►
From YouTube: 2023-08-29 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
August
29th
edition
of
the
podcast
analytics
group.
Sync,
it's
a
small
crowd
today,
so
it
could
be
nice
and
cozy,
but
as
far
as
getting
things
started,
we
have
the
Board
review
but
I
think
since
it's
just
the
three
of
us
I
think
we
can
just
do
a
quick,
stand-up
style
thing
and
then
we
can
go
from
there
just
to
not
run
through
a
bunch
of
issues
that
may
not
have
had
any
movement
yet
we'll
follow
through
async
and
see.
A
If
there's
anything,
we
need
to
follow
up
on,
but
I've
been
seeing
some
things
that
I
need
to
respond
to
Alan,
one
of
them,
including
yours.
So
hopefully
we
can
discuss
that
today
or
I
can
follow
up
with
that
after,
but
I
can
go
ahead
and
start.
So
my
my
focus
still
continues
to
be
the
same.
So
clickhouse
cloud
is
working.
I
have
a
cluster
set
up
now,
specifically
for
instrumentinggitlab.com.
A
A
So
my
my
goal
for
this
for
the
next
day
or
so
is
to
really
push
those
Mrs
so
that
it's
officially
official
as
far
as
Cloud
support
is
concerned,
but
at
the
same
time,
barring
some
Secrets
management
setup
stuff
that
I'm
trying
to
work
with
infrastructure
on,
we
should
be
able
to
connect
elab.com,
hopefully
soon
so
exciting
stuff
there.
A
In
the
meantime,
I'm
still
continuing
to
push
the
conversations
with
the
infrastructure
review
as
far
as
our
production
Readiness
I'm,
trying
to
follow
up
with
appsec
to
kind
of
see
if
they
have
everything
they
need
to
do,
conduct
full
review.
They
haven't
responded
since
I
said
that
I'd
be
happy
to
send
them
data,
and
things
like
that.
So
again,
just
continuing
to
big
figure
out
if
there's
any
blockage
or
anything
I
can
do
to
help
expedite
the
approval
process
for
us
to
start
to
onboard
external
customers.
B
I
can
go
speaking
of
external
customers
had
a
meeting
with
one.
Today
they
are
ready.
We
have
a
list
of
projects
that
we
can,
what
building
the
soonest.
We
would
do
that
would
be
next
week
and
once
we're
ready
we'll
let
them
know
that
we
flip
them
and
kind
of
talk
them
through
or
be
able
to
be
available
for
a
sink
call
to
talk
through
the
setup
but
shared
with
them.
B
All
of
the
sdks,
though
they're
going
to
start
looking
at
instrumentation,
so
are
they
setting
there
I
hanged
in
Channel,
and
my
only
issue
I
think
in
the
Milestone
is
around
brainstorming
for
customer
feedback
methods.
A
Gotcha
no
it'll
be
I,
mean
I,
guess
if
anything
and
I'll
add
it
to
the
idea
just
to
make
sure
if
we
do
end
up
having
calls
with
them
kind
of
you
know,
whatever
four
six
eight
weeks
after
they've
gotten
set
up
and
things
like
that,
it'd
be
good
to
see
if
we
can
get
engineering
involved
as
well
to
maybe
help
answer
the
questions
in
terms
of
like
what's
possible
and
what's
not,
if
there's
any
suggestions
or
feedback
that
we
can
kind
of
help
out
with
there.
A
I
did
also
forget
one
thing:
I
was
just
talking
to
bossy
about
it
this
morning,
sorry
to
jump
back
to
mine
with
all
of
the
event
pipeline
disruptions
we've
been
having
due
to
that
events
or
missing
columns,
and
things
like
that.
There
are
some
settings
that
we
should
be
able
to
tweak
in
Click
house
to
allow
it
to
skip
rows
or
pieces
of
data.
A
Erroring
on
I
have
to
validate
whether
that
solution
is
going
to
work
or
not,
because
this
is
more
about
importing
a
large
batch
of
data
through,
like
a
data
blob
and
allowing
to
skip
the
errands
rows
and
continuing
down
the
rest
of
them
and
defining
like
error
ratios
before
clickhouse
will
throw
an
exception.
A
I,
don't
know
how
that
applies
as
I
mentioned
before
how
in
cluster
clickhouse
Works
in
terms
of
it
pulling
the
events
from
Kafka,
whether
that
would
still
meet
that
would
still
respect
those
settings.
So
I
have
to
do
some
testing
to
make
sure
that
that
all
works,
but
assuming
it
does
well.
That
means
the
clickhouse
server
I
mean.
