►
From YouTube: GitLab 13.2 Kickoff - Release:Progressive Delivery
Description
GitLab 13.2 Kickoff - Release:Progressive Delivery
https://gitlab.com/gitlab-org/ci-cd/progressive_delivery/-/issues/4
A
You
hi,
my
name
is
Enrico
Lewinsky
and
I'm
a
senior
product
manager
of
the
progressive
delivery
group
of
the
release
stage
here,
I
get
left
today,
I
want
to
walk
through
what
we
plan
to
deliver
in
13.2.
We
do
have
a
planning
issue
that
you
can
find
under
the
progressive
delivery
group.
If
you
want
to
read
more
in
depth
of
more
than
what
I
want
to
explain
now.
A
So,
let's
get
started
so
the
first
issue
that
I
want
to
discuss
is
extend
ECS
template
to
support
EWS
Fairgate,
and
you
might
recall
that
we
have
been
pushing
a
lot
of
features
around
streamlining
AWS
deployments
and
we
just
want
to
make
the
experience
of
deploying
to
AWS
very
simple
and
integrated
within
get
lab.
So
we
previously
added
support
to
deploy
to
UCS
using
either
a
good
lab,
C
IMO
template,
and
then
we
connected
it
also
to
out
of
the
box
in
this
issue.
What
we
want
to
do
is
extend
the
support
to
ECS
Fairgate.
A
Up
until
now,
it
only
supported
easy
to
sew,
but
you,
we
simply
are
asking
to
do,
is
add
a
new
launch
type
called
Fairgate,
and
then
everything
else
will
work
just
as
before
when
the
deploy
target
was
ECS
ec2.
So
I'm
really
excited
about
this.
We've
gotten
a
lot
of
feedback
from
the
community,
including
some
contributions,
so
excited
to
see
that
this
is
a
really
needed
feature
and
we're
happy
to
be
working
on
in
his
liver
on
the
next
feature.
A
That
I
want
to
discuss
is
ad
cancel
pipeline
button,
2
environments
page-
and
this
is
a
continuation
of
what
we
started
to
work
on
around
post
deployment,
monitoring
an
idea,
post
deployment
monitoring.
Is
that
once
you
deployed
something
to
production-
and
you
are
monitoring
it
and
some
kind
of
alert
comes
in
that
something
is
not
behaving
as
it
should.
You
can
basically
take
a
decision
to
either
automatic
or
manually,
stop
the
rollout
or
even
rollback.
A
So
in
our
previous
muscle,
we
were
working
on
just
adding
the
alerts
into
the
environments
page,
so
it's
more
visible
and
now
what
we
want
to
is
some
kind
of
cancel
button
so
that,
if
you
see
an
alert
present,
you
can
either
manually
press
it
and
that
will
stop
the
rollout
and
additional
to
that.
We're
also
adding
an
auto
cancel
deployment.
So
if
the
alert
manager
detects
critical
alert
on
the
pipeline,
it
will
just
stop
the
rollout.
A
A
Another
really
interesting
issue
we've
discussed
this
before,
but
there's
so
much
feedback
and
discussion
around
it
that
we
keep
changing
according
to
the
feedback
that
we're
getting
from
community.
So
thank
you
for
that
and
that's
allow
parent
project
developers
to
create
pipelines
for
merger
quests
for
for
Clemens.
A
So
basically,
what
we
want
to
do
in
this
issue
is
allow
users
that
are
forked
to
request
from
someone
who
has
high
enough
permissions
in
the
parent
pipeline
to
run
the
pipeline
for
them,
so
that
you
can
see
that
nothing
breaks
and
this
even
helps
us
internally
get
lab
with
community
contributions,
and
the
first
decision
that
we're
going
to
see
is
on
the
merge
request.
Widget
itself,
they're
gonna
see
this
new
message
saying
that
this
pipeline
is
run
in
the
fork
pipelines
context
Aska
maintainer
to
trigger
those
pipeline
as
part
of
the
projects.
Context.
A
If
you
are
running
from
the
fork
tamar
and
in
the
parent
project,
the
maintainer
or
reviewer
will
actually
see
a
different
different
note.
That
says
the
pipeline
is
run
from
the
fork
projects
context.
So
please
make
sure
that
you're
doing
a
proper
code
review
so
really
excited
also
to
move
this
forward.
This
is
a
one
of
our
most
popular
issues
and
one
that's
been
around
for
a
really
long
time
and
we're
really
really
happy
to
be
working
on
this.
A
Another
feature
that
we
have
related
to
progressive
delivery
theme
around
feature
flags,
which
is
you
know,
the
highlight
of
progressive
delivery
and
something
that
we
actually
need
really
badly
internally,
when
we
started
dogfooding
and
we're
a
new
feature
flag
subset.
This
came
up
so
we're
adding
a
new
strategy
called
flexible
rollout
strategy
which
will
actually
replace
our
current
percent
rollout
and
the
way
that
it
works
is.
It
gives
maximum
flexibility
in
terms
of
what
you
can
actually
define
in
your
percent
rollout.
A
So
today
and
present
rollout,
you
can
set
the
strategy
to
be
percent
robot
and
you
can
set
the
percentage
so
you
can
set
it
set
it
to
50
percent
100
percent
25
percent,
whatever
you
choose,
and
now
we're
also
adding
the
way
that
that
it
handles
stickiness.
So
we're
have.
We
have
a
new
type
called
flexible
rollout
and
you
can
still
set
the
percentage,
but
you
can
also
set
what
is
the
stickiness
going
to
be
based
on?
A
A
So
those
are
our
top
deliverables
for
13.2.
We
do
have
two
plans:
community
contributions
that
I'm
also
really
happy
to
see
the
first
one
is
notifications
for
when
pipelines
are
fixed,
so
this
is
actually
already
active
on
Comm,
but
we're
checking
it
out
and
and
we
plan
to
continue
working
on
it
and
basically,
when
a
pipeline
fails,
you
get
notification
that
it
fails,
but
when
you
fixed
it-
and
it
succeeds,
there
is
no
notification,
so
we're
adding
this
new
notification.
A
That
tells
you
that
now
your
pipeline
has
been
I'm
just
like
any
other
notification,
I
get
double.
You
can
decide
to
disable
this
notification
if
it
bothers
you-
and
here
is
a
simple
sample
email
that
is
shown.
That
shows
you
that
the
pipeline
has
now
been
fixed.
So
really
great
community
contribution
always
happy
to
see
that
and
the
next
community
contribution
that
we
have
is
actually
related
to
the
EWS
faragut
that
issue
that
I
mentioned
before.
A
So
this
is
a
really
really
nice
improvement
again
by
the
community
that
it
extends
our
current
deploy,
CCS
functionality
by
making
sure
that
when
you
create
a
new
revision
of
the
testa
finition
in
dcs,
it
makes
sure
that
it
makes
sure
that
all
that
tributes
that
were
configured
before-
let's
say
memory
or
CPU
or
other
things
in
the
task
and
cluster
configuration
in
AWS-
are
not
forgetting
to
forgotten.
And
there
are
persistent.