►
From YouTube: GitLab 16.4 Kickoff - Database
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
So
in
general,
our
team
continues
to
operate
on
initiatives
that
impact
the
availability
and
reliability
of
gitlab.com
I.
Think
one
of
the
ways
this
manifests
is
that
our
work
is
often
maintenance
related
and
very
fun
foundational
and
fundamental.
So
we
want
to
take
our
time
making
sure
we
get
it
right,
but
it
often
takes
us
multiple
Milestones,
to
make
some
big
progress,
cumulatively.
So
a
lot
of
the
themes
that
you're
seeing
in
our
top
priorities
for
16.4
are
repeated
priorities
for
the
past
several
months.
A
In
16.4
we
have
our
table
size
reduction
effort.
This
falls
in
line
with
our
previous
work
in
partitioning
the
main
idea
being
you
know,
we've
got
big
databases,
big
tables,
and
this
makes
data
in
and
out
very
challenging,
so
reducing
table
sizes
will
bring
us
a
whole
host
of
sustainability
and
reliability
benefits.
A
B
A
B
I,
it's
actually
sasa's
complete
and
right
we're
sorry.
This
didn't
get
updated,
but
SAS
is
actually
complete
and
we
are
currently
in
the
process
of
wrapping
up
self-managed.
So.
B
Yeah
there
there
were
some
bugs
with
the
initial
self-managed
migrations
of
a
couple
of
them
and
so
we're
working
through
making
sure
those
customers
are
in
good
States
and
then,
but
for
ones
where
we
had
not
yet
started.
Aka
notes
and
events
were
we're
plugging
right
along
getting
those
out
so
cross
said
it
looked
like
there
had
been
good
progress
when
I
chat
with
him
about
that
earlier.
A
Nice
cool
and
then
the
last
big
theme
Here
is
automated
database
testing.
I
know
we've
been
looking
at
this
for
some
time
and
and
I
believe
we've,
you
know
done
a
whole
bunch
of
stuff
with
AI
which
sort
of
competed
with
this,
but
at
the
end
of
the
day
I
know
we
started
looking
at
some
queries
and
Alex
did
you
want
to
touch
on
the
bug
that
we
needed
to
fix
that
came
up
recently
on
this
one.
B
Yeah,
so
there's
sorry
I.
The
progress
update
here
is
is
a
little
dated
and
I
I
meant
to
get
that
updated.
But
what
we're
trying
to
sorry?
We
had
an
issue
with
the
initial
there.
B
There
was
a
bug
that
caused
the
analysis
not
to
run
by
default
on
default
branches,
so
we
reintroduced
default,
Branch
running
for
generating
the
auto,
explain,
queries
and
then
that
got
reverted
because
it
cost
another
unforeseen
side
effect
and
so
they're
working
on
now,
reverting
that
back
again
and
we're
hoping
to
have
captured
artifacts
both
from
branches
and
from
default.
Sorry,
merge,
requests
and
from
default
branches
in
order
to
start
doing
comparisons.
B
This
Milestone
and
extracting
not
just
all
of
the
queries
in
bulk,
but
actually
extracting
what
looks
like
new
queries
so
that
we
can
keep
track
of
what's
new
and
and
alert
on
that
which
would
be
very
snazzy,
yeah
I.
A
Think
definitely
once
we
have
some
some
tangible
queries
that
we
have
to
compare
against.
I
think
that
that
makes
our
database
review
job
a
lot
more
robust
and
I.
Think
that
will
help
a
lot
in
terms
of
just
making
sure
we
have
a
resilient
system
as
we
build
towards
a
bigger
and
bigger
amount
of
information
that
we
have
all
over.
B
B
It's
it's
difficult
to
parallelize
that
work,
but
once
we've
got
the
the
added
the
list
of
added
queries
we'll
be
able
to
Branch
out
a
lot
more
in
terms
of
analysis,
and
we
should
be
able
to
pull
more
people
in
building
various
analyzers
which
which
should
accelerate
this
work
pretty
substantially.
A
Think,
there's
there's
a
few
remaining
items
to
close
out,
but
I
think
the
plan
here
is
to
document
what
we've
done,
wrap
this
up
and
then
broadly
socialize
this
perhaps
with
the
engineering
group,
so
that
other
people
can
start
taking
advantage
of
what
we've
built
here
to
improve
background
processing
across
our
entire
fleet
cool
and
then.
Lastly,
this
is
this
is
one
of
those
like.
A
We
know
that
upgrading
is
very
important
to
gitlab,
but
you
know
making
sure
our
migrations
run
in
a
good
ordering
would
help
make
sure
upgrades,
don't
accidentally
break
for
a
lot
of
customers.
So
I
think
this
is
something
that
we
did
a
proof
of
content,
Spike
and
and
now
we're
just
really
just
digging
into
some
of
the
work
here
and
I
won't
go
into
the
Super
specific
details
because
there's
a
bunch
of
bits
here
but
Alex
is
there
anything
you
wanted
to
highlight
from
here
as
well.