►
From YouTube: Helm Project 2022: How You Can Benefit, Ho... Scott Rigby, Matt Butcher, Martin Hickey, Andrew Block
Description
Don’t miss out! Join us at our upcoming hybrid event: KubeCon + CloudNativeCon North America 2022 from October 24-28 in Detroit (and online!). Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Helm Project 2022: How You Can Benefit, How You Can Help - Scott Rigby, Weaveworks; Matt Butcher, Fermyon; Martin Hickey, IBM; Andrew Block, Red Hat
Helm is a widely used and stable CNCF project that manages packages for Kubernetes, and will continue to dependably deliver what users have grown to love. So, which direction will Helm go in 2022 and beyond? In this session, Helm maintainers will show you why you should continue to be excited about Helm, and what lies ahead. We’ll highlight key features that you can take advantage of – both tried and true features users have grown to love, as well as newly matured features from over the past year. For the dynamic future of Helm, we’ll cover how interested users like you can keep the momentum going by participating and even contributing to the project. We’ll preview what this looks like, from simple tasks that most users can get started with, to helping shape the vision of Helm 4.
A
Okay,
folks
welcome
to
this
evening's
cabaret,
so
today's
talk
we're
going
to
go
around
is
it's
the
helm,
maintainer
track
and
we're
looking
at
where
we
are
now
where
we're
trying
to
go
next,
how
people
can
get
involved?
So
before
I
kick
off,
can
everybody
stand
up?
Please
we're
going
to
go
with
this
whole
party
trick.
Okay,
now
we
we've
been
away
a
long
time,
and
this
is
my
first
time
on
the
road.
It's
my
first
time
back
at
cubecon
since
2010
hit
me
out
here:
19
yeah
2019.!
A
A
A
So,
let's
grab
a
seat.
My
name
is
martin
hickey.
I
work
over
at
ibm
and
my
fellow
co-presenters
today
are
andrew
block
who's
over
at
red
hat
scott
rigby
who's
over
at
we've
worked
and
matt
butcher
who's
over
at
furman
and
the
ceo
of
hermann
so
welcome
right,
so
basically
jinder
what
we're
going
to
do
today.
A
So
let's
kick
it
off.
So
I've
put
up
here,
two
sentences
taken
from
the
helm
website
called
helm.sh
you're
looking
for
us,
and
I
suppose
everyone
has
that
definition
of
that
helm
is
the
package
manager
for
kubernetes
and
I
think
that's
a
really
really
good
definition,
but
what
does
it
mean?
So
I
suppose
what
it
means
is
it
allows
you
to
take
these
packages
or
charts,
which
is
the
package
format
and
for
that
helm
uses
to
deploy
apps
out
into
the
cluster.
So
it
allows
you
to
do
that.
Crud,
your
you
know.
A
Your
credits
up
to
its
elites
of
the
particular
apps,
but
also-
and
probably
it's
one
of
the
key
parts
of
it
is-
is
that
ability
to
share
this
ability
to
be
able
to
deploy
an
app
and
we've
talked
about
this
a
few
times.
Man
haven't
we,
where
you
know
people
often
take
helm
to
understand
what
the
name
of
goodness
is
kubernetes.
Don't
underneath
the
hood,
you
know,
because
you
can
just
take
a
chart
and
we
would
see
it
in
a
while.
You
can
just
go
helm,
install
and
bang.
A
A
Now,
if
you're
a
chart
maintainer
and
all
that,
yes,
you
understand
that
okay,
so
what
I'm
going
to
do
is
I'm
going
to
do
a
little
example
that
a
former
colleague
of
mine
did
oh
a
good
few
years
ago
and
he's
a
lovely
video
online
you'll
see
the
link
down
there.
It's
a
nice
little
nine,
nine
minute,
video!
A
Now
it's
slightly
out
of
date,
because
he
was
more
pitched
for
him
too,
when
the
architecture
was
slightly
different
and
the
commands
are
slightly
different,
but
it's
still
the
same
concept,
but
I'm
going
to
grab
his
id
of
this
e-commerce.
So
he
the
idea
of
an
e-commerce
on
the
holidays.
You
might
you
know
you
might
have
you
might
have
scaled
it.
You
might
increase
the
replicas
on
it.
So
for
the
amount
of
traffic
that's
going
on
now,
my
one
isn't
going
to
be
that
fancy
we'll
see
in
a
minute.
A
What
is
the
image
of
this
app
called?
