►
From YouTube: 2020-06-11 Ops Cross Stage ThinkBIG!
Description
02:55 - GitLab Kubernetes Agent and its implications to various product groups by Viktor Nagy
24:45 - Increase adoption by providing “deployment target” options on project creation by Viktor Nagy
A
All
right,
I
am
super
excited
to
welcome
everyone
to
the
ops
cross
stage.
Think
big
session.
This
is
our
session
number
four
and
I'm
excited
to
see
that
participation
is
not
dropping,
would
be
nice
to
have
more
people,
but
we'll
talk
about
that
later
yeah
for
those
who
may
be
joining
this
session
for
the
first
time,
I
just
wanted
to
spend
like
two
minutes
on
some
housekeeping
and
goals
behind
this
session.
A
Of
course,
we
are
also
open
for
anybody
who
just
wants
to
join
in
and
listen
or
those
people
who
have
like
a
topic
that
has
a
dependency
on
other
stage
groups
and
maybe
something
that
other
groups
have
been
sold
in
before
and
your
stage
is
about
to
go
into
that
direction.
That's
a
good
place
to
discuss
such
kind
of
topics,
or
maybe
you
clearly
know
that
that
it's
not
only
your
stage,
who's
going
to
be
yeah
thinking
about
a
certain
feature
or
about
solving
a
certain
problem
for
the
user.
A
So
these
topics
are
perfect
for
these
sessions
to
be
discussed.
We
are
using
this
for
brainstorming
together.
The
goal
here
from
the
sessions
is
to
have
a
couple
of
ideas
being
sparkled,
and
then
we
creating
an
issues
out
of
issue
out
of
the
sessions,
and
then
we
are
assigning
the
thing
big
label
that
we
have
created
to
track
the
sessions.
A
So
from
the
previous
experiences,
we
learned
that
we
are
trying
to
limit
the
session
to
the
two
topics,
so
we
have
an
agenda
in
a
document
where
each
session
we
are
collecting
two
ideas
that
we're
going
to
be
discussing
and
yeah.
There
is
also
a
retrospective
issue.
That's
at
the
top
of
the
agenda
document
feel
free
to
leave
your
feedback,
it's
very
important
for
us.
We
want
to
improve
and
make
sure
that
we
keep
the
sessions
interesting.
So
today,
our
star
4
topic
is
Victor
4
topics.
A
B
Wanted
to
ask
for
that.
So
thank
you.
Everyone
for
for
joining
this
session.
So
the
first
topic
that
I
added
is
about
the
cubensis
good
luck.
Identities,
agent
I'll,
give
you
as
much
background
as
I
can,
because,
probably
you
are
not
really
familiar
with
it
so
to
get
started.
Let's
see
what
is
command
system
and
how
does
it
integrates
with
grid
lab
command
is
basically
a
computer.
That's
it
to
put
it
simple,
pretty
complex
computer
that
you
can
manage
only
through
API
calls
and.
B
Why
it's
it's
a
pretty
awesome
computer!
That's
why
it's
so
famous,
let's
say,
and
basically
you
can
currently
even
create
commands
clusters
from
within
gate
lab,
but
we
are
not
talking
about
that
part,
but
what
it
requires
to
to
manage
to
have
any
kind
of
integration
with
clusters.
Today.
Our
support
integration
actually
requires
you
to
allow
administrative
rights
to
the
cluster,
which
is
something
people
don't
like
to
give
and.
B
Yeah
I
just
saw
the
comment,
so
so
one
of
these
shortcomings
we
have
with
the
current
current
approach
within
bit
lab
is
that
it
requires
administrative
rights
and
the
other
is
that
it
requires
you
to
open
up
your
cluster
to
the
Internet.
Actually,
if
you,
if
you
want
to
use
github
calm
with
your
your
cluster,
so
this
is
how
we
started
to
idea
that
how
could
we
solve
these
two
really
big
problems
with
the
current
integration,
because
we
received
the
feedback
from
basically
all
of
our
customers.
B
This
is
a
huge
problem
and
many
people
are
not
really
happy
to
to
provide
us
these.
