►
From YouTube: Delivery Fast Boot: Day 3 discussion with Product and Development representatives: Part 1
Description
Delivery team fast boot Day 3: Discussing the process from development perspective with Jason Lenny from Product team and Douwe Maan from Create team.
The issues listing the work done are contained in epic https://gitlab.com/groups/gitlab-org/release/-/epics/17
A
So
I
think
the
tool
you
will
be
perfect
guests
here,
because
we're
gonna
listen
to
this
from
other
stations
whose
nothing
the
topic,
and
then
you
can
interrupt
me
and
like
to
tell
me
like
him,
but
we
can
do
this
set
your
club.
We
announce
tips
like
this
month
or
something
like
that
and
now
you're
gonna
have
to
tell
me
about
how
this
is
not
or
is
going
to
work.
For
me.
A
Recap
what
we're
currently
doing
here
so
first
of
all,
this
is
just
from
the
release
management
managers
perspective.
This
is
what
we
currently
have,
where
we
use
chat.
Ups
commands
to
trigger
creating
stable
branches,
so
this
is
yeah
manual
right,
so
developer
would
create
the
merge
request
with
the
label
applied
peeking
through
whatever
and
then
yeah.
That's
their
tasks
done
after
that
release
manager
does
issue
a
check,
check
ups
command
and
we
automatically
pick
up
image,
requests
that
have
the
label
passed
and
so.
B
A
Then
this
is
again
manual
because
Arabs
need
to
trigger
the
command
from
check
ups
when
we
are
ready
to
take
the
release.
So
when
this
happens,
we
wait
for
pipelines
to
pass
for
see
any
manual
emerge
c3
and
then,
when
we
are
ready,
we
take
a
relation
and
that
disk
loans,
Cee
tags
correctly
applies
versions,
tags,
omnibus
and
then
waits
for
the
pipeline
to
pass.
Once
the
pipeline
builds
the
package
it
automatically
release
automatically
gets
deployed
to
staging.
A
If
you
don't
mind,
did
you
close
the
window,
because
it's
again
really
so
I'll
go
to
these
things?
Currently,
the
only
automated
thing
where
it's
like
unattended
is
basically
just
released
to
staging
and
1b.
What
we've
been
doing
the
past
two
days
here
and
a
bit
of
prep
work
beforehand
was
automate
all
of
this,
so
that
we
don't
actually
have
to
do
any
of
these
things.
So,
right
now
we
are
working
on
or
we
already
have
a
job
that
creates
auto
deploy
branch
automatically.
A
C
A
A
D
C
A
Every
time
you
make
a
process
change,
it
takes
a
while
to
propagate
through
15.
So
there's
another
difference
now
that
we
are
going
to
be
automatically
picking
this
on
a
schedule
as
well.
So
you
as
a
developer,
for
example,
you
merge
something
that
gets
deployed
to
staging
in
Canary.
It
breaks
something
or
you
need
improve
you
as
a
developer,
will
only
need
to
apply
this
label
and
know
that's
almost
a
jewel.
This
is
going
to
be
picked
and
the
message
is
going
to
be
left
on
the
merge
request.
A
D
A
Action,
so
it's
working
on
messaging
right
now.
Basically,
the
the
message
is
going
to
be
becomes
automatically
pickets.
You
are
going
to
have
to
go
to
this
track,
Emma
or
stable
branch
or
whatever
and
target
that
branching
stuff.
So
yes,
this
was
written
to
master,
but
you
were
going
to
manually
have
to
like
take
that
branch
or
comics
from
that
branch
and
try
to
send
the
request
that
way.
D
A
A
We
are
not
appointment
so
yeah
once
that
gets
in
it
gets
deployed
into
staging
and
when
he's
deployed
to
staging
canary
and
production
for
now
are
going
to
remain
manual
step
that
we
are
going
to
be
handling
right.
