►
From YouTube: GitLab 15 Pre-launch demo 20220615
Description
William Arias provided an overview demo on GitLab 15.
A
All
right
welcome
everyone,
thanks
for
joining
us
for
cs
skills,
exchange
a
very
exciting
topic
today,
with
good
lab
15
coming
out
soon
we
have
tai
and
william
here
with
us,
I
think,
ty.
You
said
you
were
going
to
go
ahead
and
kick
us
off
so
take
it
away
and
I'll
put
the
notes
document
in
the
zoom
chat
for
anybody
who
has
questions.
B
All
right
welcome
everybody,
so
this
is
going
to
be
williams,
call
I'm
just
getting
on
here
to
introduce
him,
but
at
the
same
time
I
kind
of
set
up
what
the
demo
or
the
present
the
technical
presentation
that
he's
going
to
be
giving.
This
is
something
that
we
are
currently
doing
with
analysts
in
terms
of
showing
him
what's
new
since
gitlab
14,
all
the
way
up
to
15
and
the
progress
that's
made,
there's
he's
right
in
the
middle
of
a
current
road
show.
B
So
if
you
were
able
to
take
a
look
at
the
presentation,
deck
prior
even
watch
his
demo,
this
is
what
he's
going
to
be
giving
today
it's
going
to
open
it
up
for
questions.
The
presentation
part
is
not
going
to
be
part
of
this
cs.
Skills
exchange,
that's
something
that
you
can
take
a
look
at
offline
and
go
into
that
a
little
bit
more
now
with
that
presentation.
That
is
also
that
has
a
marketing
aspect
to
it,
then
also
a
product
management
aspect
to
it
with
forward
forward.
B
I
just
forgot
the
forward
thinking
forward
statements
that
are
progress
in
going
forward
with
gitlab
and
what's
going
after
15..
So
with
that
gonna
hand
it
off
to
william
who's,
done
an
amazing
job
and
is
doing
an
amazing
job
briefing.
All
these
analysts
right
now
with
get
lab
15
and
what's
been
going
on
since
14
all
right
go
ahead,
william.
C
C
C
And
what
are
the
things
here?
You
will
see
when
you
go
through
the
presentation,
but
I
will
just
do
a
quick
introduction
here
that
the
main
topics
on
the
main
theme,
the
key
themes
that
we
have
introduced
for
gitlab15
they
go
from
enterprise,
agile
planning
and
workflow,
and
then
we
move
into
workflow
automation,
continue
security
and
compliance.
C
Here
we
make
emphasizing
all
the
features
around
compliance,
also
audit,
auditing
and
value
stream
management.
So
the
story
that
we
tell
and
that
we
have
been
discussing
with
the
analyst
is
that
we
said
these
are
the
key
themes
from
gitla
15
and
now
we
will
show
a
demo.
That
is
not
a
features
based
demo,
but
it's
a
demo
that
shows
and
goes
through
each
one
of
those
fillers
and
we
can.
C
We
can
use
them
as
a
story
that
illustrates
how
these
things
can
come
to
life,
so
we
start
with
enterprise,
agile
planning
and
workflow,
and
here
I
want
to
show
to
you
first
of
all,
what
is
the?
What
is
the
the
problem
that
I
want
to
solve
and
what
is
the
solution,
the
partial
solution
that
we
have
today
so
as
part
of
the
technical
marketing
team,
we
are
always
trying
to
create
content
that
is
super
relevant
for
for
our
communities.
C
C
Where
is
a
space
where
all
these
are
already
consolidated
for
us?
So
it
looks
to
me
that
it
would
be
a
good
idea
to
instead
of
thinking
that
I
believe
this
topic
to
to
write
a
blog
about
should
be
cool.
Why
don't?
We
just
go
to
stack
overflow
and
we
bring
all
those
thousands
of
questions
and
I
apply
some
text
processing
and
we
try
to
learn
from
that
data.
C
So
in
that
way
I
will
extract
the
topics
where
the
community
is
struggling
and
these
topics
are
are
derived
from
the
questions
that
real
people,
our
community,
is
asking
on
a
stack
workflow.
So
that's
a
challenge
how
we
can
create
educational
content,
but
we
will
use
data
from
a
stack
overflow
to
drive
the
creation
of
content
to
choose
the
topics,
at
least
so
many
of
the
things
I
show
here
they
come
from
real
life,
so
it
happened.