These
things
likewise
to
admin,
rights
and
opening
up
their
clusters,
and
we
go
to
conclusion
what
is
not
new
thing
in
the
industry,
which
is
to
have
an
agent
running
in
the
cluster,
and
actually
the
agent
reaches
out
to
get
Lancome
and
not
the
other
way
around.
So
that
was
the
good
luck
event.
B
This
agent
is
basically
it's
our
spy
inside
the
cluster,
who
can
talk
to
get
Lancome
and
actually
it
has
a
continuous
connection
to
be
dot-com
and
it's
already
sitting
inside
the
private
network
of
the
of
the
user.
So
it
can
do
everything
that
it
has
rights
to
do
inside
the
cluster,
but
once
we
have
it,
we
overcome
the
problem
of
authorization.
B
B
The
agent
approach
and
I
think
this
is
my
question
number
three
where
I
would
not
even
three
okay,
these
were
just
the
notes
from
now
yeah
I
would
start
with,
will
actually
pressure
number
five,
which
is
that,
as
we
have
the
agent
inside
cluster,
actually
we
have
an
active
presence
inside
the
cluster,
and
this
opens
a
huge
space
of
new
opportunities
like
right.
Now
you
have
to
either
install
parameters
from
within
get
lab
in
order
for
get
up
to
know
about
it
or
you
have
to
manually
set
up
your
parameters
inside
gitlab.
B
Forget
that
to
know
about
it,
and
actually,
if
we
have
the
agent
in
the
cluster,
we
can
teach
the
agent
to
look
for
the
most
common
spaces
where
providers
installation
can
be
found
and
automatically
set
up
promoters
integration.
So
this
is
great
for
getting
metrics
and
and
actually
supporting
other
stages
and
other
groups
inside
good
lab.
Similarly,
there
might
be
possibilities
for
doing
security
scans
inside
the
cluster
on
deployed
things,
which
is
another
possibility
inside
of
it
lab.
B
So
there
are
a
few
new
area
of
opportunities
that
the
agent
actually
opens
up,
and
if
you
have
any
ideas
around
this
that
in
your
top
in
your
stage,
something
was
discussed
where
you
think
it
would
be
good
to
be
more
close
to
the
to
the
environment
where
code
code
is
running,
then
these
dados
to
me,
because
that
could
be
interesting,
there's
a
use
case
for
the
agent
in
the
future.
So
this
is
what
number
question
number
five
is
about.
B
Please
start
talking
if
you
have
anything
to
say,
because
otherwise
I
will
just
continue
with
other
questions.
I
think
this
is
the
most
important
question
to
to
all
of
us.
Basically,
because
this
is
something
totally
crossed
stage
question
are
our
group
itself
can
can
come
up
with
our
own
needs,
but
but
really
support
everyone
else.
The
media
inside
there
so.
B
C
Ahead,
one
question
that
I
always
run
into
whenever
I
look
at
our
stages,
I
think
about
the
work
that
you're
doing
as
infrastructures
code
and
maintaining
like
the
infrastructure
base
for
customers,
whereas
release
is
focused
on
application,
deployment
and
I'm
wondering
if
there's
ways
that
we
can
leverage
these
look,
these
agents
for
application
development
and
not
just
like
infrastructure
as
code.
What
are
you,
what
are
your
go,
and
how
can
I
reconcile
this
in
my
head,
like
I,
feel,
like
sometimes
yeah,
compete
in
that
area?
I.
B
Totally
agree
so
there's
a
huge
overlap
between
our
areas,
actually
because
I
used
to
say
it
is
that
from
one
side,
if
you
there's
the
two
sides
of
the
same
coin,
from
one
side
you
deploy
software
to
cluster.
This
is
your.
This
is
your
approach
to
releases.
My
approach
stories
is
that
you
modify
your
infrastructure
and
we
speak
about
the
same
thing.
Actually,
so
there's
a
is
definitely
a
it's.
A
totally
well
I
think
that
there
is
a
huge
overlap
between
our
approaches
and
and
I'm
discussing
it.
B
So,
basically,
given
this
situation,
clearly
there
are
many
opportunities
for
deployments.
