►
From YouTube: Istio User Experience Working Group July 23, 2019
Description
Istio User Experience Working Group meeting held July 23, 2019
A
Hello,
everyone
how's,
the
sto
user
experience
being
the
I'm
showing
document
for
dinner,
but
right
one
thing:
we
had
planned
to
talk
about
a
cute
man
from
Walmart
labs,
but
we
may
not
be
getting
our
demo
for
that,
but
it's
in
the
notes.
If
anyone
wants
to
check
it
out,
there
are
some
screenshots
for
it
and
it's
got
some
SEO
specific
stuff
in
it
is
to
your
interest,
recipes
and
stuff
here.
I
haven't
got
a
lot
of
playing
with
it.
A
The
next
one
I
want
to
talk
about
is
the
structural
idiom,
because
a
lot
of
sub
demands
have
been
composed
for
a
steel,
puddle
and
Lois
pointed
out
that
there's
sort
of
different
ways
of
doing
these
commands.
So
we
all
introduced
to
in
cube
kernel,
cube,
color
get
pod,
but
docker
is
sort
of
different
OpenStack
and
OpenStack
compute
service
create.
There
were
lots
of
sort
of
different
patterns
for
that
and
I
was
wondering
so
we
have
this
was
it
to
comment
on.
A
This
is
to
put
a
mesh
if
I
and
we
were
talking
about
it
there.
It
seems
like
a
bigger
thing
than
a
cut.
There
should
be
a
bigger
comment
than
just
a
comment
on
this.
If
I,
maybe
we
should
document
I
sort
of
suggest
to
people
writing
commands
how
these
should
all
go.
I
had
originally
talked
about
is
to
cuddle,
add
to
mesh
or
meza
fie.
What
do
we
think
should
be
the
structure
for
steel
curls
up
tonight.
B
A
We
have
for
the
stuff
like
getting
delete.
The
cube,
cuddle
stuff
works,
great
I
love.
It
keep
cuddle
somewhat
falls
down
with
things
like
cube,
cuddle
create,
can
cuddle
court
forward.
These
commands
are
like
completely
different
things.
There's
no
notion
of
our
sub
command,
so
it
is
to
yo
cuddle.
We
have
a
sub
finland,
experimental
or
I.
Think
there's
an
authorization
and
I
think
that
one
of
the
questions
is
sort
of
how
house
should
we
have
some
commands?
How
should
we
structure
them?
A
If
it's,
if
it's
just
you
know,
verb
type
cube
cuddle
is
great.
I
prefer
cube
cuddle
to
the
open
stack
way,
but
I
remember
when
I
switched
from
open
stack
to
kubernetes
I,
didn't
confuse
the
good
things
wearing
a
board
where
there
was
not
compute
subdomain
cube
girl
they
create
was
sort
of
after
type,
so
I
want
to
stick
as
opposed
to
creep
cuddles
possible.
What
do
we
think
about
some
of
these
operator
commands
where
they
belong
so.
C
I
think
you
cuddles
cindex
works
if
you
have
credence
in
on
resources
and
that
those
operations
are
common
across
all
those
resources,
it's
not
clear
given
to
the
examples
I've
seen
with
the
operator
of
the
mesh.
If
I
the
the
verbs
there,
how
well
you
could
standardize
on
the
set
of
verbs
I
add
to
mesh,
would
we
add
too
much
all
the
resources
that
we
support?
What
are
the
standard
set
of
resources
that
sto
is
composed
of,
and
in
does
that
combination
of
verbs
and
nouns
compose
it's
a
reasonable
way.
C
A
B
B
A
C
E
B
A
C
This
Lee
yeah
just
the
examples,
I,
don't
know
how
much
additional
you
know
textual
description
of
what
those
commands
do
is
necessary,
which
seems
like
it
should
be.
Just
the
first
pass.
Looking
at
those
two
side
by
side,
if
there's
some
common
verbs,
does
it
make
sense
to
add
to
mesh
a
pod
and
every
other
type,
and
then
what
are
the
types
that
we
have.
C
We're
full
up
for
cube
cuddle
because
you
have
a
standard
set
of
resources
defined
by
the
API
conventions
and
there's
common
set
of
verbs
that
operate
on
those
whose
resources
I'm,
not
sure
that
is
necessarily
true
with
this
deal.
But
but
maybe
that's
the
case
we
haven't
so
examples
will
help
clarify
that.
A
A
A
A
A
Clearly,
we
have
a
problem:
most
all
of
the
sto
cuttable
tests
is
tested.
It
prints
it
off.
With
this
expected,
there
were
no
tests
to
make
sure
nobody
added
any
options.
So,
of
course,
when
someone
added
50
of
our
20
options,
it
just
didn't
fail,
but
this
is
this
is
clearly
log
needs
to
check
on
who
did
that
so.
A
B
A
A
A
Port
8080,
but,
of
course,
the
service
defines
both
80
80
and
80
79,
and
when
I
was
testing,
it
I
had
all
sorts
of
problems.
The
sto
caudill
TLS
check
reported
the
879
who's
doing
MPLS,
but
it
wasn't.