C
It
was
something
like
I
started
planning
this
application
and
then
the
editorial
team
came
to
me
and
they
told
me
we
would
like
to
increase
the
impact
of
the
content
created
and
we
have
an
epic
created
for
that.
So
I
said:
okay,
I
am
already
working
in
a
some
initiative
that
is
consist
on
developing
this
application.
C
So
this
is
where
we
can
illustrate
that
this,
this
epic,
that
we
can
say
was
created
by
this
person.
I
added
a
content
manager,
the
editorial
manager,
it's
something
that
comes
from
the
business
side.
Let's
say
where
they
just
have
this
business
objective
of.
How
can
we
increase
the
the
impact
of
the
content?
C
And
I
am
saying
that
there
is
another
epic
that
is
called
main
topic
visualization
and
analytics,
which
is
this
one,
the
one
where
I
have
all
the
the
work
that
I
have
to
do
to
build
this
application
that
at
the
moment,
is
blocking
this
one.
So
this
is
one
of
the
new
features
that
we
have,
that
we
can
link
epics
and
the
one
of
the
cool
things
about
doing.
This
is
that
from
here?
What
we
do
is
that
we
are
giving
the
the
the
business
side.
C
We
are
giving
visibility
or
what's
going
on
in
the
development
and
operation
side,
because
now
the
editorial
team-
they
know
they
can
come
to
this
epic
and
they
can
see
all
the
work
that
is
being
done.
Building
such
application
that
eventually,
we
will
use
to
find
a
way
to
increase
the
the
impact
of
the
content
and-
and
this
mvp
here
that
you
see
in
this
epic
right.
You
have
something
sub
child,
a
child,
epic
that
I
created,
that
is
called
mvp
deployment.
It
can
be
summarized
in
this
architecture.
So
what
is
this?
The
architecture
here?
C
C
First
of
all,
I
create
a
group
because
I
want
to
implement
a
multi-project
pipeline
where
I'm
sharing
artifacts,
and
I
am
using
compliance
framework
also
to
make
sure
that
I
I
bring
extra
attention
to
the
loader
project,
the
one
bringing
the
questions
when
I
was
working
on
this.
I
was
reading
the
the
law
at
this
year
in
europe
and
they
say
that
anything
that
can
be
used
to
identify
a
person
is
part
of
personal
data
and
it
shouldn't,
and
it
should
have
some
treatment.
C
The
gdpr
and
I-
and
there
are
people
on
the
stack
overflow,
that
their
profile
image
is,
is
not
a
it's,
not
a
super
hero
or
something
is
there
actual
picture?
So
I
said,
let's
make
sure
that
when
we
bring
this
data,
I
am
not
bringing.
I
am.
I
am
cleaning.
I
am
filtering
out
the
profile
image
just
to
make
sure
that
I
am
not
storing
anything
that
can
be
used
to
identify
them
and
one
of
the
the
in
this
this
project,
the
loader
one.
C
This
is
a
compliance
pipeline.
I
am
making
sure
that
it's
only
using
the
image
that
I
set
is
going
through
executed
scanners,
that
I
decided
and
and
part
of
the
story
here
is
saying
that
the
developer.
They
cannot
change
this.
They
they
just
need
to
do
what
they
they
are,
the
best
of
doing
they
have
to
develop,
and
these
other
steps
that
help
them
to
remain
compliant
and
secure
are
taken
care
of
by
our
compliance
framework.
C
One
of
the
the
steps
that
are
part
of
the
local
gitlab
ci
yaml
file
is
to
get
the
questions
from
from
stack
overflow
and
then
apply
the
gdpr
step,
which
is
an
assertion
that
I'm
doing
just
to
make
sure
that
I
was
able
to
properly
clean
the
json
and
not
bring
in
personal
data,
because
this
is
a
multi-direct
pipeline.
So
what
I
do
here
is
that,
once
the
loader
is
has
finished,
it
will
trigger
the
visualizer
project.
That
is
the
one
in
charge
of
building
the
front.
C
So
two
things
here
that
are
important
to
highlight
and
that
we
have
been
highlighting
so
far-
is
that
this
already
is
a
data
science
workload
or
at
least
a
entry-level
one,
because
here
what
we
are
doing
is
that
when
we
build
an
application
like
this
one,
this
is
the
output
of
the
of
the
of
this
application.
When
this,
when
this
multi-project
pipeline
finishes-
and
I
manually
deploy
my
application,
this
is
a
mvc.
C
This
is
one
of
the
main
things
that
I'm
just
testing
the
the
the
concept,
but
I
already
can
see
different
things
that
are
important.
