►
From YouTube: CNCF SIG App Delivery 2021-04-21
Description
CNCF SIG App Delivery 2021-04-21
B
B
A
Yeah,
we
got
just
one
question
from
kartik
who
wants
to
give
us
a
brief
update,
only
five
minutes
to
get
started
today.
If
this
is
fine
for
everybody,
but
just
be
really
brief
because
they
wanted
to
mention
the
chaos
engineering
white
paper,
they
were
just
very
highly
high
level.
Given
update
that
it
actually
exists.
It
will
really
be
pretty
short,
but
he
has
to
drop
off
early.
So
I,
if
okay
for
everybody
but
just
given
the
chance
to
just
say
this-
is
it
what
we're
working
on.
D
A
A
A
A
A
E
A
A
A
So
cornelia,
do
you
want
to
give
any
updates
on
the
github's
working
group?
You
don't
have
to
necessarily
sure
I
can
do
that.
Yeah
I'll,
put
you
at
the
end
of
today's
agenda.
I
know
this
is
not
like
we're
a
gentleman
like,
but
the
reason
is
we
have
two
project
presentations
and
we
skipped
one
already
the
last
time
and
otherwise
we
could
do
the
github's
working
group
in
one
of
the
next
meetings
as
well.
A
Good,
having
said
that,
let's
actually
get
started.
I
know
kartik,
you
have
a
hard
stop,
so
everybody
agreed
on
letting
you
go
first
discussing
the
chaos
white
paper
that
you're
planning
to
work
on
so
yeah
I'll.
Just
let
the
group
you
give
a
brief
introduction,
what
you're
up
to
where
you
can
find
the
current
draft
and
how
they
can
participate.
A
G
We're
trying
to
create
a
white
paper
for
the
state
of
chaos,
engineering
in
cloud
native
and
one
of
the
reasons
I
think
the
motivations
for
this
is
kiosk.
Engineering
is
gathering
speed
as
a
discipline,
and
the
adoption
of
kubernetes
has
just
increased
the
approach
of
chaos,
a
lot
of
folks
migrating
to
communities.
They
want
to
do
a
lot
of
fault
injections
and
they
want
to
do
chaos
engineering
in
ci
cd.
They
want
to
do
it
before
they
go
and
deploy
their
applications
into
production.
G
G
This
is
try
and
gather
folks
from
the
cncf
projects
that
are
working
on
chaos,
engineering,
litmus
chaos,
which
is
one
of
these
projects
and
also
chaos
mesh,
which
is
another
chaos
engineering
project
which
is
under
the
sig
network,
and
we
sort
of
got
together
and
exchanged
some
ideas
and
thoughts.
We
wanted
to
further
establish
what
we
mean
by
cloud
native
sub
category
of
chaos,
engineering.
G
What
are
the
general
architectural
approaches
that
we
see
in
cloud
network
house
engineering
and
also
take
a
look
at
what
problems
users
are
facing
and
what
is
the
reasons
why
they
are
not
in
chaos,
engineering
and
how
they
are
doing
it?
There
are
people
who
are
still
doing
a
game
day,
oriented
freestyle
manual,
execution
of
chaos.
There
are
folks
using
it
in
continuous
delivery
pipelines
and
the
people
who
are
integrating
it
with
detox
pipelines.
There
are
a
lot
of
ways
they're
using
it.
G
So
we
basically
are
going
to
capture
our
parts
in
this
document.
There
is
a
working
draft.
It
basically
contains
a
set
of
sections
here.
You
can
see
we
will
be
defining
the
goals
that
I
just
spoke
about.
We
will
be
trying
to
introduce
the
concept
of
chaos,
engineering
and
then
define
the
subcategory
of
cloud
native
chaos,
engineering
and
its
current
stage.
How
practitioners
are
looking
at
it.
G
We
will
summarize
the
available
cnc
for
projects
in
the
landscape
today
and
try
and
take
a
brief
look
at
the
approaches
that
we
are
taking
to
solve
this
problem
and
also
learnings
around
that
we've
got
from
the
end
user
community
some
recommendations
around
how
they
can
adopt
starch
for
our
fields,
engineering.
There
are
some
well-known
practices,
emerging
people
doing
chaos
for
observability
infrastructure
in
the
beginning
has
a
low
entry
barrier
when
they
take
into
production.
I
mean
they
take
it
to
the
other
business
apps.
G
G
It
has
books
from
the
geosphere
and
litmus
chaos
communities,
and
I
really
appreciate
really
encourage
you
to
take
a
look
at
this
and
see
if
you
would
like
to
include
more
topics
around
chaos.
Engineering
here
and
you
can
also
join
the
slack
channel-
and
this
is
open
for
comments
and
we'll
be
happy
to
incorporate
your
feedback
and
come
up
with
the
white
paper.
G
There's
also
an
issue,
so
you
can
go
ahead
and
your
thoughts
has
issue
comments
here
as
well.
We've
not
got
much
in
terms
of
content.
We
just
brought
together
last
week,
so
we
expect
that
there's
going
to
be
more
content
as
we
go
in
the
following
weeks.