The
security
that
we
thought
was
turned
on
was
not
turned
on,
because
the
IP
tables
never
sent.
The
traffic
through
envoy
so
know
is
to
rule
said.
Any
effects,
though,
do
take
it's
a
serious
problem
that
if
you
don't
define
a
container
port,
you
don't
get
your
security
stuff,
respected.
A
C
C
F
B
A
F
A
No
I
had
planned
it
for
the
beginning,
but
I
will
put
you
at
the
end,
so
maybe
in
another
20
minutes.
Okay,
so
right
now
we're
doing
a
bug
scrub
trying
to
see
all
the
user
experience
bugs
that
we
need
here.
So
we
can
get
someone
signs
to
them
on
Dartmouth
control,
plane,
starting
after
full
restart.
It's
tagged
in
a
bunch
of
areas.
So
I
will
ignore
this.
One
deployment
manager
template
does
not
work
with
the
latest
releases.
Does
anyone
know
what
this
is.
A
Would
flick
act
like
this
is
hard
for
me
to
get
serious
about
commanding
someone
to
fix
it
thing
with
this
handle
envoy
exit
when
killed
by
accident.
This
one
seems
to
be
just
us.
A
C
C
Well,
I
mean
I,
think
it's
a
user
experience
area,
but
it
kind
of
follows
into
the
functional
domain
of
a
networking
group
I.
Imagine
a
lot
of
these
issues
would
would
be
similar,
so
I
don't
know
what
the
right
flow
for.
That
is,
because
it
sometimes
we're
all
responsible
for
user
experience
and
I.
Think
it's
hard
for
this
group
alone
to
to
solve
a
networking
usability
problem
or
a
telemetry
usability
problems
on
their
own.
That's.
A
A
Okay,
chair
I'm,
sorry
hard
time
hearing
you
but
I
will
not
try
to
get
an
owner
for
it.
Pilot
log
levels
are
not
correct.
This
is
just
user
experience,
probably
got
Auto
tagged,
I've
noticed
since
the
Auto
Tiger
has
been
labeling
things,
I
haven't
been
able
to
or
haven't.
I
have
been
forgetting
to
triage
things
as
quickly
as
before,
when
sidecar
rejects.
Envoy
configs
is
tagged
as
info
level.
A
So
he
reports
the
proxy
status
and
open
TLS
checks,
do
not
work.
I
will
assign
myself
just
to
triage
it
and
figure
out
if
it's
correct
or
not
list
of
applications
incompatible
with
sto
is
a
I
can
start
to
be
an
epoch,
although
it's
not
officially,
as
in
of
epoch,
I've
added
to
it
this
week,
the
problem
that
we've
been
having
with
JME
XJR
MP.
A
So
in
this
item
I'm
trying
to
capture
all
of
the
other
issues
where
we
have
problems
with
compatibility.
So
what
I
noticed
was
that
people
were
fixing
compatibility
problems
with
STL,
but
it
was
hard
to
know
if
we
were
making
progress
or
not.
So
this
sort
of
captures
middleware
that
I
think
either
doesn't
work
or
is
almost
impossible
to
configure
with
this
deal
and
where
the
hope
is
that
as
we
these
things
we
cross
them
off
from
here,
and
so
we
can
sort
of
keep
proud
of
our
progress.
A
A
In
other
cases,
some
people
claim
they
work
like
cockroach
DB,
but
I
couldn't
get
them
to
work
and
some
cases
they
simply
don't
work,
putting
them
here
at
least
let
people
Google
them
and
see
it,
but
it
would
be
nice
if
looking
people
spend
three
hours
trying
to
get
couch,
bangs
work
or
elasticsearch
that
they
sort
of
got
some
hints
here.
How
should
I
make
check
like
this
is
doc
or
what
should
I
do?
I.
B
D
G
A
Makes
sense
so
commander
requested
a
semantic
difficult
for
the
proxy
config.
A
A
A
Have
some
tricks
getting
around
the
keys,
I
use,
JQ
and
not
being
sorted,
I
use
JQ
and
things
when
I'm,
comparing
this
myself
but
I
have
certainly
found
it.
Difficult
I
spend
a
lot
of
time,
comparing
config
domes
when
I
was
trying
to
people
of
the
issue
that
are
not
to
be
no
container
for
being
exposed.
Sorry
things
that
we
have
for
this
would
be
good,
but
I.
Think
no
one
had
a
good
idea
of
a
predictive
tool
to
use
with
this.
H
E
H
E
A
Okay,
so
I
think
I
think
most
of
this
stuff
as
an
owner
or
is
not
exclusively
musics
games.
I
will
try
to
go
back
and
on
the
next
call
and
try
to
focus
on
his
name.
Please
I
music
evil
football.
So
maybe
we
should
move
on.
Oh
and
the
Cuban
Amazonia
I
am
expecting
to
come
from
Walmart
labs,
who
has
released
cube
man
and
I
thought
that
it
would
be
interesting
for
people
to
see
how
Romer
Labs
thinks
users
are
going
to
be
interacting
with
kubernetes
and
sto
Frank.
