►
From YouTube: 2023-02-28 Product Analytics Group Sync
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
Hello,
everyone
and
welcome
to
the
February
28th
product
analytics
sync
a
couple
of
items
today,
but
of
course
we
will
start
with
the
build
Board
review,
so
I'm
going
to
go
ahead
and
share
my
screen.
Hopefully
everyone
can
see
that
thumbs
up.
Thank
you
all
right,
so
we
got
a
couple
things
here:
I'm
going
to
refresh,
because
I
just
was
trying
to
edit
some
of
the
issues.
A
It's
deploying
hot
analysts
to
new
public
handbook.
This
is
formally
deploying
to
the
gitlab
project.
The
description
itself
actually
still
needs
to
be
updated,
but
weird,
there's
not
that
many
opt-in
users
for
that
for
the
project
for
it
to
make
sense.
So
we're
going
to
be
pivoting
that
over
the
Milestone
planning
issue
is
just
there
for
excuse
me
reference
and
then
we've
got
the
Pod
analytics
application
setting
for
the
collector
host,
which
is
assigned
to
Khalil
but
I,
believe
he's
out
this
week.
A
So
we
can
check
in
another
time
you.
B
Run
that
plan
by
me
last
week
and
it
looked
pretty
pretty
solid
for
when
he's
back.
A
Great
awesome,
then,
hopefully,
we
can
move
that
into
and
Dev
next
week,
when
he's
back
cool
we've
also
got
a
new
issue
here
in
severity
for
bug
called
Product
analytics
breadcrumbs
do
not
return
to
the
right
page
who's
who
created
this
one
or
we
can
open
it
up.
C
I
mean
I
made
this
one
looks
like
two
or
three
three.
A
Weeks
ago,
I
apologize,
that's
confidential,
yep.
D
C
A
It
all
right,
we
can
triage
that
and
figure
out
what
we
want
to
do
there
and
then
Alan.
You
have
the
backend
issue
for
adding
an
API
to
retrieve
the
state
of
a
product
project
analytics
stack.
A
Is
there
anything
you
want
to
call
out
here
with
this
issue.
A
B
Yeah
I
might
close
this
issue
after
the
call
I'll
check.
It
was
mainly
just
a
sort
of
spark
conversation
between
myself
and
busty
about
how
they
can
start
moving
to
using
snow,
plow
and
product
intelligence,
and
it
looks
like
an
MRS
open
to
that
effect.
So
leave
it
for
now,
but
I
may
close
it
afterwards,
but
there's
nothing
else
for
me
to
do
on
it.
Right
now
sounds
good.
A
And
then
we've
got
an
issue
for
removing
the
default
date
range
limit.
If
that
base
to
have
a
discussion
from
last
week,
there's
no
particular
reason
right
now
to
have
the
date
range
limits
there
and
then,
as
we
do
some
load
testing
as
we
go
through
our
customer
release
phases
we
may
want
to
reintroduce
those
to
some
degree.
A
I
also
opened
up
a
discussion
issue
that
we
can
bring
up
later
in
the
agenda
as
far
as
how
we
want
to
handle
our
pre-configured
dashboards
as
well
as
the
ones
that
customers
are
able
to
build.
So
we
can
discuss
that
sync.
If
we
have
time
moving
over
to
the
planning
breakdown
issue,
we
have
present
call
to
actions
to
configure
dashboards
in
the
shared
dashboard
listing.
A
E
If
they
present
call
to
actions,
that's
with
me
Tim's
right.
Some
questions
that
I
need
to
get
back
to
I
was
going
to
do
it.
Friday
about
as
I
feel.
A
No
worries
great,
should
I
assign
this
to
you
then
or.
D
A
Cool
and
then
you've
got
the
next
blocked
issue
here
for
removing
product
analytics
specific
dashboard
listing
anything
to
call
out
here.
A
Good
moving
on
to
ready
for
development,
then
we
have
the
cube
query,
rendering
visualization
column
chart
which
is
assigned
to
L1
who's,
not
present,
given
the
time
zones.