We
just
created
a
rough
draft
of
how
we,
when
and
how
we
engage
and
what
might
be
some
significant
milestones
with
regards
to
this
white
paper.
So
you
can
take
a
look
at
this
and
share
your
comments.
G
A
Thanks
for
sharing
and
yeah,
as
kartik
mentioned,
this
actually
came
out
of
me
asking
him
for
the
upcoming
kubecon
europe
talk
about
what
they
are
seeing
from
a
chaos
engineering
perspective,
because
with
is
it
called
the
litmus
chaos
hub
or
or
yeah
what
they
are
seeing,
which
which
kills
experiments?
People
are
running
what
they're
doing
they're,
actually
quite
some
interesting
insights,
and
why
don't
you
team
up,
and
which
I
think
is
also
great,
with
the
chaos
mesh
team
on
on
civic
networks,
to
to
bring
this
together
into
a
more
condensed
format.
Yeah.
A
I
know
it
was
pretty
much
last
minute
that
I
told
you
you
should
bring
it
up
to
people
over
here
so
yeah,
obviously
as
soon
as
you
have
something
to
share,
feel
free
to
always
book
a
couple
of
minutes
on
on
the
white
paper.
I'd
also
recommend
you
think
with
the
operator
white
paper
team
on
best
practices.
What
worked
well,
what
did
not
work
well
from
them,
so
they
have
obviously
also
learned
a
couple
of
things
along
the
way.
What
did
work
better
than
than
other
things?
So
I
think
it's
jennifer,
thomas
or
omar.
A
I
think
they
also
help
they
can
also
help
you
hear
how
to
best
structure
it
and
get
it
aligned.
They
got
a
lot
of
praise
by
coping
what
the
security
team
should
do.
They
did
a
great
job
by
themselves
as
well,
but
I
think
it's
also
well
received
by
the
cncf
the
way
they
they
actually
structured
this,
so
I
also
would
recommend
reaching
out
to
them,
especially
when
you
talk
about
timelines,
to
get
feedback
on
what
worked
and
what
what
didn't
work.
Obviously
kubecon
eu
is
out
of
the
time
horizon.
A
B
B
I
B
C
B
We've
seen
with
a
lot
of
users:
they're
they're,
getting
tired
of
hearing
about
digital
transfer,
vendors
they're,
also
getting
tired
of
hearing
about
transfer
easy,
just
modernize,
your
apps
modify
your
processes,
change
your
culture
and
the
actual
practitioners
doing
this
work
are
really
just
still
asking
me:
what
does
that
practically
mean?
I
need
to
do
in
my
day-to-day
life
and
in
my
application,
architectures
and
things
like
this,
and
so
in
addition,.
C
B
Vendor
has
a
different
methodology
for
how
to
do
this,
how
to
re-host,
re-platform
refactor
your
applications
and
all
the
thing
that
they
bring
oftentimes
is
proprietary,
and
so
what
we've
found
in
in
the
in
the
community,
as
we've
become
faulted
with
this
community,
is
that
what
people
the
practitioners
are
actually
hungry
to
learn
about
is
how
people
are
doing
this.
How
they're?
Actually
you
know,
for
example,
string
strangling
a
monolith
or
car
office
on
a
car
container
or
you
know,
containerizing
their
their
applications.
B
So
karthik,
thanks
for
presenting
on
the
chaos
engineering
topic,
but
really
this
is
about
trying
to
drive
more
of
an
ongoing
conversation
from
practitioners
to
other
practitioners
about
how
they
are
going
through,
taking
their
applications
and
getting
them
to
kubernetes
and
and
to
use
the
cncf
projects
more
regularly.
So
that's
kind
of
the
mission
statement
is
bringing
together
community
people.
Let's
build
these
tools,
share
and
share
knowledge
and
best
practices
on
how
to
re-host
re-platform
and
refactor
your
apps
to
kubernetes
inside
that
conveyor
community.
B
We
have
really
five
projects
right
now
that
there
are
there's
active
development
on
they.
You
can
see
them
here
from
right
to
left,
so
we
have
the
forklift
project,
which
is
essentially
about
re-hosting,
your
versions
to
cube
verb.
So
this
allows
you
to
actually
import
virtual
machines
and
maps
into
the
cube
project.
We
have
the
crane
project,
which
is
about
re-hosting
our
applications
between
kubernetes
clusters.
B
It
actually
leverages
the
valero
project
to
do
a
lot
of
the
backup
and
restore
of
the
kubernetes
objects,
and
then
it
also
helps
with
moving
persistent
data
between
clusters,
move
to
cube.
This
was
recently
open,
sourced,
actually
open
sourced
by
ibm
research
and
placed
into
the
community.
It
is
a
tool
that
helps
you
re-platform.
Your
applications
from
you
know
other
content
orchestrations.
B
So
you
know
cloud
foundry,
swarm
and
other
technologies
where
you're
already
containerized,
but
you
want
to
move
to
kubernetes
the
tackle
project.
This
is
a
new
project
that
has
just
been
kicked
off
and
is
is
scheduled
to
have
their
first
community
release
in
the
june
time
frame,
and
it
is
about
helping
people
assess
and
analyze
their
applications
for
container
suitability.
