating
to
the
cloud
and
that
most
things
are
still
handled
from
the
cloud,
and
it
also
includes
support
either
for
a
w
se
es.
We
can
also
find
support
on
the
internet
and
also
its
implementation
is
completely
based
on
the
cloud,
and
it
also
includes
a
supervision
solution.
A
So
we
are
going
to
talk
a
little
about
the
advantages
that
graph
anna
has
and
well.
First
of
all
is
that
it
can
run
in
tb
mode,.
You
could
say
that
is,,
the
way
in
which
we
can
see
the
reports
in
this
tool
is
quite
innovative
and
we
can
customize
them
a
lot,
and
it
is
also
quite
useful
to
use
shortcuts
on
the
keyboard
and
in
graph
ana
to
be
able
to
facilitate
us.
The
handling
of
these
then
and
continuing
I
will
also
share
how
the
graphical
interface
looks
the
same.
A
And
here
it
is
the
same:
the
installation
is
quite
simple
and
to
lift
it,
consult
it
and
see
it
in
our
browser.
We
use
the
local
host
address
and
3000
and
it
opens
fna
there.
This
is
the
interface.
Obviously,
if
I
try
to
create
a
dashboard
here
and
I
think
they
had
mentioned
it
in
one
of
the
previous
conferences.
C
Thank
you
very
much
from
here
and
well.
We
are
going
to
continue
with
what
is
prometheus
operator
first.
He
promised
super
editor
for
the
govern,
&
expression,
Simple,
monitoring
of
government
services
in
their
implementation
and
management
in
instances,,
the
use,
well,.
We
can
do
it
to
install
and
configure
government,,
etc.
C
Implement
yourself
in
an
internal
prometheus
stack
such
as
prometheus
server
alert
manager,
rafa
nadal
host
nou
de
exportar
club
state
metrix
These
are
some
of
the
ones
that
we
will
be
able
to
observe
later
and
I
will
present
them
to
you
in
another
of
the
uses:.
The
definition
of
the
self-configuration
of
the
metric
endpoint
using
the
service
entities
monitor
customization
and
scalability
of
the
services
using
crd
of
the
confirm
operator,
making
our
configuration
completely
portable
and
declarative.
We
are
also
going
to
have
here
a
bit
of
what
the
architecture
used
by
prometheus
operator
is,.
C
C
The
library
plugins
are
the
various
ones
in
charge
of
exposing
the
internal
metrics
of
the
application
to
monitor
the
average
of
prometheus,
either
a
cpu
in
memory
and
the
cetera
so
that
they
can
be
collected
by
promise
av
server,
and
these
send
them
to
what
It
is
true
that
the
aller
manager
and,
as
I
said
before,.
He
manages
the
alerts,,
sends
them,
promises,
observes,
groups
and
notifies
the
alarms
in
different
media
defined
as
email,
player
Gianni,.
Among
others,
is
also
in
charge
of
silencing
and
inhibiting
the
alerts.
C
Sometimes,
we
must
also
remember
that
the
director
of
commit
inside
the
operator
repository
contains
default
configurations
and
services,
because
you
get
not
only
the
prometheus
operator
itself,
but
a
complete
configuration
that
you
can
start
using
and
customizing
right
from
the
start.
And
now,
let's
start
with
what
is
yes,
what
is
the
theme.
C
This
workshop
is
going
to
be
provided
through
a
hip,
hop
repository
that
the
people
in
charge
of
the
type
conference
will
be
giving
them,,
and
here
is
a
bit
of
what
What
is
the
basic
workshop,
and
how
is
it
that
we
can
implement
it
from
the
wta
monitoring?
Then
we
are
going
to
start,
well,
by
having
a
cluster
in
my
case,
well,
I,
put
it
to
crush
first,
since
I
already
have
it
up,
well,.
This
is
it,.
C
C
C
And
we
are
going
to
create
what
is
an
area,
because
this
will
contain
what
is
the
version
of
rulers
of
how
many
replicas
will
they
be
talking
about
where
it
will
be
contained
in
this
case,
because
I
am
going
to
contain
it
in
a
registry
container?
