►
From YouTube: Monitor:APM Weekly Meeting 2020-04-22
Description
Weekly meeting for the Monitor:APM group
A
A
All
right
I
can
go
ahead
and
get
started.
Hey
everybody.
Welcome
to
the
monitor
a
p.m.
weekly
meeting
first
agenda
item
I
had
was
just
in
case
you
all
haven't
heard.
There
is
friends
and
family
day
coming
up.
It
was
announced
in
the
company
fYI
select
channel.
Basically,
we
want
to
take
we're
having
our
first
gitlab
holiday,
where,
if
you
don't
celebrate,
if
you
don't
have
a
holiday
at
your
country
for
May
1st,
you
are
strongly
encouraged
recommended
to
take
that
day
off.
A
Everyone
at
get
lap
is
going
to
be
recommended
to
take
off
unless
there's
essential
work.
That
needs
to
be
done,
like
support,
team
and
so
forth,
and
if
you
already
have
that
day
as
a
public
holiday,
you're
encouraged
to
take
another
day
off,
preferably
around
that
time,
because
we
want
to
share
kind
of
what
we're
doing
outside
of
work.
I.
Think
there's
a
slack
channel
as
well,
so
I
will
definitely
be
taking
off
and
I
will
try
to
share
what
I'm
doing
I
haven't
decided
yet,
but
I
encourage
you
all
to
do
the
same.
A
The
idea
is,
you
know,
to
give
us
a
chance
to
remember
and
focus
on
our
family
and
friends
during
this
time,
where
it's
really
easy
to
to
just
focus
on
work
or
work
too
much.
So
there's
that
and
then
there's
virtual
contribute
tomorrow,
I'm
not
sure
if
all
of
y'all
have
heard
as
well.
So
this
is
the
first
time
they've
done
this
I've
been
here
for
a
while
and
so
as
Jose,
and
you
know
this
will
be
very
fascinating.
We're
trying
a
new
platform.
So
please
do
you
check
it
out.
A
I
encourage
you
to
do
it.
You
don't
have
to
attend
everything
at
least
attend
the
keynote
attend
a
few
sessions.
I
think
it'll
be
a
great
way
to
connect
the
networking
event
thing
that
they
have
seems
really
cool,
so
at
least
check
out
one
thing:
it's
my
recommendation
and
then
the
other
thing
is
I
want
to
remind
everyone
to
comment
on
the
retrospective
for
12.10.
There
hasn't
been
much
activity.
It's
kind
of
just
me
talking
to
myself.
So
please
join
in
the
conversation
because
I'm
sure
there's
things
that
we
could.
A
B
Is
one
of
questions
/?
If
anybody
has
any
comments,
so
the
annotations
feature
that
we
showed
in
the
previous
milestone
is
behind
a
feature
flag.
So,
while
I
was
talking
to
Clement,
this
topic
came
up
and
we
were
thinking
if
there's
any
use
and
putting
it
behind
a
feature
flag.
So
the
idea
is
to
lift
the
feature
flag
in
this
release
and
can
I
had
all
the
necessary
documentation
and
all
the
screenshots,
if
necessary.
So
I
want
to
see
if
anybody
has
any
comments
or
concerns
about
that.
C
My
comment
is
I'll,
just
verbalize
it.
We
need
to
make
sure
this
is
properly
documented,
ideally
with
some
examples
and
screenshot,
because
this
is
that's
it
that's
a
big
feature.
It's
a
big
deal
and
I
would
like
to
make
sure
it
is
accessible
by
anyone
that
is
leaving
out
the
communication.
So
you
can
take
an
extra
I,
don't
know
couple
of
hours
to
make
sure
this
is
properly
documented.
B
Yeah,
especially
given
that
annotations
can
only
be
created
through
the
backend
or
through
books,
I
think
I
agree.
We
need
to
add
some
more
documentation
now
and
that's
Michael
I
pointed
out.
There
is
some
documentation
talking
about
the
API,
endpoints
and
other
things,
but
I
think
we
need
to
add
few
more
details.
Yeah.
C
Based
on
the
information
that
is
coming
from
snowplow
and
a
usage
pin,
so
just
just
to
recap,
usage
king
is
the
information
that
we
are
collecting
from
our
own
plan
customers.
So
customers
that
actively
enable
usage
paying
they
are
sending
telemetry
information
back
to
us.
So
it's
approximately
twenty
to
thirty
percent
of
our
users,
so
usage
P,
the
numbers
that
you
see
this
P
probably
would
be
larger
because
they
traditionally
we
have
a
large
installed
base.
C
That
is
using
our
own
plan
much
more
than
they
want
data
using
corn,
but
it
is
in
a
queue
late
because,
as
I
mentioned,
twenty
to
thirty
percent
of
the
information
is
getting.
His
information
is
simply
blocked
on
a
snowplow,
and
this
is
where
we
are
tracking
our
comm.
We're
actually
collecting
all
the
information
and
we
can
get
high
resolution.
For
instance,
we
can
usage
field,
we
cannot
distinguish
a
session.
C
So
if,
like
this,
one
user
click
on
a
button
that
we
are
tracking
information,
1000
time
will
count
it
1000
time,
but
it
is
in
snowplowing
calm.
