►
From YouTube: Kubernetes? But I'm a Product Manager… - Matthew McNeeney, Independent & Miguel Luna, VMware
Description
Don’t miss out! Join us at our upcoming events: EnvoyCon Virtual on October 15 and KubeCon + CloudNativeCon North America 2020 Virtual from November 17-20. Learn more at https://kubecon.io. The conferences feature presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects.
Kubernetes? But I'm a Product Manager… - Matthew McNeeney, Independent & Miguel Luna, VMware
Good product teams care deeply about their users and the experience they provide to them, yet many Product Managers shy away from conversations about Kubernetes, believing that it’s something for their engineering counterparts to think about. But this is a critical mistake; many of the benefits Kubernetes brings can be harnessed to deliver better products, faster. The Kubernetes.io tagline is ‘Production-Grade Container Orchestration’, so it’s easy to see why Product Managers tend to leave it aside as an engineering concern. But much of the innovative work happening in the ecosystem can be used to ship software faster and deliver great user experiences. This talk will cover how Product Managers building different types of software can utilise features in Kubernetes to build extensible, scalable and cloud-agnostic products that drive better customer outcomes.
https://sched.co/ZenX
A
A
B
Hey
everyone,
my
name's
matt.
I
started
my
product
management
career
in
2014,
working
at
different
consumer-facing
startups,
and
then
I
switched
over
to
the
enterprise
software
world
joining
vmware
and
spending
the
last
few
years
working
on
developer,
marketplaces
for
tools
like
cloud
foundry
and
kubernetes.
A
Great:
let's
talk
about
goals
for
this
presentation
when
we
first
landed
into
kubernetes,
it
took
us
some
time
to
react
to
what
we
were
seeing
given
the
first
definition
of
kubernetes
being
is
a
contrainer
orchestration
solution
being
a
product
manager.
The
first
question
was
a
solution
to
which
problems.
Why
do
we
need
this?
A
A
B
What
we
do
want
to
do
today
is
take
a
hypothetical
product
through
each
of
these
stages
and
share
some
ideas
of
how
we
think
kubernetes
can
help
product
teams
building
products
in
each
of
each
stage.
So
let's
get
going
we're
going
to
start
work
on
a
brand
new
product
or
initiative,
and
so
we're
in
this
introduction
phase.
What
should
we
be
thinking
about
here?
B
Product
teams
and
software
companies
need
to
spend
as
much
time
as
possible
iterating
on
their
applications
that
actually
deliver
value
to
their
customers.
This
is
especially
critical
when
you're
working
on
something
brand
new,
either
at
a
startup
or
larger
enterprise,
and
you
need
to
find
product
market
fit.
B
The
only
way
to
know
you're
getting
closer
to
this
is
to
work
out
what
you
want
to
learn
plan,
how
you're
going
to
measure
that
and
then
build
and
ship
the
smallest
thing
that
will
help
you
get
that
learning
for
those
familiar
with
the
lean
startup.
This
build
measure
learn
loop
will
be
very
familiar.
This
is
one
of
the
most
critical
concepts
of
that
movement
and
your
goal
as
a
product
manager
in
a
product
team
is
to
minimize
the
amount
of
time
it
takes
you
to
get
through
each
of
these
stages.
B
B
I
co-founded
a
startup
in
2015
and
I
remember
spending
more
than
half
my
time
just
trying
to
get
the
basic
infrastructure
up
and
running.
Our
users
never
saw
this
work
and
there
were
many
things
that
we
ended
up
over
optimizing
for
because
we
were
scared
of
having
to
go
back
and
recreate
lots
of
things
later
on
paul
ingalls
wrote
a
great
post
on
medium
about
uswitch's
experience
after
they
adopted
kubernetes
their
cto
said
teams
don't
realize
how
much
they
haven't
had
to
do.
B
B
If
you've
worked
in
a
large
enterprise,
you'll
no
doubt
have
face
long
delays
in
getting
your
software
deployed
and
available
to
your
users
in
many
organizations.
You've
spoken
to
this
is
often
weeks,
if
not
months.
So
this
is
one
of
the
slowest
aspects
of
the
build
measure
learn
cycle
for
many
product
teams.
You've
spoken
to
historically
gift
gas
mattias,
who
leads
their
site.
Reliability,
engineering
and
automation.
Teams
told
us
that
adopting
kubernetes
has
helped
them
reduce
this
time
down
from
two
weeks
to
half
a
day.
B
Kubernetes
has
a
growing,
thriving
ecosystem
right
now,
and
many
of
these
open
source
projects
can
help.
You
increase
your
speed
through
the
build,
measured,
learn
loop
using
things
like
bitnami's
application.
Catalog
and
operator
hub.io,
you
could
feasibly
get
a
brand
new
feature
or
application
built
deployed
and
into
your
users
hands
in
a
matter
of
hours.
