►
From YouTube: 2020 04 06 Database Team Weekly
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
A
A
So
if
you
hear
a
constant
tapping
sound,
it's
that
one
stupid
bird
that
just
smashes
its
face
up
against
the
crate
non-stop
all
day
long
so
it'll
be
kicked
out
of
the
office
soon
enough,
and
we
also
have
a
cat
that
just
sits
and
stares
that
they
create
it's
not
smart
enough
to
look
over.
The
crate
in
at
the
bird
is
just
staring
at
the
creek.
It's
making
noise
so.
A
Alright,
okay,
our
review
so
I
think
both
of
these
okay,
ours
were
created
before
Yanis
and
pat
we're
on
the
teams
is
probably
the
first
time
you've
seen
these
from
the
hit
the
hiring
target
and
create
an
onboarding
plan.
We'll
close
this
one
out
at
100%
easily
be
honest.
Already
has
some
ideas
on
how
to
update
the
template,
onboarding
template
for
hiring
more
folks.
A
This
was
written
I
think
before
we
had
the
timeline
for
the
dot-com
rollout,
so
we
might
have
to
change
the
wording,
a
bid
to
make
sure
that
it
properly
reflects
where
we
are,
because
we
can't
really
roll
this
out
on
comm
until
the
Postgres
11
upgrade
it's
done,
and
that
might
not
happen
before
this
quarter
is
out.
So
we'll
keep
working
on
that
one.
A
Any
questions
on
okay,
ours,
since
it's
fairly
new
for
folks
next
time
when
we
create
them
it'll,
be
it'll,
be
collaborative
since
you're
all
here
and
on
the
board.
Now
one
of
the
things
that
we'll
do
for
next
one
is
the
new
thing
that
we're
doing
across
the
company
for
iteration
training.
So
it's
a
fairly
simple
template.
A
For
now
the
e
you
can
read
it's
just
a
bunch
of
information
to
read
and
some
things
to
think
about
as
you're,
going
through
issues
and
planning
as
a
team
and
how
to
iterate
and
work
on
the
smallest
thing
possible.
So
that'll
be
one
of
our
ok
hours
for
an
exporter
and
then
we'll
work
on
the
other
one
or
two.
As
a
team.
A
B
C
A
B
C
Yeah
thanks
for
thanks
for
copying
over
Craig
yeah,
so
essentially,
as
some
of
you
may
have
seen,
and
if
not
thanks
for
it
for
linking
this
announcement,
we
have
decided
to
slow
down
hiring
for
across
the
company.
Do
the
essentially
unknown
economic
impact
to
our
business?
Well,
the
of
the
ongoing
pandemic,
and
so
while
we
are
hiring
for
some
highly
critical
roles
there.
That
means
that
we
are
putting
the
pause
button
on
roles
that
don't
meet
that
really
high
threshold
of
which
the
memory
and
database
p.m.
C
is
one
of
them,
and
so
for
now
I
will
be
the
p.m.
for
the
foreseeable
future,
and
hopefully,
once
we
get.
Clarity
wants
to
get
a
few
months
into
some
of
this
new
normal
for
a
while
and
get
a
sense
of
how
much.
In
fact,
there
is
to
our
business
as
well
as
how
long
this
might
go
on
for
then
we
can,
you
know,
get
a
better
handle
on
kind
of
how
long
this
will
happen,
but
that
we
had
hopefully
can
reopen
it
soon.
C
A
A
So
we're
frozen
on
team
members
too,
so
I
feel
really
lucky
that
we
got
patent
Yanis
in.
But
this
is
the
sides
and
the
team
for
now,
so
the
other
we
had
to
open
headcount
and
those
are
cause
for
the
time
being
cube
again.
That's
just
it
until
we
figure
out
what
the
overall
impact
is
going
to
be
and
we
can
reprioritize
as
we
come
out
of
this.
A
So
this
is
the
team,
but,
as
Andrea
said
in
our
team,
retro
we've
got
a
strong
team
or
making
some
headway
and
some
great
momentum
and
having
the
expanded
number
of
database
maintain
is
gonna
help
greatly
as
well
so
well,
hiring
would
have
been
great,
I
feel
really
good
about
team.
We
have
now
so
any.
A
B
A
Don't
think
it
has
an
impact
on
the
prioritization
and
charting
and
partitioning
our
number
to
a
number
one
priorities.
If
I
do
it
in
the
right
order,
it
might
certainly
have
an
effect
on
how
fast
we
can
deliver
right,
having
a
smaller
team
and
can
plan
on
having
a
bigger
team
throughout
the
year,
because
it's
going
to
take
a
while
to
actually
fully
implement
charting
but
I,
don't
think
it
impacts
priorities
and
most
Josh.
You
have
a
different
opinion.
No.
C
I
think
our
priorities
are
the
most
important
ones.
We
don't
say
I
wanna,
make
sure
that
it
continued
make
sure
we
can
scale
our
database.
So
there
are
other
things
that
we
should
be
doing,
that
are
more
important.
