►
From YouTube: 2020 07 07 Database Team Meeting
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
B
Yeah
I
was
I
was
realizing
that
I'm
most
of
the
time
I
was
pretty
late
to
putting
comments
on
the
retrospective
issue,
and
it
led
me
to
realize
that
I
think
harder
part.
That's
partly
because
I
feel
that
the
we're
not
making
a
lot
of
use
of
the
retrospective
anyways
for
our
group
so
far
and
I
just
thought
back
to
previous
experience
with
retrospectives,
and
we
typically
anthers
quite
targeted
to
the
group
and
then
sort
of
taking
the
results
and
putting
that
more.
B
C
B
D
C
A
We're
not
the
only
team
that
does
it.
Trust
me
right
memory
team
did
the
same
thing
last
couple
retro
so,
and
the
other
thing
is
they're.
The
retro
bot
didn't
work
this
month
for
some
reason
for
memory
and
database
team.
It
didn't
ping
everybody
when
the
miles
don't
close
to
remind
everyone
to
enter
info
into
the
retro.
That
being
said,
I
mean
the
the
13
to
retro
has
been
open
for
a
little
while
right
and
I
used
to
be
okay
at
entering
things
as
they
came
up,
I
haven't
done
it
for
a
while.
A
So
it's
it's
only
something
we
can
add
in
during
the
month
of
the
milestone
like
hey.
This
was
a
really
cool
thing
and
just
put
a
note
in
there,
but
I
like
the
idea
of
the
synchronous.
One
and
it'll
it'll
actually
help
me
to
transfer
themes
over
to
the
engineering
wide
one.
So
I'll
take
a
look
at
my
calendar
and
figure
out
how
the
how
the
time
that
so
that
it
happens
right
after
the
end
of
the
milestone
and
make
sure
that
we
use
office
hours
and
speaking
of
office
hours.
A
These
are
kind
of
turned
into
like
team
meetings,
instead
of
being
like
just
casual
most
of
the
time
which
I'm
getting
a
thumbs
up
from
be
honest,
at
least,
and
not
getting
people
shaking
their
head
from
everybody
else.
So
do
in
the
interest
of
reducing
the
number
of
meetings,
deal
want
to
get
rid
of
the
Monday
meeting
and
just
assume
that
we'll
do
team
meetings
during
these
office
hours
I.
D
B
D
B
A
D
D
A
No
didn't
anybody
watch
yesterday's
meeting
or
tender
okay,
so
you
have
an
idea
of
the
direction
I
passively
watched
it
when
I
came
back
in
the
afternoon.
I
didn't
completely
consume
it,
but
it
seemed
like
the
theme
of
that
was
like
the
initial
goals
are:
how
do
we
quickly
set
up
an
environment,
a
targeted,
large
environment,
I?
Think
the
minimum
boundary
was
a
thousand
users
right
using
an
entirely
kubernetes
and
helm
set
up.
A
I
think
was
the
initial
goal,
so
I
feel
free
to
attend,
but
I
don't
think
we
are
like
direct
stakeholders
in
this
and
agreed
be
more
observers.
I,
don't
think,
there's
gonna
be
a
lot
of
work
needed
from
us
until
we
are
specifically
called
out
and
hey.
We
need
these
changes
to
better
support.
The
goals
of
this
working
group,
so
I
have
gotten
entirely
out
I'm,
not
the
DRI
or
the
facilitator
anymore.
E
B
D
E
It
the
goal
is
to
be
able
to
replicate
multiplicative
comms
without
linearly
increasing
the
number
of
support
staff
that
we
need
to
run
them.
So
that
is
the
main
goal,
and
we
want
to
do
that
by
deploying
the
kubernetes
and
leverage
automation
through
an
operator
which
can
automate
more
on
the
tasks
of
you,
know,
replica,
rebalancing,
giddily
and
other
kind
of
operational
tasks
like
coming
run
by
head.
E
I'll
borrow
sounds
like
Congress
is
making
an
operator
for
approach
Chris,
but
I
have
not
heard
of
any
short-term
plan
to
use
set
operator
for
Postgres
I
I
know.
Zalando
also
has
one
like
the
I
think
it's
a
good
comic
shop,
pretty
pretty
big
and
they
have
a
pretty
famous
operator
but
I,
don't
think
there's
any
intent
to
have
us
be
like
bleeding
edge
and
the
biggest
line
of
things
you.