B
B
A
Here,
thank
you,
matt
we're
not
growing,
and
it's
great,
let's
assume,
you're,
even
thinking
about
expanding
your
business
to
other
geographies.
Someone
talked
to
you
about
all
the
great
benefits
of
containers
and
you
think
great.
I
just
have
to
put
my
application
in
a
container
and
it's
fully
portable.
A
A
A
And
because
you're
in
a
stage
of
growth,
is
the
perfect
time
to
set
the
foundation
of
your
portability
and
how
you
plan
on
working
with
all
these
jamal,
even
when
you
are
even
bigger,
we
asked
this
question
to
matthias
from
gift
card.
How
would
you
deploy
gift
craft
tomorrow
somewhere
else,
not
sure
why
he
picked
ohio,
but
his
answer
is
very
insightful.
A
A
A
B
B
We've
spoken
to
many
many
organizations
who
have
really
struggled
with
the
first
part
of
this
goal.
How
do
you
continue
innovating
and
delivering
more
value
when
you're
suddenly
competing
for
resources
across
multiple
projects
and
other
new
initiatives,
and
this
is
hard,
especially
if,
within
a
short
space
of
time,
you've
gone
from,
say,
10
to
10
000
developers?
B
Regarding
the
second
half
of
this
goal,
there
are
likely
many
ways
you
could
do
this,
but
one
place
we've
seen.
Kubernetes
really
help
is
in
reducing
your
infrastructure
costs.
So,
let's
take
a
brief
look
into
each
of
these
bits,
so
you've
grown
to
10
000
developers.
They
work
in
awesome
product
teams
trying
to
accelerate,
through
the
build
measure,
learn
cycle
and
they're
wanting
to
deploy
tens
of
times
every
day,
but
your
operations
and
security
teams
are
in
a
lot
of
pain.
B
B
The
best
organizations
we've
spoken
to
who
managed
to
consistently
innovate
at
this
kind
of
scale,
have
all
adopted
platform
as
a
product
principles.
This
means
that
you
treat
the
platform
that
your
product
teams
use
every
day
as
a
product
in
its
own
right
and
one
that
is
incredibly
important
in
delivering
value
to
all
of
your
customers.
B
The
destination
here
is
the
kind
of
experiences
tools
like
heroku
and
cloud
foundry
give
developers
one
button
deploys
in
a
safe
and
secure
way
that
meet
the
needs
of
both
the
product
teams
and
the
operations,
security
and
audit
teams.
Many
companies
are
building
platforms
that
help
bring
the
kubernetes
and
platforms
of
product
worlds
together.
Helping
product
teams
get
kubernetes
clusters
that
meet
the
needs
of
their
security
and
compliance
teams.
With
things
like
logging,
metrics
all
ready
to
go
in
a
sec
in
minutes
or
even
less
right
rather
than
hours
or
weeks.
B
Now,
let's
quickly,
look
at
a
couple
of
examples
of
modern
companies
successfully
using
kubernetes
to
reduce
their
infrastructure
costs.
First,
let's
jump
back
to
giftgaff
again
and
read
this
great
quote:
the
automation
team
owns
the
platform.
We
provide
a
platform
as
a
service
to
other
teams,
our
cluster
scales
up
and
down
automatically
so
at
3am.
When
services
are
idle,
it
shrinks
at
3pm
the
site's
very
busy
it
scales
up
before
kubernetes.
We
couldn't
do
this.
B
We
treated
vms
like
pets,
guessing
how
much
capacity
we
needed
for
peak
times
and
we
over
allocated
90
of
the
time
anyway.
The
built-in
features
of
kubernetes
means
that
they
no
longer
need
to
bet
what
capacity
they
need
for
peak
times
like
black
friday
and
can
instead
save
that
money
by
only
using
the
capacity
they
need.
B
One
of
the
clever
examples
the
monzo
team
talk
about
here
is
how
they
now
run
their
high
usage,
build
jobs
using
spare
capacity
and
their
existing
infrastructure,
which
is
basically
free
all
right.
So
going
back
to
the
product
life
cycle
graph,
let's
say:
we've
got
through
this
maturity
phase
and
again
the
inevitable
happens.
We
move
into
decline
miguel.
What
happens
here.
A
A
A
As
you
go
from
two
nines
to
three
nines,
even
to
the
mythical
five
nines,
each
extra
nine
comes
at
an
increased
cost.
To
give
you
an
idea,
three
nines
give
you
a
downtime
of
8.77
hours
per
year
and
the
same
allowed.
Downtime
downtime
goes
down
to
5.26
minutes
with
just
two
extra
nines
just
picture
the
investment
required
to
get
to
this
level
due
to
the
constraints
from
other
elements
in
the
chain.