I
can
have
a
high
level
overview
or
what
are
the
main
questions
where
they
are
coming?
People
are
asking
about
mercy,
quests
docker
demons,
and
how
to
set
up
environment
variables,
and
now
that
I
have
the
data,
many
cool
things
can
happen.
For
example,
I
said
I
would
like
to
to
filter
the
users
with
the
highest
reputation
and
stack
overflow
what
they
are
asking
about.
What
are
they?
C
What
are
the
main
questions,
because
maybe
this
could
be
an
this?
Could
this
could
be
a
topic
for
a
tutorial
for
a
blog
post
for
a
worship
this
could
be,
but
also
because
everyone
can
contribute.
I
didn't
want
to
leave
behind
the
people
that
just
starting
on
stack
overflow.
This
doesn't
mean
that
they
are
newbies.
This
means
that
the
users
starting
a
stack
overflow,
it's
very
difficult
to
have
a
high
reputation
on
a
stack
overflow.
C
So
I
I
also
wanted
to
bring
the
ones
that
you
just
started
and
seeing
okay,
what
are
they
working
on?
What
are
the
questions
that
they
have
so
this?
This
might
give
me
an
idea
of
things,
content
that
we
can
create
and
start
aiming
for.
The
initial
challenge-
that
is,
that
I
said
that
is
how
we
can
create
educational
content.
So
at
the
moment
there
are
different
things
that
can
be
done
with
the
data
like
finding
cluster
of
topics.
This
is
this
is
still
a
work
in
progress.
C
So
this
allows
me
to
put
all
my
focus
in
now
that
this
is
solved.
How
can
I
what
can
I
do
with
this?
What
can
I,
how
can
I
create
better
dashboards,
a
better
ranking
or
maybe
a
training,
another
language
model,
I'm
using
here
just
taking
two
tokens
together
and
trying
to
see
if
there
is
a
relationship
between
that
and
having
said
that,
so
now,
when
we
switch
hats-
and
we
say-
okay-
the
application
is
deployed-
everything
is
working.
We
have
here,
you
can
see
the
production
environment.
C
C
I
can
define
a
schedule
and
just
run
it
against
my
against
my
production
environment,
so
this
is
also
helping
me
to
to
keep
track
of
different
vulnerabilities
that
maybe
they
didn't
existed
when
I
created
a
merchant
quest,
but
they
might
be
present
now
in
the
in
the
production,
but
or
also
if
there
is
some
other
misconfiguration
now
in
the
application
server
when
the
application
is
live.
So
this
gives
me
I'm
I'm
not
sure
if
you're
familiar
with
this,
you
have
to
write.
This
gives
me
this
report.
C
C
I
created
a
merge
request
and
I
could
be
easily
another
language
processing
programmer,
not
necessarily
a
software
engineer,
someone
that
knows
how
to
implement
machine
learning
algorithms,
but
it
doesn't
mean
that
just
because
I
am
doing
something
that
is
not
a
specific
that
it
is
more
research
oriented,
it
doesn't
mean
that
I
cannot
adopt
devops
practices.
So
what
is
happening
here
is
that
when
we
configure
all
of
these
things
in
a
machine
learning
workload
by
default,
I
am
now
using
all
the
practices
of
good
practices
of
software
engineering
into
the
machine
learning
world.
C
Because
when
I
create
a
merge
request-
and
I
add
all
of
these
scanners-
I
am
using
compliance
frameworks.
I
am
using
review
apps
and
I
am
making
sure
that
the
code
quality
is
not
degrading
and
I
cannot
have
the
visibility
of
the
security
scanners
and
the
vulnerabilities
that
it
found
right
at
the
in
the
in
the
merch
request.
C
C
Before
deploying
and
and
one
thing
that
I
have
been
enjoying
a
lot
is
that
when
we
use
review
apps
there
are,
there
are
things
that
are
very
interesting
here.
I
don't
know
if
you
have
followed
in
the
news
that
it
is
quite
normal
that
when
someone
deploys
or
creates
an
application,
let's
say
like
a
chatbot
that
the
output
depends
on
of
some
probabilistic
model.
C
It
can
give
either
on
desired
results
or
things
can
go
wrong
when,
when
you
are
using
probabilistic
outputs,
so
one
of
the
cool
things
that
I've
been
enjoying
of
using
this
is
that
here
I
see
big
benefits
of
having
review
apps,
because
you
can
try
the
the
model.
