►
From YouTube: Getting started with Grafana Mimir
Description
Scale to 1 billion metrics and beyond with Grafana Mimir, Grafana Labs' new open source long-term storage backend for Prometheus.
See how you can get Mimir up and running in 10 minutes in this video, or follow the tutorial: https://grafana.com/tutorials/play-with-grafana-mimir
Learn more about this easy-to-operate, horizontally scalable, highly available, multi-tenant time series database at https://grafana.com/oss/mimir
Register for the April 26 webinar at https://grafana.com/go/webinar/intro-to-grafana-mimir
A
A
A
A
The
architecture
we
run
in
this
tutorial
is
composed
by
three
mimir
instances:
nginx
used
to
load
balance,
read
and
write
request
to
the
mimir
cluster
prometheus
configured
to
scrape
matrix
and
remote
write
to
mimir
grafana
configured
with
some
pre-installed
mimir
dashboards
and
finally,
minio
object.
Storage
used
to
store
mimir
time
series,
data
and
rules
configuration.
A
A
A
Graffana
has
been
preconfigured
with
some
of
their's
mimir
dashboards.
I
open
the
writes
dashboard.
This
dashboard
show
the
traffic
on
the
right
path,
which
are
the
remote
write
request
from
prometheus
to
mimir.
You
can
see
the
request
rate,
the
latency
and
other
information
right
now
in
our
test.
Mimir
cluster.
We
are
ingesting
300
samples
per
second
for
a
total
of
about
9000
series.
A
A
A
Proceeding
with
the
tutorial,
I
want
to
show
you
the
grafana
mimir
configuration
mimir
is
configured
to
run
with
target
all
which
means
all
the
core
mimir
components
run
in
the
same
process.
This
is
what
we
call
the
monolithic
mode.
Additionally,
mimi
is
configured
to
also
run
the
alert
manager
in
the
same
process.
A
A
A
A
A
A
A
A
Similarly,
we
can
create
a
recording
rule
to
create
a
metric
with
the
number
of
healthy
instances
in
our
cluster.
Recording
rules
can
be
created
through
the
grafana,
alerting
ui
as
well.
I
click
the
new
rule
button
and
I
insert
the
name
of
the
recording
rule
sum
column
up
this
time
I
choose
managed
recording
rule,
I
select
mimir
as
a
data
source
and
I'm
going
to
use
the
same
namespace
and
group
I
created
previously.
A
A
A
A
A
In
the
top
right
corner
of
my
screen,
you
can
see
the
remaining
two
healthy
instances
are
complaining.
They
can't
talk
to
the
third
instance
anymore.
This
is
expected,
however,
since
garfanamimir
runs
with
a
replication
factor
of
3.
It
can
tolerate
an
instance.
Failure
to
show
you
mimir
is
still
working
properly.
I
open
the
writes
dashboard.
A
As
you
can
see,
mimir
is
still
ingesting
the
matrix
from
prometheus
and
serving
the
queries
to
populate
this
dashboard.
I
can
also
show
you
the
reads
dashboard,
which
is
all
green,
since
all
queries
are
still
succeeding.
Even
if
one
of
the
three
instances
is
unhealthy,
I
want
to
go
back
to
the
ring
status
page.
I
click
the
refresh
button,
as
you
can
see
now.
The
third
instance
is
unhealthy,
but
the
remaining
two
are
still
healthy
and
active
in
the
cluster.
A
A
A
A
A
To
sum
up
in
this
tutorial,
I
run
a
cluster
of
three
grafana
mimir
instances.
I
created
a
recording
and
alerting
rule
using
the
grafana
alerting
ui.
Finally,
I
simulated
a
notation
and
show
you
how
mimir
kept
working
successfully.
Even
if
one
of
the
three
instances
was
unhealthy,
you
can
learn
more
about
grafana
mimir
at
the
link
you
can
find
on
the
screen.