►
Description
Join us for Kubernetes Forums Seoul, Sydney, Bengaluru and Delhi - learn more at kubecon.io
Don't miss KubeCon + CloudNativeCon 2020 events in Amsterdam March 30 - April 2, Shanghai July 28-30 and Boston November 17-20! Learn more at kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy, and all of the other CNCF-hosted projects
State of the UI: Leveraging Kubernetes Dashboard and Shaping its Future - Dan Romlein & Spencer Sugarman, Google
Over 60% of Kubernetes users take advantage of some sort of UI for Kubernetes. Some unique benefits of a UI over a CLI include: monitoring and troubleshooting, sharing information with less-technical stakeholders, and onboarding new users. In this talk, Spencer (UX Researcher) and Dan (UX Designer) will share research and best practices around what makes a successful Kubernetes UI, and offer recommendations on maximizing the value UIs can offer for your team. They’ll demo some of the updates included in the latest release of Dashboard, and present a future vision of Dashboard based on new research.
To learn more click here: https://sched.co/FuKi
A
B
Hey
everyone,
I'm
Dan
I'm,
a
user
experience
designer
working
on
Google
kubernetes
engine
I've,
been
involved
in
the
open-source
kubernetes
project
for
over
two
years
now
and
spoke
at
cube,
con
North,
America
and
Seattle
about
dashboard
and
2016.
So
it's
fun
to
be
back
here
at
Q,
Khan
speaking
about
dashboard
again,.
A
A
So
what
is
the
kubernetes
dashboard?
The
dashboard
is
a
general-purpose
web-based
UI
for
kubernetes
clusters.
It
allows
users
to
manage
and
troubleshoot
applications
running
in
the
cluster
as
well
as
manage
the
cluster
itself.
This
screenshot
up
here
is
the
overview
page
from
the
latest
version
110,
which
was
released
about
three
months
ago,
so
developments
still
ongoing,
but
it's
actually
released
1.0
in
2016.
A
This
is
a
screenshot
from
there,
and
so
dashboards
have
a
long
time
to
mature
and
stabilize
so
don't
ask
it's
definitely
production-ready
and
like
many
things,
many
other
tools
I'm
sure
you
forgot
today.
All
this
work
is
thanks
to
the
open
source
community.
It
was
posted
on
github
and
the
crew
Bernays
project,
and
since
the
project
started,
there
have
been
over
2,000
commits
from
143
contributors.
A
These
are
just
some
of
the
active
contributors
today
and
I
want
to
give
some
special
thanks
to
Sebastian
Florrick,
who
is
a
co-lead
as
well
as
Dan
ROM
line.
There's
a
co-lead
and
Pankow
who
helped
us
create
the
great
demo.
You'll
see
you
later
and
I
should
say.
The
special
interest
group
that
is
responsible
for
dashboard
is
cig
UI.
So
if
you
want
any
more
information
about,
say,
UI
or
the
dashboard
I
can
go
on
dashboard
on
github
and
look
up
SiC
UI
to
learn
more.
B
All
right,
thanks
Spencer,
so
before
we
go
any
further
I
want
to
delve
into
this
question
of
why
we
should
have
a
UI
for
kubernetes,
there's
cube
CTL
and
you
have
api's
and
other
ways
to
interact
with
kubernetes.
So
what
function
does
a
UI
provide
we're
gonna
get
into
some
evidence
now
of
the
usefulness
of
dashboard.
So
when
that
survey,
that
Spencer
mentioned
we
conducted,
sixty
percent
of
respondents
said
that
they
found
that
they
use
a
UI
to
either
monitor
manage
their
kubernetes
resources.
So
this
doesn't
get
to
the.
B
Why
question
of
why
they're
using
the
UI,
but
it
does
suggest
that
there
is
a
reason
there.
So
what
are
some
of
those
reasons?
Well,
firstly,
UI's
help.
You
understand.
What's
going
on
with
your
system
fast
in
this
screenshot
of
dashboard
I
can
see
this
particular
pod
has
an
error
that
it's
failing,
and
that
gives
me
a
starting
point
to
then
go
investigate
and
dig
in
deeper
into
what
might
be
causing
that
issue.
B
Monitoring
graphs
are
a
big
part
of
this
understanding.
What's
going
on
in
your
system,
so
here
I'm
looking
at
an
individual
deployment,
I
think
yeah
so
select
a
deployment
and
then
I
can
see,
even
though
there
aren't
any
errors
yet
that
there's
spiking
CPU
and
maybe
this
is
gonna
cause
some
issues
and
I've
want
to
investigate
that
and
the
next
main
strength
of
a
UI
that
we'd
like
to
call
out
is
onboarding
new
kubernetes
users,
which
we
think
is
the
biggest
value.
