►
From YouTube: Visualizing Kubernetes: The Power of Dashboard by Dan Romlein, Apprenda & Piotr Bryk, Google
Description
Visualizing Kubernetes: The Power of Dashboard - Dan Romlein, Apprenda & Piotr Bryk, Google
Dashboard is the official web UI for Kubernetes. It allows users to visually manage applications running in a cluster and troubleshoot them, as well as manage the cluster itself. Since its release earlier this year it has seen rapid improvements and additions of new features. This talk will open with providing a mission statement for Dashboard and what problems it aims to solve, followed by a demo of some of those recent improvements and features. As part of their demo, Dan and Piotr will provide context by way of describing the rationale behind UX decisions made. Next they will frame known challenges remaining to be tackled and incorporate the others speakers in explaining currently proposed solutions. Finally, they will extend an invitation for the audience to weigh in and ask questions. Other key Dashboard developers and project managers will be in attendance to answer questions.
About Piotr Bryk
Piotr Bryk is a software engineer at Google and leads development of Kubernetes’ Dashboard, the official UI. He runs the SIG UI with Dan Romlein.
About Dan Romlein
Dan Romlein heads up user experience design at Apprenda. He is the primary UX and visual designer for Dashboard, and leads design critiques around new features. Dan runs the SIG UI with Piotr Bryk.
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! 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.
A
A
I
got
my
start
on
the
dashboard
project
in
May
of
this
year
spent
the
first
month
trying
to
figure
out
how
to
draw
freehand
heptagon
and
then
on
how
how
to
pronounce
Peters
name
right
and
once
I'd
master
those
I
moved
on
to
contributions
around
the
one
for
release
of
dashboard,
which
will
detail
later
on
and
currently
I'm
co-lead
of
cig.
You
I,
along
with
theater
Hey.
B
A
Are
here
to
talk
with
you
today
about
why
dashboard
matters?
If
that's
the
one
thing
you
take
away
from
this
talk,
it
will
have
been
successful
and
fundamentally
that's
because
anytime,
you
want
to
know.
What's
going
on
to
your
cluster,
dashboard
should
be
the
first
place
you
go
and
we're
going
to
be
building
a
case
for
that
throughout
this
talk
by
describing
what
it
is
our
intended
audience.
The
recent
features
we've
added
will
go
kind
of
past
present
and
future
and
talk
about
our
roadmap.
A
And
finally,
if
that
inspires
you,
how
you
get
involved
good,
so
this
is
dashboard
for
a
screen
shot
of
it.
It's
the
official
web
UI
for
kubernetes
its
production
ready.
Our
claim
is,
if
you're
using
coop
CTL,
you
should
be
using
dashboard
it
at
its
core.
It
helps
you
understand,
what's
going
on
with
your
cluster
fast,
so
to
do
that,
it
uses
monitoring
and
troubleshooting
tools
like
graphs
and
notifications.
In
addition,
you
can
also
create
read,
update,
delete
resources
within
dashboard.
A
It's
based
on
material
design,
as
a
disclaimer
I
am
designer,
so
I
might
geek
out
on
that
side
of
things
a
little
bit.
It's
one
of
the
reasons
Peters
here
to
keep
more
technically
focused,
do
my
best
and
it's
open
source.
So
this
is
where
the
magic
happens.
You
can
see
our
contributor
stats,
and
this
is
the
core
team
responsible
for
building
dashboard.
There's
me.
A
All
right,
so,
let's
jump
right
into
this.
Why
have
a
UI
for
kubernetes?
Why
should
dashboard
exist?
The
number
one
reason
to
have
a
UI
for
any
product
as
I'm
sure
many
of
you're
familiar,
is
unequivocally
donut
grass.
Nothing
quite
compares
with
the
majesty
of
a
well-designed
donut
graph.
You
can
kind
to
do
something
in
command
line
with
like
a
ski,
but
it
just
it
it
doesn't.
It
doesn't
come
close
to
the
glory
of
a
well-designed
done'
graph.