Let's
talk
about
them,
but
I
think
right
now
the
they
seem
to
still
be
the
most
important
things,
and
so
we'll
keep
an
executing.
We
are.
C
The
product
management
team
is
in
conjunction
with
the
engineer
managers
we'll
start
to
look
at
what
our
plans
were
for
the
year
kind
of,
as
as
we
go
and
then
determine
what,
if
anything,
we
should
be
cleared
to
the
business
on
that
will
likely
not
make
it
that
we
thought
would
and
that
way
we're
all
clear
on
our
we
end
up
and
in
some
cases
we
can
then
determine
whether
we
should
try
to
do
a
global
optimization,
because
there
are
some
teams.
Some
are
good.
C
There
was
like
ivory
shuffling,
we
open
up
a
bunch
of
roles,
people
moved
around
and
then
we
kind
of
think
I
pause
things
and
the
question
is,
is
what
are
the
impacts
to
the
business
of
the
current
structure
and
then
does
that
make
sense
from
a
global,
optimization
standpoint
I'm
like?
Are
we
still
allocated
intelligently
or
the
best
way
we
possibly
can
be?
So
you
know
I
think
we
don't
want
to
ask
anyone
to
do
a
thing
they
don't
want
to
do,
but
the
same
time.
C
We
also
make
sure
that
we
just
take
a
look
at
like
what
are
we
investing
in
and
is
at
the
right
investment
strategy
for
our
sort
of
current
situation,
so
I
don't
most
with
any
change
the
database
team.
It's
really
important
is
one
of
our
top
teams
for
the
year
of
enterprise
readiness,
so
I
don't
expect
any
changes.
The
database
team
and
the
conversation
there
would
be.
A
C
Thanks
so
just
toss
the
last
one,
if
you
have
concerns
about
our
ability
to
achieve
our
our
kind
of
near
term
goals
with
with
with
the
current
staff
and
then
once
let
me
know
once
again
patient
they
have.
But
moving
on
to
number
two
here,
we
did
do
an
MTS
survey.
So
if
you
weren't
from
that
NPS,
it's
essentially
a
way
to
serve
your
users.
You've
probably
filled
out
a
number
of
MPs
or
rates
in
the
past.
It
is
the
classic
rate.
C
Your
experience
on
a
one
to
ten
scale
with
I
believe
one
through
six,
the
detractors
seven
to
eight
being
neutral
and
nine
to
ten
being
I
forget
the
term,
but
the
positive
one
and
you
basically
subtract
a
negative
from
the
positives.
And
then
you
end
up
with
like
we
were
scores.
We
have
a
twenty-one
which
isn't
too
bad,
don't
compare
with
other
companies,
because
many
companies
mislead
on
their
public
excited
MPs
numbers.
So
don't
compare
them
what
we
completed
the
survey.
The
results
are
in
that
presentation.
C
C
Think
in
my
mind,
are
two
questions
are:
do
we
have
the
visibility?
We
need
to
know
what
the
performance
photons
are.
Should
we
invest
more
into
observation
and,
to
you
know,
are
we
still
focused
on
the
highest
ROI
things
to
drive
performance
forward?
So
I
think
those
are
the
two
questions
but
I
wanted
to
also
make
it
seem
aware
of
sort
of
what
the
data
was
coming
back
from
our
yesterday
as
well.
B
B
Is
you
know
we're
talking
about
partitioning
and
charting
and
that's
all
all
good
strategies,
and
this
is
truly
going
to
help
us,
but
also
have
different
problems
that
we
can't
all
fix
with
charting
partitioning
and-
and
some
of
these
are
very
basic-
and
sometimes
it's
just
really
about
using
a
feature
that
we
already
have,
and
you
said,
I'm
calm,
it's
the
first
example
in
the
list
and
other
examples
include
like
deciding
retention
strategies.
So
at
the
moment
we
keep
all
the
CII
data
forever.
B
Basically,
and
of
course,
that's
growing
over
time-
and
that
makes
our
site
slow
some
some
aspects,
I
just
wanted
to
point
it
out
and
I
think
we
I
think
we.
We
should
invest
more
into
finding
those
problems
and
and
figuring
out
if
you
know
how
we
can
solve
that,
if
that's
on
our
team
or,
if
maybe
I,
maybe
we
coordinate
with
another
team.
C
A
Okay,
yes,
we
can
you
talk
about
going
forward,
I
mean.
B
B
We
want
to
invest
everything
into
a
sharding
solution
that
we
may
pick
up
in
a
few
years
so
and
that
on
that,
but
I
think
the
growth
that
we've
seen
currently
on
kanakam,
particularly
is
we
can't
neglect
that
right
and
business
the
solution
that
we
have,
and
there
are
a
lot
of
problems
with
that
solution
and
we
can
also
try
to
fix
them.
What
we
perhaps
we
should
try
to
fix
them.
A
Some
of
those
can
be
parallel
efforts
right,
like
the
CI
builds,
we
can
find
a
good
retention
strategy
to
buy
us
time
now
and
I
mean
we
need
retention
strategies
regardless.