B
That
dashboard
can
bring
the
larger
kubernetes
community
and
it's
the
reason
that
we
classified
the
skill
level
of
this
session
as
beginner,
so
that,
hopefully,
if
you're
new
to
goober
Nettie's
you're
still
trying
to
figure
out
what
all
these
different
things
are.
Then
you
can
walk
away
at
the
end
of
this
talk
with
some
strategies
to
help
learn
about
work
more
about
kubernetes.
B
In
that
same
survey
that
we
conducted
over
half
of
our
spawn.
It's
said
that
dashboard
they
found
was
very
useful
or
extremely
useful
for
learning
about
kubernetes
and
a
quote
at
the
start,
it
was
great
to
be
able
to
visually,
see
the
resources
and
understand
how
they
are
grouped.
This
helps
someone
like
me
with
a
hands-on
visual
learning
style.
B
So,
as
someone's
learning
kubernetes,
we
know,
kubernetes
can
be
difficult
to
learn
and
there
are
a
lot
of
questions
that
come
up
like
what's
a
pod.
What's
a
deployment
what's
a
service
and
in
dashboard
I
can
see
those
resources
and
I
can
also
see
the
relationship
between
the
connections.
So
looking
at
this
replica
set,
I
can
see
the
pods
and
the
services
that
are
related
to
that
replica.
So
and
finally,
it's
also
possible
to
create
read,
update,
delete
resources
right
from
within
the
dashboard
UI.
B
B
So
describing
how
great
dashboard
is,
you
might
be
wondering,
do
we
think
that
it
replaces
cube
CTL?
What
are
we
arguing
for
here
well
to
answer
that?
No,
it
definitely
does
not
replace
cube
CTL.
We
think
dashboard
and
keeps
UTL
in
api's
all
have
their
own
strengths
and
some
of
the
strengths
we
talked
about
for
dashboard
already
are
providing
an
overview
so
that
single
pane
of
glass.
B
Dashboard
can
be
a
more
visually
compelling
way
to
show,
say
the
applications
that
your
team
is
responsible
for,
running
and
finally
infrequent
tasks.
So,
even
if
someone
is
an
advanced
kubernetes
user,
if
they
don't
do
a
particular
task
very
frequently,
they
might
be
less
confident
in
doing
that.
So
a
UI
can
assist
them
in
that.
B
So
we'd
encourage
you
to
try
out
a
UI
if
you
haven't
done
so,
our
claim
is,
if
you're
using
cube
CTL,
you
should
be
using
a
kubernetes
UI.
It
could
be
dashboard.
It
could
be
your
kubernetes
cloud
providers
UI.
There
are
a
lot
of
other
excellent
UI's
out
there,
but
we
think
that
there
is
place
for
UI
all
right,
so
dashboard
had
its
last
update,
AQ
Khan
as
I
mentioned
2016,
and
since
that
time
there's
been
a
lot
of
changes
and
a
lot
of
work
done.
B
So
these
are
all
features
that
have
been
a
small
selection
of
features
that
have
been
released
since
the
2016
update
and
I
won't
get
into
these
too
much
because
you'll
see
quite
a
few
of
them
in
our
demo.
In
just
a
few
minutes,
but
if
you're
interested
the
change
log
on
github
is
the
best
place
to
learn
about
what
new
features
have
been
released
and
as
far
as
current
work,
what
we're
focused
on
is
dashboard
see.
Ui
team
is
the
angular
migration
migrating
from
angular
1
to
angular
2.
B
The
current
version
of
angular
is
being
sunset,
so
this
moves
the
project
to
a
more
easily
supportable
place
and
hopefully
we'll
get
more
contributors
after
we've
completed
that
that,
hopefully,
will
be
finished,
q1
2019,
so
the
next
couple
months
and
we're
also
migrating
the
metrics
server
from
heap
serve
to
the
kubernetes
api.
This
is
to
maintain
our
very
popular
feature
of
displaying
nodes
and
pod
metrics,
and
the
metrics
server,
if
you're
not
familiar,
is
provides
benefits.
B
The
ecosystem
through
a
very
lightweight
and
scalable
method,
to
view
cluster
wide
metrics
and
dump
them
into
sinks
such
as
Prometheus,
which
is
another
CNCs
project.
And
this
screenshot
is
just
a
dark
theme
which
we've
heard
from
a
lot
of
people
as
a
so
that
will
be
possible
after
the
angular
migration
and
then
we're
also
working
to
make
that
customizable
so
that
if
an
organization
wants
to
change
the
branding
colors
that
they'll
be
able
to
do
that.
B
B
Someone
else
on
the
team
noticed
that
they
weren't
seeing
this
new
version
they're
still
seeing
the
old
version,
so
we've
got
a
problem
that
needs
investigation
and,
as
a
technical
note,
this
new
design
can
be
switched
on
or
off
via
a
environment.