A
So
thanks
for
asking
that
Yanis
and
in
parallel
there
were
questions
about
since
we're
not
doing
charting.
What
are
we
doing
to
ensure
that
the
database
doesn't
just
explode
right?
We
still
have
that
rolling
12
month
window
and
the
things
I
kept
pointing
to
were
the
reduced
total
size
and
growth
of
gitlab
compost
grass
database.
So
we
have
that
epic
out
there
and
we
can
talk
to
Josh
now
about
enabling
the
CI
retention
strategy.
That's
something
I
was
supposed
to
follow
up
on
last
week
and
I
don't
have
undressed,
you
have
that
issue
available.
A
B
A
There
was
an
issue
or
a
conversation
about
the
fact
that
we
already
have
a
retention
strategy
available
for
CI
builds,
and
we
just
need
to
turn
it
on,
but
we
needed.
We
need
to
talk
to
product
about
it.
You
remember
that
thread
or
that
conversation
I
do.
A
A
B
Go
ahead
terms
of
things
we're
doing
so.
The
database
doesn't
explode,
I
think
what
we,
what
we
should
do
after
finishing
the
audit
events,
migration
helpers
for
petitioning
is
basically
coming
up
with
a
plan
and
suggestions
which
tables
and
or
features
will
benefit
from
partitioning.
So
there's
a
couple
of
big
tables
that
we
should
tackle.
Probably
some
have
issues
with
the
primary
key,
so
that
might
be
motivation
tackling
them.
So
maybe
we
can
compile
a
similar,
a
bigger
list
of
things
that
we
want
to
motivate
for,
for
that
and
again
I
think
it's.
A
B
A
Okay,
yeah
so
I
mean
that'll,
be
our
roadmap
for
the
next
several
months
right.
What?
What
can
we
do
to
continue
to
fine-tune
our
database
so
that
we
are
managing
growth
and
we
don't
hit
another
wall?
There
was
a
good
analysis
on
where
we
can
vertically
scale
in
one
of
the
issues
and
the
impact
of
switching
different
types
of
machines
GCP
and
whether
we
can't
there's
well,
we
can
still
vertically
scale.
There
are
some
region
issues
where
some
of
these
machines
are
not
available,
so
it
it
gets
complicated
pretty
quickly.
A
If
we
do
need
to
get
more,
cpu
I
think
is
was
the
main
issue.
I
need
to
find
out
analysis.
We
can
link
back
here,
but
we
have
room,
it's
not
simple,
so
we
should
continue
to
figure
out
where
we
can
curtail
the
growth
of
the
database
and
maybe
even
cut
it
back
in
some
places,
and
also
how
do
we
enable
other
team
to
start
implementing
those
partitioning
strategies,
though.
B
B
Ok,
access
that
the
production
Chrome,
so
it
might
be
tempting
to
sort
of
well
I
can
I
can
use
my
development
environment
and
hook
that
up
to
the
production
database.
It's
actually
very
easy
to
do
that
and
we
might
want
to
make
a
call
on
whether
or
not
that
is
acceptable.
Or
what
can
you
do
with
that
port
forwarding
solution
and
what
is
something
that
you
shouldn't
be
doing?
Some
concerns
kasnia
raised
was
well
you're,
booting
up
the
development
environment
on
production
data.
B
Does
that
send
any
emails
or
where
does
where
does
the
production
data
land?
What
do
we
do?
What
do
we
need
to
do
to
sort
of
make
safe
environment
because
it's
ultimately
I
think
it's
beneficial
having
that
ability,
because
you
can
test
migrations
easily
and
it's
really
convenient.
You
just
use
your
local
development
environment.
You
could
up
to
the
production
clone
and
well.
You
run
migrations
on
top
of
that,
that's
great,
but
at
the
same
time
it
has
some
security
implications.
B
B
A
And
in
general,
I
need
to
send
out
a
request
for
feedback
on
how
much
value
this
is
given
folks
that
actually
are
using
it
right
now,
because
we
are
what
are
we
a
month
and
a
half
into
the
evaluation,
and
it
seems
like
some
people
haven't
even
been
able
to
hook
up
to
it.
Yet
so
there's
some
some
delays.
I
know,
there's
some
benefits
to
it,
but
getting
periodic
feedback
was
one
of
the
goals
of
this.
So
I
will
reach
out
to
folks
and
see
what
they've
what
they
think.