A
Obviously
we
still
want
to
validate
things
and
properly
categorize
things
in
the
bad
Topic
in
Kafka
and
deal
with
it
appropriately
there,
but
in
the
case
that
it
does
get
to
clickhouse
and
it
does
cause
issues,
click
that
should
hopefully
be
able
to
then
move
on
past
those
bad
events
and
we'll
have
a
little
bit
more
resiliency
there.
So
I'm
going
to
investigate
that
and
maybe
see
if
I
think
khalil's
jumping
back
in
a
couple
days
from
from
his
time
off.
So
maybe
he
can
look
into
that.
A
If
I
don't
get
to
it
cool,
then
Alan
I
know
you
already
kind
of
put
one
in
the
in
the
async
chat.
But
would
you
like
to
share
anything
as
far
as
what
you're
working
on
or,
if
you
have
any
blockers.
C
I,
don't
have
any
blockers
right
now,
I,
sorry,
the
my
daughter
started
back
to
school,
so
the
first
I
got
the
first
bug.
The
the
main
thing
I
wanted
to
bring
up
was
the
the
re
setup
issues
once
someone
sets
up
product
analytics
at
the
instance.
Level
then
goes
to
set
up
at
the
project
level
then
goes
back
to
the
incidence
level
that
doesn't
work
right
now.
We
want
to
fix
it.
Obviously,
my
main
question
is
and
you're
you're
more
familiar
with
the
production
infrastructure
setup
here.
C
C
So
I
wanted
to
get
an
answer
on
that?
That's
the
only
thing
that's
like
kind
of
blocking,
but
otherwise
I'm
just
kind
of
working
on
beta
bugs
so.
A
For
sure,
no
I
think
it's
possible
actually
I.
Think
one
of
the
issues
that
I
don't
analytics
instrumentation
is
working
on
right
now
is
actually
being
able
to
store
that
or
actually
keep
track
of
the
IDS,
because
one
of
the
concerns
I
had
was
what
I
mean
we're
generating
random
strings
as
the
application
IDs
or
that's
great.
But
at
some
point
you
know
you
know
this.
This
kicks
off
wildly
and
we
have
so
many
people
using
it.
A
There's
potentially
a
possibility
for
collisions
and
so
I
wanted
to
make
sure
that
you
know
we
can
keep
track
of
that,
so
that
we
can
avoid
that
problem.
A
And
then
to
that
to
your
point,
then
we
should
be
able
to
just
query
that
and
say
hey
for
this
mainspace
ID:
here's
the
application
ID
that
we're
using
that
would
support
further
future
use
cases
eventually,
if
for
some
reason
they
need
to
reset
it
or
rotate
it
or
anything
like
that.
So
I
think
they're
already
working
on
that
I'll
find
the
issue
and
and
paying
you
on
it.
But
if
that
simplifies
things
for
you,
then
I
I
don't
see
why
we
couldn't
kind
of
take
that
approach.
B
C
A
C
Oh,
the
other
thing
I
would
say
is
that
James
I'm
happy
to
work
on
that
graphql
rename
issue.
That
seems
like
a
pretty
straightforward
place
so.
A
Yeah
I
know
I
guess:
I
can
jump
to
that
real
quick
since
the
on
topic,
but
there's
a
Beyond
has
called
out
an
issue
for
renaming
our
current
graphql
resources
from
product
analytics,
something
more
generic,
given
that
we
have
optimized
onboarding
and
using
those
endpoints
as
well,
and
then
hopefully,
others
will
will
follow
suit
since
we're
behind
a
feature
flag
and
don't
have
to
respect
the
you
know,
versioning
roles
in
terms
of
what
we
can
break
in
terms
of
API
changes.
A
I
do
think
it's
potentially
something
we
want
to
prioritize
sooner
rather
than
later.
Of
course,
we're
still
trying
to
get
out
to
external
customers
and
this
and
that
but
James
I
wanted
to
get
your
thoughts
on
that
before
yeah.
B
I'm
happy
to
pull
this
one
into
Beta
release
and
even
16
4.,
and
then,
if
we
want
so
just
the
issue,
we
were
talking
about
before
the
project
level
and
the
instance
level
cluster.
All
of
that
kind
of
is
going
to
be
at
and
it's
a
double
cluster.
A
Yes,
they
shouldn't,
so
it's
any
time
yeah.
They
need
to
change
connected
clusters
right.
So
it's
not
just
bring
your
own.