Variable
called
app
underscore
mode,
so
you'll
see
that
in
the
demo
and
the
environment
variable
is
mapped
to
a
config
map.
Please
guys
thinking
all
right,
listen
go.
B
All
right
so
I'm
going
to
investigate,
see
what
the
problem
might
be.
I
go
to
the
website,
refresh
it
a
few
times
sure
enough.
Still
the
old
design
so
I
come
over
to
dashboard,
and
here
I
can
see
that
there
are
some
issues
that
there's
a
failed
pod:
a
failed
deployment,
so
I'm
going
to
go
up
to
the
global
search
bar
and
I'll
search
for
swag,
which
will
return
any
resources
that
have
swag
as
part
of
the
name
and
I
noticed
that
the
deployment
for
this
app
kubernetes
swag
has
an
error.
B
B
So
what
you're
going
to
see
is
an
example
of
in
the
UI
I
can
take
an
action
which
is
access
from
this
menu
and
then
I
can
view
the
ammo
for
this
deployment.
I
can
edit
it.
So
if
you
add
a
ml
ocaml,
editor
I
can
scroll
down
to
where
the
image
error
is
and
update
that
correctly,
click
update
and
then
I'll
visit
the
detail
page
for
that
deployment
that
I
just
updated
and
now
I
see
the
replica
set
and
it's
referencing
the
correct
image.
So
kubernetes
swag
new.
B
B
B
Still.
No
luck,
though,
still
not
seeing
the
update
so
I.
Remember:
oh
yeah
I
need
to
delete
the
pod
to
get
the
updated
environment
variable
so
I'm,
sorry
that
comes
next.
First
I'm
gonna
exact
into
the
pod.
So
this
opens
a
terminal
page
for
the
pod
and
this
I'm
going
to
check
the
check
the
environment
variable
here
and
still
set
to
old.
B
B
Now
we
finally
see
the
new
theme
right
so
and
then
I
just
go
back
to
the
UI
and
confirm
that
all
these
changes
are
correct.
So
I'm
gonna
skip
ahead
a
little
bit,
so
things
are
going
good,
but
then
we
have
users
that
are
reporting
they're,
seeing
500
errors,
so
I
go
to
the
UI
as
the
on-call
still
responsible
for
this
and
I
see
a
500
error
as
well,
so
back
to
trusty
dashboard
to
see
what
might
be
going
on.
B
I
go
to
the
deployments
page
and
here
I,
see
looking
at
the
overall
CPU
and
memory
usage
for
all
these
deployments
that
there's
a
spike
increasing
around
the
time
when
the
change
was
made.
So
this
is
a
this
is
a
case
of
where
dashboard
can
really
excel.
It's
showing
that
overview
of
what
might
where
the
problem
might
be.
B
B
B
B
So
issues
are
resolved.
I
can
finally
go
back
to
sleep
and
call
this
a
success.
So
this
is
just
one
set
of
use
cases
about
how
dashboard
could
potentially
provide
value
and
be
used
in
a
real-world
scenario,
all
right
so
now,
I'm
gonna
turn
it
back
over
to
Spencer
to
talk
a
little
bit
more,
a
little
bit
more
about
our
research
that
we've
conducted.
A
A
So
we
ran
a
survey
looking
into
how
do
people
use
dashboard
today
and
how
do
they
want
to
use
it
tomorrow
we
heard
from
163
people
of
all
different
mounts
apples
and
from
all
over
the
world,
and
we
heard
some
basic
feature
has
come
out
on
top,
so
the
three
things
that
we
heard
as
a
top
of
course,
it
changes
where
third-party
plugins
or
integrations
multi
cluster
management
and
improved
security
and
to
look
into
third-party
plugins
or
integration.
Specifically
again,
three
came
out
on
top
as
the
most
popular,
so
Gravano
was
definitely
the
most
popular.
A
This
makes
sense.
It's
a
charting
tool
and
very
popular
in
general,
but
also
fits
really
well
with
one
of
the
main
ways
you
can
use
group
dashboard,
which
is
for
monitoring
your
resources.
Next
was
helm,
which
is
an
application
package
manager
and
then
third
was
cloud
provider
resources.
So
this
was
things
like
load,
balancers
or
other
resources.
You
might
provision
from
your
cloud
provider
and
I
think
this
last
bit.
A
Kahn
talks
also
to
another
great
need
for
kubernetes
dashboard,
which
is
to
provide
this
single
pane
of
glass
to
show
you
all
the
resources
you
might
be
using
in
kubernetes
from
one
of
these
respondents,
we
heard
if
they
wanted
extensibility
modularity
an
API
for
libraries
to
integrate
with
dashboard.