You
can
see
the
this
probabilistic
output
before
replying
to
production,
so
you
still
have
one
one
hand
one
hand
understeer
where
you
can
see.
C
Let's
see
if,
for
example,
my
predictions
are
not
super
misleading
or
or
one
thing
one
one
instance
where
I
have
used
this
heavily
is
deploying
chatbots
that
sometimes
you
want
to
test
what
will
be
the
type
of
answer
that
the
chatbot
will
give
when
someone
is
when
someone
is
is
chatting
with
it.
So
this
is
a
golden
opportunity
using
review
apps
in
order
to
put
a
human
to
test
these
probabilistic
models
and
make
sure
that
the
the
it's
not
giving
an
undesired
result.
C
So
this
is
now
how
we
continue.
We
continue
moving
in
the
in
this
theme
ladder
and
illustrating
each
one
of
the
key
themes
from
gitlab,
15
and
and
the
last
one
that
is
visibility
and
observability
here.
What
we
do
is
that
we
show
different
things,
but
one
of
them
being
being
the
main
one,
the
analytics
under
my
project
or
at
a
group
level.
Where
now
I
can
see
different,
I
have
more
now
more
metrics.
Now
we
have
dora
metrics.
C
C
How
is
it
that
with
this,
this
is
the
demo.
This
is
the
the
real
life
situation
that
we
are
working
in:
the
tmm's
team,
understanding
our
community
and
stacker
flow
and
using
it
lab
end-to-end
and
illustrating
these
different
key
themes
that
we
have
been
discussing
and
highlighting
with
analysts
for
gitlab
15.,
so
I'm
open
to
questions
feedback
ideas.
What
else
can
we
add
to
this?
This
is
a
work
in
progress,
but
happy
to
hear
what
else
can
we?
C
What
ideas
that
occurred
to
you
every
time
that
I
showed
this
someone
tells
me
why
don't
you
do
this?
Why
don't
you
add
that
it
would
be
cool?
Someone
told
me
there
are
there,
but
I
don't
know
if
it's
dangerous,
because
in
my
data
set
I
already
have
the
approved
answer
to
some
certain
questions
stack
overflow.
C
D
D
D
It
has
nothing
to
do
with
data
science.
Sorry
I
don't
know,
that's
the
focus
of
the
call.
But
oh
I
have
another
question
about
you.
You
mentioned
like
when
you're
using
probabilistic
models
like
behind,
like
a
chatbot,
for
instance,
in
the
review
app
you
can
kind
of
test
the
output
there
is
that
sort
of
like
a
touring
test
like
pre,
pre-deployment
employment,
touring,
test
yeah.
C
That's
trendy
these
days
because
of
the
sentient
chatbot.
No,
no,
it's
not
a
terrain
test.
It's
say
that
you
say
I
would
like
information
about
devops
certification.
I'm
making
sure
that,
for
example,
the
language
generated
is
inclusive,
that
if
you
curse
to
the
chat
bodies
is
ignoring
these
type
of
things
because
there
are
there
are.
There
are
examples
where
the
chatbot
learns
from
the
data
and
they
it
learns
how
to
curse
back
so.
E
C
D
C
I
want
to
add
to
elaborate
to
my
answer
to
your
previous
question,
this
project
that
I'm
sharing
right
now
here.
This
is
actually
a
real
deployment
of
a
chatbot
that
are
that,
where
I
have
implemented
what
team
one
thing
that
we
are
presenting
and
we
have
already
available
in
gilda
15,
which
is
using
gpu
runners.
So
you
can
see
here.
This
is
a
pipeline
from
a
chatbot
and
when
I,
when
it
comes
to
trying
the
model,
this
is
a.
This
is
a
big
neural
network
and
it
uses
a
lot
of
data
transformers.
C
And
what
I'm
doing
here
is
that
only
for
this
step,
I
am
using
a
runner
that
is
using
gpu
and
for
the
rest
of
the
steps
I
am
using
the
share
the
share
runners,
so
this
is
one
of
the
things
that
is
also
available
for
the
typical
question
is
how
can
I
because,
when
it
comes
to
data
science
workloads,
they
have
to
train
algorithms,
and
this
requires
a
high-end
hardware.
So
one
thing
that
we
know
and
that
we
have
available
now
is
using
gpu
runners.
C
There
is
a
question
here:
can
we
use
the
third
party
ml
via
an
api?
Call?