Look
at
that
Husky
I
think
we
need
more
Huskies
in
in
dashboard.
A
If
you've
ever
been
a
dribble.
You
know
done'
graphs
are
where
it's
at
when
it
comes
to
UI
design
so
kidding,
but
I
mentioned
that
to
make
the
point
that,
with
each
feature
we've
included
in
dashboard,
we've
tried
to
make
sure
that
it's
not
just
something
that
should
be
any
UI.
So
we've
tried
to
dig
down
to
find
the.
Why
behind
us
doing
things,
adding
new
features
so
getting
serious?
Now
the
the
real
reason
for
dashboard
is
it's
critical
to
understand?
A
What's
going
on
to
your
cluster
and
relying
exclusively
on
the
command
line
to
do
that
can
be
limiting?
So,
where
dashboard
comes
in,
is
it
provides
you
a
single
pane
of
glass
for
your
cluster,
so
you
can
look
in
see.
What's
going
on
respond
appropriately,
it
simplifies
monitoring
and
troubleshooting
through
visualization.
You
eyes
are
good
at
this.
We
know
this
they've
got
27
inches
of
screen
real
estate
whatever
to
take
advantage
of,
and
it
helps
answer
questions
like
is
my
app
running
correctly.
How
do
I
d
bug
or
troubleshoot
my
app?
A
What's
the
architecture
of
my
app
so
less
about
deployment
and
more
once
your
app
is
in
the
wild,
keeping
it
running
and
helping
you
understand
its
state
so
less
about
liftoff
and
more
when
your
app
is
at
cruising
altitude,
keeping
it
there
and
to
secondary
functions
that
we're
trying
to
accomplish
with
dashboard
R.
It
helps
to
explain
the
value
of
kubernetes
and
serve
as
an
evangelism
tool
for
non-technical.
A
A
You
know
what
I'm
talking
about
so
there's
an
asterisk
there,
while
that
is
a
focus
of
the
product
and
initially
was
more
of
a
focus,
its
kind
of
been
pushed
to
the
backburner,
because
the
team
realized
that
there
was
still
core
functionality
of
dashboard
that
we
wanted
to
focus
on
first,
and
so
that's
currently
what
we're
working
through.
But
we
do
have
this
on
a
roadmap
to
optimize
that
onboarding
experience
for
new
users.
A
A
So,
as
I
mentioned
before,
if
you're
using
coops
ETL,
we
expect
you
to
be
using
dashboard
and
we
expect
you
to
be
using
them
together.
In
tandem,
so
I've
outlined
a
couple
of
the
strengths.
Generally
speaking
of
a
command-line
versus
a
UI
and
I'm
sure,
I
don't
have
to
sell
you
on
the
command
line
strengths.
Most
of
you
are
developers
and
familiar
here
in
terminal
workflows.
You
have
a
centralized
location
that
you
can
execute
from
you,
don't
need
to
waste
time
clicking
around.
A
You
can
switch
between
multiple
tools
all
from
one
window
and
don't
need
to
change
context
or
environment,
and
you
can
hook
up
coops
ETL
into
your
continuous
integration,
workflow
and
create
team.
Specific
scripts
where
you
I,
like
dashboard
comes
in,
is
visualization
showing
stuff.
So
dashboard
includes
debugging
and
troubleshooting
tools
like
logging
and
monitoring
views.
Allow
you
to
look
in
and
see
give
you
that
pane
of
glass,
that's
single
pane
of
glass
and
react
appropriately
to
what
you're
seeing
in
your
cluster.
A
So,
for
example,
you're
on
call
an
incident
comes
up.
You
need
to
find
the
app
figure
out
what
it's
problem
is
by
looking
at
its
components
and
then
react
and
take
that
action.
Kubernetes
is
all
about
automation,
but
sometimes
you
do
need
to
step
in
and
take
those
one-off
manual
tasks
and
that's
where
dashboard
really
shines.
B
A
You
that,
at
a
glance,
confidence
about
what
you're,
seeing
and
therefore
guide
your
actions
you're
going
to
take
and
finally
exploration
and
discovery.
