►
From YouTube: GitLab 13.3 Kickoff - Enablement:Database
Description
What the Database group is working on in 13.3
A
The
first
thing
we're
looking
at
doing
is
continuing
our
work
on
partitioning
our
audit
events
table.
This
is
the
first
table
that
we
are
partitioning,
and
so
we
are
like
a
lot
of
the
foundational
concepts
and
tools
and
documentation
so
that
we
can
better
enable
other
teams
to
do
this
and
replicate
our
work.
A
To
that
end,
we
are
working
on
again
documenting
our
design
decisions.
We
are
looking
to
enforce
the
partitioning
key
in
queries,
because
when
you
don't
include
the
key
in
the
query,
it
can
take
a
significantly
longer
period
of
time
to
complete,
and
so
you
want
to
make
sure
that
happens
and
then
we're
also
working
on
some
ui
ux
changes.
A
This
is
important
because
again
because
we
actually
want
to
make
sure
we
have
that
query
and
the
query
the
key
in
the
query.
The
key
here
is
time,
and
so
I
make
sure
that
in
the
ui
there's
always
a
time
bound
being
shown
and
then
finally,
we
also
want
to
have
a
metric
that
we
can
use
to
track
the
state
of
partitioning.
A
We
hope
to
have
the
partitioning
being
rolled
out
on
gitlab.com
database
in
the
next
couple
of
days.
Here
it
will
take
a
few
days
to
complete,
because
the
auto
events
table
is
quite
large
and
so
we'll
be
monitoring
that
kind
of
to
the
end
of
13.2
and
to
the
beginning
of
13.3.
A
A
It's
gotten
quite
large,
as
you
can
imagine,
we've
actually
had
a
proposal
to
drop
some
of
the
content
off
of
it
because
of
how
logic
got,
but
instead
we
are
looking
to
partition
it
now
and
that
way
we
can
still
maintain
performance
of
this
table
and
still
not
have
to
delete
some
of
the
activity
of
our
users
and
how
it's
tracked
so
that
will
be
kicking
up
here
in
13.3
and
from
there.
We're
also
looking
to
add
some
instrumentation
to
the
our
usage
ping.
A
To
help
us
better
understand
the
number
of
queries
that
the
database
is
seeing
per
week,
and
this
is
our
primary
product
indicator
for
the
database
team
and
so
we're
very
excited
about
having
that
instrumented,
and
so
we
can
get
better
understanding
of
the
performance
of
our
kind
of
widely
deployed
self-managed
instances,
and
so
we
can
further
target
our
work
going
forward.
That's
it
for
that
first
team.
Thank
you!
So
much
and
we'll
see
you
next.