►
From YouTube: 2022-02-16 AMA about GitLab releases
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
Okay,
so
let's
get
started
so
welcome
everyone.
This
is
the
february
16th
2022,
the
ama,
from
the
delivery
team
about
get
lab
releases
and
deployments.
A
So
we
do
have
a
few
questions
so
elena.
Thank
you
so
much
for
putting
in
our
first
question,
john
to
verbalize.
B
Yeah,
I'm
hopefully
in
the
right
place,
I'm
moana,
I'm
the
senior
product
manager
for
the
monitor,
respond
group
and
we
just
created
a
new
category
called
continuous
verification.
So
that's
monitoring
what's
happening
before
and
after
deployments,
and
I'm
hoping
this
is
the
right
place.
But
I'm
looking
for
two
to
three
people
to
interview
internally
kind
of
about
what
the
process
is
today
that
gitlab
is
using.
A
Fantastic
yeah
you're,
absolutely
in
the
right
place
perfect.
I
am
very
happy
to
volunteer
like
if
any
like.
If
anyone
else
on
delivery
also
wants
to
right,
feel
free,
like
you
know,
I'm
sure
we'll
all
have
different
perspectives
and
different
things.
I
would
also
suggest
that
the
engine
is
on
call
from
reliability.
They
also
have
some
interesting
perspective
because
they
are
perhaps
so
in
delivery.
I
think
we're
the
people
who
we
kind
of
control
the
timing
of
the
deployment
so
we're
the
ones
who
kind
of
are
able
to.
A
We
have
checks
and
things
like.
Is
this
a
good
time?
Is
it
healthy
and
do
we
deploy
now,
but
the
engineering
call
kind
of
comes
at
it
from
the
other
side,
because
what
they
may
have
is
oh
here's
an
incident
and
then
is
there
a
deployment
or
am?
A
I
is
there
anything
to
indicate
that
the
deployment
may
have
been
related
to
this
incident,
so
sort
of
look
at
it
from
the
other
angle,
so
we
have
similar
kind
of
dashboards
and
things
we're
using,
but
we
certainly
probably
have
slightly
different
requirements,
but
I'm
happy
to
help
find
you
someone
from
reliability.
Who
can
also
take
part.
B
Okay,
I
can
also
post
on
sometimes
I
just
post
in
the
infrastructure,
lounge
slack
channel
and
there
seems
to
be
some
good
response
there
and
then
I'd
love
to
pencil
in
like
30
minutes
on
your
calendar
in
the
next
week
or
two
yeah.
A
Absolutely
that's
yeah
sounds
great
awesome.
Thank
you,
cool
thanks,
so
much
and
I
suppose,
like
selfishly,
can,
can
we
also
ask
then
like
what
I
guess,
what
are
what
are
you
hoping
to
like
where
to
show
like
roadmap,
look
like
what
can
we
hopefully
adopt
internally.
B
Yeah,
so
I
think
initially
we're
going
to
start
with
just
alerts
in
gitlab
we're
still
validating
we're
kind
of
using
this
opportunity
to
validate
the
mvc
but
long
term.
We
hope
to
do
some
machine
learning
around
this
data
and
maybe
automate
some
of
the
things
that
you
do
repetitively
or
help
you
better
make
decisions,
so
I
did
link
the
direction
page.
That's
the
first
link
in
that
agenda
item
so,
hopefully
that'll
that'll
kind
of
give
you
a
little
bit
more
of
some
of
the
spoilers
to
me
brilliant
great
thanks.
So
much
yeah.
C
Sure
just
wondering
how
goes
the
investigation
into
reliable,
redis
instances
and
gates,
specifically
reliable
redis,
just
at
scale.
A
Thank
you
for
asking
for
that,
so
we
don't
have
scomac
with
us
today.
I
will
start
off,
but
anyway,
I
know
you've
been
kind
of
close
to
this
project
as
well,
so
feel
free
to
jump
in
if
there
are
things
that
you
want
to
add
or
correct,
but
I'll
give
a
kind
of
overview
from
where
I
am
so
far.
A
So
our
our
motivation
behind
starting
to
migrate
all
that
we'll
try
and
do
as
many
of
the
reddest
instances
as
we
can
it's
a
little
bit
experimental
so
we're
starting
on
rate
limiting,
which
is
a
relatively
straightforward
instance,
compared
to
our
sort
of
slightly
more
complicated
ones,
and
what
we're
trying
to
achieve
here
is
move
it
into
a
kubernetes.
It
joins
a
lot
of
our
other
services.
It
will
reduce
kind
of
overhead
of
operating
so
yeah.
We
get
sort
of
lots
of
benefits
there.
A
So
that's
the
kind
of
behind
the
scenes
of
like
why
we
started
this
work,
so
it
is
progressing
well
we're
taking
it
a
little
bit
steadier
than
some
of
our
recent
migrations.
This
is
our
first
stateful
migration.
So
that's
new
for
us.
So
we're
doing
a
lot
more
testing
around
just
making
sure
we
understand
like
how
old
day
to
behave.