A
If
we
eventually
offer
more
regions
or
I
know
private
a
far
way
out
but
like
anytime,
they
have
to
move
off
of
what
is
the
default
option
in
the
instance
level
and
I
know
that
will
look
at
more
region
capabilities,
probably
as
the
more
the
one
that's
more
most
likely
to
happen
in
the
near
or
the
medium
term
that
that's
where
things
can
potentially
go,
go
Haywire
as
well,
so
by
being
able
to
query
the
application,
ID
I
think
we'll
be
able
to
just
be
basically
be
able
to
keep
it
up
to
date,
based
on
whatever
cluster
you're
connected
to
at
whatever
level
yeah.
C
I
would
say
that
the
only
possible
way
I
could
see
somebody
getting
into
this
on.
Gitlab.Com
is:
if
we
roll
out
bring
your
own
cluster
for
them,
and
then
they
want
to
turn
that
off
then
they
will
be
in
this
problem,
but
that
obviously
that's
an
exception.
We
don't
want
to
that's
not
like
a
happy
path.
B
I
would
then
enroll
this
instead
of
in
the
beta
Epic
into
the
bring
your
own
cluster
epic
or
into
GA.
If
we
decide
that
part
of
GA
is
going
to
be
additional
regions
for
data
storage
right
when
it
would
be
a
use
case
that
we
need
to
support
for
that
functionality,
but
for
beta
I,
don't
think
we
do
things
we're
only
offering
the
managed
stack.
A
What
I
would
normally
be
doing
in
terms
of
like
resetting
up
that
so
again,
yeah
I
I
agree
like
it's
not
something
that
people
need
right
now
that
they're
not
they're
not
going
to
be
asking
for
it,
but
it
was,
and
I
only
encountered
it
because
I
was
switching
clusters
at
like
such
a
frequent
rate
than
probably
more
people.
Most
people
should
be,
but
I
think
it
will
help
make
our
lives
easier
if
we
can
kind
of
address
it
sooner
rather
than
later,
but
like
GA
like
that's,
that's
fine.
A
That
sounds
good,
and
that
also
gives,
because
that'll
give
us
time
to
check
in
with
analytics
instrumentation
group
as
well
and
see
where
they're
at
with
like
tracking
app
IDs
and
what
we
would
need
in
terms
of
updating
a
set
of
project
endpoint
to
like
not
produce.
The
database
already
exist
there
and
just
say:
there's
already
an
application
ID
here
here
you
go
so
there's
a
little
bit
of
buffer
in
terms
of
dependencies,
so
cool
all
right,
then
James.
You
have
a
next
couple
points.
We're.
B
Jumping
around
the
point
two
yeah,
the
product
analytics
product,
analytics
usage
quota,
NPC,
hoping
to
get
some
Engineers
eyeball
automatically
start
to
look
at,
especially
what
is
our
source,
the
truth
or
event
volume
and
how
we
can
then
that'll
help
us
map,
then
for
the
Fulfillment
side
of
things,
so
we're
going
to
get
that
over
into
helmer's
dot
or
Zora
or
wherever
it
needs
to
go
as
we
get
to
that
billing
component.
B
So
mostly
that
is
what
I'm
interested
in,
but
overall,
looking
at
the
existing
open
questions.
The
assumptions
that
we've
made
within
that
epic
description
and
calling
out
where
there
are
gaps
that
we
need
to
get
filled
in
before
we
can
go
ahead
and
move
forward
with
the
NPC.
A
Yeah
and
I
might
be.
This
is
probably
because
we
have
the
difference
between
the
usage
quota
and
the
billing
side
of
things.
I
know
we
had
a
little
bit
of
some
prior
investigation
from
Max
and
I.
Think
Khalil
as
well
and
I.
Don't
know
that
that's
transferred
back
to
this
issue,
so
really
I,
don't
know
if
it's
just
front
end
that
I
should
be
pinging
here.
It
should
just
be
everybody
but
I.
Think.
B
I
tagged
in
The
Proposal
the
link
to
that
discussion
yeah,
but
I
didn't
explicitly
call
out
hey.
This
was
already
discussed
here
so.
A
C
A
A
Yeah
I
mean
I
think
this
comes
out
of
our
previous
discussion,
James
about
like
moving
that
conversation
back
to
you
should
quota
because
that's
going
to
be
the
first
thing
we're
going
to
need
to
work
on
first
before
we
and
that
that
will
argument
to
your
point
outline
what
needs
to
be
done
when
we
work
with
fulfillment
and
get
getting
the
billing
side
of
things
sorted
but
like
we
still
need
to
show
Event,
Event
usage
and
things
like
that.
So
I
know
that
conversation
started
so
I.
Basically,
we
should
bring
that
back.