But
the
idea
here
is
this
is
kind
of
the
design
you're
looking
at
so
we'll.
We
will
just
imagine
for
this
that
the
mongodb
has
been
deployed
into
the
database
already
and
that's
for
your
persistence
and
what
we're
going
to
deploy
here
is
two
replicas
of
our
node.js
app
and
then
we're
going
to
use
a
service
object.
So
we
can
talk
to
our
app
basically
so
that
it
can.
We
can
reach
it
from
outside
your
cluster.
A
So
if
you're
going
to
do
this
normally
using
kubernetes
and
most
of
your
experts,
here,
I
would
say-
and
if
you're
not
it's
around,
these
manifest
files
so
right
near
yaman,
foils,
whether
you
like
camel
whether
you
don't
like
yaml.
That's
probably
a
lot
of
discussions
on
this
one.
Some
people
do
some
people,
don't.
D
A
So
if
any,
you
have
kids
out
there,
you
might
have
to
go
to
the
center
every
year
and
see
where
santa
is
at
that
time
of
year.
So
we'll
put
that
out,
there
we've
two
replicas
on
it.
With
the
the
service
object,
we're
going
to
go
with
northport,
you
normally
wouldn't
do
that
you'd
probably
do
a
load
balancer
if
you're
on
on
a
particular
cluster.
That
would
support
that
and
it's
going
to
be
exposed
on
port
to
8080.
A
So
I
won't
put
the
question:
what's
the
command
your
own?
Obviously
it's
a
cubectl
or
whatever
you
want
to
call
it
yourself
and
we
deployed
we,
the
files
out
they'll,
get
handled
by
the
kubernetes
api
server
and
it'll,
deploy
the
different
objects
or
resources
into
the
cluster
and
the
next
thing
one
then
your
signed
app
is
up
and
running
people
can
access
it
from
a
particular
ip
and
the
port
88.
All
right.
A
But
what
happens
if
we
say
to
ourself
here?
Okay:
this
is
very
simple.
This
is
very
easy
all
right,
but
if
we
want
to
deploy
another
instance
of
of
the
santa
app
a
separate
instance
in
the
cluster,
we
now
need
to
go
into
those
manifest
files
and
change
them.
Okay
or
if
we
want
to
change
the
replicas.
A
You
know
and
really
what
is
the
common
denominator
and
all
this
is
your
configuration
you're
changing?
Okay.
So
what
happens?
What?
What
would
it
be
like
if
it's
something
where
we
can
take
that
configuration
pull
it
out
of
the
manifest
files
and
just
make
those
changes
on
the
flight?
You
know,
while
we're
doing
while
we're
doing
the
apply
or
whatever
we
can
just
make
that
change,
that's
kind
of
what
what
you'd
be
looking
for
and
that's
where
the
value
is
in
it.
So
that's
where
helm
comes
into
it.
A
So
if
we
were
going
to
deploy
this
with
helm,
we
now
end
up
with
this,
so
you're,
probably
sitting
there
going
oh
great,
I
too
little
small,
manifest
files.
You
know
you're
after
sticking,
another
yama
find
open
top.
You
know
you're
you're,
just
increasing
my
workload
here
or
my
maintenance,
but
you're.
Not
what
we've
done
here
now
is,
and
this
is
where
what
helm
brings
you
four
things
that
helen
brings?
A
You
is
putting
that
extra
bit
of
templating
on
top
of
the
manifest
files,
so
that
bit
of
logic
around
the
the
go,
templating
language-
and
I
know
matt-
who
is
one
of
the
co-creators
of
helm-
talked
about
in
a
talk
there
this
week
about
one
of
the
decisions
he
made
at
the
time
was,
if
you
know
could
go
back
in
time,
he
would
be
not
to
use
gold
temptation
because
there's
a
lot
of
discussion
out
there.
Template
won't.
Do
this
for
me
that
won't
do
that.
For
me.
Yes,
there
are
restrictions
to
it.
A
Well,
really,
the
key
behind
it
was
to
give
you
that
extra
layer
where
you
could
put
in
configuration
and
small
bits
of
logic.
Now
we've
seen
people
put
crazy
logic
into
us
and
then
go
well.
Why
doesn't
it
work
in
that
situation?
So
there
are
limits
to
how
you
extend
it,
but
it
does
give
you
the
basic
logic
on
it,
and
we
can
see
here
now
our
configurations
put
into
a
value
of
spoil
okay.