B
B
A
There
was
some
feedback
that
folks
felt
like
the
requirements
for
doing
database
review,
we're
changing
and
that
we
need
to
better
communicate
and
also
help
to
explain
the
explains
and
I
think
that's
where
a
session
sharing
can
come
in
and
helping
folks
out
that
don't
necessarily
understand
how
the
plan
is
explained
and
we
need
to
think
about
how
we
can
help
improve
communication
around
of
the
review
process,
especially
if
something
changes
then
I
don't
have
any
examples
on
what
was
change.
It's
a
fairly
sparse
issue,
but
it
was
brought
up
during
the
last
retro.
A
D
A
D
D
This
portion
that
we
have
this
page,
which
is
the
top
page,
that
the
entry
page
for
for
anything
regarding
the
database
guides
and
I'm
going
to
move
up
more
things
there,
and
there
is
something
left
pane
notes
of
the
top
things
and
also
a
very
quick
intro
on
the
top.
At
least
this
is
an
entry
point
and
we
are
going
to
iterate
on
that
again,
whole
thing's,
a.
C
C
D
D
B
A
It's
both
right.
There
were.
There
was
some
confusion,
lack
of
knowledge
about
where
the
existing
documentation
was
and
also
how
do
we
better
broadcast
changes
when
they
come
up
and
how
do
we
help
with
explaining
it's
a
pretty
broad
question
and
I
think
it
started
with
the
push
to
get
more
database
reviewers,
so
helping
spread
the
knowledge
helping
to
understand
helping
to
support
those
that
are
jumping
into
being
a
database
reviewers.
So
all
of
this
helps-
and
all
of
this
applies
and
I'm
going
to
talk
about
it
in
the
retro
today.
A
D
B
B
A
Point
for
database
office
hours
thanks
for
the
suggestion,
hungriest,
so
every
long
list
of
things
we
talked
about
about
helping
folks,
understand,
database,
your
news
and
where
to
get
information.
So
that's
awesome
and
I
will
actually
and
will
add
a
lot
of
that
back
into
the
issue
to
birth,
folks
that
are
watching
that
tracking
issue.
So
thanks
to.
C
Yeah,
the
post
migration
helper,
so
testing
that
they're
sort
of
highlighted
that
the
query
doesn't
perform
that
loan
so
created
a
new
mr
made.
Some
changes
to
the
background
migration.
Well
that
works.
So
we
can
track
that
a
little
better
which
then,
hopefully
that
will
prevent
us
from
having
to
do
according
the
table
in
the
same
way,
so
that
mr
I'm
gonna
try
to
get
so.
It
kind
of
pushed
out
the
timeline
for
that
a
little
bit.
A
B
B
B
D
Reports
of
the
debates
and
we
run
the
partition
migration.
You
know
the
tube
and
undress
and
I
can
say
that
everything
seems
to
work
with
that.
We
had
some
issues
with
the
background
jobs
not
running,
but
that
was
a
setup
issue
of
our
server.
The
good
news
is
that
the
pattern
of
migration
run
for
240
million
events.
They
Bassett
the
partitions
were
created,
the
triggers
work.
All
the
data
point
to
the
proper
positions
and
I
have
run
some
experts,
and
we
can
see
that
the.
D
Data
set
at
least
everything
without
any
other
musicians,
all
the
the
query
times
for
a
status
update
filters
from
them
go
down
significantly,
some
who
is
go
up,
but
at
least
everything
is
stable.
So
we
see
stable
three
times
between
at
around
500
M
min
seconds
up
to
two
seconds
when
most
of
those
queries
were
timing
out
and
our
time
right
now
in
production.
So
I
think
that
this
is
a
success,
and
this
shows
that
without
doing
anything
else,
the
partition
strategy
works
and
the
queries.
A
E
We
we
paint
him
and
one
of
the
issues
already
I
believe
and
the
he
was
like.
The
response
was
kind
of
like
yeah
sure
we'll
do
that
and
so
I
think.
There's
kind
of
my
people
was
waiting
for
us
to.
Let
them
know
one
nation
to
do
that,
but
somewhere
there
is
a
discussion
in
one
of
those
issues
around
the
date
picker
and
that
that
there
bored
with
doing
it
I
will.
A
A
B
A
C
C
B
We
think
it's
feasible
to
submit
the
partitioning
migration
this
week
and
what
do
we
need
fill
out?
The
partition
management
is
one
of
those
things,
the
background
migration
tracking
that
better
and
we
should
be.
We
should
be
confident
that
we're
not
breaking
things
anything
else
missing
in
us
again.