The
first
thing
we
want
to
do
is
to
provide
deployments
into
the
cluster
using
the
agent.
So
that's
how
we
want
to
start,
but
even
after
that,
there
might
be
even
other
things
like
javi
do.
Monitoring
today
is
is
I,
think
rather
limited
and
having
the
agent
opens
up
again
new
possibilities
for
that.
C
B
C
Of
one
of
my
thoughts
that
I've
been
kicking
around,
is
this
idea
of
environments
being
proxies
for
deployments
for
customers
and
how
do
we
create,
like
a
distinction
of
an
environment,
that's
connected
to
a
production,
cluster
or
production
state?
And
how
do
we
treat
those
as
like
different
first-class
objects,
because
production
states
are
I
heard
higher
level?
I've
heard
a
bunch
of
different
words
to
tribute
that
this
is
a
glorified
environment
and
deployment
right,
but
yeah
it's
interesting,
because
clusters
can
run
you
can
have
you
know
thousands
of
clusters
running
production
environments.
C
D
B
We
don't
want
to
support
this
use
case
at
first,
but
we
know
about
it
and
they
we
are
waiting
for
customer
feedback,
whether
it's
it's
needed
or
not,
to
gain
to
the
point
of
environments.
I
think
there
is
this
really
basic
everything
everybody
cause
different
thing:
an
environment.
What
I
see
is
like
another
topic
that
I'm
responsible
for
is
Institute's
code
and
not
kubernetes
management
and
they're.
B
Basically,
now
I
see
that
I
set
up
my
kubernetes
cluster,
my
production
cluster,
using
terraform
and
I,
call
that
my
production
environment,
which
is
actually
not
the
cluster
but
everything
around
the
cluster
like
the
Google
project
and
the
buckets
and
all
those
things.
That's
my
production,
environment
and
I
run
my
production
cluster
in
my
production,
environment
and
I
deploy
my
production
code
into
my
production
cluster
and
basically,
during
a
single
interview,
the
into
participants
call
all
of
these
production
environments.
B
E
B
C
E
C
You're
needling
anew
so
for
a
release
management.
We
talk
about
legacy
deployments,
so
legacy
deployments
could
be
a
server.
It
could
be
a
mainframe.
It
could
be
a
random
bucket
in
AWS.
It
can
be
anything
so
an
environment
is
the
target
for
that
in
cloud
native
installations
and
kubernetes
first
situations.
That
target
is
like
a
particular
cluster,
but
the
environment
has
a
bunch
of
other
information
that
creates
those
relationships
between
microservices.
C
So
in
the
why
there's
no
difference,
you
have
to
say
URL
specify
as
a
cluster,
and
it
doesn't
really
take
into
context
that
there's
a
bunch
of
stuff
around
that
cluster.
That
really
makes
up
that
production
experience
for
a
customer,
whereas
in
other
instances
it
really
is
just
a
URL
to
an
AWS
environment,
and
that
is
the
only
thing
that
the
application
is
running
on
right.
It's
a
single
tenant
application,
for
example,
to
SAS
application.
B
Actually,
just
this
morning,
I
create
an
issue.
I
have
I've
only
shaded
it
Maryann's
here
now,
but
and
I'm
not
sure
that
it's,
it
would
be
a
solution
for
this
topic,
but
at
least
it's
it
could
be
a
starting
point,
which
is
that
we
want
to
provide
some
visibility
into
your
infrastructure
environments,
lower
level
than
kubernetes.
B
Let's
stay
like
that
as
well
at
first,
my
idea
would
be
just
to
have
a
simple
listing
of
the
turf
from
projects
that
we
are
aware
of,
which
is
kind
of
like
a
kubernetes
cluster
and
and
we
could
try
to
and
basically
with
Maria,
we
started
to
discuss
this
topic
like
in
the
past
week.
We
should
have
something
really
generic,
which
just
says
that
infrastructure
and
and
then
inside
of
the
infrastructure,
you
might
mean
virtual
machines,
you
might
mean
even
mainframes.
