►
From YouTube: GitLab 13.1 Kickoff - Configure:Configure
Description
Planning issue: https://gitlab.com/gitlab-org/configure/general/-/issues/57
A
A
A
Second
goal
is
to
develop
a
proof
of
concept
around
the
twenties
management,
2.0
direction
that
we
came
up
with
in
the
past
month,
which
will
either
validate
or
invalidate
the
technological
approach
we
take
here.
But
let's
get,
let's
see
the
details
before
I
move
into
the
details
of
these
two
issues.
I
would
like
to
start
with
a
bit
different
type
of
issues.
We
have
actually
two
confidential
security
issues
that
are
either
pure
security
issue
or
are
related
to
the
availability
of
gitlab
comm
that
we
want
to
fix
and
we
had
a.
A
We
have
a
third
issue
as
well,
which
is
related
to
to
the
performance
of
github.com.
We
want
to
avoid
these
kinds
of,
let's
say,
attacks
or
situations,
so
we
are
going.
We
are
going
to
fix
these
issues.
First,
let's
see
the
trap
formulated
ones.
What
we
have
on
our
plans,
the
first
issue
we'd
like
to
have-
is
to
provide
the
ability
to
support
multiple
from
repositories
inside
single
git
repository.
A
This
is
an
extremely
common
pattern
in
truffle
and
even
at
gate,
lab,
actually
good
lab
host,
a
single
git
repository
and
all
our
turf
or
modules
are
either
single
repository
with
all
the
environments
that
we
actually
deploy
into.
We
want
to
support
this
set
up
in
the
nice
and
easy
to
use
way.
We
do
get
lab
bag
data
from
state
storage.
A
Second
rated
to
perform
is
specialized
CI
token
to
be
used
for
telephone
runs.
We
this
would
be
similar
to
the
currently
existing
see
a
job
token
that
that
exists
in
CI
runs
where
a
given
job
can
run
with
the
authorization
rights
of
the
person
who
invoked
that
run.
This
is
extremely
important
in
the
case
of
telephone,
where
we
have
the
problem.
We
have
the
two
user
stories
here.
If
I'm
a
junior
developer,
I
want
to
be
able
to
run
this
welcome
actor
for
plan,
but
I
should
not
be
allowed
to
run
her
from
reply.
A
A
Stuff
from
apply
and
plan
command
is
crucial,
because
if
you
once
run
the
tough
home
run
with
a
higher
version,
you
will
never
be
allowed
to
move
back
to
the
previous
version,
or
it's
actually
quite
cumbersome
to
do
so.
We
that
uses
often
stake
lockdown
version
that
they
want
to
run
using
the
specific
docker
tab
and
right
now
our
users
either
have
to
use
the
official
platform
docker
images,
or
we
only
provide
single
version.
A
We
want
to
change
the
situation
and
add
support
all
of
our
nice
features
in
that
custom
image,
so
these
were
outer
formulated
developments.
Few
commands
management
created
developments
as
well.
We
sighs
the
covetous
agent
development
that
I
will
speak
a
bit
more
about,
but
basically
one
thing
we
want
to
ship
is
to
provide
a
way
quick
overview
of
the
kubernetes
instances
added
to
get
lab
by
providing
a
total
memory
and
CPU
Z
of
each
cluster
on
the
cluster
listing
page
and
that's
see
the
most
exciting
one
which
is
actually
our
kuberdas
agent
development.
A
Here,
it's
more
like
a
spike
that
we
are
going
to
run
to
see
if
we
can
set
up
an
ice
agent.
Setting
that
can
proxy
comment
between.
Oh,
it
was
just
removed.
Okay,
there
was
a
nice
image
industry
pool
yesterday
so
that
that
we
proxy
a
request
form
get
lab
to
kubernetes
and
we
are
just
figuring
this
out
the
technical
possibilities.
Rounding.
A
A
The
application
layer
Nate
needs
access
to
some
of
the
outputs
of
the
lower
layers
like
the
IP
address
or
database
URL
or
something,
and
if
all
of
these
are
in
separate
software
state
files.
This
is
called
remote
state
file
data
source
when
the
upper
layer
can
access
the
output
over
slower
layer.
We
would
like
that
support
to
the
gate.
Let
Mary
stop
or
set
as
well,
so
it
our
users
will
actually
have
a
full
stack
top
for
support
and
they
can
use
it
in
that
projects
as
well.
A
Finally,
again
waited
to
transform
is
that
heshie
Corp
is
currently
moving,
tries
to
find
owners
for
the
providers
because
they
just
realize
they
can't
support
all
of
them,
and
often
we
yourself,
where
is
requests
to
add,
support
to
this,
to
the
turf
for
mitla
provider.
We
are
currently
trying
to
take
owner
over
the
ownership
from
hash
record
so
that
we
can
provide
better
support
to
terraform
inside
of
the
plan.
And
finally,
one
issue
we
started
to
work
on
a
bit
in
the
past
month
is
to
upgrade
out
to
DevOps
to
hand
free.