So
with
this
setup
now,
that's
all
wrapped.
A
A
He
was
involved
in
a
lot
of
chat,
maintenance
and
a
lot
of
chat
creation
over
the
years
and
and
what
and
the
power
that
brought
to
a
lot
of
people
to
deploy
all
sorts
of
apps
from
wordpress
to
whatever
you
want,
ingress,
etc,
and
when
we
go
to
deploy
that,
though,
we
could
see
here
very
similar
to
cubectl
and
you're,
probably
saying
where's
the
value
in
this.
No,
it's
just
a
different
command.
It's
helm,
install
you,
give
it
the
name
of
your
release
and
then
the
particular
chart,
which
is
called
my
chart
again.
A
The
same
thing
is
happening:
you're
talking
to
kubernetes
server
or
the
kubernetes
api
server.
But
what
happens
before
it
sinks
at
the
kubernetes
api
server?
Is
it
renders
all
this
logic
pulling
all
the
configuration
into
a
manifest
file
and
the
manifest
files
are
sent
to
kubernetes
api
server?
Exactly
like
how
cube
ctl?
Does
it
and
a
lot
of
the
logic
around
your
authentication
orders,
whereas
authorization
is
the
same,
it
uses
the
whole
id
of
the
of
the
config
files
etc.
Like
cube,
ctl
does
so
you're,
probably
saying
yeah.
A
A
Yes
you're
right,
but
let's
say:
if
there
was,
I
don't
know
if
there's
far
more
configuration
in
it,
your
options
are
to
either
pull
down
the
apps
that
you've
run
in
and
redeploy
which
changes
the
manifest
files
or
go
away
and
edit
and
manifest
files
on
on
the
flight
heads
up
here.
Anyone
who's
ever
edited
man
deployed
manifest
floyd's
on
the
fly
hand,
keep
your
hands
up
if
you've
made
mistakes
while
doing
that
and
made
a
mess
of
your
cluster.
A
What
kubernetes
is
it
gives
you
or
sorry
what
hen
gives?
You
is
two
commands.
First,
one
is
upgrade,
which
means
you
want
to
operate
or
change
the
deployment
you've
already
done.
So
how
does
hell
know
about
the
deployments
it's
already
made
because
it
keeps
that
metadata
in
the
cluster?
It
keeps
it
in
the
secret
in
the
cluster,
and
it
knows
then,
when
it's
going
to
do
an
upgrade
here,
it's
going
to
do
a
delta.
So
I
come
along
now
and
I
say
I
change
our
replicas
all
right,
santa's
gone.
The
whole
house
is
federal
percent.
A
At
this
stage.
Let
just
one
replica
run
in
so
that
if
someone
wants
to
go
and
check
out,
did
santa
arrive
or
not
okay,
so
we
change
it
to
one,
and
then
we
use
the
helmet
grid.
Again
we
give
it
the
release
name
and
the
chart.
Okay.
Now
the
release
name
is
important
because
they're
different
instances
of
the
deployment.
So
with
now
with
helm,
you
can
install
at
different
times
using
the
one
chart
and
you're
just
changing
your
configuration
on
the
fly.
A
Right
so
we
mentioned
as
well.
One
of
your
key
parts
was
sharing
is
caring.
I
don't
know
if
you
had
that
phrase
growing
up.
None
of
us
wanted
to
share
our
suites
with
any
siblings,
tell
anybody
else,
but
there
you
go.
We
were
all
start
sharing
this
karen
and
here
it
really
really
helps
okay.
So
here
you
have
the
ability
to
deploy
a
particular
app
out
without
knowing
anything
about
the
chart,
and
sometimes
people
do
that
and
they've
great
fun
with
it.
A
But
a
lot
of
maintainers
have
created
charts
over
the
years
and
they're
now
being
hosted
on
their
own
in
their
own
repositories,
etc,
and
the
ability
to
be
able
to
take
you
know
whatever
you're
looking
to
deploy.
You
know
it's
used
for
operators,
it's
used
for
deploying
operators,
it's
used
for
deploying
wordpress
ingress.
A
I
don't
know
whatever
you're
looking
to
deploy
you
can
deploy
from.
So
this
is
where
the
key
comes
in.
Here
is
the
value
here
that
these
charts
can
be
shared
now
down
here
in
the
bottom.
This
was
how
we
always
share
charts
up
to
I'll,
say
this
last
release
that
we
know,
which
is
three
point.