F
Just
give
a
very
quick
update
at
cube
mass.
It
came
from
the
situation
where
we
want
to
investigate
some
complicated
and
use
cases
bug.
Actually,
when
we
deploy
a
steal
and
user
sto
and
kubernetes.
We
notice
that
some
challenge
of
like
a
gathering,
informations
and
and
also
search
for
the
information
that
we
want
to
search.
It
was
actually
a
UK's
idea
to
initiate
this
this
project,
and
he
is
also
would
be
the
person
who
demo
comparing
today
so
good.
My
alert
you
yeah,
he
said.
B
F
H
Yeah
I
was
going
to
start
with
that.
Actually,
what
Frank
just
said
that
just
for
record
15
minutes
is
not
enough.
You
cannot
see
cube
men
in
15
minutes,
ok,
so
that
is
one
second
thing.
I
wanted
to
get
out
of
the
way
cube.
Man
is
a
superhero,
so
we
laid
that
to
rest.
So,
having
said
that,
so
what
is
you've
men,
as
you
can
see,
cumin
is
really
an
application.
It's
a
tool.
I,
don't
know
if
many
of
you
have
visited
the
git
repo,
but
I
have
a
kind
of
you
know
exhaustive
documentation
there.
H
So
you
could
get
an
idea.
You
know
all
the
features
of
human
later.
Also
after
this
presentation,
as
fifteen
minutes
means
that
you
know,
I'll
probably
have
to
rush
through
it
in
first
half
and
then
I
will
leave.
You
know
at
least
five
minutes
for
some
discussion,
questions,
etc.
That
I
could
answer
so
cube.
Man
really
is
a
tool
primarily
to
investigate
to
Banaras
and
sto,
and
let
me
just
quickly
show
you
what
all
it
involves.
So
basically
the
way
it
works
is
you
see
the
the
basic
application
starts
with
nothing.
It's
empty
all.
H
Right
well,
I
see
there
is
a
toggle
button
here.
What
you
do
is
you
connect
to
a
cluster?
So
when
you
say
select
cluster,
it
will
look
at
your
local
cube.
Config
read
a
list
of
all
the
click
clusters
that
you
make
you
have
connected
to
in
the
past,
and
that's
that's
where
this
list
is
coming
from
and
it
will
also
use
your
saved
context.
H
Your
your
auth
that
you
use
to
connect
to
a
vessel
which
was
saved
in
your
tube
config
to
also
then
you
know,
use
the
same
arc
to
connect
via
API
is
so
here
it
is
a
cluster
as
I'm
just
going
to
go
to
my
doctor
desktop
and
ignore
the
fact
that
it
shows
up
twice
it's
a
it's
a
problem
in
my
config
file.
So
what
you
could
do
is
you
could
select
a
cluster
and
you
could
just
stop
there.
H
You
know
if
I
say
okay,
then
it
will
connect
with
a
cluster,
the
kubernetes
cluster
I
selected
and
now
it
gives
me
a
bunch
of
recipes,
but
before
we
go
into
that,
if
I
go
back
to
the
selection
box
again
one
more
thing
I
could
do.
I
could
go
to
namespaces
and
now
just
showing
all
the
namespaces
that
are
present.
I
could
search,
for
example,
sq
right
since
we
are
dealing
with
sto,
the
audience
is
two-dimensional.
H
This
whole
meeting
is
all
STS
I'm
going
to
you
know,
focus
primarily
on
SK
related
stuff,
so
I
could
limit
my
scope.
So
what
is
happening
here
right?
What
am
I
selecting?
I
am
selecting
my
context.
Basically,
I
am
saying
I
want
to
work
with
this
one
cluster
and
optionally
I
want
to
work
with
these
selected
namespaces,
and
then
we
will
see
you
know
how
it
plays
a
role
when
you
select
namespaces
to
limit
your
context
going
back
to
clusters.
One
of
the
things
I
could
do
is
I
could
work
with
multiple
clusters.
H
So,
for
example,
I
could
you
know
select
two
three.
It
allows
up
to
three
clusters
to
be
selected
for
now
why
a
limit
on
three
are
basically
for
screen
real
estate,
as
you
will
see
as
when
we
go
through
some
of
the
features
that
you
know.
The
screen
real
estate
is
always
limited,
and
hence
you
got
to
draw
a
line
somewhere.
So
I
drew
a
line
on
three,
so
let
me
go
ahead
with
that.
H
I
change,
my
selection,
so
now
you
see
it
is
showing
me
on
the
table
that
I
am
connected
to
my
daugher
desktop
I
have
selected
sto
system,
namespace,
all
right
bunch
of
the
things
that
I'm
not
going
to
go
through
all
the
features
in
a
fifteen
minutes
like
Lindgren,
it's
not
shows
menu,
so
you
get
an
idea,
and
you
know
most
all
of
you
are
pretty
much
overly
already
aware
of
all
the
aspects
of
culinary's
and
Sto.
So
you
see
you
can
you
you
can
do
a
bunch
of
things
around
clusters.