B
So,
there's
a
there's
actually
a
manual
data
entry
point
for
entering
the
characteristics
of
the
application,
there's
a
shared
application
inventory
and
then
there's
a
static
code
analysis
piece
that
can
actually
determine
you
know
how
well
your
applications
are
to
be.
You
know
confirmed
to
the
follow
factor,
app
kind
of
mindset
and
then
all
the
way
over
to
the
right
is
polaris.
This
was
actually
developed
by
some
folks
in
a
consulting
organization
to
help
measure
the
dora
metrics.
B
So
you
know
the
you
know
the
mean
time
to
recover
your
change,
failure
rate
and
so
on,
and
this
is
a
project
that
actually
instruments
across
various
other
projects
running
on
top
of
openshift
to
measure
your
actual
software
delivery
performance.
So
those
are
kind
of
the
the
five
projects
that
we're
working
on.
You
know
it's
very,
I
would
say
early
days,
they're,
all
in
different
stages,
move
to
cube
has
releases
tackle,
will
have
its
first
community
release
in
june.
As
I
mentioned,
crane
exists
and
is
available.
B
There's
a
there's
a
link
at
the
bottom
I'll
share
these
slides.
You
can
look
back
and
provide
feedback
if
you're
interested
and
then
there's
also,
you
know
a
contributor
ladder
approach
as
well.
That's
in
there
as
well,
so
I
I
did
want.
I
just
want
to
keep
it
really
brief
and
kind
of
give
you
guys
an
introduction
and
take
actually
open
it
up
to
more
feedback
or
questions
or
comments.
We're
on
the
kubernetes
slack
channel
so
pound
conveyor
on
slack
that
kubernetes
dot
io.
B
If
anybody
wants
to
ever
propose
a
meetup
we're
very
open
to
it,
we
have
some
pretty
simple
rules
which
are:
don't
don't
contain
proprietary
technology
in
your
demonstration.
You
know
make
it
demonstration
based
so
that
you
can
show
people
things
and
yeah,
that's
pretty
much
it
and
then
joining
the
quarterly
planning
meeting.
So
we're
we're
starting
to
have
quarterly
planning
meetings
for
all
the
projects.
B
At
the
same
time,
just
to
kind
of
get
everybody
engaged
so
that
if
there
are
integration,
points
or
sharing
that
we
can
do
between
the
teams,
that
happens
and
it's
open
to
anybody.
So
we
have
you
know
some.
Some
financial
services
institutions
that
are
gonna
be
joining
us
in
the
june
time
frame,
as
well
as
I'm
from
red
hat,
so
the
red
hat
engineering
team
will
be
there
ibm
research
and
some
others.
But
with
that
maybe
I
can
stop
my
screen
share
and
ask.
B
B
H
C
B
Yeah
yeah,
so
that's
essentially
kind
of
a
it
has
an
application
inventory
that
applications
can
be
added
to
whether
it's
you
know
through
human
interaction
or
an
automated.
You
know
machine
to
machine,
and
so
that
application
inventory
is
then
good.
There's
an
assessment
piece
that
that
is,
like
you
know,
asking
questions
around.
You
know
the
human
factors
around
the
application
that
you
would
answer
and
then
there's
an
tackle
analyze,
which
is
like
the
piece
that
hooks
into
that
application
inventory
to
do
code
analysis.
B
Is
that
the
point
so
yeah?
So
the
idea
is
to
make
the
recommendations.
We
eventually
want
to
get
to
the
ability
to
start
to
decompose,
but
that's
a
really
difficult
problem,
so
we're
starting
with
like
assess
and
analyze
and
then
we're
actually,
the
the
ibm
research
team
is
actually
open
to
open
sourcing.
Several
other
tools,
they're,
actually
working
on
that
right
now
to
open
source
a
couple
of
tools.
One
of
them
is
called
application
container
advisor.
It
uses
natural
language
processing
to
see.
B
C
H
A
A
We
just
said
we're
not
initially
going
to
work
on
it
because
we
didn't
have
frankly
quite
an
idea
of
what
what
we
could
provide,
but
this
looks
interesting
from
from
a
tooling
perspective.
So
is
your
plan
then
also
to
have
conveyor
as
a
cncf
project,
or
do
you
want
to
just
keep
the
community
closely
related
to
what
the
cncf
is
doing?.
B
Yeah,
it's
a
it's
a
really
good
question.
I
I
mean
the
the
guidance
I've
been
giving
to
all.
The
teams
in
the
community
is
like,
let's,
let's
get
a
lot
of
people
using
these
tools
and
I
think
we
could
sort
out
most
the
other
things
later
right
like
it's,
not
necessarily
something
that
needs
to
you
know
happen
right
away,
but
of
course
I
think
I
I
also
don't
know
the
cncf
well
enough
and
so
josh
burke
is
you
know
he
works
closely
with
with
a
lot
of
the
cncf.
B
Sigs
has
been
giving
me
guidance
on
this.
I
you
know,
but
I'm
open
to
you
know.