B
G
E
C
Review
environments
are
very
challenging
for
people
in
both
the
legacy
deployment
style
to
use
because
they
may
have
different
environments
per
stage,
and
it
could
be
multiple
layers
that
they're
going
out
and
they
have
to
specify
that
in
there
am
well
as
a
separate
job
and
that's
really
hard.
Whenever
you
have
a
bunch
of
micro
services.
That
you're
also
deploying
an
alongside
like
a
single
production
environment.
C
So
I
would
say
that
people
who
aren't
deploying
with
gitlab
that's
their
biggest
complaint
is
that
environments
are
inflexible
and
they're
unable
to
accomplish
all
the
things
they're
wanting
to
do
with
monitoring
deployment
and
like
more
more
progressive,
rollout
either.
So
incrementally,
there's
a
couple
things
that
we're
doing
to
make
this
easier.
One
like
allowing
the
sharing
of
group
environments
across
groups,
and
then
we
do
need
to
kind
of
nail
down
like
how
do
we
want
to
treat
clusters
differently
than
environments,
and
maybe
they
become
a
different
object.
But
we
should
consider
stuff
that
tracks.
B
You
mentioned
here,
you
have
to
share
environments
across
groups.
That's
actually
I
think
it's
a
very
interesting
question.
What
is
the
right
level
of
abstraction
here
better?
We
should
elevate
things
that
the
group
level
instance
level
project
level,
because
that's
that,
that's
again
something
very
specific
like
what
I
run
into,
for
example,
is
that
I
was
thinking
about
cluster
cost
management
and
all
our
metrics
pages
are
at
the
project
level
and
I
would
like
to
see
metrics
as
part
of
the
metrics.
The
cluster
cost
as
well,
but
actually
I
have
a
group
level.
C
C
The
original
MVC
was
going
to
be
aggregating
project
level,
metrics
that
are
in
the
CI
CD
analytics
tab
at
the
group
level
and
then
providing
more
granular
like
drill
down
and
I
was
hoping
to
replace
the
current
environments
dashboard
with
that
view,
so
that
you
can
filter
any
project
or
group
inside
of
your
instance
right
so
I'll
link
back
here.
Oh
some
beer
got
it.
C
G
B
B
B
It's
we
are
currently
having
a
proof-of-concept
around
that,
but
actually
I
would
say
that
it
could
support
anything
even
perhaps
open
ships,
so
not
just
the
vanilla
kubernetes
like
GK,
aks,
etc,
but
we
might
be
able
to
support
OpenShift
clearly
with
openshift,
as
it
has
much
tighter
security
and
and
a
different
different
model
of
authorization.
We
have
to
figure
it
out,
but
it's
true,
but
we
might
be
able
to
support
it
as
well.
For
the
open
policy
agent
there
is
no
overlap
at
all.
B
Actually
open
policy
agent
is
about
is
about
policías
code,
so
it
it
specifies
various
things
what
you
can
do
inside
the
cluster,
given
a
certain
action
as
far
as
I
understand
and
with
the
agent,
the
agent
is
actually
really
a
spy
inside
the
cluster.
So
that's
how
you
should
think
about
it
that
the
agent
acts
on
bitless
behalf
inside
a
cluster.
It
can
run
open
policy
agent
in
open
policy
configurations
inside
a
cluster
as
well.
B
C
A
A
B
B
Why
the
to
get
lab
users
didn't
understand
at
all
our
storyboards,
and
we
had
to
realize
that
they
have
no
idea
that
we
have
kubernetes
integration
or
they
just
know
that
gitlab
does
source
code
management
and
CI.
That's
it,
and
it's
great
for
that,
and
this
was
really
shocking
how
badly
they
were
not
aware
of
anything
related
to
our
stage
and.
B
I've
had
a
few
ideas,
but
this
was
the
first
one
that
I
actually
brought
down
that
you
can
see
in
the
link
issue,
which
is
about
specifying,
where
you
want
to
deploy,
which
might
not
really
be
a
deployment
because
it
might
be
just
a
package
registry.
But
what
is
the
target
of
your
of
your
project
and
I
thought
about
this?
B
B
One
of
my
questions:
what
are
the
other
other
ways
to
also
attack
targets?