The
sharding
and
partitioning
know
we're
still
in
exploratory
stage,
but
I,
don't
know,
I
can't
think
of
an
alternative
right
now
and
how
much
time
it
would
take
to
find
an
alternative,
then
to
migrate
to
that.
It
seems
like
that
would
be
super
expensive,
so
I
think
we're
on
the
right
path.
I
think
you're
right.
A
Speaking
of
it's
a
good
segue
database
strategy,
so
there
was
an
issue
that
was
brought
up
was
linking
to
it.
What
is
our
database
strategy
and
we
need
to
document
it
and
I
get
shared
out
with
the
rest
of
the
engineering
org?
It
was
linked
to
this
issue,
so
the
origins
of
this
was
that
being
feature
was
being
considered
with
a
different
data
store
other
than
Postgres
sequel,
and
it
was
a
fleeting
conversation
about
hucking
about
a
no
sequel
database
or
some
other
storage
strategy,
and
it
was
quickly
changing
directions
that
yeah.
A
We
need
to
stick
with
Postgres
sequel
and
the
question
about
single
datastore.
What
does
that
mean
single
application?
So
the
request
was
that
we
publish
a
database
strategy
with
guidance
about
when
these
questions
or
decisions
come
up.
I
put
together
an
M
R.
It's
definitely
working
progress.
I've
asked
for
some
feedback
to
make
sure
we're
heading
in
the
right
direction.
A
A
A
We
have
the
working
group
coming
up
next
I'm
sure
they're
gonna
ask
so
we'll
talk
about
situs
and
where
we
are
and
with
licensing
I
know.
If
everybody's
caught
up
on
that
one,
but
enterprise
licensing
seems
like
a
non-starter.
If
they
go
with
the
perdón
hood
licensing,
it's
gonna
be
over
a
million
dollars
a
year.
They
do
have
some
wording
on
their
site
about
a
site
license
I'm,
not
sure
what
that
means,
I'm
happy
to
ask
them
about
it,
but
it
seems
like
enterprise
is
too
expensive.
A
D
No,
the
data
said
she
has
are
they're,
mostly
just
like
a
single
large
project
and
they're,
not
really
what
we
need,
which
would
be.
You
know,
really
a
bull
sight
kind
of
thing.
They
do
have
scripts
to
populate,
but
it
uses
the
API
which
we
don't
have
set
up
on
our
cluster
to
populate,
so
I.
Think
really,
you
know
I
feel
like
we
probably
need
a
better
strategy,
which
is
something
you
know.
Production
data
set
that
we
can
anonymize
kind
of
build
on
demand
that
we
can
load
I.
D
Just
don't
know
that,
there's
enough
real
data
out
there
that
we
can
realistically
test
and
not
just
for
situs
but
just
partitioning
like
anything
going
forward.
You
know
right
now
we're
kind
of
using
production
backups,
which
is
not
really
not
ideal.
We
probably
should
not
be
doing
that
yeah.
A
solution
for
that.
That
is,
you
know,
take
their
production
data
I
clean
it
in
some
way,
so
that
we
can
use
it
safely
and
be
able
to
reload
that
on
demand
for
testing.
B
That
would
help
us
in
so
many
other
places
as
well.
We
used
to
have
a
Nanami
zation
script,
maybe
two
years
ago,
for
staging
that's
why
we
created
staging
I
could
rid
of
all
the
relevant
data
that
is
obviously
broken.
It
hasn't
been
maintained
at
all
and
getting
that
back
would
really
help
us.
In
many
places,
infrastructure
is
looking
to
get
these
tests
datasets,
and
we
can't
do
that
because
it
has
all
the
all
the
customer
information
there's
a
high
risk
from
from
us
whenever
we
try
to
use
something
like
this.
You
know.
B
D
D
You
know
I
have
to
run
it
somewhere
else,
or
we
have
application
that
we
can
build
something
but
I
think
still
whatever
we're
looking
at
generating.
Is
it's
just
not
really
on
the
volume?
We'll?
Probably
we
want
to
test
using
situs
because
we're
only
using
a
database
that
this
tenth
the
size
of
production?
Is
that
really
a
meaningful
test
at
that
point,
especially
when
we're
really
testing
for
future
growth
of
10x
or
something
you
know
using
a
500
gig
database,
even
which
is
pretty
large,
is
still
not
really.
A
D
A
A
D
The
pre
populated
test
said
she
has
I,
don't
think
will
be
useful
because
they
would
look
like
their
single
project.
There's
one
forget
lab.
There's
one
for
Linux
I
really
helpful.
This
script
does
generate.
You
know
we
could
write
a
script
around
their
script.
To
generate
sort
of
multiple
groups
with
then
would
kind
of
more
realistically
model
the
data
that
is,
you
know
if
we're
looking
for
something
to
do
quickly.
That
is,
you
know,
probably
going
to
be
the
quickest
route.