This
when
we
talked
to
kubernetes
users,
is
a
common
thing.
They
enjoy
that
ability
to
just
click
around
and
find
things
in
their
cluster
and
learn
things
about
their
cluster
that
they
wouldn't
have,
if
relying
on
their
own
expertise-
and
this
obviously
is
more
relevant
in
the
new
kubernetes
user
use
case.
So
we
are
trying
to
close
the
gap
between
command
line
and
UI.
A
We
don't
like
to
use
the
word
parity,
because
that
implies
the
same
thing
just
in
UI
form
and
that's
not
what
we're
trying
to
do
things
like
taking
down
your
cluster
you're,
not
gonna,
be
able
to
do
in
the
UI
anytime,
soon
more,
instead,
focusing
on
these
strengths
as
a
designer
I'm
not
going
to
try
to
learn
go
because
it
would
be
a
waste
of
my
time.
I'll
instead
focus
on
my
typography
skills,
for
example.
A
So
that's
similar
to
what
we're
trying
to
do
and
work
within
these
strengths
of
a
UI
and
designing
features
for
these
strengths.
We've
really
tried
to
keep
you
kubernetes
users
at
the
forefront
of
the
design
process
and
that
might
sound
cliche,
but
we
put
a
lot
of
effort
into
it
and
to
talk
about
that
now,
I'm
going
to
turn
it
over
to
pewter.
Yes,.
B
Thanks
Dan,
but
so
we
made
this
statement
of.
Why
do
we
do
that?
But
why
do
we
make
that?
Fourth,
but
now
I
want
to
explain
to
you
who
the
you
are
is
for
so
before
I
jump
into
this?
Let
me,
let
me
explain
you
what's
the
process
that
the
team
has
for
explain
for
finding
out
who
is
the
user
and
watch
picture.
B
Talk,
yes,
this
is
Donna
graph,
so
the
most
important
thing
here
is
that
we
listen
to
kubernetes
users
to
UI
users
and
want
to
understand
them.
So
last
year
we
run
the
user
survey.
It
was
really
successful.
You've
got
like
300
responses
or
even
more,
and
then
we
compiled
that
and
if
this
was
like
the
most
important
source
of
our
for
our
project
resolution
tasks.
So
we
took
things
that
were
most
important
to
use
earth
there
and
compiled
that
into
the
roadmap.
B
So
we
take
that
all
that
user
feedback,
user
studies,
user
researches
and
we've
compiled
that
into
target
users
of
the
UI.
So
the
there
is
like
three
main
users
that
we
want
to
cutter
for,
which
is
cluster
operator,
that
kind
that
the
user
is
kind
of
overseeing
the
cluster.
Checking
its
capacity
in
states
of
everything
does
not
necessarily
look
at
every
single
application,
but
wants
to
have
a
global
view
of
everything's.
There
is
a
also
application
operator,
which
is
a
person
who
checks
application.
Health.
That's
loading,
rolling
updates,
creates
new
things.
B
Basically,
that's
all
the
operation
work
and
what
we
want
also
to
make
the
UI
work
for
application
developers.
Supplication
developers
have
kind
of
different
workflow.
They
often
do
one
of
tasks
they
edit
tricks
through
the
UI.
They
debug
the
SSH
into
that
art
and
do
similar
things
that
operators
and
cluster
admins
not
necessarily
do
so.
These
are
three
users
and
there
is
secondary
secondary
level
of
user,
which
is
basically
we
call
them
beginners
to
kubernetes.
But
we
don't
explicitly
mention
them
here,
because
it's
like
when
you
are
good
for
all
these
three
people.
B
B
So
I
told
you
who
the
UI
is
for
and
now
and
there
will
be
a
brief
walkthrough
to
the
history
of
the
of
the
project.
So
we
started
the
UI
like
a
year
ago
and
we
did
initially
release
on
cue
cone
in
London
in
March
2016
and
the
goal
there
was
simple
and
well-defined.