Probably
there
are,
but
if
we
also
attack
targets,
we
won't
know
about
it.
That
was
Jackie's
question
also
like
also
DevOps
in
a
sense.
Currently,
that's
what
or
it
is
working
on
they
try
to
alter,
detect
it.
It's
not
really
out
detection,
because
you
they
go
with
a
very
simple
logic
that
if
you
have
a
kubernetes
cluster
set
up,
then
we
try
to
deploy
to
that.
B
B
B
The
other
is
just
pushing
my
code
two
if
I
have
started
already
locally
I,
never
use
the
second
one,
because
it
it's
usually
faced
for
me,
I,
don't
know
why,
but
so
I
always
go
for
the
UI
and
who
do
you
I
if
I
would
be
presented
with
this?
That
I
can
help
you
with
deploying
your
code.
I
will
be
really
interested
at
that
point
when
I'm
committing
my
first
real
commit
to
the
repple.
B
It
might
be
just
me
who
is
working
with
CI
tools
for
like
a
long
time
since
he
started
to
code
basically,
but
but
that's
something
that
I
would
be
super
happy
about
and
if
I
would
have
known
about
how
to
DevOps
before
I
would
have
been
happy
about
that
as
well.
But
I
wasn't
so
I
always
created
all
kinds
of
deployment
scripts
initially.
B
But
that
was
the
basic
idea
and
something
very
simple:
that
I
just
could
say
that
yeah
I
want
to
deploy
to
kubernetes
and
then
first
we
can
just
show
a
documentation
link
and,
at
the
same
time,
save
users
choice.
So
we
get
already
dots
out
of
the
deployment
targets
that
our
users
actually
have
and
later
we
can
expand
it.
E
So
what
I
would
like
to
do
is
go
through
the
user
journey
and
see
in
which
parts
of
the
journey
it's
relevant
for
us
to
inject
features
of
our
stage.
Buttons
links
any
sort
of
functionality
that
would
drive
adoption
and,
just
as
a
side
note,
the
girls
team
is
doing
explorations
of
first-time
user
experiences,
so
we're
working
with
a
girls
team
to
improve
by
cig
3d.
First,
you
user
experience
and
still
with
the
adoption
side
note.
B
I
don't
know
if
system
I'm
here
thinking
of
many
things
they
are
the
most
important
I
have
in
my
mind
that
it
would
be
really
great
to
have
something
integrated
throughout
lead
lap.
So
I
decide
that
I
had
my
own
button
and
then
someone
else
adds
very
similar
button,
just
a
bit
different
place
and
a
bit
different
way,
and
and
that's
why
I
think
that
that
that's,
why
I
actually
I
like
this
proposal,
because
this
could
be
throughout
the
product
used
by
many
many
stages,
I
think.
G
On
the
package
site
to
kind
of
raise
that
awareness
of
what
we
can
offer
is
trying
to
detect
when
they're
moving
away
from
get
lab,
and
so,
if
they're
deploying
like,
for
example,
they
have
a
valve
and
a
pipeline
that
deploys
to
a
third-party
registry.
Maybe
there's
an
opportunity
in
the
UI
to
say:
hey.
We
saw
you
deploy
out
to
here.
Here's
an
advantage.
E
D
B
D
Yeah
no
I
get
the
part
I
think
what
I
was
wondering
is
if
people
can
also
I
see
all
these
options.
You
know
like
deployment
targets
and
everything
and
the
other
way
you
know,
I,
don't
care
about
these
I,
not
interesting
story
in
curing
aids,
yet
or
like
doing
these
I'm
just
gonna
use
what
I
came
for.
You
know,
which
is
probably
using
just
like
racial
control
features
and
like
the
basics,
you
know
you
know
like
they
just
stayed
there,
because
they're
conversing,
so
I
think
what
I'm
asked
is
work
two
three
four
somewhere
on.
D
J
Sorry
Jackie
just
real
quick
to
that
point,
one.
The
capstone
project
that
will
read
is
she
she
found
out
from
those
who,
don't
you
see,
I
or
had
just
started
using
sky?
One
of
the
barriers
was
yeah.