A
You
know
there's.
Actually
this
is
your
needs
eyes
from
both
front
and
backhand,
so
I'll,
just
I'll,
just
ping.
Everybody
on
that
one,
so
cool.
A
I
mean
that's,
that's
actually
a
good,
that's
a
good
point.
We
should
try
that
for
those
watching.
The
call
later
on
the
the
general
idea
is
that
we
need
some
way
to
query
all
the
clickhouse
databases
to
see
like
how
many
events
total
have
we
processed
effectively.
Anything
in
clickhouse
has
been
processed,
and
where
do
we
store
that
data?
A
Ideally
I,
think
one
of
the
core
principles
is
that
we
want
gitweb.com
or
GitHub
the
getlab
instance
to
be
the
source
of
Truth
for
that,
and
then
that
can
then
push
too
eventually
once
we
get
into
the
billing
NBC
the
Fulfillment
flow
of
reconciling
like
billing
and
things
like
that,
but
we
need
to
put
things
in
place
and
I
think
the
configurator
is
going
to
become
more
than
just
a
configurator
in
that
case,
where
it's
like
kind
of
a
cluster
Management
Service.
A
If
you
will
where
yes,
it
can
manage
the
clickhouse
databases
and
the
server
itself,
but
also
needs
to
do
some
reconciliation,
because,
ideally
you
know
if
we
can
get
deeper
into
it
once
we
actually
have
some
Engineers
looking
at
at
the
age
you,
but
we
want
to.
Maybe
do
some
batch
processing
or
do
some
aggregating
so
that
when
gitlab,
the
gitlab
instance
does
call
or
that
usage
data
that
we
can
very
efficiently
send
that
over.
But
we'll
we'll
have
some
good
discussions
out
of
that.
A
Just
just
wanted
to
recap
that
cool
back
to.
B
You,
my
next
point
is
read
only
but
I
will
say
it
out
loud
anyway.
Take
a
look
at
the
issue
for
brainstorming,
thumbs
up
ideas
that
you
like
or
add,
new
ones
and
I.
Think
we've
talked
through
Yan's
issue
around
the
current
graphql
stock
yeah.
A
A
No
I
think
we're
I
think
we're
good
as
far
as
everything
else
up
to
up
to
point
five
I
had
an
issue
at
a
point
just
to
kind
of
get
some
clarification
of
infrastructure.
A
I
think
you
know
accepted,
but
it's
not
here,
but
anyone
from
infrastructure
to
be
able
to
maybe
you're,
watching
the
recording,
just
to
kind
of
better
understand
like
what
needs
to
be
done.
I
know:
we've
had
this
infrastructure
support
issue.
I
know:
we've
got
something
like
we
want
to
get
the
everything
organized
and
set
up
so
that
we
can
deploy
the
staging
and
production
all
that
kind
of
stuff.
But
it's
not
clear
what
is
what
what
is
like
the
top?
What
do
we
actually
have
to?
A
What
are
the
work
items
that
will
help
us
get
to
that
definition
and
done
I've,
pinged,
Pierre
and
Mel's
on
that,
and
then
just
trying
to
get
some
clarification
on
that
just
in
terms
of
so
so
we
understand
what
has
to
be
done
and
we
can
also
make
those
calls
and
maybe
kind
of
work
with
figuring
out
if
certain
things
can
be
scoped
out
to
be
able
to
get
it
to
help
better,
facilitate
a
production
launch
sooner
rather
than
later,
so
we
can
onboard
some
external
customers.
A
So
if
you're
watching
the
recording
or
the
agenda
would
appreciate,
ideally
in
the
issue,
a
response
in
terms
of
just
like
what
what
actually
needs
to
get
done
and
how
I
could
potentially
or
anyone
else
could
potentially
help
with
that
as
well.
A
So
yeah,
that's
that's
that's
my
point
for
that
and
then
the
last
item
is
I
think
we
already
spoke
about
it.
Alan.
Is
there
anything
you
wanted
to
speak
to
about
that?
Or
did
we
cover
the
reset
up
issue.
A
Awesome
then
we
are
at
the
end
of
the
agenda:
I,
don't
really
have
anything
for
show
and
tell
but
I'm
hoping
by
next
week.
I
can
show
and
tell
getwell.com
data
being
coming
through
and
hopefully
not
blowing
up
our
new.
Our
second
excuse
me
cluster,
but
it'll
be
exciting
to
see
how
that
that
all
works
out,
but.
B
A
See
you
too
have
a
good
rest
of
your
days
and
weeks
and
see
in
the
next
one
see.