The
real
benefit
to
your
customers
reduces
exponentially
until
it
reaches
almost
zero.
A
A
Your
customer
base
versus
investing
in
the
reliability
and
scalability
that
will
keep
existing
customers
happy
is
difficult.
This
is
where
the
automation
of
kubernetes
controllers,
for
example,
can
help
us
releasing
some
of
that
engineering
time.
You
could
even
decide
to
go
beyond
an
extended
kubernetes
using
the
operator
pattern
which
will
allow
you
to
automate
many
of
the
human
operations
and
releasing
engineers
from
them.
A
So
as
product
managers.
It
is
your
responsibility
to
bring
to
the
team
enough
information
from
what
is
valuable
to
customers
and
help
them
make
a
decision
as
to
what
is
the
best
way
to
invest
their
time,
striking
the
right
balance
between
managing
reliability
and
improving
your
product
right.
So
that's
it
for
the
product
life
cycle
matt.
Why
don't
you?
Let's?
How
about
why?
Don't
we
do
a
quick
check
on
the
agenda
topics.
B
All
right
thanks
miguel,
so
if
we
quickly
jump
back
over
to
our
agenda,
then
hopefully
we
can
mark
this.
The
product
life
cycle
and
kubernetes
card
is
done,
which
is
awesome
so
next
up
on
our
agenda
is
miguel's
going
to
talk
to
us
about
how
we
can
apply
some
product
practices
to
this
world
over
to
you,
miguel.
A
A
A
Timing
is
important,
as
you
might
have
the
perfect
solution,
but
if
it's
not
ready
when
your
customers
need
it,
then
it
will
be
of
no
value
to
them,
but
before
we
make
some
uninformed
prioritization,
let's
set
a
scenario
which
might
help
us
do
more
than
just
guessing,
so
your
product
is
web
hosting
for
geocities.
I
hope
someone
remembers
jessica.
It
was
a
long
long
time
ago,
so
your
company
has
positioned
your
value
proposition
as
being
the
fastest
hosting
reliable
note.
A
The
three
nines
here
and
also
a
big
differentiator,
is
one
pound
a
year,
after
speaking
with
stakeholders
analyzing
the
data
from
customers.
You
want
to
take
some
actions
and
avoid
that
product
decline.
So
you
hold
the
workshop
and
start
rank.
How
pressing
the
improvement
of
these
outcomes
is
for
your
customers.
A
A
A
A
A
We
want
to
do
an
ideation
session
to
answer
that
question
of
how
might
we
release
some
engineering
time,
you
allow
your
team
to
diverge
and
have
some
personal
thinking
to
come
up
with
potential
solutions.
In
this
case,
your
team
has
suggested
different
options.
Like
someone
said,
oh,
why
don't
we
do
canary
releases
to
reduce
the
risk
of
breaking
up
breaking
up
things
upon
release.
A
Another
person
said:
let's
build
a
kubernetes
operator
and
automate
these
operations,
which
take
a
lot
of
time.
Furthermore,
let's
re-platform
x,
application
to
kubernetes,
and
that
way
we
leverage
on
the
kubernetes
automation
instead
of
having
to
be
on
call
to
restart
this
app
every
time
it
crashes.
A
And
finally,
someone
said:
oh,
why
don't
we
just
create
a
tooling
team,
so
from
all
these
solutions,
your
team
can
start
experimenting.
It
does
not
mean
that
you
go
heads
down
for
six
months
and
build
something.
I
encourage
you
to
take
the
lean
approach,
a
lean
approach
into
it,
experiments
that
might
give
you
leading
indicators
of
whether
the
solution
should
work
will
work
and
just
keep
iterating
on
it.
B
Yeah
so
hopefully,
we've
managed
to
achieve
our
goals
of
demonstrating
how
we
believe
kubernetes
can
help
product
managers
and
product
teams
no
matter
what
stage
of
the
product
lifecycle
they're
working
in
and
also
where
some
useful
product
practices
can
help
to
finish
off.
We
want
to
quickly
go
through
a
getting
involved
section
then
filing
a
q
a
so
we
promise
we
talk
about
how
product
people
can
get
more
involved
in
the
community,
but
to
be
honest,
this
is
where
we
think
we
need
your
help.
B
There
used
to
be
a
product
management,
special
interest
group,
but
that
no
longer
exists
and
we're
not
really
sure
where
product
folks
tend
to
hang
out
in
this
community,
but
we'd
love
to
find
out.
We
really
do
think
that
kubernetes
is
a
transformative
technology
and
this
applies
to
product
management.
Equally,
as
it
does
to
software
engineering,
and
we
think
some
awesome
things
could
be
done
by
bringing
these
worlds
close
together
and
so
we'd
love
your
feedback
on
this.