Is
it
conveyor
that
would
belong
or
is
it
the
each
sub
project?
You
know
that
would
become
one.
You
know
a
sandbox
project
in
the
future
or
something
if,
if
that
should
come
to
be
right,
so
I
don't
know
if
it's
like
the
whole
community
or
if
it's
the
projects
right
individually,
but.
A
So
we
had
obviously
project
portfolios
in
the
past
from
like
especially
project
maturity
time,
some
kind
of
gets
hard.
If
you
have
a
lot
of
individual
projects
that
you
have
to
kind
of
all,
more
or
less
right
do
the
due
diligence
on
together,
especially
if
they
are
in
a
different
phase
of
maturity.
It
it's
not
impossible,
but
the
more
they're
getting
the
harder
it
obviously
is
having.
A
A
C
G
F
B
So
yeah,
I
think
I
think
you
know
what
the
reason
I
we
kind
of
like
created
this
conveyor
umbrella
was
because
we
recognized
that
migration
tools
aren't
necessarily
long
lived
like
platform
technologies,
and
so
you
know
one
you
know
everybody's
going
to
re-host
their
applications
to
you
know
a
kubernetes
or
cncf
model
or
what
or
what
have
you?
And
then
you
know
they're
going
to
have
a
continuous
refactoring
still
going
on,
but
those
tool
sets
might
change
over
time.
B
So
we
wanted
kind
of
the
community
brand
of
this
area
that
is
designed
to
for
people
that
are
practitioners
that
are
constantly
modernizing
apps.
But
we
know
that
you
know
the
tools
inside
may
end
up.
You
know
they
might
have
a
maturity
or
a
curve
to
them
where
eventually
they're
not
used
to
three
years
down
the
line
and
there's
new
tools
that
are
available.
If
that
makes
sense,
so
yeah.
F
B
Yes
yep,
so
it's
specifically
it's
ibm
research
and,
and
then
the
red
hat
team.
That's
why
we
put
that
contributor
ladder
in
place
in
governance
and
now
we're
starting
to
go
out
and
engage
with
some
of
our
even
our
customers
and
users
to
encourage
them.
To
start
to
so,
there's
going
to
be
some
press
coming
out
around
kubecon
around
this
community
and
like
us,
trying
to
kind
of
catalyze
this
community
and
get
more
contribution
into
it.
A
It
is
a
topic
that
many
people
are
concerned
with
this,
maybe
also
other
groups
within
the
cncf
that
they're
interested
in
is
a
dedicated
group.
That's
more
targeted
towards
this
business
and
digital
transformation
focus
areas
so
yeah.
Let
let
better
think
a
bit
about
it,
but
it's
definitely
interesting.
I
think
it's
definitely
worth
while
having
more
of
a
discussion
about
this.
B
That's
a
good
question.
I
I
don't
think
at
this
point.
I
think
this
is
really
the
polaris
project
is
really
kind
of.
Was
a
skunk
works
project
kind
of
created
by
some
consultants
trying
to
measure
this?
For
specific,
you
know
users
and
customers,
so
I
think
it
could
definitely
relate
to
it,
and
I
don't
know
if
there's
anything
is
there
a
specific
project
inside
of
the
cncf
is
that
the
four
keys
project.
B
E
B
A
B
I
Hello,
everybody,
I
guess
everybody
can
see
my
screen
perfect,
all
right,
yeah,
I'm
here
today
to
present
to
you
lagoon
well,
what
is
the
mission
of
lagoon
so
lagoon
tries
to
drastically
reduce
the
time
required
to
bring
your
application
to
kubernetes
and
not
only
that,
but
also
with
all
the
configuration,
tooling
and
insights
to
actually
run
your
application
at
scale
and
securely
now?
Why
did
we
create
lagoon?
And
I
will
shortly
explain
later
who
we
are?
But
let's
look
at
what
are
the
actual
problems?
I
Then
there's
also
no
automation
out
of
the
box,
meaning
that
if
you
maybe
have
a
cli,
you
need
to
write
like
a
github
actions
or
a
git
lab
or
a
jenkins
integration
to
deploy,
and
that's
just
very
hard
for
some
and
for
many
people
out
there,
then
also
there
is
no
real
good
best
practices
around
base
images,
meaning
which
images
to
actually
use
to
deploy.
Yes,
there
is
dockerhop,
but
they
are
maybe
not
secure
or
they're,
not
maybe
building
the
best
practices.
I
I
So
we
as
a
company
we
work
with
a
lot
of
developers,
and
so
we've
built
the
tool
lagoon
and
lagoon
is
fully
developer
focused.
What
do
we
mean
and
first
of
all,
you
don't
need
any
kubernetes,
yaml
knowledge
or
not
even
kubernetes
access
directly.
You
can
learn
it.
You
can
look
at
what
actually
lagoon
creates,
but
there's
no
need
out
of
the
box
to
learn
some
yammer.