Yes,
and
this
is
one
of
the
things
that
I've
been
also
using
in
in
my
projects.
There
is
this
one
that
is
called
dvc
for
data
version,
control
and
cml,
and
this
is
this.
What
it
does
is
that
we
can
create
a
model
model
registry,
this
all
of
these
things
we
eventually
will
have
in
our
in
our
platform,
but
these
these
are
other
things
that
we
can.
C
We
can
use
so
one
of
them,
for
example,
by
using
one
of
those
third-party
apps
for
email.
It
creates
automatically
whenever
you
create
a
merge
request.
If
this,
for
example,
chatbot
whatever
plots
are
meaningful
for
you
to
measure
the
performance
of
the
model,
you
can
configure
that
to
put
them
in
the
merch
request.
F
B
E
How
do
you
go
about
choosing
you
know
what
ci
scans
to
to
to
run
to
enforce
like
the
example
that
you
were
giving
was
very
specific
to
your
project
around.
You
know
images
from
from
stack
overflow
profiles,
but
that's
not
like
secret
detection.
You
know
that
was
actually
quite
you
know
one-off.
So
I'm
just
wondering
how
you
go
about.
How
would
you
recommend
us
go
about
telling
customers
hey?
This
should
be
your
gdpr
scan
secrets
container
sas
or
you
know
what
I
mean.
C
Okay,
yeah,
I
don't
have
the
expert
answer
to
that
question.
What
I
can
say
from
my
experience.
For
example,
before
gitlab
I
used
to
work
as
a
consultant
for
api
management
and
when
we
had
to
comply
with
banking
regulations,
what
we
were
doing
was
basically,
first
of
all
understanding.
What
is
the
regulation
and
what
are
we
supposed
to
fulfill
and
many
cases
translating
what
they
are
saying
there
with
what
is
available
in
the
in
the
security
scanners
or
in
the
in
the
different
policies?
C
If
for
indicate
in
this
case
that
I
was
using,
I
did
more
or
less
the
same.
I
said
how
can
I
make
sure
that
I
am
fulfilling
and
I
am
and
I
am
being
compliant
with
gdpr
and
then
I
had
to
read.
I
went
to
to
read
the
gdpr
law
and
I
said,
and
then
I
found
this
paragraph
when
it
says
anything
that
can
be
useful
to
identify
someone.
You
shouldn't
store
it
or
you
should
ask
for
permission.
So
I
say:
okay,
have
you
said
that
I
will
enforce
that?
E
Okay,
yeah,
I
see
that
in
like
samir
it
depends,
and
so
you
know
that's
true
and
I
think
that
gitlife
has
docs
out
there.
I've
looked
at
like
a
few
months
ago
that
kind
of
spells
out
the
the
common
ones
and
it
summed
like
summed
it
up:
it
looked
like
pretty
much.
Every
single
security
scanner
would
be
required
or
very,
very
useful,
so
it
becomes
quite
heavy-handed.
You
know
if,
if
you
are
collecting
pci
data,
you
know
credit
card
data
or
pii
data.
So
it's
almost
like
all
eight
security
scanners.
E
G
Yeah
and
calmly
one
one
more
thing
I
want
to
add
to
that
is
remember
that
when
you're
talking
compliance
security
scanning
is
just
one
small
portion
of
it,
there's
also
audit
trailing,
there's
also
other
aspects
that
you
have
to
consider.
So,
while,
in
the
context
of
what
you're
asking
yes,
you
can
define
specific
kinds
of
security,
things
that
meet
the
compliance
requirements,
it's
never
enough
to
say:
oh,
we
met
the
entirety
of
the
pci.
G
A
All
right
short
and
sweet
today,
thank
you
so
much
william.
That
was
a
great
demo.
Definitely
if
anybody
has
any
follow-up
questions
you
know
you
can.
I
think
you
can
put
them
in
customer
success
or
we
can
reach
out
a
different
way,
but
I'll
have
this
recording
available
after
this
call,
but
thank
you
all
so
much
for
joining
and
we
look
forward
to
all
the
demos
that
are
coming
out.
Yes,
final.
H
Announcement,
I
missed
the
first
couple
minutes
so
if
this
was
already
said,
another
shameless
plug,
please
please,
please
tell
your
customers
tell
your
partners.
Tell
your
friends
about
the
gitlab
launch
party
coming
up
on
june
23rd.
We
are
up
to
3
600
folks
registered,
but
let's
go
for
five,
so
we
can
do.
I
think
everybody
should
be
excited.