►
From YouTube: 2020 12 25 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
Welcome
to
this
version
of
the
database
team
meeting,
it
is
december
15th
and
there
are
a
few
non-verbalized
topics
in
there
fairly
full
agenda,
so
we'll
jump
right
into
it.
From
last
week
we
were
going
to
spin
off
asynchronously
and
have
everybody
score
topics
in
the
rice
framework
in
the
spreadsheet
and
people
are
supposed
to
make
their
own
copies
and
do
it?
Did
anybody
actually
get
to
that?
Last
week?.
A
Yep,
I
don't
think
it's
super
critical
for
today.
So
we'll
just
add
that
to
the
list
of
things
to
do,
we
have
plenty
for
the
remainder
of
this
milestone.
So
we'll
just
put
the
follow-up
here.
A
All
right,
we
have
a
bunch
of
new
and
incoming
issues.
We
need
to
triage.
A
A
It
looked
like
there
was
just
a
missing
step
within
a
mile
migration
somewhere
because
there
are
there's
a
couple
fixes
noted
in
here
where
they
just
dropped
either
a
view
or
a
table.
I
can't
remember
the
specific
details,
but
there's
there's
a
fix
noted
in
the
upgrade.
A
A
I
was
hoping
jose
would
join
us
today,
but
he's
not
here
yet
sometimes
he
joins
later.
Then
we
got
a
bunch
from
infradev
and
they're
all
I
think
they're
mostly
open
here,
so
we
start
from
the
top.
Most
of
them
look
like
this,
where
it
selects
something
from
a
table
with
somewhere
clause,
and
I
tried
to
go
through
as
many
as
they
could
and
find
the
this
is
the
marginalia
data
that
comes
along
with
it,
correct
to
make
sure
it
was
included
on
the
individual
issues.
A
Otherwise,
you
have
to
go
back
through
the
list,
find
it
and
get
an
example
from
it.
So
these
are
all
marked
as
sub
one.
I'm
going
to
join
the
intro
call
on
thursday
and
talk
about
how
we
can
better
categorize
these.
These
don't
seem
like
severity
ones
to
me,
but
maybe
infradev
has
a
different
severity
scale
than
we
do.
Yeah.
D
D
A
D
Spoke
about
that
earlier
with
janice
and
I
am
operating
off
the
handbook
definitions
of
what
these
labels
mean,
and
I
I
mean
you
know
I
I
know
the
least
about
these
things
in
the
room
here,
but
at
least
to
my
understanding.
Those
issues
are
you
know
they?
Don't
they
don't
qualify
as
a
as
a
p1
s,
one
according
to
those
definitions,
so
I
think
we
need
to
align
on
how
important
this
really
is,
because
I
I
jokingly
said
to
yanis.
D
If
all
of
those
are
really
p1s
ones,
we
can
re-record
our
kickoff
video
for
the
next
release
and
just
do
this,
and
that
may
be
entirely
fine
right
if
that
is
actually
the
case,
but
it
doesn't
feel
like
it
right
now.
So
clarification,
I
think,
would
be
would
be
good.
A
Yeah-
and
I
started
this
discussion
on
the
same
topic
here-
haven't
gotten
a
lot
of
feedback
yet,
and
it's
not
to
minimize
the
fact
that
these
are
concerning
for
infradeb
we're
seeing
some
spikes
and
some
growth
that
needs
to
be
mitigated,
but
again,
I
think
the
severity
in
the
audience
that
are
being
hit.
A
We
need
to
have
a
conversation
about
it.
So
comments
on
that
thread
would
be
great.
I
have
a
note
to
follow
up
on
later
on
today,
so
my
impression
from
these
is
that
we
need
to
figure
out
where
it's
being
called
from,
and
these
marginalia
tags
will
help
right
and
then
maybe
send
it
back
to
the
team
that
owns
that
area
and
see
if
there's
something
they
can
do
to
reduce
the
frequency
of
calls
or
excuse
me,
look
into
caching
right.
It's
not
looking
through
mostly
select
statements.
It
didn't
seem
like
there's
anything
immediately.
A
The
database
group
could
do
but
feel
free
to
correct
me
if
I'm
wrong
or
we
can
go
through
these
one
by
one
and
triage.
B
I'm
a
bit
on
the
fence
with
that
one,
for
one
reason,
is
that
the
I
think
the
the
observation
was
made
before
we
deployed
the
query.
Caching
change
so
only
after
we
made
that
sort
of
measurements
of
the
query
caching
kicked
and
then,
presumably
that
helps
to
reduce
exactly
those
kind
of
queries
where
you
have
like
n,
plus
one
queries
and
those
get
cached
now
from
from
rails.
B
That
would
one
type
of
problem
that
we
can
fix
here
perhaps
is
looking
into
these
and
sort
of
figure
out
if
those
are
actually
n
plus
one
type
of
problems,
because
those
are
the
ones
that
we
can
fix.