H
You
can
get
a
quick
overview
of
you
know.
What
does
my
cluster
involves
in
this
case?
This
cluster
has
only
one
node,
we're
just
showing
you
the
annotations,
the
labels.
You
know
how
many
CPUs
etc
etcetera.
Just
getting
a
cook
quick
overview
of
the
cluster
I
could
get
a
usage
of
the
CPU
memory.
This
is
going
to
fire.
Cupid
I'll
come
and
get
collector
the
data
and
come
back
I'm
going
to
just
move
on
I'm
not
going
to
wait
for
it.
H
We
get
the
HPA
status,
one
of
the
interesting
things
that
you
could
do
is
between
out
of
my
own
experience,
and
many
of
these
features
is,
you
know,
came
to
my
mind
as
I
was
giving
playing
to
Vanaras
and
sto.
Then
then
you
have
to
deal
with
some
issues,
then
the
pain
that
you
go
through
the
ideas
cube
meant
primarily
the
philosophy
is
that
you
know
what.
H
If
we
we
capture
those
pains,
those
learnings,
but
the
wisdom
make
it
actionable,
and
one
of
the
things,
for
example,
is
many
times
you
run
into
the
issue
where
you
see
you
have
an
IP
address
in
logs
of
one
part
right.
This
board
is
trying
to
connect
to
another
part,
it's
getting
a
5
or
3,
maybe
from
ingress
pH
way
right,
you
see
five
or
three
and
the
five
or
three
that
the
log
message
has
some
IP
address
and
you
want
to
know
who
was
there
other
other
party
that
responded
with
five
or
three?
H
So
one
of
the
things
this
plugin
with
this
recipe
here
does,
is
you
can
type
an
IP
and
then
it
can
find
the
components
that
matching
that
IP?
Now,
since
I
don't
have
an
IP,
handy,
I
will
come
back
to
this
I'll
look
at
some
other
stuff
and
then
pick
an
IP
come
back
to
this.
You
can
look
at
events,
not
that
interesting.
You
can
look
at.
Are
you
guys
able
to
see
the
screen?
Do
you
want
me
to
zoom
in
a
little
bit?
Maybe
we
can
see.
D
H
Can
zoom
in
I'm
in
cumin
allows
you
to
zoom
in
also
by
the
way
so
so
resources.
So
these
are
top-level
groupings
right,
cluster
recipes
event,
recipes
resource
recipes,
so
under
resources
I
will
find
everything
related
to
resources.
But
at
some
point
you
will
start
saying
Vedas
again,
there
are
very
too
many
menus.
I
cannot
deal
with
this
well,
I
can
start
typing.
I
can
say,
show
me
everything
you
have
about
service
okay,
so
it's
too
many
service
logs
you
see.
H
Now
it
is
as
it
has
a
this
quick
search
thing
here:
I
can
find
recipes
based
on
you
know
some
keywords:
now
it
is
telling
me-
or
you
have
a
bunch
of
things
of
related
to
service
log,
so
you
can
check
the
logs,
which
is
at
a
point
in
time.
I
can
tale
them
and
you
know,
keep
object.
So,
for
example,
if
I
go
there,
then
then
just
to
show
you
the
cumin.
You
know
general
usage
pattern
if
there
is
a
recipe
that
requires
one
or
more
services
right.
H
So
here
this
recipe
requires
one
service
to
be
selected
so
that
it
can
show
logs
of
their
service.
Now
it
is
telling
you,
oh
I
have
seven
items.
Seven
services
pick
one.
Why
seven,
because
now
we
go
into
the
the
thing
I
was
talking
about
earlier
I
limited
my
scope
to
system
namespace
right
so
now
it
is
only
looking
at
services
which
are
defined
in
the
Association
namespace.
H
If
I
remove
the
namespace
scope
and
I
keep
the
whole
cluster
is
my
scope
and
it
would
have
found
you
know
five
five
hundred
services
and
it
would
have
given
me
a
huge
list.
It
would
take
more
time
it
will
be
slower,
but
I'll
have
more
options.
I
can
then
search
here
as
well.
I
can
say:
oh
I
was
looking
for
ingress
gateway
logs
now
what
what
it
will
do.
So
what
is
happening,
and
this
I'm
just
gonna,
give
you
just
quick
examples
of
the
usage
patterns.
What
did
happen
here?
H
H
Are
there
backing
goes
in
in
those
parts?
Let
me
collect
and
combine
the
logs
from
all
of
those
parts,
and
then
it
is
giving
you
a
unified
view
of
that
all
the
logs
for
that
service
good
in
some
cases
useful
other
cases
when
you
want
just
one
part
slug,
you
could
say
I
want
pod
log,
and
now
you
see
the
same
pattern
you
can
find
check
logs
of
selected
pods.
Instead
of
you
know,
selected
services.