So
we
wanted
to
have
a
good
user
journey
for
kubernetes
newcomers.
B
The
UI
serves
that
user
journey
well,
but
we
quickly
realize
that
this
is
not
enough,
so
the
beginner
and
a
newbie
users
were
happy,
but
once
you
wanted
to
some
real
work,
they
were
is
limited.
So
we
in
the
next
video
1.1
release,
we
added
support
for
all
user
workloads
that
are
there
are
in
the
cluster.
So,
instead
of
showing
only
a
curated
view
of
replication
controllers
and
services,
we
started
showing
replication
controllers
deployment.
B
Replica
set
and
all
other
controllers
that
are
there,
so
it
was
did
was
a
huge
step
forward,
but
still
we
realized
that
people
interact
with
not
only
controllers
and
services
in
the,
but
basically
everything
that
is
there
in
kubernetes
world,
so
fast-forward
to
1.4
release
that
was
released
a
couple
of
weeks
ago,
and
it
includes
a
set
of
interesting
features
to
cut.
Therefore,
the
gaps
that
we
had
and
Dan
will
walk
you
through
them.
Now,
thanks.
A
Peter
so
I
have
this
sorry
I
have
this
image
up
here,
because
there
should
be
an
alignment
between
the
features
that
I'm
about
to
describe
and
those
user
types
and
use
cases
and
their
needs
that
Peter
just
outlined.
We've
tried
to
make
sure
one
that,
with
everything
with
every
feature,
we
add
we're
solving
a
real
kubernetes
user
need
and
two
that
the
solution
we
come
up
with
is
backed
by
research
and
testing.
So
donut
graphs
could
be
a
perfectly
valid
thing
to
put
in
and
I
want
to
knock
Nona
graphs.
A
But
if
we
use
that
pattern
we
want
to
make
sure
that
it's
the
best
one
for
our
use
case
and
that
there's
not
some
better
alternative
out
there.
So
if
you
see
that
there's
something
out
of
alignment,
that's
great
because
it
gets
us
a
chance
to
potentially
build
a
better
product.
So
please
note
down
anything
you
see
or
if
you've
seen
it
previously
in
the
one
for
release
of
dashboard
and
there'll,
be
an
opportunity
to
to
share
that
feedback
coming
up
all
right,
so
jumping
right
into
it.
First
up
was
showing
more
objects.
A
This
was
kind
of
a
no-brainer.
We
knew
people
needed
to
be
able
to
interact
with
all
the
objects
that
they
could
see
in
coops
ETL
in
dashboards.
So
this
chunk
is
what
made
it
in
to
the
one
for
release
you
can
see.
We
have
just
a
couple.
Things
left
coming
up
four
one
five,
and
this
is
about
getting
to
showing
everything
in
dashboard.
You
can
see
in
coops
ETL
next
up
was
navigation,
and
this
was
my
first
contribution
for
the
product.
When
I
started,
Peter
mentioned
yeah
I've
got
a
design
issue.
A
If
you
want
to
pick
it
up
should
be
fairly
trivial
like
all
right
great,
it
turned
out
to
be
significantly
harder
than
either
of
us
had
anticipated.
For
a
few
reasons,
one
being
navigation
has
cascading
effects
on
just
about
every
part
of
the
product,
so
it
wasn't
like
some
widget
that
we
could
put
over
in
the
corner
or
find
later
we
really
want
to
nail
the
first
time
around
and
secondly,
we
wanted
to
know
what
the
problem
was,
that
we
were
solving.
A
It
was
kind
of
like
well
duh,
we
need
a
navigation,
but
we
wanted
to
figure
out
the
specific
needs
and
the
use
cases
that
would
be
served
by
it.
So
the
problem
was
fairly
acute.
You
can
see
in
the
previous
release.
We
didn't
have
a
navigation,
so
you
could
drill
down
and
then
go
back
up
to
that
top
level,
but
it
involved
a
lot
of
extra
clicking
around
and
it
wasn't
fair
to
users
and
thirdly,
by
design
kubernetes
defies
and
neatly
nested
hierarchy.