We
want
our
manual
verification
that
anything
goes
to
that
goes
to
production
like
we
can
control
it
for
now,
we'll
see
whether
we
want
to
automate
canary
in
production
fully,
but
until
we
test
this
process
and
iron
out
the
kinks.
We
want
to
be.
D
A
Nothing
so
to
revert
it,
it's
gonna
be
pretty
much
silent
until
someone
picks
up
on
exiting
that
is
gonna
make
the
brand
screen,
and
we
were
going
to
be
responsible
for
making
sure
that
we
echo
that
Hey
deploy
branches
threads.
We
need
help.
We
need
to
figure
this
out
and
I
hope.
That's
gonna
also
like
force
us
to
spend
a
bit
more
time
in
disciplining
ourselves
with
what
kind
of
specs
we
have
guaranteeing
a
bit
more
or
you
know
doing
something
with
them
as
well,
because.
D
C
C
A
C
C
A
D
A
No,
we
are
on
purpose
defaulting
to
on
and
the
only
reason
why
I
think
we
need
to
consider
future
prices
to
be
able
to
quickly
react
to
any
changes
right
on,
especially
when
we
have
auto
deploy,
testing
and
I
think
we
should
still
keep
it
defaulted
to
on.
But
then,
if
you
as
a
developer
on
the
shore,
why
I,
like
your
I'm,
not
really
sure
whether
this
works
properly
or
there
are
some
like
scaling
concerns
I
would
I
would
like?
A
C
A
C
A
So
yeah,
that's
that's
the
difference
from
our
perspective
like
what
changes
here,
but
the
common
uncover
what's
what's
happening
currently
in
our
regular
process.
Right,
like
we
have
the
seventh
gnome
sevens
at
that
point,
we
call
that
the
feature
freeze,
we
create
the
stable
branches.
That's
the
manual
part
with
you
right
now
and
then
anything
that's
in
stable.
We
creates
an
RTO.
A
This
whole
process
is
this
process
just
sped
up
like
instead
of
having
this
like
in
week's
time
with
more
than
two
weeks,
I
guess
it
was
compressed
a
bit
more
and
we
have
an
ambition,
ambitious
plan,
whether
that's,
doable
or
not,
is
a
different
question
of
actually
running
all
of
this
per
day.
So
every
day
you
would
anything
that
marks
a
certain
way
would
be
baked
into
the
auto
deploy
branch.
That
same
day,
we
would
stack
and
deploy
to
staging
and
then
promotes
anything
that
we
bean
okay
promoted
to
Canary
and
production.
D
D
B
A
Know
front
end
cares
about
from
pixel
affair
on,
but
if
it
actually
affects
people
who
cannot
click
the
button
button
disappeared
or
something
like
that,
then
you
can
apply
all
of
that.
So
it's
basically
you
know
the
things
that
you
would
want
to
see
resolve
some
problem
that
was
introduced
and
you
want
to
see
it
there
faster.
Yes,.
B
A
Is
going
to
become
a
go-to
place
for
you
to
verify
your
changes
as
well?
You're,
not
gonna,
have
the
QA
manual
check
list
anymore,
but
you
are
still
as
a
developer,
I'm
going
to
change
the
definition
of
done.
That
says
until
you
saw
your
feature
to
production
you're
not
going
to
have.
This
is
not
done.
A
A
Us,
instead
of
having
checklist,
it's
now
going
to
have
chapters
anymore
and
that's
gonna,
be
like
our
first
step
into
informing
people
hey.
This
is
now
you
state
your
example.
We
like
we
talked
a
bit
a
bit
about
life
post
in
the
comments
and
all
of
that
and
definitely
some
keywords
doing
uses
a
time.
Yeah
right.
My
first
step
is
going
to
be
like
having
this
process
somehow
automated,
so
that
we
can
then
start.
A
For
anyone
to
check
off
something
or
anything,
it's
gonna
be
information,
hey
this
is
what's
deployed,
and
that
also
allows
us
to
play
with
some
of
the
compliance,
but
I
just
want
to
quickly
also
explain
this.