We
can
distinguish
by
by
session
ID.
So
there
is
some
mix
and
match
that
I.
Try
to
collect,
based
on
both
points
of
both
source
of
information,
with
your
help
that
you've
done
the
instrumentation
and
I
was
able
to
collect
I.
C
Think
all
the
information
that
we
can
get
and
we
can
get
to
a
higher
resolution
is-
will
continue
to
create
the
implementation
issue
to
collect
some
of
the
some
of
this
information,
so
I
just
want
to.
Obviously
this
dashboard
is
accessible.
Everyone
can
use
it.
If
you
have
any
issues
around,
you
know,
but
just
to
go
over
it,
because
it's
it's
quite
interesting.
C
So
the
first
point
of
information
is
users
that
are
fetching
information.
Basically
looking
at
dashboards,
it
doesn't
matter
if
it's
deaf
default,
dashboard
or
custom
dashboard.
As
long
as
someone
is
looking
at
the
dashboard
and
the
dashboard
does
return
something,
so
it's
not
looking
on
an
empty
dashboard.
C
This
is
the
this
is
the
metric
here
and
it
is
broken
down
into
users
that
are
using
the
custom,
dashboard
and
the
default
dashboard.
I
know
it
does
not
add
up,
because
it
depends
on
the
way
we
advocate
the
information,
but
it's
like
it's
almost
okay.
It's
not
it's
not
accurate,
but
you
can
see
the
breakdown.
How
many
users
are
using
our
default
dashboard
and
users
that
are
using
custom.
Dashboard,
I'm
I
have
to
say
that
I'm
looking
at
trends,
I
think
this
is
I.
C
Don't
look
at
the
numbers,
especially
not
on
usage
King,
but
I
like
the
fact
that
we
can
see
that
more
users
are
starting
to
use
our
custom,
dashboard,
I
think
tied
in
to
the
fact
that
we
are
adding
more
and
more
enhancement
to
our
dashboard,
and
you
can
see
that
the
trend
is
going
up
and
it's
great
amount
of
user
there's,
not
a
lot
of
money,
users
that
are
using
the
adding
metric
using
our
models.
So
if
they
are
clicking
on
the
add
metric
a
button
and
by
the
way
the
data
is
aggregated
by
month.
C
Not
a
lot
of
users
are
doing
that
I
think
mainly
for
two
reasons:
one
the
add
metric
button
is
in
we
made
and
we
move
it
to
cool
it's
it's
custom
attics,
basically,
and
we
move
it
to
call
only
in
this
release.
So
I
expect
to
expect
this
to
go
up
and
also
the
fact
that
you
can
add
geometric
using
a
Amon
facts.
You
don't
need
to
go
to
the
model.
You
can
do
it
directly
from
the
Yemen
numbers
of
using
that
are
creating
alerts.
I
know.
Al
else
is
more.
C
The
health
team
is
more
working
on
that,
but
still
it's
it's
important
to
know.
We
don't
have
a
lot
of
traffic
there
and
probably
we
to
do
some
more
work
on
that,
and
we
know
we
are
where
the
tower
letting
is
lucky.
I
would
expect
that
most
of
the
users
that
are
using
a
lot
weave,
get
labelled
using
that
health
manager
infirmities,
but
I,
don't
think
it's
possible
to
track.
It
now
included
nah
yeah.
C
A
A
B
D
B
A
C
I
think
yeah
I
think
we
I
think
we
should
and
again
intimated
matter
of
priorities,
because
each
channel
mutation
issue
is
another
I'm
not
trying
to
overload
the
team
with
with
instrumentation
I
mean
there
are
so
many
actions
that
we
can
do.
We
can
think
about
it
if
we
think
that
this
is
a
critical
metric.
I
think
that
custom
dashboard
is
something
that
can
like
hint
on
the
amount
of
users
that
are
starting
to
add
new
metrics.
C
Maybe
we
can
think
about
collecting
this
information
through
the
yamen
file
and
yeah,
but
I
get
a
lot
of
points
of
data
that
we
can
and
we
can
collect
continue
with
that
number
of
users
that
install
parameters,
basically,
click
on
parameters,
the
install
button
and
obviously
the
number
I
think
it's
quite
impressive,
because
this
is
an
action
that
is
being
done
only
one
time.
So
it's
not
something
you
are
going
like.
You
know
watching
the
dashboards
one
time.
One
time
thing
same
with
number
of
users
that
install
elasticsearch
for
logging
were
subtracting.
C
The
number
of
users
that
are
doing
uninstall
and
I
was
able
to
find
use.
I
was
was
able
to
find
information,
but
how
many
users
are
actually
clicking
on
download
the
CSV
form
from
the
dashboard?
This
is
all
coming
from
phone
comm,
a
new
search
pane.
We
don't
track
a
lot
and
I
have
to
state
the
information
there
is
it's
not
so
accurate.
C
The
telemetry
team
is
doing
some
work
on
that.