But
on
the
other
hand,
if
they
are
not,
then
this
is
like
it's
a
primary
key
lookup
right,
so
these
happen
everywhere
and
sort
of
suggesting
that
we
start
to
catch
them.
I
think
this
is
rather
an
open,
a
more
broader
discussion.
B
We
don't
have
any
of
these
caching
mechanisms
at
the
moment,
aside
from
the
on
the
rails,
caching,
so
it's
not
that
each
of
those
issues
can
be
like
fixed
by.
Oh,
let's
add
this
to
the
cache,
because
we
don't
have
this
cache
right.
So
there
is
part
of
that
is
rather
rather
an
open-ended.
Maybe
architectural
question:
isn't
it.
A
B
My
instinctively,
I
would
say,
reach
recheck
that
this
is
actually
the
case
that
we're
still
suffering
from
those
because
there
have
been
changes
since,
since
we
made
those
observations-
and
I
I
on
a
coffee
chat
today,
I
heard
that
sort
of
a
lot
of
these
issues
that
were
created
during
during
the
last
week.
Those
were
observations,
and
some
of
them
had
like
big
question
marks
on
them.
B
A
So
what
what
are
the
specific
steps
on
checking
it?
Do
we
just
watch
the
reports
that
they
generate
on
a
daily
basis,
because
a
lot
of
these
come
out?
If
I'm
looking
at
this
one
here,
these
are
daily
reports.
Do
we
just
need
to
compare
the
tables
here
to
see
if
they've
fallen
down
in
priority
like
this
one
was
called
out
for
projects.
A
We
have
six
of
them
right.
So
do
we
check
the
daily
reports
to
see
if
the
six
that
were
brought
up
to
us
have
fallen
down
and
then,
if,
if
some
remain
like,
if
the
project's
one
remains
then
figure
out
what
the
next
steps
are,
whether
it's
caching
or
some
other
optimization
just
trying
to
figure
out
what
the
definitive
steps
are
because
yeah
we
need
to
watch
to
see
if
the
what
was
implemented
by
the
memory
team,
the
rails,
caching
actually
helps
in
figuring
out
which
ones
it
did
not
improve.
B
So
those
are
all
quite
specific
queries,
so
maybe
we
can
see
their
frequency
over
time
and
see
if
something
changed
in
the
last
last
couple
days.
That's
one
side,
the
other
side
is
sort
of
re-evaluating.
If
that
is
actually
such
a
high
severity
problem
for
us,
because
just
from
from
looking
at
those,
it
does
very
single
record
lookups.
So
might
not
be
that.
B
I
don't
know
either
what
to
do
with
them.
I
think
it's
a
little.
It's
it's
not
totally
clear
what
we
would
expect
to
do
on
those
and
if
they
are
still
relevant.
It's
also
clear
to
me.
C
A
We
have
on
thursday
there
is
the
infrastructure
weekly,
I
think.
Well,
I'm
gonna
join
for
sure
it
would
be
good,
maybe
andreas
if
you
could
join
on
that
one
too
and
kind
of
ask
them
some
questions
about
what
they
are
looking
for
for
updates
on
these,
and
can
you
give
the
database
team
perspective
on
it?
It's
at,
like
5
p.m.
Utc.
A
C
A
Okay
yeah,
maybe
we
can
get
some
clarification
yeah.
Some
changes
to
the
process
too,
because
again
the
sev
one
is
a
bit
concerning,
but
we
can
talk
about
that
on
the
discussion
issue
as
well.
D
Then
I
would
like
to
understand
very
like
very
clearly
why
that
is
the
case
and
also
if
everything
is
urgent
and
important,
nothing
is
it's
like.
There
may
be
some
preference
here
as
to
what
gets
to
happen
first
right,
and
maybe
some
things
are
not
as
bad
as
others.
I
don't
know.
I
think
it
would
be
great
to
clarify
yep.
A
Okay,
and
actually
that's
just
gonna-
talk
to
jose
about
that
very
topic
today,
but
he's
not
here
so
we'll
move
on
to
the
next
one.
B
And
that
is
is
in
a
similar
direction.
I'm
just
chatting
today
about
the
intradeath
board,
so
there
is
a
board
for
that
triaging
process
and
we
were
wondering
like
how
can
we
best
contribute
to
that
process?
I
think
looking
back
at
quite
a
few
issues
that
were
created
last
week.
What
is
I
think
that
those
are
all
very
good
observations
like
from
from
an
operational
perspective,
we
were
seeing
like
this
queries.
B
Bad
some
even
had
examples,
specific
query
examples
with
the
plan
and
marginalia
comments,
but
basically
they
lack
the
application
context.
