►
From YouTube: Monitoring Strategy Thoughts from Sid - 2019-09-20
Description
Sid's thoughts on the monitoring strategy during our product scaling meeting.
Follow up issue was created by the GitLab Monitoring Product Management team. You can track our progress here - https://gitlab.com/gitlab-com/Product/issues/461
A
So
I've
been
thinking
about
monitoring
and
get
lamp
and
I
think
right
now
our
ambition
is
like
well.
If
you
want
really
good
dashboards,
you
should
probably
views
graph
on
and
not
get
lamp
and
we're
starting
to
point
people
that
way.
That's
not
how
we
ever
got
get
map
into
like
best
in
class
and
if
we
keep
sending
all
the
people
that
want
more
away
instead
of
having
them
improve,
get
lap
that
we're
not
we're
gonna
get
there.
Good
love
is
about
a
single
interface.
We
need
to
be
the
interface.
We
cannot
have
another.
A
A
PD,
interface
and
I
know
that
it's
a
lot
of
work
to
get
there,
but
we
choose
each
charge.
They're,
pretty
good
I
think
we
can
make
them
better
and
our
I
don't
think
our
ambition
level
should
be
graph.
Anna,
honest
awesome,
but
I
think
our
ambition
level
should
be
data
doc.
Data
dark
makes
is
really
good
to
go
from
alert
to
metrics
logs.
A
A
A
What
is
happening
in
the
monitoring
market
is
that
horns
printing
money,
because
you
paper
for
storage
and
we
just
grateful
all
the
providers
and
I,
but
we
can
be
a
disrupter
here-
we're
open
source
you're,
not
gonna,
pay
for
stories.
As
long
as
you
provide
the
storage,
so
we'll
have
a
dual
business
model.
What
people
using
self
hosted
gitlab
self-managed
gitlab
they
can.
A
They
don't
have
to
pay
per
gigabyte,
I
think
that's
gonna
be
really
attractive,
because
I
think
people
are
sitting
there
seeing
their
logging
and
metrics
bills
go
through
the
roof
and
forget
about
calm.
People
can
pay
us
for
the
extra
storage,
so
it
will
be
a
way
to
kind
of
do
consumption,
pricing
and
generate
additional
income
for
get
back
home.
A
We
we
do
have
to
get
it
into
a
single
application
and
I
think
we
can
and
we
have
great
building
blocks.
We
can
base
it
on
Prometheus,
like
we
already
do
for
metrics.
You
can
base
the
logging
on
elasticsearch,
but
we
do
have
to
supply
the
interface
and
that's
a
lot
of
hard
work,
but
we
have
the
talent
and
then
the
people
to
just
start
doing
that
and
I
think
it's
really
important
to
visualize
this.
Our
ambition,
because
I
think
people
kind
of
have
a
hard
time
visualizing,
where
we're
gonna
go
and
so
I
love.
A
Some
sketches.
I'd
love
some
help
from
our
UX
team.
I
love
help
to
like
sketch
hard
is
just
wanna
kind
of
work
in
Gail
I've
been
in
get
lab.
We
can
make
it
a
more
seamless
flow
than
anywhere
else.
We
can
just
generate
a
whole
lot
of
value
for
our
users
and
customers
so
about
my
short
ambition
level
for
monitoring
any
input.
I.
B
A
I
think
I
think
in
tracing
we
should
do
the
iframe
I
think
we
know
what
to
do
there.
It's
easy
to
do
it
I
think.
There's
been
a
bit
of
confusion
about
the
security
implications
there.
We
can
just
put
a
warning
above
it
like
hey
in
the
screen
below.
You
should
never
see
a
login
prompt
if
so,
report
abuse
and
we
can
just
I
frame
Jagr,
that's
something
we
should
do
it's
not.
It
doesn't
seem
like
a
lot
of
work
but
other
than
that
I'm.
A
Okay,
with
focusing
on
logging
like
we
need
to
start
aggregating
the
logs
right.
Now
we
have
people
go
out
to
the
pods
need
to
start
sending
them
back
by
default,
need
to
start
having
elasticsearch
as
a
good
lab
manager
application
to
do
that,
but
also
lots
of
inter
free
interface,
work
to
be
done.
Okay,.