Nine.
Am
I
right
three
point.
Eight.
A
Yeah,
sorry,
I've
jumped
a
gun.
Three
by
nine
is
underway.
It's
an
rc
okay,
so
we
didn't
so.
This
is
how
you
normally
did
it
before,
and
these
were
called
helm
chat
repositories.
They
were
essentially
http
servers
within
index
in
it
you
could
put
it
on
github.
You
could
use
github
pages,
there
was
chat,
museum,
etc.
You've
probably
used
it,
but
no
oci
support
has
gone
ga
in
3.8
yep
and
you
now
have
the
ability
to
do
it
from
an
oci
registry
hands
up
anybody
who
knows
what
an
oci
registry
is.
A
Okay
for
those
who
may
not.
These
are
a
standard
of
what
were
originally
for
docker
style
registries.
How
you,
how
you,
how
you
store
your
images,
your
container
images,
docker,
helped
set
up
the
oci
or
put
the
four
spec
and
then
eventually
that
spec
went
1.0
last
last
year
in
november,
maybe
sometime
or
maybe
the
year
before,
I'm
losing
a
sense
of
it.
A
So
what
I
gave
you,
then,
was
the
ability
to
put
different
artifacts
into
what
would
traditionally
have
been
a
container
registry
or
an
image
registry
you
know,
can
put
him
charts
into
it.
The
command
is
slightly
different
because
you
want
to
tell
her
what
the
protocol
is
and
also
down
here
at
the
bottom.
You
can
see
here.
My
repo
is
an
alias.
You
can't
put
an
alias
in
for
oci
registers,
but
folks
will
go
through
that
in
a
minute.
A
So
so
you
can
see
here,
you
know
the
ability
to
install
anything
without
knowing
exactly
what
it's
about,
but
you
can
get
your
app
in
there
and
you
don't
have
to
know
about
kubernetes
resources
etc
right.
So
where
are
these
charts,
these
magical
charts
and
if
you
haven't
seen
them
before
so
up
to
about
a
year
about
two
years
ago,
there
was
a
main
chart
repository
to
himself
repository
you'd,
still
find
it
out
there.
It's
now
going
to
end
of
life.
A
Dash
initially
was
started
as
a
sample
id
for
folks
to
be
able
to
use
traffic
and
go
what
happened
was
it
became
a
huge
piece
of
the
thing
with
mass
amount
of
charts,
and
there
was
a
small
group
of
people
trying
to
maintain
it,
paul
being
one
of
them:
scott,
matt,
farina
and
others.
Okay,
it
became
unmanageable
and
unmaintainable.
A
So
the
decision
came
that
look.
If
you're
going
to
create
a
chart,
you
need
to
host
that
chart
yourself
and
you
need
to
maintain
that
chart.
So
what
has
happened
is
all
these
charts
and
and
vietnamese
have
been
great
at
the
front
of
this
they've
taken
all
their
charts.
I
don't
know
what
they
call
vietnamese
they're.
A
A
This
is
just
a
search
engine
essentially
or
a
market
or
a
marketplace
tool
where
you
can
go
in
and
you
can
check
for
different
artifacts.
No,
it
started
out
with
the
helm
hub,
which
was
just
for
him
charts
and
then
through
dan,
who
passed
away.
His
idea
was
to
make
it
for
all
artifacts
that
are
out
there,
so
you
can
see
there's
a
list
of
them
down.
There,
you'll
probably
find
a
few
of
your
favorite
ones
in
there.
A
Okay,
so
you
can
go
in
and
have
a
search
there
and
when
you
search
for
your
particular
chart,
you're
looking
for
it
will
give
you
all
details
about
the
chart.
It
also
does
scanning,
et
cetera,
on
the
chart
and
you'll
feel
the
nice
things
that
keep
being
added
to
it.
So
it's
always
ongoing
yeah.
So
if
you
want
it,
you
go
and
look
it
for
there,
you
get
url
and
you
can
use
the
type.
A
So
to
conclude
on
this
is
so
we've
looked
at
a
you
know,
I
suppose
a
kind
of
a
a
typical
scenario
you
might
have
today
about
deploying
apps
out
into
a
cluster
or
whatever,
and
we've
looked
at
the
ability
that
helm
brings
you
to
extract
out
those
configuration
files
and
then
the
ability
to
share
true
charts
after
that,
so
I'm
going
to
just
hand
over
now
to
andy
okay.