That
would
make
designing
a
navigation
so
much
easier
right.
A
So
there's
this
quote
from
a
technical
overview
of
kubernetes
by
Brendan
burns.
He's
talking
about
the
design
architecture
of
kubernetes
and
I
heard
this
and
I
was
like.
Ok,
that's
probably
the
contributing
factor
to
why
this
is
so
hard
and
we
actually,
we
were
talking
with
the
Google
cloud
platform,
UX
team
yesterday
about
this
very
problem.
How
do
we
balance
some?
A
The
freedom
and
customizability?
That's
a
word
of
being
able
to
slice
and
dice
your
cluster
any
way
you
want.
That
is
an
asset
of
kubernetes
with
some
defined
structure
so
that
people
can
understand
just
what
this
thing
is
in
visual
form.
So
this
is
us
starting
to
explore
that,
with
an
information
architecture,
diagram,
putting
all
the
objects
on
a
on
a
canvas
and
playing
with
what
groupings
and
what
relationships
make
most
sense
to
express
any
UI.
So
you
can
see
we've
added
cluster
operators
and
application
operators
and
developers.
A
This
is
something
that
came
largely
out
of
our
experience
at
aprender
and
working
with
enterprises.
How
much
they
like
to
have
roles
in
products
that
they're
using.
So
you
can
see
we've
added
an
admin
section
for
the
operator
use
case
and
then
kind
of
DevOps.
Everybody
else
use
case
below
that
to
bring.
Hopefully
a
we've
struck
the
right
balance
between
that
freedom
and
defined
structure.
A
Something
we
didn't
include
in
navigation
was
icons
and
I
mention
this,
because
it's
an
example
of
us
stepping
back
and
saying
is
this
the
right
investment
of
our
time
right
now
and
we
icons
are
fun.
They're
kind
of
like
donut
graphs
are
easy
to
talk
about
and
have
discussions
around
in
github,
but
when
we
step
back
and
evaluate
it,
if
it
was
the
right
investment
of
our
time,
we
decided
that
we'd
rather
really
nail
the
core
functionality
of
navigation,
so
icons
are
still
hopefully
on
their
way.
This
is
our
early
discussion.
A
Stay
tuned
for
for
Windows
make
it
into
the
product
the
new
admin
section
in
one
for
as
I
mentioned,
we
have
this.
We
don't
have
the
concept
of
roles
or
role
based
access
control,
yet
we
realize
how
important
that
is
for
enterprises,
especially
so
that's
on
our
roadmap
and
this
kind
of
lays
the
foundation
for
that
by
adding
those
sections,
monitoring
or
troubleshooting
graphs,
as
Peter
mentioned,
and
that
beautiful
donor
graph
displayed.
These
are
really
hot
item
at
last.
A
A
Which
counts
as
a
feature
because
it
affects
usability
and
if
you're
not
familiar
material,
is
Google's
official
design
system
for
all
their
products
they
open
sourced
at
about
two
years
ago,
it's
used
in
over
a
million
apps
and
the
team
building
dashboard
had
familiarity
working
with
material,
so
it
made
sense
to
use.
We
did
have
to
do
some
improvisation
to
make
it
work
for
our
use
case.
A
We
have,
by
necessity,
a
denser
information
architecture
that
we're
showing
to
people
and
finally,
materials
focuses
on
mobile
devices,
and
while
we
want
a
full
functionality
to
mobile
size,
we
recognize
most
use.
Cases
are
going
to
be
on
a
desktop
monitor
and
the
Google
cloud
platform.
Ux
team
was
really
instrumental
in
helping
us
make
that
customization
from
the
material
design,
spec,
Keith
and
I
would
email
back
and
forth.
It
was
like
Scripture
would
be
like
well.
A
So
most
of
the
things
you
wouldn't
even
notice,
but
the
the
logo
is
now
the
official
kubernetes
logo,
rather
than
just
arial
helvetica
whatever
it
was,
which
is
driving
me.