They
need
to
convince
their
management
to
let
the
music,
so
they
may
have
used
it
prior
in
a
different
place,
but
they
couldn't
use
it
at
their
current
place
because
everybody
was
pulled
afraid
or
quote
I.
Don't
know
what
to
do
with
that
unquote.
So
it
was
like
where
the
the
biggest
takeaway
was
in
documentation,
the
documentation.
J
You
need
to
start
scaffolding,
people
stop
assuming
that
they
already
know
how
to
do
this
and
show
them
how
to
do
a
new
fancy
feature
actually
take
them
from
the
beginning
and
and
walk
them
through
how
to
use
it.
So
then
they
can
walk
their
management
through
why
they
need
to
start
using
Auto
DevOps
in
their
company.
That
was
one
of
their
really
big
barriers.
D
C
Okay,
so
I
think
another
opportunity
that
I
that
I
started
to
see
is
bringing
people
from
the
merger
quest
into
environments
and
into
pipelines
and
into
other
areas,
the
application
so
bringing
awareness
to
secure
or
monitoring
from
a
view
that
most
customers
are
in
and
after
kind
of
validating
an
assumption
with
users.
It's
a
it's
a
challenge
because
of
all
the
different
personas
are
an
organization
that
are
purchasing
gitlab
using
it
lab
and
sort
of
advocating
for
gitlab,
and
when
you
don't
have
the
end
state
metrics.
C
The
buyers
and
the
managers
of
the
organization
are
interested
in
implementing
additional
features
because
they
can't
render
the
ROI
or
they
can't.
They
can't
point
and
show
the
ROI
when
they're
able
to
do
that
more
effectively
in
an
outside
tool,
due
to
the
way
that
their
systems
have
been
configured
so
kind
of
in
line
with
that
same
ethic.
That
Laurie
was
was
discussing
that
they
need
permission
to
extend.
C
D
C
D
Make
sense
I
said
that,
because
I
also
heard
in
a
couple
calls
that
people
said
like
you
know,
I
like
get
up
a
lot,
but
sometimes
awesome
what
I'm?
What
it's
hard
for
me
is
to
just
like
to
go
on
justify
like
the
cost
rubric.
You
know
like
we
go
up,
one
tier
or
like
we
are
spending
more
in
CI
minutes
or
like
if
we
need
like
this
comes
particularly
from
the
idea
of
runners,
like
the
context
of
runners,
which
runners
by
design
need
a
different
machine.
D
You
know
like
a
different
host
environment
and
because
of
that,
like
yeah,
the
decision
makers
are
like
well
we're
paying
for
gridlock.
Why
are
you?
Why
do
we
need
to
pay
for
like
this
runner
thing
as
well?
And
you
know
like
it's
that
type
of
idea,
so
yeah
I
just
was
trying
to
find
some
some
common
ground.
That
was
the
case,
but
that's
just
another
problem.
B
I
would
have
a
question
that
actually
is
rooted
in
in
yes
comment
how
you
try
to
show
people
that
they
could
use
get
lab,
which
is
that
what
one
thing
I'm
thinking
lot
and
I
don't
have.
An
answer,
though,
is
that
I
would
like
to
offer
these
possibilities.
What
Jackie
spoke
as
well
and
like
that
they
they
could
use
cube
and
disintegration,
etc
at
a
place
which
is
the
most
convenient
for
that
specific
topic
like
I,
don't
know
for
what
we
did
a
few
months
ago.
B
B
That's
what
I'm
thinking
a
lot
that
what
might
be
the
right
point
of
contact
where,
where
the
user
is
really
open
to
to
our
suggestions
to
our
changes,
mr
might
be
interesting
for
some
features
for
sure,
but
but
in
many
areas
I
don't
see
this
point
actually,
I,
don't
know.
If
you
understand
my
question.
D
B
G
So,
on
the
package
side,
when
it
comes
to
that
MMRC
file,
we
have
to
demonstrate
why
it's
more
valuable
to
include
it
in
gitlab,
and
so
this
is
where
the
cross
age
opportunities
really
come
up
for
us.
Where
we
can
say,
we
can
include
these
kind
of
standings
and
we
can
have
better
management
and
you
don't
have
to
pay
for
that.