So
if
there's
anything
to
call
out
from
anyone
else,
please
do
otherwise
we'll
just
move
on
to
Yan's
issue
for
the
404
for
handling.
If
a
dashboard
is
not
available.
F
Yeah,
so
I
haven't
been
able
to
look
at
that
just
yet
the
same
for
the
following
issue:.
A
G
No,
this
is
basically
waiting
now
that
the
load
is
getting
merged
and
then
the
safe
is
next
and
then
new
and
adding
a
dashboard
is
the
third
one
in
line
got.
A
It
I
have
a
question
about
that
load,
one
but
I
guess
that's
part
of
the
pre-configured
and
then
customer
build
one.
So
we
can
get
to
that
later.
A
Sam,
you
have
the
next
one
for
creating
a
post.
A
blog
post.
Excuse
me
about
product
analytics
internal
preview.
Anything
you
want
to
call
out
here.
C
Yeah
I'm
working
with
Sandra
on
this
one
so
feel
free
to
check
out
the
the
issue.
That's
linked
there.
If
you
want
to
see
the
draft
comments,
feedback
welcome
on
it
as
always,
hopefully
we'll
get
it
published
and
it'll
be.
You
know
a
good
way
to
get
people
excited
about
what
we've
been
calling
the
internal
preview
and
what
we're
going
to
be
building
and
shipping
externally
to
them
in
the
near
future.
A
Awesome
I
have
the
next
two
connecting
Dev
and
staging
gitlab
environments
with
respective
clusters.
The
dev
clusters
available,
but
the
staging
one
I
need
to
work
on.
I've,
publish
the
work
to
use
different
environment,
specific
values
on
the
helm
charts.
So
the
staging
one
should
be
a
lot
easier
to
set
up
at
this
point,
but
once
though
they're
set
up,
then
we
can
actually
connect
Dev
and
staging
dot
gitlab
to
those
clusters
to
be
fully
operational.
A
As
far
as
the
Jitsu
data
persistence
on
this
one
just
has
been
was
blocked,
but
now
that
we're
moving
to
snow
plow
I'm
just
going
with
a
quicker
Band-Aid
just
to
prevent
failover
from
the
configurator.
So
hopefully
we
can
get
I
can
get
to
that
this
week
and
then
yeah
and
you
have
the
next
one
for
the
GL
line
chart
for
height.
F
This
is
a
bug
actually
from
gitlab
UI
that
we
don't
respect
the
height
Auto,
setting
I've
spoken
to
the
foundation
team
and
we've
agreed
on
a
solution
to
this,
and
the
solution
is
basically
to
add
some
classes
to
a
jail
line
chart.
This
bug
is
also
present
on
everything
that
uses
GL
chart
inside
galab,
UI,
so
I,
while
I'm
there
I'm
going
to
try
and
fix
it
for
all
the
charts
which
isn't
that
big
of
a
change,
but
it
is
going
to
be
a
lot
of
validation
because
I
need
to
validate
everywhere.
F
A
chart
is
currently
being
used
inside
gitlab.
So
that's
why
I've
bumped
the
weight
from
a
one
to
a
two
I've
also
split
out
the
work
that
was
being
blocked
by
this
in
this
milestone,
which
is
fixing
the
chart
height
from
all
the
other
UI
bugs
that
I'm
fixing.
So
that's
I
can
like
unblock
all
of
that
stuff
and
created
a
new
one-weight
issue
for
just
fixing
that
once
this
GitHub
UI
issue
is
solved.
D
E
Yeah,
so
the
initial
big
front
end
Mr,
is
currently
set
to
M
merge
when
pipeline
succeeds
and
then
I've
got
a
couple
more
Mrs
that
need
to
roll
out
to
continue
that
moving
process.
Obviously,
as
people
add
work
into
the
existing
product,
analytics
I'm
also
monitoring
and
having
to
Port
that
over
too
so
there
may
be
an
additional
Mr
at
the
end
of
all
this
before
I
wipe
everything
to
make
sure
everything's
ported
over
properly.