Nuts,
the
font,
yeah
and
the
spacing
on
things
has
now
been
refined,
fine-tune
color
changes
so
effectively
making
for
a
more
cohesive
user
experience
and
as
an
aside.
A
Material
mirrors,
the
elegance
of
kubernetes,
in
that
it's,
these
reusable
modular
components
which
render
Mearns
talks
about
that
technical
overview
of
kubernetes
talk
and
I
thought
was
a
neat
parallel
between
the
two
that
you've
got
this
rather
than
starting
from
Ground
Zero
with
a
drop
down.
I
can
have
something
to
work
from
and
customize
that
to
my
workflow,
all
right
so
to
show
this
just
shows
an
insight
into
how
we
work.
A
A
So
I
mentioned
that
to
encourage
you,
even
if
you're,
not
a
designer,
feel
free
to
ask
questions,
make
design
critiques
I
think
anyone
can
have
valuable
input
about
what
this
thing
looks
like
and
how
it
behaves
all
right
so
to
show
off
those
features,
I'm
going
to
turn
it
over
to
Peter
with
a
use
case.
Yes,.
B
B
So
anyway,
so
we
are
a
company
that
does
ecommerce
application.
It
is
a
simple
application
for
selling
t
shirts
underneath
kubernetes.
It
uses
deployment
and
services
to
to
cover
everything.
So
this
is
a
simple
thing,
but
I
am
application.
Devops
and
I
got
paged
that
something
is
wrong
with
that
application.
It's
the
middle
of
the
night.
So
what
do
I
do
I,
try
to
open
the
application
and
figure
out
what's
there,
whether
it
is
actually
a
problem,
so
I
refresh
the
application
a
couple
of
times
and
oops.
B
There
is
500
internal
server
error,
so
this
is
actually
a
problem.
Let
me
try
to
reproduce
that.
Yes,
this
is
visible.
Whenever
I
refresh
a
couple
of
times
there
is
a
500
internal
server
arrow.
So
this
is
a
problem
with
the
application
is
on
production.
So
let
me
try
to
fix
that.
But
so
first
thing
I
notice
is
that
we
are
running
a
be
experiment
in
the
e-commerce
app,
so
every
other
request
shows
material
design,
redesign
of
the
e-commerce
app
or
the
other
el
questro,
the
old
version.
B
B
So
what
I
do
I
open
it
try
to
figure
out
what
was
wrong
with
it.
So
I
see
it
state
that
there
is
a
replica
set
one
port.
Everything
is
fine,
so
far,
I
open
the
replica
set
to
figure
out.
What's
the
problem,
nothing
wrong.
So
far,
memory
usage
is
okay,
no
events.
So
let
me
dig
more
into
the
problem.
So
I
see
a
port
that
is
running
this.
This
ecommerce,
app
I,
see
its
containers.
I
see
that
a
be
experiment
is
turned
on.
B
It
has
a
value
and
coming
from
a
config
map
that
I
have
here
and
says
config
map
a
be
experiment
turned
on
its
sure.
Okay,
does
a
data
point
and
that's
all
I
know.
Let
me
show
the
logs
and
here
I
see
a
red
line
here.
Locks
say
that
material
design
that
CSS
file
is
not
found.
We
have
we've
got
the
structures
so
yeah,
that's
that's
a
data
point
and
now
I
know
I.
B
Recreate
my
part,
so
that
it
picks
up
the
value
so
I
did
it
that
bad,
it's
good,
it
gets
recreated.
The
new
part
is
here
and
there's
no
metrics
for
it.
Yet,
but
I
try
to
refresh
my
life
up.
There
is
no
500
at
all.
So
that's
fine!
That's
perfect!
I
fixed
the
problem
through
the
UI
and
Vista
gated
it,
but
I
know
that
there
is
something
wrong
with
that
application.
So
I
asked
my
colleague
in
the
meantime
to
do
a
hot
fix
for
it.
They
did.
B
The
hot
fix
and
I
have
a
new
version
of
the
application,
so
I
will
do
a
rolling
update
and
of
the
application.