Third
party
you're
already
having
to
deal
with,
and
that's
usually
the
selling
point
is.
We
can
actually
help
reduce
your
costs
because
you
don't
have
to
pay
for
one
of
our
competitors.
G
B
Understand
I
just
understand
generally
argument:
it's
it's
an
interesting
one.
I
could
argue
it,
but
I
don't
want
to
just
still
I'm
thinking
like
what
other
groups
have
have
similar
ideas
like.
What
do
they
see
is
the
best
point
like
because
in
the
case
of
Amar's,
it's
really
I
can
see
security
there
to
be
shown-
and
that's
that's
great
if
you
know
about,
but
what
other,
because
we
just
trying
to
get
some
more
impulses
from
you.
So
so
they
can
come
up
with
ideas
as
well.
Ourself
I,
wonder.
C
Like
select
a
persona
and
drill
into
that
user
journey,
and
then
we
could
find
triggers.
That
would
make
more
sense.
So
if
we're
thinking
about
the
developer,
where
they
most
at
when
they
go
in
to
get
labs,
they're
going
to
be
in
the
project
repository,
they
might
be
in
the
merger
class
they'll
likely
be
in
a
comment
like:
where
do
we
pull
relevant
information
into
those
feeds
so
that
they
know
oh
I
can
go
to
AWS
from
here?
I
can
manage
a
cluster
from
here.
I
have
a
jump
off
point
right.
C
This
is
very
different
than
from
like
a
release
manager
where
they
might
be
at
the
group
view,
and
they
may
be
looking
at
a
milestone
because
they're
trying
to
coordinate
a
version
from
multiple
different
repos
that
are
being
pulled
into
a
single
view.
So
we
would
want
to
be
able
to
surface
information
on
the
issue
rather
than
on
the
merge
request
about
configurating
could
configurating
what
words
that
I
guess:
configuring
things
like
clusters
and
AWS
targets
and
other
sort
of
pieces,
but
there's
also
like
people
who
are
only
using
the
API
and
get
lab.
C
So
we
have.
You
know
probably
two-thirds
of
our
population.
That
would
not
necessarily
be
cognizant
or
aware
of
these
targets,
because
those
aren't
surfaced
in
the
API.
So
we
need
to
render
something
in
the
documentation,
and
this
is
where
things
like
blog
posts,
partner
talks,
engagement
like
creating
a
marketing
campaign
around
this
information
becomes
more
useful
and
we're
actionable
because
we're
not
going
to
necessarily
meet
those
people
with
this
information
right.
A
B
A
You
want
to
continue
any
of
the
conversations
on
any
of
the
topics
feel
free
to
create
an
issue
or
use
additional
one,
and
we
appreciate,
if
you
that
the
label
to
that,
so
we
can
follow
that
later
and
track
the
progress
as
well
and
yeah.
Thank
the
people.
Who've
been
involved
in
the
discussion.
We
took
very
little
note,
so
hopefully
that
couldn't
help
you.
Thank
you.
Thank
you
very
much.
A
I
would
like
to
round
up
this
session
with
st..
They
were
saying,
thanks
to
everyone
who
attended,
I,
really
love
that
people
are
coming
and
being
walked.
All
this
is
what
really
helps
us
it's
nice
to
see
how
many
topics
we
actually
have
in
common
between
the
stage
groups
and
I'm
sure
there
are
much
more.
If
you
have
any
idea
for
next
sessions,
please
leave
it
in
the
agenda.
We
have
a
table
there.
We
have
two
sessions
for
each
of
two
topics
for
each
session.
A
The
next
session
is
fully
booked
already,
which
is
great
so
but
yeah.
If
you
need
any
help
scheduling
any
of
those,
please
feel
free
to
contact
me.
I
would
love
to
help
you
and
if
you
have
any
feedback
about
how
we
could
make
the
sessions
better,
there
is
a
issue
at
respective
issue
at
the
end
of
the
today's
notes,
and
also
at
the
top
of
the
edge
in
the
document.
Thanks,
everyone
was
nice.
Having
you
and
I
see
you
give
weeks.