G
F
These
are
all
interview
with
robes,
so
once
the
Rope
has
a
chance
after
working
on
the
dashboard
listing
to
review,
this
would
be
nice,
but
these
are
all
relatively
minor
and
small
changes.
I
went
bug
hunting
while
improving
the
panel
responsiveness
and
also
solved
to
other
issues,
while
I
was
working
on
that,
including
the
panel
titles
should
be
truncated
and
also
that
single
stat
visualizations
shouldn't
have
decimal
places
when
the
value
is
zero,
it
was
pretty
aware
that
it
was
like
0.00.
F
A
Awesome-
and
that
concludes
our
young
and
Rob
column,
then
going
for
close
issues,
I
believe
our
last
one
was
the
internal
preview
of
dashboard
behavior
issues.
So
then,
we've
got
I.
Think
if
I
just
counted
that
correctly,
six
I
don't
know,
maybe
seven
close
the
shoes
we'll
just
go
in
order,
so
we
have
the
dashboard
lists,
retained,
fluid
container
after
visiting
a
dashboard
or
the
panel
designer
Tim
looks
like
you
fixed
this
one
anything
you
want
to
call
out
with
this
issue.
G
It's
simply
we
have
so
this
is
basically
reaching
out
to
the
overall
container
and
it's
basically
adding
and
removing
classes,
and
you
come
and
basically
go
out
of
that
page
again
and
simply
makes
the
dashboard
itself
fluid
completely.
And
as
soon
as
you
go
back
to
the
list
or
to
something
else,
then
it
basically
goes
back
to
the
limited.
With
again.
A
Good
stuff
there,
this
is
just
a
discussion
issue
regarding
the
now
visualization
designer.
So
as
I
mentioned
last
week,
we
had
already
done
naming,
but
then
I
just
wanted
to
cover
the
overall
strategy
and
how
we
planned
for
potential
visual
designers
in
the
future,
so
I
added
that
after
a
call
last
week.
So
if
you
haven't
seen
that,
then
you
can
catch
up
on
that
there.
A
We
have
a
couple
issues
from
Max
here
for
clickhouse
and
funnel
analysis.
Anything
you
want
to
call
with.
B
These
yeah,
so
we've
made
some
documentation,
changes
to
make
it
possible
to
export
rural
data
from
cube
in
Json
format.
So
that's
in
documentation
today
and
then
funnel
analysis
is
now
working.
You
can
Define
funnels
in
yaml
and
query
them
using
the
API.
So
it's
available
for
front
end
to
work
its
magic
at
some
point
but
yeah
you
can
now
Define
funnels
as
long
as
they're
only
made
out
of
page
View
events,
because
we
don't
support
any
others
yet,
but
yeah
that
works
well.
Yeah.
That's
all.
C
B
C
Yeah
that'd
be
helpful
for
me.
I'd
appreciate
if,
if
it's
not
going
to
take
too
much
time
and
then
another
question
about
the
spike
issue,
part
of
that
issue
is
getting
issues
created
for
the
the
front
inside
of
the
feature
to
be
implemented
later,
were
those
issues
created
in
the
Epic
as
well
or
do
they
still.
A
No
I
think
we
need
to
create
those,
because
Diablo
wants
some
for
how
we
want
to
handle
custom
events
in
the
future
and
then
front
end
needs
to
start
thinking
about
as
well
as
ux.
We
need
to
start
thinking
about
the
actual
workflow
for
for
creating
those
funnels
similar
to
our
visualization
designer.
So
so
we
have
a
couple
issues
to
create
there.
C
B
A
I
think
if
you
just
create
a
new
one,
that's
that's
fine,
but
yeah.
We
should.
We
just
definitely
want
to
make
sure
we
kind
of
hand
that
off
to
the
next
party,
to
continue
on
great
and
then
it's
our
sixth
issue,
which
is
also
our
first
Community
contribution.
A
So
I
can't
really
speak
to
it
as
far
as
the
author,
but
we
got
some
internet
internationalization,
externalization
work
done
to
consolidate
to
use
the
same
namespace
so
that's
great
to
see
and
it
knocks
out
one
of
our
things
targeted
for
the
first
GA
release.