So
I
just
just
change
the
patch
version,
because
I've
got
a
new
patch,
so
I'm
doing
the
running,
update
I,
see
that
the
new
replica
set
is
is
coming
up
and
all
the
replica
set
is
coming
down.
That
was
the
state.
Let
me
refresh
the
page
and
yeah.
The
running
update
was
fast
because
the
application
is
small,
as
I
have
got
the
new
replica
set
and
everything
is
green.
B
Yeah,
whenever
I
request
something
new
logs
appear,
so
that's
fine,
and
but
let
me
wait
a
minute
to
check
the
CPU
usage
and
memory
of
the
of
the
pad.
So
so
far
so
good.
The
memory
usage
is
low,
like
70
kilobytes,
trees,
700
kilobytes.
Previously
it
was
also
loud,
but
you
know
when
you
are
doing
a
hotfix.
You
better
look
at
things
as
they
change.
B
So
while
we
are
waiting
for
that
for
the
for
the
metrics
to
to
be
collected,
I
can
show
you
how
the
how
can
I
change
the
a
B
experiment
for
for
this
for
this
application.
So
I've
got
these
all
my
kinetic
mass.
Currently,
it's
very
simple.
It's
one
config
map
for
my
1a
B
experiment,
but
it
is
here
and
I
got
I
can
go
to
the
same
config
map
through
my
deployment
through
my
pod.
So
so
the
UI
is
all
about
showing
connections
here.
B
So
whenever
there
is
a
connection
between
two
kubernetes
objects,
we
tend
to
link
or
shut
the
show
richly
them
like.
Here
there
is
a
connection
between
config
map
and
environment
variable,
and
here
there
is
a
creator
of
this
path
but
and
I've
waited
a
while
and
I
see
that
the
memory
usage
of
my
pod
is
like
400
megabytes,
which
is
a
lot
compared
to
what
was
previously
previously.
It
was
only
1
megabyte
so
clearly
there
is
a
memory
leak
somewhere
in
my
application,
so
that's
a
problem
and
that
I
didn't
anticipate
I.
B
B
And
now
let
me
talk
about
the
roadmap,
so
we've
seen
things
that
we've
added.
But
what
are?
What
are
the
things
that
we
want
to
add
today?
Why?
So?
There
is
a
couple
of
tracks
for
for
the
for
the
project.
The
the
most
important
thing
out
there
is
showing
more
kubernetes
objects
in
a
better
way.
So
that's
our
our
core
work
that
they
spend
most
time
on
so
in
this
year
we
want
to
support
all
of
kubernetes
objects
that
you
might
want
interact
with
in
a
sane
way,
but
in
the
following
year
in
following
quarters.
B
We
want
to
do
this
in
a
better
way
so
that
you
can
spend
your
daily
routine,
interacting
with
your
kubernetes
everything
inside
the
UI,
maybe
not
necessarily
edit
things
in
the
UI,
but
view
them
and
understand
the
state
of
of
your
kubernetes
applications.
There
is
other
track
which
is
called
ease
of
use,
so
we
want
to
make
the
UI
the
entry
point
for
new
kubernetes
users.
While
this
is
not
the
case
currently
for
the
UI,
you
need
to
be
cool.
B
You
need
to
know
kubernetes
in
order
to
use
the
UI
when
to
change
that
and
make
the
UI
entry
point
for
new
kubernetes
users,
because
browser
is
a
new,
is
a
nice
place
for
explaining
things
and
linking
lots
of
information
together
in
an
easy,
manageable
way.
There
is
the
third
track,
which
is
monitoring
troubleshooting
in
debugging.
So
whenever
we
ask
users,
what
would
they
like
to
do
in
a
UI
like
this
or
what
do
they
do
in
similar?
B
You
eyes,
it's
like
troubleshooting
and
and
being
able
to
fix
things
through
the
UI,
and
so
that's
a
large
track
that
we
spend
time
on,
and
these
include
things
like
exacting
into
your
path,
having
a
better
view
for
your
logs
across
different
resources,
enhanced
monitoring,
maybe
some
UI
intelligence
and
suggestions.