They
should
follow
the
philosophy
of
kubernetes
and
should
be
the
foundation
of
which
we
can
build
our
custom
command
center.
A
Earlier
this
year,
researchers
found
that
there
were
over
16,000
installations
of
dashboard
that
were
exposed
and
that's
a
huge
number,
and
it's
can
very
much
lead
to
your
application
or
your
company
being
attacked
through
the
specter.
But
running
dashboard
securely
is
possible
and
really
just
takes
man
the
same
things
you
would
do
to
run
Kublai's
securely
in
general.
A
A
So
this
is
talking
to
is
the
fact
that
security
is
obviously
important
for
so
many
reasons,
but
it
also
is
potentially
a
blocker
to
more
people
using
dashboard,
including
areas
in
general,
as
Dan
mentioned
earlier,
being
able
to
share
what's
happening
inside
your
cluster
with
non-technical
stakeholders
is
a
great
thing.
You
can
do
a
dashboard
and
providing
better
and
easier
ways
to
do.
That
would
mean
that
more
people
would
be
able
to
use
dashboard
and
what
people
will
be
able
to
be
involved
and
understand
what's
happening
inside
the
cluster.
A
But
there
were
just
a
few
of
the
many
things
we've
heard
and
there
are
dozens
more
even
in
github
as
issues,
so
dashboards
come
a
long
way
since
2016
and
all
the
new
features
that
are
being
added
are
great,
but
there's
still
a
really
bright
future
and
things
we
can
do
to
make
dashboard
even
better
than
the
best
kubernetes
UI
that
it
can
be
so
I'm
just
gonna
turn
it
over
to
Dan
one
more
time,
who's
going
to
talk
about
how
you
can
get
involved
with
the
projects.
We
can
learn
more
information,
all.
B
Right
thanks,
Spencer
yeah.
If
all
this
sounds
exciting
and
you
want
to
try
out
dashboard
or
get
involved,
we'd
love
to
have
that
happen.
So
the
dashboard
repo
is
the
best
place
to
start
probably
for
overall
knowledge
about
dashboard
and
you
can
download
it
and
learn
more
mini
cube
if
you're
not
familiar
is
a
really
easy
way
to
run
kubernetes
locally
on
your
machine
and
it's
got
a
great
integration
with
dashboard.
That's
includes
dashboard
by
default,
so
you
can
just
run
dashboard
with
a
command.
B
It's
super
easy
and
if
you
want
to
contribute
or
get
involved
further,
the
slack
channel
is
probably
the
best
place
to
go.
There's
over
600
people
in
there
I'm
the
slack
Channel.
There
are
people
that
are
smarter
than
me
and
the
slack
channel
looking
to
answer
your
questions
and
that's
my
twitter
handle
if
you're
on
twitter.
B
This
is
the
spencer
mentioned
cui,
is
the
special
interest
group
in
kubernetes
for
dashboard
the
owns
dashboard.
So
on
this
page
we
have
all
this
information
plus
our
mailing
list.
You
want
to
be
part
of
that
and
a
link
to
our
weekly
zoom
cost,
but
every
week
we
have
a
video
chat
meeting
that
anyone
is
welcome
to
join,
and
that
is
all
we
had.
Thank
you
very
much
any
questions.
B
C
B
D
B
B
E
My
question
is
related
to
the
previous
question
like,
like
you
say
that
you
want
to
dashboard,
is
also
for
like
end
users.
Is
there
a
way
like
in
your
demo?
You
were
like
you're
deleting
the
part
like,
let's
say,
if
I'm
in
the
end
user
I
can
just
only
view
the
status
of
the
CPU,
but
there's
no
way
I
can
edit
the
part
like
that.
We
want
to
limit
the
the
end
users
from
you
know,
creating
them
each
turtley.
B
A
E
A
A
B
A
G
Twelve,
you
are
both
designers
and
I
have
laid
hold
I
want
to
ask
about
designer
when
you
design
the
dashboard.
Do
you
have
compliance
or
some
guidelines
to
like
the
font,
size
or
the
lines
and
space
of
the
between
lines,
because
we
design
our
is
very
crazy
about
this
kind
of
stuff?
So
do
they
will
also
have
this?
These
compounds
I
was.
A
B
So
dashboard
is
built
in
angular
and
there's
a
curio
library
that
it's
pulling
from.
So
a
lot
of
those
components
are
are
set
and
we
can
just
kind
of
pick
and
choose
from
them
and
those
are
set
to
be
accessible
for
different
types
of
users.
But
then
also
we
have
designers
and
involved
in
the
project,
so
they'll
be
providing
input
on
those
things
like
font,
size
and
color,
and
all
those
look
and
feel
details.