So
thank
you.
Thank
you,
sir.
E
C
C
It's
less
infrastructure.
I
don't
want
to
maintain
another
source
if
you're
producing
content,
you
probably
are
using
a
container
registry.
Why
not
use
it
for
something
else?
So
that's
exactly
what
we
tried
that
we
try
to
do
here.
Is
we
now
give
you
an
easier
vehicle
to
be
able
to
share
and
consume
helm
charts
for
those
of
you
who
are
not
familiar
with
it?
It
is
a
way
that
you
can
store
it
within
a
container
registry.
C
However,
very
important
to
note
that
there
are
some
commands
that
are
not
really
applicable
to
oci,
charts,
the
helm,
repo
and
the
helm
search.
They
really
don't
work
very
well,
because
they're
mainly
meant
to
maintain
helm
repositories,
and
this
whole
effort
took
a
long
time
play
three
full
releases.
C
In
addition,
there
were
members
of
the
auris
project
who
helped
us
get
over
that
finish
line
with
a
couple.
Final
final
features,
so
really
a
great
open
source
success
story
here.
One
note
that
I
wanted
to
call
out
is
that
if
you're
using
oci
registry
or
oci
charts,
there
is
a
dependency
on
what
registry
you
use.
C
C
E
C
But
that's
the
number
one
thing
is
that
if
you
start
out
with
docker
and
containers,
you
probably
had
a
docker
account,
you
may
have
gone
somewhere
else
for
you
something
else
now,
but
I
know
I
have
one
still
I
try
pushing
my
chart.
I
get
a
nice
friendly
air.
What
what
I
screw
up
so
that
really
it
comes
down
to
it
doesn't
support
it
quite
yet.
It's
coming
other
features
that
came
out
recently
is
validation,
time
enhancement.
C
C
In
addition,
we
added
some
maintenance
features
number
one.
We
upgraded
the
libraries
for
kubernetes
to
124.
Obviously
there
is
dependencies
on
what
apis
you're
leveraging
as
well
as
what
target
cluster
you're
using.
I
have
some
friends
of
mine
who
are
using
a
very,
very,
very
old
version
of
kubernetes.
C
A
It
also
happens
sometimes,
if
you've
deployed
in
a
cluster
with
objects
that
were
being
deprecated
and
have
been
removed,
and
you
upgrade
your
cluster.
You
get
caught
as
well
for
that.
But
we
have
a
tool
around
that.
That's
something
we
can't
it's
out
of
our
control
and
it's
all
down
to
kubernetes
doesn't
obey
the
cement,
but.
A
C
Yep,
I
think
that
was
it,
and
one
thing
I
wanted
to
note
is:
if
you
want
to
know
about
some
of
the
features,
deprecations
etc.
The
kubernetes
blog
always
has
a
very
detailed
blog
about
the
release,
and
especially
when
there's
a
deprecation
you'll
see
a
special
blog
specifically
on
deprecations,
I'm
going
to
turn
it
over
to
you.
E
Okay,
so
getting
involved.
The
first
thing
to
say
is
we
really
do
want
you
to
be
involved,
and
I
just
wanted
to
first
start
out
by
saying
that
you
know
when
helm
started
before
I
got
involved,
but
even
when
I
got
involved,
it
was
a.
It
was
a
bit
of
a
smaller
group.
You
know,
and
kubernetes
was
also
a
bit
younger.
E
Kubernetes
is
very
mature
and
helm
is
a
very
mature
project
at
this
point,
and
so
it
can
also
be
you
know
I
can
understand
for
for
folks
that
might
be
newer
to
the
community
as
well
as
people
that
have
been
around
for
a
while,
but
haven't,
haven't
really
contributed
in
those
areas.
It
can
be
a
little
intimidating.
Please
don't
allow
that
to
be
the
case
for
yourself,
because
it
really,
it
really
doesn't
have
to
be
first
of
all
by
the
maintainers.
E
E
E
If
you
can't
read
slash
helm,
slash,
helm,
dash,
www,
that's
where
all
the
you
don't
really
need
to
know.
Well,
you
don't
actually
need
to
know
how
to
operate
kubernetes
in
order
to
to
contribute
to
the
project.
You
probably
do
want
to
know
a
bit
about
helm,
or
at
least
try
to
figure
out
a
bit
about
home
in
order
to
contribute
to
the
documentation,
but
fresh
eyes
are
really
important
too.