H
Oh,
you
have
STS
and
proxy
alright
proxy
and
same
thing,
so
you
get
the
idea
at
what
is
happening,
so
the
pattern
that
few
men
is
really
implementing
is
that
it
takes
the
knowledge
the
knowledge
of
given
it
is
the
knowledge
of
Sto,
and
then
it
tries
to
present
them
in
easy
to
consume
way
so
that,
as
an
end
user,
I
don't
have
to
deal
with
command
lines
and-
and
you
know,
trying
to
connect
the
dots
as
much
manually.
Many
of
these
recipes
can
do
it.
H
For
me,
okay,
let's
move
on
what
are
some
of
those
examples
of
how
you
can
connect
dots
I
want
to
skip
some
of
the
kubernetes
thing,
because
I
guess
you
guys,
might
be
more
interested
in
sts
stuff,
that's
already
stuff,
so
as
to
endless
recipes.
There
are
a
whole
bunch
of
them.
I
can
just
look
at
ingress
details.
It's
going
to
build
a
quick
summary
of
you
know.
H
H
So
my
investigate
way
on
this
local
docker
is
running
these
two
containers,
one
part
I
have
for
gateways
deployed
for
virtual
services
deployed
and
then
I
can
look
at
the
Y
ml
of
the
service
and
deployment
as
well
all
right,
good
I'm
going
to
skip
through
some
of
those.
But
you
can
you
can
take
a
quick
look
at
these
recipes
here
on
the
left
and
you
can
imagine
what
it
might
be
doing.
Some
of
the
more
interesting
things
find
overlapping
gateways.
For
example,
this
came
out
of
our
our.
H
You
know
real
experience
here
that
at
some
point
we
ran
into
this
bug
in
sto,
where,
if
you
have
two
gateways
which
are
overlapping
in
the
sense
of
you
know
the
same
hostname,
then
then
you
know
when
the
pilot
sends
against
the
convoy
it
breaks
it
get.
It
gives
this.
You
know,
processes,
error
and
noise
stops
starting
then
edit.
This
recipe,
this
recipe,
will
look
at
your
pleasure.
Are
there
any
gate
waves
which
have
host
overlap?
Now
the
STI
issue
has
been
fixed
since
then,
but
nonetheless,
the
recipe
lives
Gateway's
bit
missing
secrets.
H
It
will
look
at
all
your
bigot
is
defined
in
the
cluster
and
find
if
there
are
any
gateways
that
reference,
a
secret
SD
a
secret,
but
the
secret
is
missing,
in
which
case,
of
course,
you
know
it
will
not
work
right.
So
this
is
like
giving
you
an
opportunity
to
preemptively
look
at
issues,
certs
report.
It
will
look
at
the
all
the
certificates
that
are
deployed
on
investigate
way
and
then
it
is
giving
a
summary
of
how
many
gateways
are
using
SDS
search.
H
How
many
gears
are
using
mounted
search
are
more
interesting
or
more
advanced
stuff,
for
example,
is
you
can
you
know
now
you
can
use
given
to
also
investigate
envoy
proxy
stuff
right.
So,
for
example,
I
can
look
at
all
the
Envoy
clusters
like
in
the
ground,
where
listeners
route
etc.
Then,
here
I
can
search
give
me
all
the
clusters
which
have
nginx
dot.
One
run:
none
okay,
engine,
X,
dot,
service,
dot,
SC,
no
something
like
that
right.
So
you
can
basically
search
which
points
out
to
another
feature
of
cube.
Man
is
basically
any
recipes.
H
Output
becomes
searchable
and
then
it
will
highlight
it.
There
is
also
you
it
is
filtering
the
results
based
on
the
match,
so
it
gives
you
a
quick
way
for
you
to
look
at
on
were
configuration.
For
example,
right
I
can
look
at
on
requesters
fine
based
on
you
know,
my
service
name,
I
can
look
around
wipe
listeners
and
then
I
want
to
find
listener.
Ford
Ford
70-74
here
is
the
listener
for
port
70
sandy.
Okay,
then
I
can
go
into
the
overview
mode.
How
many
filter
change
it
has?
H
It
has
one
filter
chain
which
has
one
filter.
It
is
actually
pointing
to
the
already
about
HDTV
sandy
sandy
I
can
go
to
the
routes.
So
you
know,
I
was
investigating
something
some
connectivity
issue,
then
I
could
use
this
tool
and
look
at
what.
How
is
the
Envoy
listener
configured?
How
is
on
without
considered
alright
for
a
certain
port
and
then
I
can
I
can
try
to
connect
the
dots
so
the
more
interesting
stuff.
So
just
exactly
what
I
showed
you
just
now
right
what
I
did
is
I.
H
My
goal
was
to
find
on
my
configuration
for
ports
9070
a
one
of
the
advanced
recipes
which
I
am
working
on.
As
of
now
it's
not
released
yet
because
you
know
it's
not
done
yet
is
analyze
port,
so
you
see
now
I
want
the
analysis
recipes
and
here
there
are
a
bunch
of
advanced
recipes,
and
what
do
you
mean
by
advanced
well
analyze
port,
you
give
it
an
input
of
port,
so
cuban
has
many
patterns.
H
Usage
patterns,
one
of
the
things
is
that
recipes
can
take
input
and
you
want
to
give
an
input
to
recipe
your
to
prefix
it
with
slash.