A
We
don't
have
any
sort
of
data
loss
as
we
go
through
this,
so
current
status
is
we
have
a
planned
migration
path
and
we're
currently
doing
some
work
to
figure
out
how
we
get
host
names
working
with
that
vm.
So
our
vms
and
our
kubernetes
clusters
will
reference
things
differently,
so
figuring
out
how
we
can
actually
make
them
comparable.
A
So
as
we
do,
the
migration
we'll
be
running
in
hybrid
for
a
period
and
we
need
to
be
able
to
easily
switch
between
the
vms
and
the
clusters,
so
that
work
is
ongoing
and
we're
doing
checks
around
what
observability,
tooling
and
things
like
that.
Will
we
have
available
for
us
we're
not
too
far
away,
we're
probably
I'd,
say
perhaps
a
couple
of
weeks
from
getting
to
the
pre-environment
once
we
have
this
on
the
pre-environment
we'll
be
able
to
test
roll
backs
roll-outs
and
actually
see
how
this
starts
to
perform.
C
Nice
and
and
context
for
everybody
else
why
distribution
cares.
One
of
the
reasons
we
don't
tell
customers
right
now
to
run
the
redis
in
our
helm.
Chart
is
because
it's
a
stateful
service
and
if
you
don't
know
what
you're
doing
with
the
stable
service
kubernetes
or
you
don't
have
solid
recommended
best
practices
for
people
trying
to
deploy
that
they're
going
to
have
a
bad
day.
C
Things
are
getting
better,
but
at
the
same
time
we
generally
try
not
to
recommend
things
to
our
customers
that
we
ourselves
would
not
actually
run
so
until
we
know
how
it
works.
We
know
that
it
will
work.
We
don't
want
to
recommend
it
for
a
reference
architecture
or
a
default
deployment
for
the
gitlab
chart.
A
Yeah,
I
think
that's
a
really
good
point
and
it
sort
of
explains,
I
suppose,
some
of
the
the
the
time
it
takes
to
do
a
migration
as
we
experiment
with
different
approaches
and
figure
out
what
what
is
going
to
be
a
safe
approach
for
for
us
and
other
people
to
use
and
then
make
the
changes
necessarily
to
support
that
so
yeah.
Thanks
for
sharing
that
jason.
A
No
okay,
so
one
thing
I
just
share
just
for
sort
of
general
awareness
is
one
of
the
projects
we've
been
working
on
for
the
last
few
months
is
reordering
our
deployment
pipeline,
and
this
is,
we
have
at
the
moment
a
little
bit
of
a
kind
of
a
risk
gap
around
mixed
deployments
going
to
production
and
that's
based
on
the
fact
that
not
every
package
that
we
create
for
deployment
is
guaranteed
to
reach
production.
A
So
at
the
moment,
with
our
deployment
pipeline,
we
deploy
through
our
staging
environment
and
our
production
canary
environment,
and
then
we
head
to
production.
We've
been
working
for
the
last
few
months
to
bring
in
a
staging
canary
to
give
us
an
extra
environment
to
check
and
then
we're
very
close,
the
next
probably
the
three-ish
weeks.
Following
the
monthly
release
and
the
security
release,
we
will
be
doing
a
reorder,
which
means
that
deployments
will
hit
the
staging
canary
and
then
our
production,
canary
and
in
our
staging
environment
and
in
our
production
environment.
A
So
we're
flipping
the
order
of
the
staging
and
the
production
canary.
So
we've
been
doing,
we've
got
some
extra
tests
and
it
should
be
a
concretely
transparent
switch,
but
we
will
be
sharing
more
details
as
we
get
closer
to
making
that
change,
and
particularly
if
people
are
testing
on
staging.
A
Perhaps
what
you
need
to
be
aware
of,
as
you
will
now
be
doing
your
testing
on
the
staging
canary
instead
so
watch
for
more
announcements
around
that
stuff
in
the
next
few
weeks.
D
Sorry
joining
light,
so
you
apologize.
If
you
may
have
answered
it
before
work
with
customers.
What's
the
the
place,
they
should
go
to
to
look
for
changes
upcoming
changes.
If
someone,
let's
say
it's
in
a
security
role
and
is
only
interested
in
security
changes.
Where
do
they
go?
How
do
they
find
that
information.
A
For
what
has
for
worldwide
security,
it
will
be
what
has
been
released.
So
the
security
blog
post
released
blog
posts.
The
changelogs
will
give
you
the
most
complete
view
and
then
we
for
the
patch
and
monthly
releases.
We
have
the
same,
I'm
not
sure
does
any
is
there's
anyone
here
from
product.
Maybe
who
can
say?
Is
there
somewhere?
You
can
see
what's
coming
up
in
the
next
in
the
next.
D
E
There's
a
monthly
a
milestone
kick
off
on
the
18th
of
every
month.
That's
live
streamed
on
youtube
and
there
is
as
well,
of
course,
a
written
list,
that's
auto
generated.
I
will
try
to
quickly
find
it
and
post
it
in
the
doc
in
case.
I
can't
find
it.