So,
even
folks
that
don't
know
a
lot
about
that.
E
E
I'm
not
signing
up
for
this
yeah,
you
know
I
do
as
well.
Others
do
so.
So
please
do
and
feel
free,
at
least
for
me,
to
reach
out
to
me
on
slack
or
twitter
or
whatever
I'm
easy
to
find
right
and
other
maintainers.
I
think
when,
if
folks
are
generally
like
pleasant
is
perfectly
fine.
The
one
note
I
want
to
make
about
this
is
that
one
thing
that's
not
super.
E
Welcome
are
pr's
that,
let's
say
just
sort
of
it's
like
one
spelling
character
off
or
maybe
there's
like
one
period.
That's
changed
or
like
the
order
of
a
list
is
changed
slightly
a
lot
of
those
pull
requests.
Are
we
do
want
you
to
contribute?
We
want
your
name
in
the
the
main,
the
list
of
people
who
have
contributed,
but
we
only
want.
We
want
you
to
contribute
something.
We
don't
want
just
a
name,
so
we
don't
want
it.
It.
E
A
In
and
you're
trying
to
get
going
and
what
scott
is
saying
as
well
is
when
you're
going
through
the
documentation,
that's
a
nice
place.
You
know
if
you're
trying
to
figure
out
how
to
deploy
your
environment
and
you
find
issues
then
that's
a
great
place
for
you
push
phr
pr
and
don't
worry
if
you
think.
Oh,
I
can't
push
that.
That's
that's
just
a
simple
pr
in
front
of
me
go
for
it.
You
know
what
I
mean,
because
you
are
contributing
and
you're
learning
as
you
go
along
as
well.
Thanks.
E
Martin
yeah
absolutely
well
said
yeah,
so
so
yeah
and
we
and
we'll
work
with
with
with
anyone
who
wants
to
make
a
pr,
regardless
of
what
you
find
to
make
that
of
some
substance
for
yourself,
and
perhaps
the
next
person
that
follows
you
so,
okay,
that's
where
that's
what
you
can
do,
there's
that's
how
you
do
it
and
you
know
I
mentioned
reaching
out
on
slack.
Here's
the
slack
stuff
so
inside
kubernetes
slack
is
where
you
know.
E
I
know
we're
at
cloud
native,
con
and
and
kubecon,
but
you
know
helm,
as
many
of
you
know
was
part
of
kubernetes
or
early
early
on
part
of
the
kubernetes.
Organ
was
split,
split
out
to
be
its
own
project
in
order
to
for
for
kubernetes
to
to
maintain
the
focus
it
wanted.
Helm
is
always
a
big
part
of
the
humanities
community
right
so
from
from
helm's
inception
anyway,
and
that's
why
it's
there,
instead
of
in
the
cncf
slack
so
there's.
E
E
E
If
for
those
of
you
that
remember,
those
is
no
longer
a
central
place,
even
though
the
the
artifacts
from
those
old
charts
are
still
posted,
you're
going
to
have
different,
different
different
distributed,
helm
repositories,
but
anything
that
relates
to
any
of
those
charts
is
still
discussable
in
there
because
there's
a
lot
of
commonality
and
yeah.
I
don't
really
honestly
check
that
as
much
as
I
as
much
as
I
used
to
so.
E
E
You
know
really
for
folks
who
do
want
to
try
doing
that.
I'm
saying
you
know
go
through
the
documentation.
Please
do
follow
that
to
go
ahead
and
you
can
run
you
can
run
helm
yourself.
You
know
like
it's
a
it's
a
go
project.
You
can
run
it
on
your
own
laptop.
You
can
test
it
out
and
that
can
really
help.
I
think
the
document
help
the
documentation
for
those
interested
again,
there's
not
a
pressure
to
do
that.
E
This
is
kind
of
me
recruiting
you
in
a
way
or
trying
to
for
folks
who
are
interested,
but
it's
not
really
you're
very
welcome
to
be
a
helm
user
and
never
contribute
anything
right.
But
but
if
you,
if
you
want
to
try
and
if
you
want
to
just
sort
of
see
how
it
works,
please
just
follow
the
docs
and
let
us
know
if
anything's
anything's
off
and
again
we
can
get
go
back
to
that
first
slide.