So
I
expect
this
to
be
enhanced,
but
for
now
what
I
was
able
to
find
is
the
number
of
users
that
are
clicking
on
potlucks,
and
you
can
see
this
increase
of
number
over
the
last
month
and
I
think
this
is
mainly
because
of
two
factors:
one
we
move
logs
from
ultimate
or
premium
to
Tuco,
so
more
users
are
starting
to
use
that
and
elasticsearch.
You
can
see
that
there
is
a
correlation
here,
number
of
users
that
are
installing
elasticsearch.
You
can
see.
C
The
number
is
going
up,
whether
it
is
this
num,
this
the
data
from
usage
being
a
bigot
on
a
weekly
basis.
This
is
on
daily
and
I'm
tracking
those
metrics
and
I
do
see
an
increase
of
we
cover.
We
can
probably
change
the
aggregation
for
a
week
here
and
to
be
able
to
see
it
and
the
last
one
is
still
working
progress.
Just
a
point
of
information
I
was
able
to
track
project
that
has
active
tomatillos,
but
the
numbers
here
I
think
a
bit
wrong.
Those
spikes
30
millions,
70
million
and
6
million
I.
C
Don't
think
that
accurate
were
probably
around
this
number
like
the
20k
or
30k.
But
the
folks
on
telemetry
team
are
working
on
making
sure
is
that
I
would
be
accurate
because
it
really
doesn't
help
me
a
lot
with
looking
at
the
metrics
because
of
those
spikes.
So
this
is
still
walking
tour
bus
and
this
is
it
for
now
and
as
we
continue
with
the
with
the
release,
I'll
create
more
implementation
issue
for
instrumenting
and
collecting
telemetry
information,
which
I
think
is
important
and
interesting.
C
C
A
If,
like
90%,
if
we
realize
90%
of
our
users
are
creating
the
dashboards
through
the
yamo
file,
then
maybe
storing
the
database
might
be
I
mean.
Maybe
we
have
to
figure
out
technical
solution
to
mirror
it
or
something,
rather
than
just
remove
that
support
I'm,
not
sure
if
that
was
the
intent
of
the
idea
of
moving
to
database,
but
something
to
think
about.
C
C
B
Yeah,
so
more
suggestions,
I
thought
it
would
be
nice
to
see
how
many
metrics
our
charts
are
added
to
custom
dashboards
because
looks
like
there's
an
upward
trend
in
the
usage
of
custom
dashboards.
So
it'd
be
nice
to
know
how
how
many
metrics
users
in
general
have
in
those
custom,
dashboards
and
another
thing
I
could
think
of
was
it
would
be
nice
to
also
track
if
projects
generally
have
manual,
Prometheus
configurations
or
automatic
configurations
yeah.
That's
all
yeah.
C
Demand
the
manual
and
external
is
something
that
we
are
tracking
I'm,
not
sure
if
it's
possible,
so
there
is
on
the
dashboard.
There
is
sorry,
no
I
thought
we
are
tracking
it.
We're
just
tracking
number
of
religious
installation
I
think
we
can
I
think
we
have
that
metric
I
need
to
try
to
fight
it
in
usage,
Bing
or
snow
plowed.
Soil
yeah
I
think
we're
talking
that
I.
Remember
that
we
had
an
implementation
issue
for
that.
D
C
Yeah,
the
question
is:
can
we
collect
this
in
in
snowplow
because
it
needs
to
be
very
superior?
Yama,
fine,
yeah?
Okay,
so
tell
me
amplify:
we
can
do
it,
not
per
user
I'm,
fine
yeah.
We
need
to
I
think
we
should
start
with
like
just
number
of
charts
and
then
like
walk
away
as
an
MVC.
Let's
first
count,
the
number
of
child
chief
makes
sense
and
then
see
if
we
can
get
like
a
high
resolution
of
number
of
child
they'll
pay
homage.
Fine,
it's
been
interesting,
I
agree.
A
They
also
had
a
random
of
whether
we
think
it's
possible
to
sense
snowplow
data
and
to
Prometheus
just
think
of
you
really
cool
use
case
for
us
to
dog
food
in
a
way.
That's
not
necessarily
meant
for
oversee
and
trying
to
do,
but
an
interesting
way
to
to
you
know
push
the
limits
of
what
we
can
do.
Is
there
charts
not
not
necessary
to
do
right
now,
but
it's
just
something
to
think
about.
C
E
Could
try
using
up
in
telemetry,
we
can
check
out
there.
They
have
any
anything
that
connects
to
snow
globe,
Oh
like
as
a
middleware
yeah.
That
way,
we
don't
have
to
rewrite
all
of
our
snow
blow
code,
which
you
can
just
have
the
up
in
telemetry
stuff
and
have
have
it
send
the
data
to
both
the
snow
blow
and
Prometheus.
C
A
C
A
B
Quick
question
so
do
be
for
these
snow
plow
events.
Do
we
by
any
chance
filter
out
certain
projects
because
those
five
snow,
plow,
sorry,
those
five
CSV
downloads
I,
feel
like
I
just
did
three
of
them
when
I
was
just
testing
on
one
or
two
sandbox,
so
do
we
filter
the
internal
testings
projects
within
gitlab
org
out
of
those
I
wonder
if
you.