Then
it
becomes
an
input
if
I
just
type
without
slash,
then
I'm
searching
in
the
output
right,
so
here
I'm,
giving
an
input
7070
I
hit
enter
now.
What
is
it
doing
now?
Cumin
is
going
to
look
at
the
whole
cluster
to
see
through
the
entire
cluster.
How
many
places
where
port
70
70
is
being
referenced
now
take
a
pause
and
think
about
it.
H
How
much
effort
would
you
have
to
go
through
to
come
to
to
collect
this
information
just
to
collect
this
information
and
then
your
analyze?
It
further
right
about
whether
the
information
has
any
holes.
There
is
anything
missing,
but
should
just
to
find
out
that
how
many
items,
how
many
configurations,
how
many
services
gateways
virtual
services
do
I
have
in
this
cluster,
which
relate
to
port
70
70?
Well,
you
all
you
have
to
do
is
types
and
assembly
and
Sherman
has
found
everything.
H
So
let
me
go
into
the
overview
mode
just
to
show
you
quickly
what
all
it
found.
It
found
that
there
are
three
gateways:
three
virtual
services,
and
there
are
five
services.
Then
it
is
also
showing
me
the
ingress
get
we'll
listen
stairs
where
it
is.
It
looked
at
the
ingress
gateway
and
it
is
making
sure
whether
in
this
gateways,
listening
on
this
port
or
not
right,
it
said
that
Oh
port,
seventy
seventy
is
forwarded
directly
and
now
I'm
talking
out
the
best
choice
of
words.
H
A
D
A
H
Me
let
me
do
that,
just
that,
because
you
know
that's
that's
an
excellent
use
of
cube
men.
So
let's
say
I
went
back
to
my
context
and
I'm
selecting
two
clusters
now
right
when
I
select
two
clusters,
actually
I
will
go
back
and
limit
it
to
its
system
namespaces,
because
otherwise,
some
of
these
clusters-
these
are
def
clusters.
They
are
not
really
very
fast.
These
are
one
of
our
internet,
Walmart
Labs
clusters
and
some
some
of
them
actually
take
a
long
time
to
you
know,
connect.
H
So
let
me
go
here
and
let
me
select
sto
system
I
transferred
acting
a
system
namespace
and
on
both
these
classes.
So
now
I
changed
my
context
to
two
clusters.
Now
there
are
a
whole
bunch
of
things
that
I
can
do
in
terms
of
comparison
right,
exactly
what
you
said
right.
So
let
me
quickly
search
what
all
do
I
have
in
terms
of
compare
I
can
compare
two
secrets:
config
man,
I
can
compare
clusters.
Namespaces
I
can
see
let
I
click
on
this
program,
person
name
spaces
just
in
just
a
quick
idea
right.
H
This
could
be
a
adding
point
in
your
investigation
churning.
You
want
to
see
which
namespace
exist
in
which
quest
cell-
maybe
you
know
as
a
part
of
our
deployment
pipeline,
we
miss
some
namespace
and
what
it
is
doing
is
it
took
a
super
set
of
all
the
namespaces
from
both
classes.
It's
showing
you
a
master
list
on
the
left
and
for
each
cluster.
It
is
telling
you
whether
the
namespace
exists
in
this
one
class
or
another
right,
and
it's
also
highlighting
it
in
red
when
it
sees
a
difference.
H
So,
for
example,
foo
is
is
not
in
f1.
It
is
an
ef3
right
above
same
thing
and
and
vice
versa.
Right
there
are
some
namespaces
here
and
there.
Okay,
let's
keep
going,
compare
names,
there's
deployments.
What
does
this
do?
It
will
look
at
in
the
selected
namespace.
It
will
look
at
all
the
deployments
and
tell
you
which
deployment
is
present,
where
all
right
so
again,
the
same
idea.
Telemetry
is
present
in
AF
1,
not
in
here
3.
It
gives
you
a
clue
as
so.
You
can
now
investigate.
Oh
wait,
a
second.
H
Why
assignment
we
not
present
in
EF,
3
I
could
go
further.
I
could
choose
two
deployments
right,
so
I
can
see
a
ingress
ingress.
Why
not
from
f1
and
f3
right
see
what
I
am
doing
now.
I
am
selecting
ingress
gateway
deployment
from
two
different
clusters
and
I
am
asking
tube
men
to
compare
it
and
now
what
is
doing
it
is
going
through
its
field.
D
H
Field
so
bunch
of
this
fields,
but
a
bunch
of
the
data
it
extracts
from
the
from
the
deployment
yeah
mel's,
and
it
shows
you
line
by
line
and
at
the
bottom.
It
also
shows
you
the
full
yeah.
No,
but
it
it
extracts
certain
field,
also
just
so
that
it
can
highlight
certain
data
for
you,
for
example,
it
is
telling
you
that
the
annotations
are
different.
Ok,
one
is
on
division.
One
revision,
three
doesn't
matter
I
know,
but
you
get
the
idea
right.
What
it
is
doing.