So
it's
great
and
that's
the
Board
review.
Unless.
A
They've
contributed
a
number
of
different
internationalization
Mrs
throughout
different
parts
of
the
product,
so
I
don't
know
how
they
found
it
because
I
don't
think
we
classified
it
as
an
internalization
internationalization
label,
but
they
somehow
dead
yeah
we
could
reach
out,
but
I
think
we
in
the
slack
Channel
they've
they've,
made
previous
contributions
similar
to
this
one
to
other
product
areas,
got
it.
Okay,.
C
A
Yep
cool
that
about
wraps
up
the
Builder
Board
review.
If
there's
nothing
else,
anyone
would
like
to
add.
A
Cool
then
stop
sharing
for
now.
A
I
have
a
point
about
a
discussion.
I've
opened
up
as
I
called
out
on
a
billboard
review
about
how
they
want
to
handle
pre-configured
and
customer
built
dashboards.
I
started
responding
to
some
of
the
discussions
there,
but
there
are
some
open
ones.
It
doesn't
seem
like
we
have
too
much
on
the
agenda.
So
if
there's
no
objections,
I
would
like
to
maybe
just
discuss
some
of
those
points,
so
we
can
perhaps
get
some
some
closing
thoughts
on
or
close
out
some
of
these
threads.
A
Just
let
me
reshare
my
screen,
though
I've
just
lost
that
window.
Give
me
a
second.
A
So
probably
have
a
weirdly
rewarded
question
for
everyone.
This
was
about
how
we
want
to
show
all
the
different
dashboard
configurations
based
on
the
number
of
sources.
The
first
thread
was
about
the
functionality
in
terms
of
how
we
currently
support
it,
but
this
thread
here
is
about
what
do
we
do
when
we
have
pre-built
dashboards
is
defined
somewhere?
A
I
would
imagine
hard
coded
in
the
in
our
code
base,
and
then
we
have
customer
built
ones
and
how
we
want
to
actually
show
all
that
in
the
listing
Max
Tim
Sam,
you
have
a
couple
different
Dots
here.
Would
you
like
to
vocalize
your
your
thoughts
generally
yeah
I.
B
Think
my
my
general
thoughts
initially
were
kind
of
they
were
based
on
my
previous
assumptions,
so
the
quote
said
they
were.
Was
it
strong
opinions,
weekly
held
and
I'm
more
than
happy,
because
I
think
my
understanding
diverges
from
from
Sam's
specifically
for,
for
my
points
to
be
broadly
ignored
and
that's
fine.
D
G
Please
go
ahead
and
summarize,
and
especially
compare
with
Sims,
because
I
haven't
read
Sims
yet.
A
Sure
so
I
think
I
agree
with
Tim's
Point
here,
basically
he's
summarizing
that
we
should
be
offering
pre-built
out
of
the
box.
We
shouldn't
have
to
like
give
them
the
option
to
opt
in,
and
so
I
think
with
the
shared
listing
and
how
we
display
audience
and
behavior
already,
and
they
have
to
already
set
up
the
cluster
or
at
least
initialize,
that
that's
the
easiest
route
to
to
show
value
and
analytics
and
get
them
instrumenting
their
applications.
A
And
so
we
should
always
show
that,
in
order
to
kind
of
just
have
an
easy
path
to
providing
value
there
and
then
the
customer
that
custom
dashboards
are
just
for
yeah.
Whatever
else
they
build.
So
correct
me
if
I'm
wrong
Tim,
but
basically
we
should.
We
should
show
both
at
all
times
right.
G
Exactly
that,
I
think
that
the
custom
dashboards
can
be
anything
at
the
end
of
the
day.
There
is
a
there's,
a
really
good
Epic
from
the
optimized
team
on
custom
dashboards,
where
you
can
really
see
how
much
customers
are
interested
really
at
the
end
of
the
day
into
those
dashboards,
you
can
see
tons
of
customer
requests
and
customer
success
teams,
basically
saying
yeah.
