►
From YouTube: Demo of the GitOps features available in GitLab Free
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
A
Gitob's
the
github's
workflow
is
finally
available
in
the
free
tier
with
the
gitlab
agent
for
kubernetes
pool
based
deployments,
also
known
as
gitops
is
becoming
the
best
practice
for
kubernetes
production
deployments
because
of
its
security
and
scalability
features.
Gitlab
shipped
pubes
deployment
support
initially
in
version
13.7,
but
until
now
it
was
a
gitlab
premium
feature
since
the
initial
launch.
We
learn
what
are
the
important
features
for
premium
and
ultimate
customers,
and
this
allows
us
to
move
the
basic
github's
functionality
to
the
gitlab
free
tier.
A
The
gitlab
agent
for
kubentus,
unlike
other
github
stores,
is
not
only
for
pool-based
deployments.
It's
the
basic
building
block
of
all
the
gitlab
commence
integrations
as
of
today.
Besides
github's
workflow,
it
supports
gitlab,
ci
cd-base
deployment
as
well
and
operational
container
scanning
in
a
single
tool.
A
A
They
are
today
right
now,
under
the
packages
directory,
we
have
a
namespace,
basically
the
agent
once
the
base
deployment
happens,
we'll
create
a
namespace
for
us
called
the
demo
agent
and
the
other
item
that
I
will
speak
a
bit
longer
as
well
is
this
config
map,
which
it's
called
agent
demo
inventory?
I
will
speak.
Why
is
it
a
special
why
it's
a
special
config
map?
I
won't
speak
much
about
drastic
about
the
airbag
in
the
third
item.
A
It's
for
other
demos
that
I
use
the
same
wrapper
for
you
can
see
here
in
the
namespace
that
there
is
no
demo
agent
namespace
in
the
namespace
list,
so
this
namespace
does
not
exist
today.
Similarly,
if
I
go
to
let's
say
all
namespaces
and
look
up
the
config
maps,
there
is
no
config
map
with
this
in
agent
demo
inventory
name.
So
these
are
the
resources
that
will
be
created
in
this
demo.
A
The
agent
is
configured
within
gitlab
in
code
itself.
You
can
see
the
configuration
file
for
this
agent
here
and
in
this
case
the
configuration
points
to
manifest
directory
and
all
the
ammo
files
from
this
directory
will
be
watched
by
the
agent
architecture
and
they
will
be
deployed
whenever
they
change
to
simplify
this
demo.
A
I
will
only
copy
the
packages
directory
where
these
three
files
exist
under
the
manifest
directory.
As
you
can
see
here,
and
we
see
how
the
agent
speaks
up,
the
files
from
under
the
manifest
repo
and
the
manifest
directory
and
applies
them
to
the
cluster.
A
A
similar
ci
job
could
run
in
theory,
a
customized
build
or
a
ham
template
command
as
well,
and
you
could
save
the
output
of
those
commands
back
to
the
repository
as
I
do
now
the
copy
to
follow
along
first.
Let
me
start
the
logs
for
this
agent.
It
will
soon
appear,
and
after
that,
I'm
going
to
commit
the
the
ci5
that
I
created
here
so
yeah
updated
fci.
That's
exactly
what
I
want
and
that's
coming
directly
to
main
branch
and
just
to
simplify
our
life
for
the
number
okay.
A
So
what
happens
first
is
that
the
ci
jobs
have
to
run,
and
by
default
it
was
set
up
to
run
on
when
the
files
under
packages
change.
So
I'm
going
to
run
it
by
hand.
A
It
will,
it
might
take
a
few
minutes
to
run
and
then
I
will
pause
the
recording
you
could
see.
The
first
loss
even
though
appeared
because
there
was
a
commit
happening
in
the
repository-
the
commit
of
this
change
that
I
made,
but
there
is
nothing
to
sing
still
the
we
said
the
objects
got
synchronized.
A
Let
me
put
in
word
wrap
a
line
wrapping
here,
so
objects
got
synchronized,
but
there
were
no
objects
actually,
at
the
same
time,
the
pipeline
is
almost
over
and
it's
over
now
and
immediately
the
agent
picked
those
up
and
seemingly
everything
worked
as
we
can
see,
there
was
a
namespace,
an
airbag
that
I
mentioned.
I
won't
speak
much
about
and
the
counting
map
as
well
somewhere
around
here
I've
seen
before
inventory
action.
No
that's
different
anyway.
There's
the
conflict
as
well,
but
let's
see
if
they
are
really
there.
A
We
are
back
to
the
namespaces
list,
as
you
can
see,
and
there
is
the
demo
agent
namespace
that
we
expect
it
to
have,
and
if
I
open
again
or
the
config
maps,
then
you
can
see
that
here
under
the
gitlab
agent
namespace,
we
have
the
config
map
from
the
repository
itself,
oops
I
this
is
one
word
you
could
say
that.
Okay,
this
concludes
the
demo,
but
actually,
I
would
like
to
give
you
a
bit
more
information
about.
Why
do
I
deploy
this
surprising
quantity
map?
A
The
agent
was
initially
built
on
shape
code
base
with
argo
cd,
but
we
decided
to
move
away
from
that
shared
code
base,
primarily
because
of
its
quality
features,
and
today
the
agent
is
built
on
something
called
cliu
tips.
Client
is
using
inventory
objects
to
keep
track
of
the
managed
resources.
This
approach
scales
better
than
simply
using
labels
on
the
applied
resources.
Moreover,
in
the
end,
the
inventor
object
is
actually
just
a
config
map
with
a
unique
hash.
A
A
If
no
inventory
object
is
available,
then
the
agent
will
actually
create
an
inventory
object
for
us
and
besides
this
really
elegant
management
of
the
resources,
click
provides
other
amazing
features.
The
most
important
one
is
very
likely.
Server
side
applies,
so
the
agent
uses
server
side
guides
out
of
the
box
without
setting
up
anything.
A
This
allows
you
to
use
the
pull-based
deployments
together
the
tools
like
horizontal
scalers,
even
though
both
resources
might
change
the
same
deployment
resource
the
fields
define
who
owns
specific
fields
and
the
server
side
of
guys
knows
how
to
handle
all
the
conflicts.
If
conflicts
arise.
A
A
We
just
finished
moving
the
basic
github's
functionality
to
get
them
free.
This
is
what
I'm
speaking
about
right
now
and
we
have
a
few
other
tools
like
adding
direct
support
to
harm
and
customize
into
the
agent
and
to
support
better
scaling
into
more
manifest
projects
like
today.
The
manifests
are
in
the
same
project
where
I
store
the
agent
configuration
itself
and
the
engine
is
registered
in
the
future.
We
want
this
to
allow
the
manifest
to
be
stored
outside
of
this
project
in
other
private
projects
as
well.