So-
and
perhaps
this
is
something
that
we
can,
we
can
contribute
to.
That
process
is
basically
helping
figure
out
like
what
is
the
functionality
that
is
causing
these
queries.
B
Perhaps
what
is
the
backend
team
that
owns
that?
Why
is
that
functionality
working
in
that
particular
way?
I
think
that
is
for
us
is
a
lot
easier
than
when
you
sit
on
the
rather
operational
side,
and
I
wonder
like
how
can
we?
How
can
we
do
that
in
a
good
way?
B
One
way
that
was
pointed
out
was
basically
looking
at
the
intro
dev
board
and
looking
at
the
open
issues
in
that
lane
and
sort
of
helping
challenge
their
severity
and
add
those
details
yeah
as
they
are
needed,
but
also,
on
the
other
hand,
if,
if
we
need
quicker
input,
then
perhaps
we
can
also
get
pinged
on
those.
It
can
be
both
ways.
A
Yeah
I
used
to
go
to
the
infrastructure
weekly
meeting,
where
we
would
review
that
board
triage
things
and
get
them
assigned
out
to
the
appropriate
groups
on
a
weekly
basis,
or
maybe
it
was
bi-weekly
at
the
time
that,
as
the
issues
fell
off
the
need
for
the
frequent
triage
was
diminished.
So
I
stopped
going
so
I'm
going
to
start
going
again
on
a
weekly
basis
to
make
sure
these
things
get
caught.
It's
that'll
work
for
me.
A
I
know
we're
supposed
to
focus
mostly
on
asynchronous,
but
having
a
synchronous
meeting
where
we
can
actually
hear
the
urgency
and
maybe
have
a
real-time
editing
of
the
priorities
and
severities
works
well
for
me.
So
I'll
start
going
on
thursdays
it's
every
week
at
the
same
time,
so
I
can
join
those.
A
A
Yeah
for
sure,
because
again,
if
every
database
problem
has
to
come
through
us,
it
doesn't
scale
very
efficiently,
especially
with
the
team
size
that
we
have
now
and
the
back
end
teams.
Perhaps
they
just
need
some
guidance
to
say
hey.
If
you
did
this
just
like
the
pat
and
I
were
talking
about
the
api
improvement
where
they
implemented
key
set
pagination
last
night,
they
got
a
was
a
20x
improvement
on.
E
A
A
All
right,
let's
switch
over
to
the
board,
so
they're
going
to
wrap
up
the
deployment
steps
today
for
release
package,
but
milestone
I'm
assuming
all
the
moving
issues
and
closing
them
up.
The
milestone
will
still
happen
on
17th,
but
we
need
to
go
through
13
7
to
see
if
there's
anything
lingering
in
there
that
needs
to
be
carried
over
so
just
start
from
the
top
be
honest,
you're
up.
First.
C
F
So
hopefully
this
is.
C
We
have
unbanned
that
from
running
the
migration
for
events,
so
this
is
only
the
migration
developers,
so
the
new
plan
is
if
we
can
merge
that
at
some
point
and
then
we
will
around
the
events
party
at
the
events,
migration
next
and
thirty-four
days,
because.
E
C
Be
too
dangerous
to
run
it?
You
know
two
days
before
cut
off.
C
Yeah
address
we
take
it,
hopefully
yeah,
but
it's
better.
A
To
be
safe
than
sorry
so
yeah
all
right,
andres
grafana,
annotations.
B
I've
been
working
on
manual
re-indexing
today,
so
there
is
a
request
to
repack
a
few
indexes
and
I
suggested
to
use
the
application
feature
for
that,
and
that
needs
some
manual
oversight,
and
today
I
managed
only
to
do
one
third
of
the
indexes.
So
that's
gonna
happen
tomorrow,
again,
perhaps
even
on
a
different
day,
so
yeah
that's
rather,
I
think
the
annotations
is
going
to
carry
over
all
right.
I'm.
A
Just
moved
over
now,
giannis.
C
Yeah
this
one
will
move
over
because
there
are
a
lot
of
indexes
there,
but
I'm
I'm!
Oh
okay.
I
was
keeping
that
there
it's
my
one
thing
that
yeah
to
remind
you
something
more
okay,.
E
Yeah
that
one,
I
think
it
should
be
good
for
this
milestone.
We
were
blocked
on
it.
For
a
few
days.
We
were
waiting
for
input
from
the
verified
team,
but
we
went
forward
with
it
anyway,
so
I
think
it'll
merge
soon.
Hopefully.
D
A
Right
this
one
doesn't
have
anybody
assigned
to
it.
E
Yeah
that
one
will
carry
over
that
one
they're
waiting
for
a
response
from
me,
but
I
have
to
reorganize
some
files
to
make
it
easier
for
them
to
add
the
changes
to
ci
to
catch
there's
specific
files.