G
This
company
XYZ
has
already
built
a
dashboard
in
the
RPI
tool,
with
the
data
from
optimize,
but
at
the
same
time,
with
the
error
rates
and
so
on
and
I
think
that's
also
what
what
we
want
to
be
able
to
enable
our
customers
is
at
the
end
of
the
day
that,
like
we
will
think
about,
it
is
okay.
We
have
a
team
which,
let's
not
make
it
too
complex,
so
that
it
doesn't
confuse
everyone.
G
Let's
say
the
plan
project
management
team
has
a
dashboard
at
the
end
of
the
day
to
see
how
often
are
there
features
used
on
daily
level?
How
often
do
errors
happen
in
production?
How
good
is
their
error,
but
that
how
good
is
their
throughput
rate?
How
many
Mrs
have
been
done?
This
Milestone,
and
so
on,
so
really
anything
that
is
relevant
to
the
to
the
team
and
the
more
flexibility
they
have
to
mix
and
match
what
they
are
actually
using
and
what
they
need.
D
A
Cool
I
do
want
to
clarify
the
questions
which
again
I
could
afford.
It.
Better
is
for
customer
built
ones,
and
so
we
have
different
sources
potentially
right
now
we're
looking
at
the
project
repository
itself,
but
we
have
the
ability
to
define
a
repository
elsewhere,
which
contains
this,
which
can
be
used
at
a
higher
level
group
level.
A
Perhaps-
and
so
this
question
specifically
the
reason
why
we're
talking
about
whether
we
show
pre-builds
or
not,
is
because
some
of
the
discussions
as
a
coming
out
of
this
question,
but
the
question
just
to
clarify
is
around:
we
have
customer
built
dashboards
in
multiple
places.
How
do
I
want
to
actually
like
show
all
of
them,
or
do
we
if
the
repository
is
defined?
Do
we
not
do
we
ignore
configurations
in
the
project
itself?
A
So
just
with
that
clarification,
Sam
is
there?
Does
that
help
in
terms
of
understanding
the
question
or.
C
Yeah
I
think
it
does
so
from
my
perspective,
if
a
customer
has
defined
a
project,
a
project
pointer
to
say,
configuration
is
done
over
in
that
repository.
That
is
the
place
that
we
should
look
for
configuration
in
the
sole
place.
We
should
look
if
we're
going
to
look
at
a
place
where
the
project
pointers
point
and
in
the
local
repository
that's
going
to
introduce
weird
edge
cases,
it's
going
to
make
our
information
architecture
harder
to
understand
for
people.
C
C
Okay
and
I
wrote
a
little
bit
more
about
this
in
my
response,
but
really
one
of
the
key
goals
of
those
is
to
make
it
so
that
customers
have
the
shortest
path
from
the
idea
of
I
want
to
add
product
product
analytics
to
my
project,
to
my
application,
to
I,
see
something
that
I
find
interesting
and
I
get
value
out
of
the
the
whole
process.
We're
going
to
ask
users
to
take
of
reading
our
documentation
grocking
it
instrument
their
application.
Sending
us
data
that's
very
difficult.
C
If
we
then
add
more
steps
to
okay,
you
have
the
data,
but
you
have
lots
of
additional
steps
before
you
can
even
see
it.
That's
going
to
give
customers
even
more
opportunities
to
say
you
know
what
this
is
too
frustrating
I'm
going
to
go,
do
something
else,
and
so
that's
really
a
big
part
of
why
we
want
to
always
be
showing
these
pre-configured
dashboards
so
that
customers
can
get
to
Value
very,
very
quickly.
C
I
do
think,
there's
an
argument
to
be
made
for
providing
a
way
to
disable
those
dashboards.
If
customers
don't
want
them
once
a
customer
reaches
a
given
point
of
maturity.
I
imagine
they'll
mostly
have
customized
dashboards
and
maybe
won't
want
our
predefined
ones
anymore.
But
that's
that's
a
future
issue.
We
can.
We
can
deal
with
as
a
feature
later.