That
already
brings
it.
As
far
as
google
is
concerned,,
then
this
is
already
automatic,
and
well,
the
name
of
the
project,
and
if
it
is
not
going
to
be
called
my
application,,
the
ports
are
going
to
be
port
80.
C
C
I,
don't
know
what
the
friction
is
going
to
be:
called,
likewise,
well,,
the
service
nou,
where
the
new
content
is
going
to
be
and
well,
I'm
going
to
show
you
this,
I
already
had
it
here
to
be
able
to
optimize
a
little
time,.
So
as
you
can
see,
the
basic
English
and
the
service
nov
were
on
the
cam,.
We're
also
going
to
paste
what
It
is
a
directory,
and
this
is
going
to
contain
what
is
a
profile
that
is
going
to
obtain
the
data
to
be
able
to
raise
what
this
server
is.
C
C
C
B
C
C
And
as
a
first,
it
is
a
fairly
simple
service
and
well,.
We
can
raise
it
and
have
the
data.
True
for
this,
the
data
was
taken
from
the
jason
plays
holder
page.
It
is
really
very
useful
when
you
want
to
have
data
quite
quickly
if
they
are
already
there
through
a
through
an
application
of
a
true
service,
and
it
can
be
sent
to
the
request-
and
we
are
already
with
that.
We
are
going
to
do
the
last
one.
C
There
we
are,
this
part,
is
already
generated.
Well,
we
can
go
here,
the
income
and
the
services
that
they
are
up,.
Well,
we
can
see
that
that
part
is
already
up,.
We
are
going
to
go
with
what
is
the
gra
funk
and
for
this
case
are
going
to
obtain
from
this
one
from
this
repository
that
is
also
here,
right,.
It
is
this
one,
we
are
going
to
download.
C
What
is
that
that
he
is
going
to
have,
because
he
pointed
out
that
it
is
him
he
promises
us
operator
so
once
downloaded,
it
is
in
this
repository.
Well,
we
can
enter
what
you
want
to
what
the
code
is
and
we
can
already
create
them
in
the
manifests
and
all
the
data
What
can
be
done
to
be
able
to
enter.
C
C
C
C
B
C
C
And
as
we
can
see
well,
there
are
different
types
of
dashboard
of
what
our
posts
is,
what
our
entire
cluster
contains
in
itself,
the
data
that
is
going
to
be
displayed
in
this
case
I
am
going
to
show
you
what
the
alert
manager
is
like
how
it
is
going
to
be
graphed
by
means
of
those
of
our
bots,
well,.
Here
you
can
see
different
ones,,
it's
missing
as
you
just
raised,
it,
well,.
They
won't
be
able
to
be
seen,,
so
we
can
also
see
what
our
posts
are
without,.
We.
C
C
The
use
of
clusters,
as
for
hands,
as
we
can
see,
is
what
our
cluster
of
posts
contains.
These
of
the
use
of
7
truth
that
each
one
contains
and
well
with
this,
we
would
finish
what
is
this
this
part
of
what
we
have
promised:
keep
working
hard
I
hope
it
has
been
very
useful
to
you
and
we
hope
it
is
very
useful
for
you.
Thank
you
very
much.
Thank
you.
B
A
Well,,
look
at
the
difference
between
the
monitoring
of
the
black
box
and
the
white
box.
In,
the
white
box,.
It
is
already
verified,
well
as
your
composite,,
responding
to
what
you
are
doing,
that
is,,
if
you
are
making
a
super
application.
Obviously
your
computer
is
going
to
be
affected
by
the
ram,,
the
disk
and
all
these
components.
A
Obviously
the
performance
will
go
up.
So
this
is
what
the
white
box
monitor
and
the
black
box
monitor
are
for,
and
it
is
more
than
anything
for
tests
that
They
are
used
to
check
the
operation
of
the
ports,
among
other
things,.
So
I
would
say
that
the
one
that
helps
you
the
most
to
verify
that
the
component
is
still
dying
is
the
monitoring
of
the
white
box.
In
fact,.