E
We
love
issues
and-
and
so
this
last
one
is,
you
know,
kind
of
what
I
was
alluding
to
before
that
you
know
one
of
the
a
and
and
and
what
I
think,
andy
andy
mirrored
back
when
he
said
bandwidth
right
like
we
have
a
lot
of
people
that
want
to
contribute
to
helm.
There's
always
several
hundred
pull
requests
open
at
any
given
time,
no
matter
what
we
do,
if
we
go.
C
E
And
spend
like
a
whole
week
triaging,
we
could
like
basically
all
get
together
for
a
whole
week
and
sit
down
and
just
like
keep
going.
We
do
that
and
we'll
we'll
we'll
either
will
either
improve
close
merge
or
whatever
needs
to
happen
like,
except
in
those
pull
requests
right
and
we
could
close
like
50
of
them
close
100
of
them.
What
happens
is
the
very
next
day,
the
very
next
week
that
sort
of
emboldens
folks
to
be
like
all
right?
You
know
the
reviewing.
Pr
is
more.
E
E
B
And
a
lot
of
times
building
the
pr
and
testing
it
and
just
saying
hey,
I
built
this
manually
and
tested.
It
is
helpful
likewise
with
responding
to
issues
one
of
the
most
frequent
time
sinks
for
the
developers
is
trying
to
figure
out
how
to
reproduce
some
of
these
issues.
So
just
going
hey
yeah
this
broke.
For
me,
too,
this
was
my
system.
This
is
how
I
set
it
up.
This
is
what
happened:
here's
the
output
oftentimes,
even
if
it
appears
to
be
a
duplicate
of
previous
ones.
We
can
then
go.
B
C
And
if
you're
a
developer,
you
may
be
a
no
js
developer,
not
a
goaling
developer.
Just
building
the
project
just
gets
your
system
yourself,
understanding
the
workflow
of
goaling
development.
So
it
really
helps
not
only
for
this
project,
which
is
what
you'd
be
focusing
on,
but
also
other
projects
in
the
kubernetes
space,
just
because
many
of
them
are
going
based.
A
And
I
suppose
one
other
aspect,
then,
is
that
if
you
have
kubernetes
experience,
then
in
the
issues,
cool
is
quite
good
as
well,
because
we
get
a
lot
of
kubernetes
questions
and
issues
that
aren't
actually
held.
So
people
jumping
in
I
know,
there's
a
person
joe
at
the
moment.
That's
doing
that
and
stuff,
and
probably
the
last
thing
to
say
in
it
as
well,
is
helm,
isn't
just
to
see
a
lion
and
use
them
directly.
E
Cool
and-
and
I'm
going
to
pass
it
off
to
in
a
moment
for
what's
next
but
really
right
before
we
wrap
up
this
section,
we've
got
about
five
minutes
left
total
here,
so
we
we
have.
We
have
a
process
for
a
contributor
ladder
process
here.
This
says
maintain
our
ladder
because
we're
focusing
on
the
maintainership.
E
Whatever
you
want,
you
know,
any
emoji
you
want
is
what
I'm
saying
mostly,
but
if
you
make
one
contribution
you
know
there
have
been
a
lot
of
people
that
have
contributed
the
project
only
once
over
the
years,
and
you
know
why
not
try
another
one.
You
know
what
I
mean
like.
Why
not?
Why
not
stay
for
a
bit?
Why
not
you
know,
kick
your
shoes
off
a
little.
E
Get
a
second
round
right:
yeah
yeah,
whatever
analogy
you
want
and
and-
and
it
just
helps
us
and
to
help
you
and
each
person
can
help
each
other
as
they
get
more
familiar
with
the
project.
So
once
that
happens,
though,
sustained
contributions
of
of
of
whatever
kind
we
we
want
more
maintainers
for
home
that
helm
has
different
sub
projects
and
helm
has
a
core
maintainership.
There's
you
can
start
off
by
being
a
triage
maintainer
that
helps
with
anyway.
That's
that
that's
I'm
just
gonna
fly
through
them,
real,
quick!
E
That's
folks
that
have
more
permissions
to
work
on
issue
to
to
to
change
labels
and
work
on
issues
and
pr's.
The
core
maintainer
is
people
that
sustain
the
helm,
helm
project
and
the
org.
Maintainer
are
people
that
work
for
governance,
so
there's
a
whole
wide
open
field
for
you,
I'm
gonna
pass
it
off
to
matt
thanks.
B
So
I'm
quickly
going
to
just
talk
about
where
we've
been
where
we're
going.