I
We
also
have
local
development
included,
meaning
you
can
run
the
same
environment
locally
and
as
also
in
production
or
in
development,
meaning
you
have
the
same
containers,
the
same
images,
the
same
system,
which
is
hugely
important
for
application
developers
that
need
to
recreate
problems
that
they
face
in
the
kubernetes
clusters
on
their
local
and
it's
fully
built
around
githubs
and
infrastructures
code.
What
does
that
mean?
There's
no
cli
that
you
need
to
run
or
install
locally.
I
All
you
do
you
just
push
your
code
in
git
repository
and
from
the
git
depository
automatically
environment,
for
each
git
branch
for
each
pull
request
will
automatically
be
created
in
your
kubernetes
cluster.
We
have
a
full-fledged
ui
that
has
all
the
same
functionalities
as
the
api
and
the
cli.
So
also
people
like
stakeholders,
project
managers-
they
all
have
also
access
to
it.
They
can
see
what
is
running
and
not
only
the
developers
that
used
to
have
a
token
that
have
the
cli.
I
Then
lagoon
also
includes
space
images
for
all
different
type
of
applications
that
are
focused
on
security
ease
of
use.
They
have
best
practices
in
them
that
you
usually
don't
find
on
the
regular
docker
hub
images
and
lagoon
is
multi-kubernetes
cluster
capable
meaning
it
can
deploy
in
a
lot
of
kubernetes
clusters
at
the
same
time,
and
it's
only
fully
end-to-end
tested.
That
means
we
actually
start
up
a
kubernetes.
We
deploy
different
applications
in
it.
We
test
all
the
functionalities
before
we
release
new
images,
new
new
lagoon
versions
and
logging
and
backups
are
automatically
included.
I
I
Now,
how
does
lagoon
fit
into
a
whole
hosting
stack
and
well?
In
the
end,
we
are
fully
based
on
kubernetes,
so
lagoon
just
has
the
requirement
to
have
a
kubernetes
somewhere
installed
and
that
somewhere
can
be
pretty
much
anywhere.
As
of
today,
we
have
lagoon
running
in
eks,
gk,
aks
and
k3s,
but
in
the
end
there
is
no
specific
requirement
on
the
infrastructure
so
and
we
have
people
that
are
working
and
working
on
getting
and
running
on
digitalocean
alibaba
cloud
wherever
you
want
to
do
it
and
lagoon
right
now
is
focused
on
web
applications.
I
So
it
assumes
a
little
bit
that
you
actually
want
to
show
a
website,
but
from
a
technical
point
of
view,
lagoon
just
deploys
containers,
so
you
could
also
run
some
machine
learning
or
any
other
application
and
that
you
can
run
in
containers
and
we
have
existing
templates
for
very
well-known
website
frameworks
like
drupal
present
and
we're
in
the
progress
to
creating
much
more.
This
means,
if
you
have
one
of
these
sites,
you
basically
just
copy
paste
a
couple
of
configuration
files
in
your
repository
in
your
push
and
it
deploys
automatically
if
you
use
an
application.
I
That
is
not
that
we
don't
have
templates
yet
you
need
to
config
them
a
bit
more
than
normal,
but
it
should
be
quite
easy
to
get
your
site
running
with
lagoon.
Now.
How
does
the
lagoon
architecture
work?
I
don't
want
to
do
a
full,
an
intro
into
how
or
every
single
piece,
but
basically
we
follow
a
core
and
remote
or
ancient
structure,
meaning
a
developer
pushes
into
any
kind
of
git
repository
that
can
be
github,
kit,
lab
bitbucket
or
whatever
else,
and
this
then
informs
the
lagoon
core
or
the
control
plane.
I
If
we
use
kubernetes
terminology
about
this
and
then
lagoon
core
talks
to
the
lagoon
remote
that
is
actually
running
in
a
kubernetes
cluster.
So
that's
an
agent
that
lagoon
remote
that
you
install
it's
an
operator
based
on
the
operator
framework
that
basically
connects
back
to
the
lagoon
core
meaning.
You
can
also
run
this
kubernetes
cluster
in
an
air-gapped
environment,
because
the
connection
is
done
from
from
the
lagoon
remote
to
the
lagoon
core
and
the
lagoon
core
basically
tells
the
lagoon
remote
hey.
I
There
is
a
new
deployment,
it
tells
it
which
project
which
name
space
it
should
create,
etc,
etc,
and
then
that
lagoon
remote
creates
a
lagoon
build
pod
and
the
lagoon
build
pod
will
actually
clone
the
git
repository.
That
means
the
git
repository
never
actually
goes
touches
the
lagoon
core.
So
the
security
of
the
code
itself.
A
A
A
I
Long
you
keep
talking
until
you
realize
that
nobody's
going
to
hear
you
anymore
well,
the
strangers
think
I
heard
you
saying
we
lost
him
and
I
was
like
no.
I
can
still
hear
you
anyway,
okay
yeah,
so
what
I
wanted
to
say
is
that
it's
really
important
that
the
git
repository
is
directly
cloned
into
kubernetes
cluster
and
that
lagoon
build
parts.
Then
this
is
the
actual
piece
that
then
creates
talks
with
kubernetes
cluster.
So
it
creates
all
the
objects
that
are
required
stuff
like
that.