This
date
is
important
because
we
are
planning
on
turning
off
future
freeze
and
on
the
first
day
of
the
summit,
basically
all
the
contribute.
B
A
And
then
we're
moving
into
a
transition
period.
Transition
period
is
we're
turning
all
of
this
on
and
these
branches
are
going
to
be
created
and
then
this
process
is
going
to
start.
The
reason
why
we
chose
this
date.
First
of
all.
Well
we're
hoping
we're
going
to
get
all
this
done
by
then,
so,
the
sooner
we
get
there.
A
The
better
second
of
all
I
expect
way
less
changes
to
go
in
because
everyone
is
going
to
be
traveling
around
the
world
and
we're
going
to
have
like
basically
a
two-week
window
of
people
traveling,
and
you
know
not
that
much
work
being
executed.
So
we
will
be
able
to
iron
out
any
kinks
that
we
see
and
then
come.
You
know
22nd
or
something
like
that
will
understand.
B
A
Is
also
deployed
to
production
on
22nd.
This
is
what
is
going
to
be
shipped
to
the
customers,
and
this
is
where
the
product
message
is
also
going
to
have
to
change,
because
we
are
going
to
have
this
out
in
public
on
github.com,
so
we
will
have
to
inform
users
way
before
like
as
soon
as
something
is
merged.
We
need
to
say
like
okay
in
within
the
next
week.
A
This
is
going
to
be
deployed
on
github.com,
so
breaking
changes
are
going
to
need
to
be
communicated
differently,
also
like
ensuring
that
whatever
we
announce
on
the
22nd
is
the
actual
thing
that
we
are
sure
we
are
going
to
ship.
So
it's
going
to
be
really
tough
to
say
on
a
28,
for
example:
oh
we
just
merged
it
now
put
it
on
the
22nd
blog
post
item
yeah.
D
A
B
A
C
If
you
don't
think
it's
a
problem
that
the
that
were
continuing
development
on
like
you,
we
used
to
have
a
moment
where,
like
on
this
evidence,
we
would
stop
development
I'm
wondering
since
our
developments
on
the
other,
and
we
extended
that
for
good
reasons,
the
let
things
fit
in,
but
now
that
we're
just
kind
of
continually
delivering
a
production,
not
filling
my
pattern,
where
you're
the
state,
where
you're
doing
working
on
two
releases
simultaneously.
Maybe
we
should
ask
engineers
to
switch
back
to
like
whatever,
whatever
you
guys
decide
like.
A
Mean
I'm
thinking
about
like
the
whole
development
process,
is
wrapped
into
this
milestone
the
days
from
22nd
to
22nd
and
that's
actually
what's
creating
a
lot
of
confusion.
I
think
we
might
need
to
move
away
from
milestones
that
are
like
that
and
instead
of
that
having
well.
This
is
finished
now,
where
am
I
in
the
process
of
deployment
all
right,
self-managed
customers.
We
are
here,
for
example,
that
means
you're
gonna
get
that
thing.
A
A
C
A
Is
exactly
why
also
one
of
the
reasons
why
I
chose
this
date,
because
if
you
start
working
on
something
on
the
eighth
and
merge
it
on
the
ninth
first
of
all,
it's
going
to
be
very
much
lower
volume,
so
we
can
actually
handle.
Whoever
does
that,
then
it
ends
up
being
deployed,
and
basically,
this
month
of
me
we're
going
to
like
just
offset
that's
not
working
fully
by
extending
this.
You
know
ourselves
and.
A
A
B
C
D
A
A
A
D
A
D
A
B
D
C
Will
you
be
able
to
enroll
Harry?
Will
you
be
able
to
roll
back
so
he
was
still.
Do
you
still
package?
Will
you
be
able
of
like
that
when
he's
I
think
if
you
pick
up,
you
know
whatever
came
in
that
day
and
then
we
discovered
that
that
day
has
an
issue.
Can
you
then
roll
back
to
the
previous
day
or
whatever.