I
did
a
back
of
the
napkin
calculation
a
couple
of
weeks
ago,
if
you
have
a
small,
rather
successful,
open
source
project
with
a
couple
thousand
users
and
you
introduce
a
breaking
change
given
about
the
average
for
what
it
cost
a
software
developer
or
devops
engineer,
and
it
impacts
only
around
10
percent
of
your
user
base.
Right,
it
cost
a
couple
thousand
dollars.
B
The
numbers
that
I
was
coming
up
with
were
upwards
of
145
million
dollars,
and
I'm
thinking
that
is
a
very
profound
way
to
look
at
the
cost
of
a
breaking
chain.
Very
literally
right,
even
if
it's
a
minor
breaking
change,
it
cost
a
lot
of
people.
You
know
opportunity
costs
and
time
spent
fixing
this
kind
of
thing,
frustration,
sometimes
downtime
and
money,
and
so
you
know,
helm
has
for
a
long
time
had
a
fairly
rigid
policy.
B
Now,
on
occasion,
people
interpret
that
and
go
oh,
so
the
project's
basically
not
doing
anything
interesting
anymore
right,
but
that's
not
the
case,
we're
just
trying
to
make
sure
that
we
operate
at
a
cadence
that
allows
people
to
predict.
Oh,
I
need
to
spend
more
time
preparing
for
this
release
because
it
will
have
breaking
changes.
Oh
this
one's
a
minor
or
a
patch
release.
B
I
can
be
more
confident
that
I
can
roll
it
out
quickly
and
not
have
to
worry
about
a
nasty
surprise
in
production,
so
because
of
that,
then
you
can
kind
of
frame
out.
This
slide.
Don't
use
helm2
by
the
way.
If
you
are
using
helm2,
you
are
in
security,
hot
water,
it
hasn't
been
maintained
in
over
1.5
years.
Helm
3
is
the
current
stable
major
release
for
the
most
part,
the
last
big
feature
we
wanted
to
have
in
helm3
was
oci.
B
It
took
us
a
lot
longer
than
we
wanted
and
we
introduced
it
very
slowly.
It
had
an
experimental
flag
so
that
we
could
keep
operating
on
it
up
until
we
were
confident
that
it
was
ready
for
production,
and
we
put
that
one
out
in
two
releases
ago
now
so
with
that
one
we're
planning
a
couple
minor
features
that
we'll
roll
out
largely
helm3,
we're
just
going
to
try
and
keep
it
stable,
keep
it
usable.
B
B
So
helm4
is
coming.
We
have
it.
We
we
have
begun
sort
of
the
informal
stage
of
planning
for
helm4,
but
again
when
we
think
about
the
impact
of
a
large
project,
making
big
changes
and
the
community
having
to
adopt
it
in
existing
production
clusters
and
things
like
production
environments.
We
want
to
be
very
intentional
about
it
and
very
careful,
and
so
as
we
go
into
helm4,
we're
requiring
that
all
new
major
features
and
many
new
minor
features
have
to
be
completely
written.
B
The
design
proposal
has
to
be
completely
written
so
that
everybody
can
kind
of
take
a
look
at
it
and
say
yeah.
This
is
good.
This
is
bad.
Did
you
think
about
these
cases
over
here?
So
we
have
a
process
called
the
helm
improvement
proposal
process
that
we're
asking
developers
to
walk
through
contributors
to
walk
through,
as
we
plan
out
the
new
things
for
helm4
we've
been
using
this
process
for
two
and
a
half
three.
D
B
And
so
we've
been
using
it
for
a
while
we're
getting
fairly
confident.
I
know
it
can
be
burdensome
to
write
a
document
and
then
have
to
deal
with
comments
and
things
like
that,
but
largely
the
the
benefits
have
have
way
outweighed
the
cost
of
that.
So
you
know
on
the
bottom
of
these
are
actually
the
same
link
both
times,
but
the
github.com
community
repository
has
a
directory
in
it
called
hips
and
that's
where
all
the
helm
improvement
proposals
are.
E
B
B
B
At
this
point,
we
are
sort
of
a
slow
moving
project,
but
we're
doing
it
in
a
way
that
is
intentional,
because
we,
what
we
want
to
do
is
help
build
something
that
accrues
the
most
benefit
to
the
community,
and
our
big
big
needs
right
now
are
just
more
people
helping
out
in
small
ways
and
big
ways
around
the
community.
So.