I
If,
when
the
deployment
is
done,
it
then
informs
the
lagoon
core
again
saying:
hey
I've
deployed
what
you
told
me
and
then
lagoon
core
informs
the
developer
back.
So
we
have
different
notification
channels
that
can
be
slack
rocket
chat
and
you
can
see
the
emails
things
like
that.
So
everything
that
the
developers
to
do
is
just
push
into
a
git
repository.
I
I
If
something
can
opens
be
open
source,
it
should
be
open
source.
So
we
open
sourced
it
in
august
2017
and
in
august,
2019
did
a
big
rewrite
and
with
full
role-based
access
control,
support,
meaning
that
we
could
actually
find
who
has
access
to
what
before
it
was
like.
Everybody
had
access
to
everything
and
yeah.
So
we
implemented
that,
then,
in
a
during
that
time,
we
realized
that
we
had
more
and
more
requests
for
just
kubernetes
and
not
open
drift.
So
we
then
implemented
this
in
beginning
2020.
I
I
Now,
how
is
lagoon
connected
to
the
cncf
lagoon
already
uses
a
lot
of
tools
that
are
either
cncf,
incubating
sandboxes
or
just
kubernetes
tools
and
under
the
hood
we
use
helm.
We
use
fluenty
promises,
grafana,
we
use
harbor
with
the
image
scanning
we
have.
We
use
the
open
policy
agent.
We
use
the
logging
operator
from
banzai
cloud.
I
We
use
ktop
for
backup,
so
we
really
try
to
use
everything
that
already
exists
or
if,
even
if
it's
only
like
an
80
or
70
fit,
we
use
an
already
existing
tool
and
maybe
contribute
and
make
it
better
than
implementing
it
ourselves
and
yeah
from
amazio.
We
are
fully
committed
to
donate
lagoon
as
a
sandbox
and
follow
the
guidelines
that
are
required.
That
also
includes
we
have
dedicated
lagoon
product
leads,
and
his
name
is
toby
he's
in
australia
right
now
and
it's
2
a.m.
I
So
we
decided
I'm
going
to
present
this
and
we
have
developer
advocates.
We
have
product
designers,
we
have
engineers
that
purely
solely
work
on
lagoon
and
make
sure
that
it
continues.
Amazing
always
has
a
marketing
team,
so
amazing
will
also
support
lagoon
the
dragoon
team
for
marketing
and
very
exciting.
I
What's
ahead,
I
mentioned
the
cncf
sandbox.
We
are
we're
finishing
moving
everything
into
the
use
lagoon,
github
organization.
It
used
to
be
still
inside
the
amazio
and
we
want
to
give
lagoon
its
own
space,
we're
working
on
its
own
website
and
to
also
disconnect
it
more
from
a
cio,
and
then
we
will
apply
as
a
cntf
sandbox
and
we're
looking
into
creating
more
templates
for
other
applications
and
we're
looking
into
like
communities
like
the
type
of
three
or
other
communities
to
reach
out
to
them
and
learn
how
exactly
we
can.
I
We
can
buy
build
these
templates.
There's
a
huge
security
push
right
now
as
we're
working
with
a
lot
of
governments.
They
have
really
strict
security
requirements,
so
we're
working
on
running
all
the
containers,
with
rootless
by
default.
Enforcing
network
policies
have
gatekeeper,
use
more
the
capabilities
of
harbor
like
image,
scanning
and
stuff,
like
that.
We're
also
working
a
lot
of
on
portfolio
management.
I
That's
basically,
if
you
have
400
websites
that
you
need
to
deploy
and
that
you
can
say,
redeploy
me
all
the
sites
and
it
automatically
does
it
for
you
and,
like
I
said,
we're
also
focusing
on
managing
and
maintaining
lagoon,
meaning
that
if
somebody
wants
to
use
it
that
there's
documentation
how
to
upgrade
how
to
handle
issues
and
things
like
that
and
overall
we're
just
working
on
creating
a
community
around
lagoon.
And
that's
why
we're
here,
yeah
and
if
you
wanna,
learn
more
and
lagoon.sh
is
the
website
that
currently
redirects
to
me
cio.
I
A
I
I
but
it's
it's
a
us-based
company
that
basically
does
hosting
for
a
specific
cms,
and
they
they
also
come
from
a
world
where
they
like
spin
up
ec2
instances
for
every
single
site
and
they've
only
used
containers
and
they're.
Now,
looking
into
a
way
how
to
modernize
their
infrastructure
with
kubernetes.
I
Yes,
and
no
so
we
definitely
see
some
interest
of
hosting
providers,
because
they
are
basically
in
the
business
of
bringing
people
to
their
hosting
platform
and
giving
your
customers
that
you
don't
really
know.
Access
to
your
kubernetes
cluster
is
always
still
a
bit
freaky
and
especially
if
you
have
like
a
multi-tenant
infrastructure,
kubernetes
cluster,
so
because
of
that
the
int
or
the
the
fact
that
you
don't
need
to
give
customers
access
to
the
community's
cluster
is
very
interesting
to
them.