C
D
C
Think
that's
everything
in
my
my
comment,
but
please
please
go
ahead
and
read
that
as
well.
If
you
like.
A
For
sure
so
kind
of
for
the
takeaways
for
this
discussion
then
and
I
realize
it's
taking
a
lot
more
time.
So
we
can
skip
the
next
Point,
but
ultimately
we're
we're
kind
of
it
seems
like
we
want
to
try
to
provide
pre-builds
unless
you
opt
out,
at
least
to
start
to
provide
value
and
reduce
friction
and
as
far
as
reading
customer
built
ones,.
A
If
they're,
if
they're
a
customer
about
defining
a
dashboard
repository,
we
should
read
from
there
to
avoid
edge
cases
with
additional
configurations
naming
conflicts.
Things
like
that
in
the
local
repository.
Does
that
sound
okay?
Well,
like
let's
I'll
I'll,
follow
up
on
this
thread,
and
then
we
can
get
some
some
people's
thoughts
on
that
just
to
move
that
forward
in
the
interest
of
time.
So
I'll
follow
up
on
this
issue.
A
Cool
I
just
wanted
to
call
out.
There
is
a
discussion
about
picking
a
alternative
meeting
time
for
the
a
more
APAC
friendly
time,
for
this
call
that
we
can
alternate
between.
So
please
add
your
thoughts
there.
Otherwise
Tim
you
have
a
point
for
show
and
tell.
G
G
So
pasty
already
started
to
work
on
on
Mr
or
for
our
dev
kit,
which
basically
brings
in
snowplow
so
that
we
are
able
to
collect
snowplow
data
in
the
GDK
itself.
So
right
now
the
GDK
has
a
setup
where
you
can
say:
I
want
to
run
snowplow
micro,
which
is
then
basically
collecting
those
events
like
for
that.
We
do
one.com
so
that
you
can
test
this
locally
and
what
busty
did
is
he
has
taken
this
and
basically
rewrote
it.
G
G
So
this
is
already
using.
If
we
see
edit
and
say
code,
then
it
doesn't
show
because
it's
in
the
external
files,
but
if
we
go
to
page
views
and
say
all
the
pages,
the
code,
what
is
happening
it
already
sends
basically
to
a
new
schema,
which
is
the
snow,
plow
events
and
has
there
the
page
views
count.
G
So
what
in
reality,
we
need
to
do
to
as
soon
as
our
collector
counts
and
has
snowplow
data
is
have
a
new
adapted
schema
file
for
snowplow
events
have
that
in
a
cube
and,
on
the
other
hand,
have
basically
how
all
our
dashboards
reconfigured,
which
is
pretty
easy
to
be
honest,
so
this
takes
less
than
two
hours
to
get
through
all
of
them
and
test
them
and,
on
the
other
hand,
also
the
visualization
designer
keys.
G
So
all
those
things
that
you
can
click
here
are
basically
using
table
names
and
based
on
these
table
names,
you
are
able
to
field
names,
you're,
basically
rewriting
that
query,
and
that
goes
already
to
the
snowplow
event.
So.
G
Simply
named
a
little
bit
different,
the
nicest
part
that
I
wanted
to
call
out
here
is
that
if
we
look
that's
the
problem
with
white
screens,
you
can't
share
the
full
screen,
because
otherwise
it
will
be
a
horrible
experience
for
everyone
watching
this,
but
I
need
to
switch
Windows.
All
the
time
is
what
you
can
see
here
is
that
we
get
way
more
columns
from
snowplow
I
linked
in
the
agenda.
There
is
also
a
listing
basically
of
the
enriches,
but
I
didn't
add
the
link.
Yet
I
will
do
this
afterwards.
G
So
snowplow
has
a
couple
of
enriches
already
in
place
and
I
was
discussing
with
pasty
that
we
most
probably
are
going
to
Simply
activate
all
of
them
that
snowplow
hair
spill
the
model
based
on
that,
so
that
we
get
all
the
data
and,
at
the
later
Point
look
that
basically
users
are
able
to
opt
and
opt
in
for
a
specific
enrichers
or
opt
out,
but
that
we
could
build
our
dashboards
on
top
of
those
images.