It
is
telling
you
the
templates
are
different.
H
H
Let's
see
what
it
does
so
plug-in
is
it
allows
you
to
compare
engraved
gateway
onward
proxy
configurations
across
pods
which
can
come
from
different
clusters.
Not
getting
more
interesting.
I
can
select
this
and
this
right,
two
different
parts
one
is
coming
from
EF
modem
coming
from
EF,
3
I
can
say
compare
now.
This
is
where
what
it
is
going
to
do.
It
is
going
to
load
the
unwelcome,
shakes
the
whole
config
from
one
pod.
H
Then
the
second
part
it
will
compare
them
and
then
it
will
tell
you
what
matches
what
does
not
match,
and
it
will
do
for
all
all
four
categories
in
the
bootstrap,
the
clusters,
the
listeners,
the
the
routes.
So
it
takes
time
it
is
going
to
be
not
do
it
one
category
at
a
time
present
the
output
to
you,
you
can
look
at
the
output
and
so
what
what
kind
of
output
it
produces?
Well,
it
did
not
have
a
good
started,
saying
hey
at
least
the
bootstrap
does
not
match.
H
Ok,
fine,
and
it
is
sending
you
that
I
see
this
one
bootstrap
in
in
cluster,
one
don't
see
it
in
another,
keep
going
then
it
says:
oh
I
see
this
would
step
on
the
right
side.
I,
don't
find
it
in
the
in
the
first
one.
Let
me
go
into
the
overview
mode
quickly.
You
click
on
the
top.
You
get
the
overview
mode
right,
so
let
me
start
from
the
top
again
bootstrap,
but
we
mostly
know
that
good
stress
actually
will
differ
right
because
ooh
step
is,
we
are
good
set
configuration.
H
H
Basically,
because
you
know
if
we
go
into
too
much
details
and
run
out
of
time,
but
what
this
plugin
is
doing
is
it
is
comparing
the
Anglican
sakes
and
it
will
tell
you
four
things
that
match
it
will
just
give
you
a
list
like
this
on
the
top
right.
It
was
just
80
matches
for
matches.
Hence
I
am
NOT
going
to
show
you
the
conflicts
of
80
and
443.
H
If
you
want
to
see
the
conflicts
of
those,
there
are
other
recipes,
they
will
do
that,
but
for
those
there
where
there
is
a
mismatch
either
the
same
config
is
present
in
both
places
where
they
are
mismatched
or
they
are
present
in
one
and
missing
in
another.
Then
it
will
show
you
the
country
right.
Then
you
can
go
further.
You
can
compare
pilot
can
take
bit
one
on
the
way.
What
it
will
do
is
you.
You
know.
That
means
is
just
click
on
just
so
that
you
get
an
idea.
H
It
will
load
a
list
of
Hanoi
proxies.
It
will
ask
you
to
select
one
proxy,
then
it
finds
out
which
pilot
is
feeding
that
problem
right,
because
we
know
that
I
different
pilot
instances
feed
different
on
my
proxy
instances.
At
any
point
in
time,
one
envoy
could
be
connected
to
any
one
of
those
pilots,
so
once
a
non-word
proxy
selected
and
in
this
recipe
an
ingress
gateway
is
being
treated
as
onward
proxy.
So
that's
fine
and
remember
my
context
is
limited
to
a
system
right
now.
H
That's
why
the
only
foxes
I
am
seeing
is
investigating
if
I
expend
my
contacts
I
will
see
the
side
guards
as
well
right,
then
you
select
one
on
a
proxy
now
it
found
out
which
pilot
it's
connected
to,
and
it's
comparing
the
config
from
wire
and
and
the
on
web
proxy,
and
it's
going
to
tell
you
oh
listen,
Ertz
match.
There
is
no
mismatch
in
the
listener.
It's
setting
you,
which
routes
matched
all
these
routes
are
matching
no
miss
my
sprouts,
which
is
good.
H
It
skipped
a
static
route,
some
logic
where,
if,
if
cumin
not
figured
out
how
to
compare
certain
routes,
which
is
most
people
typing
static
routes,
I
just
skip
them,
but
I
list
them
here.
Nonetheless,
at
hey,
I
did
not
compare
them.
Then
it's
saying
these
clusters
are
matching,
let's
say
if
it
found
any
mismatch.
I
hope
not.
That
would
mean
that
pilot
is
in
sync
with
this
animal
proxy.
But
now
you
see
the
the
value
of
this
rate.
No
miss
miss
Krista
country.
What.
H
D
D
D
A
G
H
Electron
I
don't
know
if
you're
familiar
with
electron
electron.
Basically,
that's
a
node.js
and
chrome
vf8
from
v8
engine
combined
electron
is
a
framework
which
basically
uses
nodejs
for
the
backend
takes
the
chrome
engine,
rendering
puts
them
in
a
single
package
which
can
be
used
to
build
desktop
applications.
So
then
the
entire
code
base
is
written
in
Thai.
Script
runs
on
in
in
the
v8
engine
for
rendering,
when
some
node.js
for
some
of
the
back-end
stuff.