I
We
do,
though,
see
also
like
web
agencies,
so
customers
that
build
a
lot
of
websites
for
a
lot
of
different
customers,
and
they
need,
like
development
kubernetes
cluster,
that
they
start
to
use
this
or
whole
companies.
That
say,
like
we
have
a
university
in
in
germany
that
uses
it,
which
said
we
want
to
give
our
all
the
different
departments.
They
all
have
websites
that
they
need
to
host
and
we
want
to
give
them
access
to
our
kubernetes
cluster
and
again
the
same
problem.
Is
they
have
no
idea
about
kubernetes?
I
A
F
F
I
was
waiting
and
I
I
I
definitely
noticed
that,
for
example,
flux
wasn't
on
your
list
of
technologies
but
you're
doing
a
whole
bunch
of
things
that
flux
does
like
cloning
get
repositories
and
sending
out
notifications
and
tooling.
All
that
up
sure
yeah
would
love
to
see
if
you're
interested
in
that.
I
Yeah,
I
mean
that's,
definitely
something
that
why
we
want
to
also
go
into
the
cncf
to
maybe
actually
rip
some
of
the
code
that
we
currently
do
rip
it
out
and
replace
it
with
existing
other
tools,
and
we
really
see
lagoon
as
like
a
tool.
That's
basically,
a
bit
opinionated
combines
existing
cncf
projects
together
into
something
that
you
can
install
and
in
one
click
deploy
and
and
so
yeah
we,
the
the
less
code
we
need
to
maintain
the
better
yeah
omar
also
mentioned
argo
yeah.
We
also
look
at
our.
I
I
A
I
Yes,
there
have
been
many
other
ideas
how
to
do
this,
also
how
to
run
like
it
and,
of
course,
like
other
docker
demons
like
potman,
came
up
or
container
d
that
didn't
exist
in
the
beginning,
and
some
of
them
also
allow
you
to
like
root
less
and
things
like
that.
So
that's
definitely
another
piece,
that's
very
interesting
to
come
to
collaborate
and
figure
out
better
ways
to
do
it
than
just
yeah
start
the
docker
in
docker
all
the
time.
A
Yeah,
because
they,
I
think,
I
think
they
are
they
more
than
follow
the
heroical
pattern
of
like
a
lot
of
those
and
the
people,
the
pattern
of
a
lot
of
those
things
yeah.
So
another
question:
how
complex
can
the
applications
be
that
you
run
with
lagoon?
You
mentioned
mostly
websites,
which
are
not
say
that
complicated
from
a
structure.
Software,
however,
run
a
complex,
multi,
complex,
complex,
microservice
architecture
that
consists
of
a
multitude
of
services.
Is
this
something
you
when
you
run
an
opinionated
approach?
A
I
But
in
the
end
there
is
no
limit
and
on
how
many,
let's
say,
parts
or
containers
you
could
deploy
inside
one
git
repository
actually
lagoon,
one
uses
lagoon
to
deploy
lagoon
and
we
wanted
to
do
a
bit
inception
and
eat
our
own
dog
food.
We
realized,
though
this
is
then
very
hard
to
tell
people
how
to
install,
because
if
you
don't
have
a
lagoon,
how
do
you
install
a
lagoon?
I
So
that's
why
we
now
actually
moved
into
helm
so
to
run
lagoon
core
and
remote.
You
just
use
helm
and
but
lagoon
itself
has
around
25
micro
microservices
that
worked
together
and
that
was
deployed
by
lagoon
and
it
worked
without
a
problem,
so
you
can
definitely
use
it
for
more
complex
and
we
have
some
customers
today.
A
A
So
it's
good
ops
in
itself,
which
is
kind
of
an
exception,
but
you
have
to
to
be
fair.
You
have
to
deploy
the
operator
via
helm,
which
obviously
makes
sense.
Yes,
yeah
yeah.
Another
project
I
would
propose
to
look
at
how
I
think
we
will
run
out
of
time
today
is
how
you
actually
model
these
applications,
because
om
is
definitely
a
project
to
look
into
thomas
who's.
Also,
here
will-
and
I
think,
the
rest
and
lx
who
look
into
like
an
enable
application.
A
Enablement
working
group
here
because,
like
there
are
lots
of
opinionated
approaches
out
there
or
that
are
getting
started
right
now,
which
I
think
is
great
and
useful.
I
think
just
some
interoperability
at
some
point
would
be
because
there's
also
a
community
to
engage
with.
They
did
not
present
today,
but
the
last
time
so
happy
to
share
things
there,
but
yeah.
I
see
it's
definitely
becoming
more
and
more
like
different
types
of
templating
mechanisms.
A
Oem
cube
velar
would
be
like,
I
think,
the
most
similar
project
that
I
could
think
of
would
be
cube,
velar
most
likely,
which
is
also
that
at
an
early
stage
to
to
maybe
engage
with
not
saying
that
you
have
to
as
a
sandbox
project.
This
is
also
obviously
to
to
to
provide
some
some
some
feedback
here,
but
I
think
it's
good
to
see
this
project
emerging
and
again.