G
To
give
you
a
little
bit
of
under
standing
of
that.
What
does
this.
D
G
Is
that
you
have
way
more
enrichments
around
the
browser
itself,
so
it
uses
way
more
modern
approaches
of
figuring
out
what
kind
of
device
it
is
the
operating
system
and
basically
can
break
it
down,
because
there
is
currently
a
trend
to
the
deprecating
reality,
the
user
agent
string.
So
this
is
the
classic
way
you
have
a
user
agent
string.
Everyone
throws
whatever
they
want
in
there
and
then
all
the
analytics
packages
figure
out
what
to
parse
out
of
it.
G
But
there
is
now
a
new
initiative
coming
since
one
or
two
years
where,
basically,
the
browser
provides
already
way
more
options
on
it,
and
there
are
a
couple
of
more
enrichments
going
from
the
the
the
referral
parser.
So
you
can
basically
get
an
understanding
of
what
kind
of
medium
it
is.
So
you
can
basically
already
have
in
the
snowplow
data.
G
Is
this
coming
from
search,
internal,
unknown,
social
or
email,
and
it
supports
all
the
classic
stuff
for
marketing
campaigns
with
what
is
really
important
for
marketing,
of
course,
and
what
else
is
I
think
you
can
do
IP
lookups,
so
geolocation
is
also
a
thing
in
different
varieties
and
we
can
do
all
the
IP
anonymization
automatically
the
pi
upside
randomization
works
also,
and
there
was
also
one
more
thing:
interesting
yeah
we
can
have
and
what
was
it
yeah,
the
iaap
enrichment?
G
G
So
that's
clearly
the
target
that
we
will
work
with
the
product
intelligence
team,
getting
this
in
Mr
and
we
will
get
a
huge
schema
in
clickhouse
that
we
will
be
able
to
work
against.
But
this
gives
us
way
more
options,
also
in
the
pre-built
dashboards
and.
D
G
A
But
you
can
keep
going
I
just
wanted
to
ask
about
the
pseudonymization,
but
that's
cool
that
it's
already
an
enricher
and
then
we
would
just
update
the
SDK
just
in
terms
of
what
these
options
are.
I
imagine
so,
but
I
will
have
to
drop
but
feel
free
to
show
it
off,
because
it'll
be
on
the
recording.
We
can
review
it
later.
Thank
you,
perfect.
G
G
Right
now,
we
will
have
definitely
the
topic
that
we
are
rather
focused
on
having
a
certain
model
and
we
reuse
it,
but
we're
also
looking
now
into
different
options
to
use
all
the
thousands
of
offers
that
came
out
of
the
came
out
of
Earth,
basically
over
the
last
couple
of
months,
and
one
of
them
is
of
course
open
AI
and
what
I
was
playing
around
with.
It
is
what
you
can
see
here.
It's
not
completely
done,
but
I
have
an
Postman
I
have
a
server
for
this.
Let
me
activate
that.
G
So
what
does
this
do
so?
This
is
basically
using
open,
Ai
and
open
AI
got
the
model
of
the
snowplow
table
from
me,
and
the
nice
thing
is:
there
is
tons
of
snowplow
code
out
there
in
the
world
so,
but
this
model
is
basically
capable
of
doing.
Is
it's
capable
of
writing
SQL
for
us,
which
we
then
can
take
and
send
to
Cube
and
basically
create
a
dashboard?
G
Found
it
let's
try
again,
and
this
works
really
really
well,
and
it
understands
so
many
things.
So
the
SQL
queries
work
by
now.
Most
of
the
time.
G
Is
what
some
people
did
is
like?
They
said
they
tried
to
send
in
the
SQL
to
whatever
database.
They
have
and
say,
works
or
didn't
work
and
basically
report
back
to
open
AI,
hey
this
card.
Didn't
work,
please
retry,
and
they
will
give
you
a
view
query
and
they
basically
do
it
three
times
until
it
actually
works.