So
basically
it's
exit,
but
it's
packaged
as
a
as
a
desktop
application.
If.
A
H
Those
will
do
more
recipes,
basically,
because
you
know
one
of
the
things
that
that
that
the
design
of
this
tool
is
that
when
you
launch
the
tool,
the
tool
will
wait
for
you
to
do
something.
But
then
there
could
be
one
recipe
like
there
are
some
of
the
recipes
like
look
at
this
right,
analyse,
service
details
and
routing
all
I
have
to
do.
Is
click
this
recipe
and
pick
a
service.
H
So
in
this
case
again,
since
my
context
is
limited
to
a
system,
I
am
only
singing
a
system
services,
but
now
what
it
is
doing
is
exactly
to
some
extent
what
you
were
thinking.
I
asked
human
to
just
run
like
a
full,
thorough
checkup,
but
just
from
the
point
of
view
of
this
service,
what
it
is
going
to
do
is
look
at
a
whole
bunch
of
things
which,
which
I
will
show
you
later,
but
it
will,
even
you
know,
go
through
every
basically
and
if
it
detects
any
issues,
it
will
give
it
to
you.
H
But
then
one
of
the
things
I
think
what
you
had
in
mind
is
you
know
you
just
launched
the
application
and
then
it
is
proactively
doing
something
right
now
it
doesn't
do
anything
proactively.
You
have
to
then
trigger
a
recipe
and
that
recipe
could
you
do
something
very
little.
It
could
do
some
a
huge
amount
of
work.
H
H
E
H
What
I
was
expecting
is,
you
know,
maybe
say:
I
was
excited
to
present
it
to
Pacific
team
shooting
in
general
is
sorted
and
once
one
you
know
the
experts
of
this
place
right.
So
once
you
start
thinking
you
look
at
the
students
start
thinking
that
you
can
probably
come
up
with
a
lot
more
ways
that
you
know
our
your
users.
Life
can
be
anything
and
more
things
that
foods
like
women
can
do
and
so
on.
So
then
we
can
keep
enhancing
so
I
hope
that
we
will
really
get
six.
Are
the
defects
I
hear
it.
H
Now
is
basically,
this
is
the
go-to
tool
inside
well,
not
for
any
issue
related
to
sto
at
least,
and
then
it
is
now
that
option
is
spreading
we've,
basically,
anything
which
is
either
onboarding.
Saudi
has
left
something
right.
We
provide
as
well
and
su
as
it
has
a
platform,
and
then
there
are
product
teams
which
were
responsible
for
bringing
their
services.
Their
applications
on
board
onto
the
net
is
an
ST
right.
So,
as
far
as
funding
goes,
you
know
they
have
been
already
using
the
student.
H
They
are
seeing
the
value
this
pretty
much
for
any
kind
of
SEO
related
issue
that
we
investigated
come
of
it
now.
The
service
teens
are
also
associated
hardening
their
applications
on
good
on
certain
areas.
That
says
they
will
probably
see
the
value
they
start
using
it
as
well.
So
that
option
is
10.
F
B
F
I
D
D
There's
a
couple
tools
that
things
out
there:
I
guess
that
kind
of
help
you
use
it
to
troubleshooting
issues
related
to
its
deal.
So
Chi
alley
will
be
the
first
one
that
comes
to
my
mind,
even
though
hi
Ally
doesn't
do
what
you
have
today.
But
people
have
been
looking
at
Chi
alley
to
do
some
of
the
troubleshooting,
the
other
thing
it
will
be
sto
CTL,
which
we
are
also
adding
operator
capability
into
it
still
sepia,
but.
D
It's
more
about
x-men
land,
but
not
in
so
much
about
that.
Food,
however,
is
still
CTL
does
have
a
command
to
allow
you
to
long
back
to
dashboard.
So
you
could
do
is
fill
CPR
dashboard,
which
potentially
launches
one
was
a
data
dashboard,
and
this
might
be
one
of
the
what
Oh
integrates
with
one
of
our
existing
ones.
So.
A
D
A
H
A
H
Run
as
a
server,
yes,
it
will
require.
You
know
the
code
itself
is
node.js
code
or
in
a
basic
types
of
code
which
can
be
reflected
out
and
then
then
you
can
make
it
as
a
hosted
server
side
offering,
but
the
current
incarnation
of
this.
The
way
it
is
set
up
right
now
is
meant
to
be
a
desktop
application.
D
H
E
Quick
quick
question:
in
my
normal
day-to-day
I
always
use
localized,
tube
config
files
or
a
merged
coop
config
there
any
way
to
get
to
man
to
use
I,
don't
adjust
my
local
global
coop
config
I.
Never
do
that,
because
there's
just
too
many
different
types
of
clusters
that
modify
that
at
weird
way,
so
I
isolate
them.
I
avoid.
H
E
D
E
E
H
E
H
A
D
I
A
A
I
Yep
no
problem,
if
you
could
take
a
look
in-
and
you
know,
we
can
certainly
do
it
offline
through
the
document
as
well.
All
right,
thank
you.
Thanks.