We
want
to
trigger
the
build
for
or
trigger
the
test
run
against
pg-12,
and
I
have
to
reorganize
this
a
little
bit
to
make
it
easier
for
them.
E
A
This
one's
been
lingering
for
a
while,
I
assume
it's
going
to
carry
over.
I
don't
think
anybody's
worked
on
this
one.
Yet
the
database
connect
timeout.
A
A
B
And
then
this
one
yeah
that
that
should
close
today
on
the
fourth
try
or
something
like
that,
but
I'm
hopeful
that
it
works.
So
it's
a
reformation
right
and
that.
E
B
A
A
A
Thursday,
all
right,
let
me
talk
about
here,
the
logos
database,
so
this
is
awesome.
I
like
this
one,
I
don't
know.
Is
there
any
preference
between
the
elephant
and
the
simplified
database
logo.
C
D
A
Don't
talk
to
ben
kochi
because
he
wants
to
look
into
cockroach,
so
yeah,
but
good
good
call
on
future
proofing.
I
know.
F
A
A
D
So
I
have
to
pull
over
and
it's
quite
nice
right,
so
it's
the
same
site.
So
my
my
review
is,
it's
pretty.
I
think
it's
a
nice
to
be.
Actually
the
quality
is
pretty
good.
Zip
up
is
probably
also
nice.
I
personally
prefer
those,
but
I
didn't
even
know
we
could
order
them.
A
A
So
it's
it's
pretty
much
choose
your
own
adventure
here
as
long
as
we're
all
using
the
same
logos,
if
you
prefer
zip
ups,
feel
free
to
go
that
route
and
share
your
design.
Here.
I'm
sorry
not
here
in
the
issue,
obviously,
and
yeah
one
question
fabian,
I
haven't
gotten
an
answer
to
how's
the
sizing.
Is
it
pretty
true
or
do
they
size,
big
or
small?.
D
I
so
I
am
one
meter
89,
so
relatively
tall
and
I
have
an
m
and
I
think
it's
still
relatively
big,
so
I
I
would
say,
like
it's
reddit,
it
is
true,
but
if
I
had
an
xl
hoodie,
I
would
wear
a
tent
right
and
I
think
this
year
already
feels
like
if
I
just
like
get
up
it's
like.
It's
definitely
not
like
super
tight
for
an
m
right.
D
A
Thanks
for
the
the
modeling
session
and
the
visual
appreciate.
A
That's
awesome
all
right,
andreas!
You
have
the
next
option.
B
Oh
yeah,
nikolai
and
I
were
chatting-
and
I
think,
he's
interested
in
a
lot
of
the
topics
that
we
work
on
like
partitioning
or
the
integer
overflow
topic
is
like
a
while
back
he's,
basically
written
a
lot
about
that
and
how
to
accomplish
it
in
the
old
version
of
chris
and
stuff,
like
that
and
he's
he's
also
has
an
update
on
migration
testing
so
what
they
they
have
as
a
feature.
B
Perhaps
to
help
us
with
that,
and
I
was
wondering
if
we
should
invite
him
on
thursday
to
our
call
and
we
all
wanted
to
sort
of
have
a
group
chat
or
if
you
should
find
another
time
for
that.
Do
we
have
something
specific
for
thursday
in
mind
for
that
meeting.
B
F
No
nothing
I
need.
I
have
got
some
reasonably
interesting
stuff
going
on
in
the
database
space,
though
so
we've
got
a
ticket
relating
to
dead
locks,
which
I'm
trying
to
get
my
head
around.
What
exactly
is
going
on
so
that
they're
generating
deadlocks
using
ci,
which
is
a
neat
trick.
F
And
another
one
relating
potentially
to
the
edb
flavor
of
postgres
misbehaving,
there's
a
recent
post
in
the
database
channel
about
that.
So
we'll
see
what
the
customer
comes
back
with.
It's
not
boring
in
database
land.
F
There
is
yeah,
let
me
stick
it
in
the
agenda.
I'm
I'm
just
working
my
way
through
now
just
to
try
and
get
my
head
around
how
one
might
go
about
reproducing
it.
That
seems
like
the
obvious
approach.
F
I
think
that's
the
right
one,
so
I'm
quite
happy
to
do
that
heavily
lifting
there
to
try
and
unpick
the
ticket,
because,
obviously
I
can
go
back
and
get
clarification
if
needed.
So
if
you
spot
anything
interesting,
then
shout
by
all
means.
F
And
that's
the
one
wrong.
Sorry,
that's
the
long
lock
as
well
one's
one
one
one
knot
would
be.
I
guess,
on
the
deployments
table
and
the
other
one,
the
ci
builds
table
and
there's
two
jobs,
locking
the
same.
F
D
Portion
just
one
quick
question
for
next
week
is:
are
people
still
around
for
before
christmas?