So
how
many
page
did
we
have
this.
G
And
this
is
basically
then
creating
automatically
down
here.
It
is
telling
me
there's.
Oh,
that's,
also
interesting
to
create
an
SQL
with
a
missing
volume,
so
someone
was
creating
that
stuff.
So
this
is
basically
already
trying
me
because
I
found
an
internal
Cube
API.
There
is
a
new
feature
coming
in
Cube,
which
is
called
SQL
Runner
that
you
can
activate
where
you
can
send
directly
as
well,
which
is
cleaned
out
which
didn't
work
so
far.
G
They
have
already
a
CBE
on
it,
but
Cube
works
in
the
in
the
way
that
they
have
an
SQL
API
to
support
all
kinds
of
business,
intelligence
and
so
on
tools,
so
that
you
can
basically
connect
in
the
way,
as
you
would
to
a
database
postgres
database,
but
in
reality
this
is
in
the
background,
is.
D
G
This
can
convert
basically,
at
the
end
of
the
day,
we
should
be
able,
through
something
like
this,
to
basically
really
create
a
dashboard
panel
to
say
that
you
simply
type
what
you
want
and
say:
okay,
I
want
to
know
all
the
page
views
that
happened
in
the
plan
stage,
and
it
should
be
able
to
understand
this
quite
clearly
and
straightforward
and
as
this
is
creating
as
well,
we
can
validate
it
against
Cube
and
convert
it
into
a
cube
query.
But
it's
pretty
nice.
G
So
it's
almost
complete
I
will
try
to
get
this
into
our
visuality
station
designer
POC,
because
we
are
currently
trying
to
collect
a
couple
of
those
topics
to
see
how
we
are
going
to
adapt.
Maybe
our
AI
strategy
in
using
rather
external
services
or
existing
models
and
then
to
write
our
own.
But
this
is
definitely
something
which
I
would
love
to
see
the
project
product.
Because,
if
you
ask
me,
this
is
a
constant
struggle
that
the
data
team
has
to
do
is
tons
of
their
work.
Simply
hey.
G
Can
we
have
all
the
events
for
this
and
that
and
that
page
and
in
plan
we
are
looking,
for
example,
also
that
you
are
able
to
that.
We
are
able
to
feed
all
the
issues
from
a
project,
and
you
can
ask
it
okay,
how
many
Bucks,
what
types
of
bugs
do
we
have
on
the
to-do
page
and
stuff
like
that
so
and
I
think
this
would
be
a
pretty
nice
addition
to
the
problem.
G
Questions
Max
did
you
have
a
question
because
you
wanted
to
know
how
this
is
done?
Well,.
B
Yeah
I'm
more
I'm,
just
wary
of
the
fact
you
don't
want
to
AI
us
out
of
a
job,
but
no.
This
is
very
cool.
E
Did
we
take
the
I
know
we're
doing
ml
models
in
gitlab?
Currently,
in
some
incubator
related
work,
could
we
take
some
of
those
models
or
do
we
have
the
models
already
for
the
gitlab
database?
So
we
could
retie
in
the
snowplow
data
with
gitlab
specific
project
related
data
all
already
without
too
much
work.
G
G
G
D
G
Language
model
in
the
background-
and
that
might
be
a
question
that
we
want
to
discuss
quite
soon
so
that
we
are
capable
of
keeping
up
with
everyone.
G
It
also
actually
understands
the
difference
between
click
house
and
postgres.
That
was
a
funky
experience,
so
you
can
tell
a
tape
rather
right,
click,
house,
SQL
and
then
it
converts
different
types
of
SQL
commands,
because
the
date
stuff
is,
for
example,
different
and
so
on.
So
but
yeah
it
was
my
fun
Friday
project
and
but
definitely
trying
to
see
that
we
get
some
capabilities
like
this
and
think
about
some
use
cases
in
a
very
near
future.
Hopefully
to
get
this
to
the
product.
D
G
Good
great
seeing
everyone
happy
start
to
your
week
as
it
is
a
forest
Monday,
take
care
everyone
and
see
you
soon.