This
is
all
in
that
track.
And,
finally,
we
want
to
make
Dartford
UI
that
platform
for
kubernetes
features.
So
whenever
there
is,
there
is
a
LAN
add-on
or
a
platform
that
builds
on
top
of
kubernetes.
B
A
Right
so
in
conclusion,
anytime,
you
want
to
know.
What's
going
on,
the
cluster
dashboard
should
be
the
first
place
you
go
because
of
its
strengths
around
visualization
and
in
showing
you
the
state
of
your
cluster,
we're
striving
to
build
a
product
so
good
that
it
comes
to
mind.
Is
the
first
stop.
You
want
to
make
and
figuring
out
the
state
of
your
cluster
if
this
has
been
inspiring
to
you
and
you
want
to
go,
maybe
see
with
fresh
eyes
that
official
kubernetes
logo
and
the
one
for
release
so
it'd
make
me
so
happy.
A
This
is
the
repo
it's
pretty
easy
to
find.
It's
included
by
default
in
most
kubernetes
installations,
cig
UI
on
slack,
that's
the
thing
Peter
and
I
are
co-leads
of.
Is
the
go-to
place
for
all
things?
Crew
Burnett
is
UI
related,
we're
stoked
to
be
here
at
coop
con
and
meet
real
live
kubernetes
users.
So
we
scoped
out
there's
the
lounge.
If
you
go
out
of
this
out
of
these
doors
and
then
hook
around
to
the
left,
there's
some
couches
and
tables
we're
going
to
be
set
up
there
from
3
to
4.
A
A
As
I
hinted
at
we're
redoing
the
survey
that
generated
that
monitoring
and
troubleshooting
donut
graph,
so,
as
evidenced
by
this
talk,
we
really
listen
to
what
you're
telling
us-
and
this
is
a
chance
for
you
to
shape
the
future
of
the
project
and
adjust
that
roadmap
on
what
we
work
on
next
we'd
really
appreciate
it.
If
you
check
out
that
survey
well,
the
slides
are
up
online.
A
So
we'll
promote
this
on
Twitter
in
a
couple
different
places,
I
think
two
weeks
we
said
we
would
leave
it
open,
then
we'll
do
up
a
nice
pretty
infographic
and
send
that
out
publicly
I'm
at
Daniel
ROM
line
on
Twitter
peter
has
one
Twitter
follower,
and
it's
me
he's
like
we
can
use
my
Twitter
account.
It's
like
I'm,
not
sure,
if
I
trust
that
really
so
brick'
do
Google
calm
feel
free
to
reach
out
we'll
try
to
be
responsive,
and
we
really
would
like
to
hear
from
you.
Thank
you
very
much.
A
B
So
basically,
we
observe
I
will
users,
so
that
was
based
on
our
observation
and
and
consensus
with
consensus
we've
had
in
the
CGI
group.
So
there
was
like
a
couple
of
different
roles
that
people
were
observing,
but
at
the
very
end
of
the
discussions,
those
three
emerged
as
the
top
users
that
we
want
to
Qatar
for,
and
you
ask
who
is
application
operator
so
in
the
detailed
discussions
and
explanations
on
github?
We
have
this
person
I
explained,
and
there
was.
There
was
a
brief
explanation
of
this
on
this
slide.
But
this
is.
A
A
B
B
C
B
Access
control
and
how
do
we
do
a
role
based
access,
control
and
I
am,
and
the
answer
is
that
currently
everyone
is
admin,
that's
the
current
state,
but
at
the
moment
people
from
our
team
work
on
on
privileges
so
that
users
can
actually
log
in
to
the
UI
and
act
on
the
bill
hard.
So
this
is
the
first
step
that
we
will
that
people
work
on
currently,
and
we
will
deliver
this
this
feature
and
after
that
we
will
maybe
work
on
this
thing.
We
just
show
you
only
things
that
you
have
access
to.