Also,
if
you
need
connections
to
those
projects,
we
can
obviously
also
help
there
as
well.
I
A
Okay,
I
think
people
just
we
are
out
of
questions
right
now
I
mean
the
criteria
for
sandbox.
Are,
I
think,
pretty
straightforward
and
you're
doing
some
of
the
right
things
like
putting
it
into
its
own
organization
right
now,
it
has
a
bit
of
a
feel
that
it's
an
amazing
project
and
it's
more
or
less
like
an
open
core
which
isn't
a
bad
thing
thing
either
or
a
bit
more
than
open
core
here.
A
I
Yeah,
no,
we
are
fully
aware,
like
that,
was
actually
we
talked
about
like
almost
a
year
ago
to
donate
it
as
sandbox,
and
at
that
point
it
was
still
like
too
much
an
amazing
io
thing,
but
now,
with
lagoon
2
like
yeah,
we
actively
have
work
where
we
remove
everything
and
make
sure
that
other
people
can
use
it
plus.
We
now
have
like
the
companies
that
I
mentioned
there
now,
starting
to
actually
use
it,
deploy
the
first
test
project.
I
So
we
have
a
real
world
example
that
other
people
beside
of
us
can
actually
use
it,
because
we
obviously
have
like
huge
imposter
syndrome
that
we
feel
like.
No,
it's
always
going
to
be.
You
can
never
run
this
outside,
but
seeing
others
actually
do
it
on
a
daily
basis
is
pretty
cool,
and
I
think
this
gave
us
also.
They
believe
that
what
we
have
built
can
be
also
used
by
others
or
is
interesting
to
others.
A
F
Sure,
yeah
and-
and
I
have
a
hard
stop
at
the
top
of
the
hour,
but
it
will
only
take
a
couple
of
minutes,
so
the
get
ups
working
group
just
a
couple
of
updates,
I
think
most
people
know
that
we
are
doing
a
get
ops
con.
The
get
ops
working
group
is
hosting
that,
and
that
is
a
day
zero
event
at
hoop
con
virtual.
F
Of
course,
we
just
discovered
yesterday
that
we
completely
dorked
up
the
times
and
we
had
it
scheduled
for
like
5
p.m,
to
10
p.m,
central
european
time,
and
so
we're
nudging
that
we're
splitting
the
difference
so
that
west
coast
people
don't
have
to
get
up
at
one
but
could
get
up.
You
know,
can
participate
at
five
or
six
in
the
morning,
so
we're
making
an
adjustment
on
that
call
for
papers
closed
on
friday.
F
F
Talk
because
we
want
to
make
sure
that
we
are
linking
these
two
communities
and
the
the
cd
con
has
a
day,
zero
event,
which
is
also
git,
ops,
focus
so
I'll
put
in
a
little
plug
for
that,
as
well
called
the
get
ups
summit.
I
think
so
that's
good
stuff
happening
like
I
said
the
agenda
will
be
posted
in
the
next
day
or
so
and
we'll
be
letting
people
know
about
that.
Those
submissions
and
acceptances.
F
The
second
thing
that
the
team
has
been
working
on
really
significantly
is
the
print
are
the
principles.
There
is
a
pull
request
in
the
get
get
ups
working
group
repository
that
defines
the
principles,
and
so
there
has
been
both
asynchronous
collaboration
through
that
pr,
and
that
is
the
primary
thing
we
have
been
hosting
a
number
of
synchronous
meetings
where
we
can
have
live
conversations.
Those
are
all
recorded
and
posted,
and
the
results
of
all
of
those
things
are
folded
back
into
the
ppr.
A
F
Oh
and
the
other
thing,
maybe
that
I
haven't
mentioned,
is
that
we
did
tease
apart
when
we
submitted
when
we
created
the
get
ups
working
group
we
submitted
as
a
sandbox
project,
which
is
kind
of
a
weird
thing
to
put
with
the
getups
working
group
and
the
toc
said.
We
don't
like
the
name
github's
working
group,
but
we're
not
going
to
worry
about
that.
They'll
straighten
that
out.
Well,
the
confusion
was
pretty
big
right
from
the
get-go,
and
so
we
have
teased
that
apart.
F
We
now
are
naming
the
sandbox
project
open,
get
offs.
The
working
group
cares
for
open
get
ops
and
that
sandbox
project
is
the
place
where
all
the
long,
lasting
artifacts,
like
the
principles
like
educational
materials,
like
any
code
samples,
things
like
that,
will
will
be
housed
under
that,
and
so
we
have
started
to
clean
up
some
of
the
mess
that
we
made
when
we
came
in.
A
F
Me
that
one
last
thing
is
that
get
at
kubecon
we
did
have
a
talk,
get
accepted
for
just
the
get
ups
working
group
update,
which
is
something
that
chris
sanders
from
microsoft
and
I
recorded
just
earlier
this
week.
We
were
the
bad
folks
that
got
it
in
way
late,
but
we
did
get
it
in
on
monday,
so
we
recorded
that,
and
it
only
was
about
a
20-minute
recording,
because
we
wanted
to
leave
lots
of
time
for
people
to
ask
questions.