►
From YouTube: Discussion to track SMAU
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Let's
see
so
we're
in
the
meeting
we're
discussing,
we
want
to
get
proper
smile
numbers
because
we
don't
have
them
and
you
need
it
for
product.
We
need
it
for
investors.
We
need
to
know
it.
We
need
to
have
it
to
spend
our
money
properly.
We've
been
trying
for
a
year
to
get
small
numbers.
We
have
a
lot
of
misunderstanding.
What
smile
is?
A
So
that's
the
problem
we're
going
to
solve
that's
what
we
should
focus
on
as
a
team
make
sure
the
email
numbers
get
reported
for
all
the
instances
that
send
the
usage
being
including
Atlanta
comm
and
then
to
simplify
things.
Technically,
we're
not
going
to
look
at,
like
suppose,
there's
three
actions
in
a
stage
we're
not
going
to
look
at
those
three
actions
together
and
then
deduplicate,
like
you
kind
of
add
them
and
then
duplicate
them
for
the
people
that
were
in
all
three.
A
A
I
can
check
and
we
can.
We
can
look
at
multiple
activities
and
see
which
is
hires,
but
it's
probably
like
for
plan
is
fully
who,
who
created
an
issue
or
who
locked
a
comment
on
an
issue
is
probably
an
even
better
one
and
then
for
merchant
requesters
who
opened
or
left
a
comment
on
a
metric
rest.
Therefore,
monitoring
I,
don't
know
who
looked
at
a
graph
I'm,
not
sure
we're
tracking
that,
if
we're
not
tracking
that
like
who
made
a
new
dashboard
or
made
a
modification
to
it,.
D
A
D
D
A
E
A
E
A
Awesome
and
alpha
do
you
have
any
ideas,
suppose
suppose
we
solve
this
as
quickly
as
it
now
looks.
Do
you
have
an
idea
on
how
to
do
up
kind
of
a
smile?
That's
an
aggregate
of
a
couple
of
things.
I
think
that's
technically
super
complex
like
say:
either
someone
created
a
merge
request
or
they
commented
on
emerg
request
or
they
I
don't
know,
enter
the
label
somewhere
and
in
all
those
cases,
and
we
could
should
take
the
aggregate
of
that
and
then
duplicate
it.
But
the
data
is
in
different
tables.
E
E
F
Have
a
thought
there,
that's
one
of
the
problems
with
and
there're
issues
where
we've
discussed
this
previously,
where,
if
you
open
ups
now
to
be
defined
in
that
way,
that
we've
traditionally
talked
about
it
being
defined
is
that
different
instances
will
be
calculating
smell
differently.
I,
like
this,
a
male
being
Earth's
now
being
the
highest
a
male
metric,
because
you
help
solve
for
that
problem,
which
is
where
different
instances
have
kind
of
different
things.
They
are
aren't
tracking.
Why.
F
A
E
A
For
now,
for
now,
for
the
purposes
of
giving
clear
directions,
that's
we're
gonna,
assume
that
calculating
like
small
over
multiple
tables
or
multiple
events.
It's
too
hard
we're
not
going
to
do
it.
It
also
is
the
problem
that
Emily
articulated
that
you
did
you
get
into
definition,
questions.
Their
goal
is
to
get
email
from
every
instance
that
have
used
its
tracking
on
fix
the
time
out
problem
make
sure
this
I
don't
know
how
maybe
improve
the
query,
maybe
make
sure
it's
run.
E
Yeah
I
mean
the
thing
with
time
out
for
me
now
from
three
months.
Experience
seems
that
we
are
using
something
called
PG
bouncer
in
the
between
the
database
and
they
can't
and
we
have
15
seconds
of
timeout
for
every
one.
We
had
in
a
large
client
the
project
import
problem
and
we
sold
it
by
moving
the
import
from
the
PG,
bouncer
and
15
seconds
time
on.
So
what
I'm
saying
this
I
think
it
should
be
possible
in
a
background
migration
to
do
it
to
get
any
number.
Can't
you
one.
Oh.
E
E
A
Not
gonna
solve
that
here,
but
I
think
we
should
for
sure.
Try
is
like
running
the
queries
in
background
jobs
and
making
sure
that
they
write
to
some
other
huge,
huge
ping
data
table
or
something
and
they
use
it
being
is
sending
the
users
being.
Data
think
should
be
like
a
tenth
of
a
second.
There
shouldn't
be
any
query
new
big
queries
involved
in
that
it
should
just
be
sending
stuff
out
it
aggregated
over
the
background
over
time.