►
From YouTube: Layer5 Development Meeting - March 11th, 2020
Description
@nveenjain delivers support for @istiomesh v1.5!
Join the community calls and learn more about Meshery, the management plane for service meshes at https://meshery.io.
B
C
A
A
E
A
A
No
kidding,
that's,
oh,
that's
fantastic,
because
you're
gonna
get
really
good
at
that
now
and
actually
I
expect.
Probably
mystery.
Ctl
will
be
probably
much
improved
or
the
course
of
your
time
at
Red
Hat,
which
would
be
great
so
yeah.
E
B
A
E
A
A
F
G
So
yesterday,
like,
yes,
really,
you
bring
me
to
issues
which
I
had
to
work
on
so
yesterday,
I
couldn't
get
time
because
of
the
holy
so
I'm,
looking
into
both
of
things
mainly
they
killed
with
these
missing
safety
ring
gke,
so
I'll
work
on
them,
if
at
all
so
I
think
I'll
work
on
both
issues
and
I'll
see
what
also
I
think
these
are
some
poor
changes.
So
I'll
do
an
erase
of
PR
so
that
you
can
approach
you.
Can
they
do
it?
Yeah.
A
All
of
them,
and
so
not
by
the
way
Sanjay
Anand,
has
been
probably
the
longest
standing
contributor
that
has
focused
on
mesh
resi
TL
specifically,
and
it's
so
good
good
that
he's
here,
good
that
you
guys
can
meet
and
talk
him
and
and
he
can
help
he'll,
be
helping
review
the
PRS
and
things
like
that.
So
we're
I'll
say
this:
that
the
CLI
is
one
of
those
areas
that
you
can.
A
A
You
know
a
point
of
human
interaction
and
for
many
administrators
a
lot
of
times
it's
the
more
comfortable
interaction
because
they
can
get
a
lot
of
things
done
quickly.
They
can
script
around
it
and
so
considering
giving
some
forethought
as
to
the
structure
of
that.
Cli
is
important
because,
just
like,
when
you
design
a
user
interface-
and
you
use
start
to
train
and
teach
the
user,
how
it
is
that
the
application
works
that
over
time.
A
So
in
this
case,
when
we
think
about
the
work
that
you
might
be
be
doing
and
the
feature
that
you
might
be
enabling
around
letting
people
create
their
cube,
you
know
export
their
cube,
config
from
mini
cube,
letting
them
more
or
less
restructuring
their
cube,
config
to
be
gke,
compatible
or
restructuring
their
cube,
config
to
be
AWS
or
eks
compatible
that
when
we
think
about
what
command
the
people.
What
the
syntax
should
be.
A
Having
said
all
that,
much
of
what
we're
that
forethought,
that
we
try
to
give
or
that
that
design
goes
into
this
document
here,
so
this
is
the
mastery
CLI
commands
and
documentation.
This
is
more
like
mesh.
We
a
bit
of
master
CLI
reference
and
an
on
did
some
I
think,
just
in
the
last
time
that
we
spoke
last
week,
we've
a
new
command
has
come
forth.
The
perf,
command
and
I
know
you've
been
giving
some
consideration
to
this
and
really
the
perf
command.
A
A
G
A
G
A
Yeah,
that's
a
great
question:
I'm
glad
that
you're
thinking
that
far
forward
the
reality
is
that
today,
inside
of
nursery,
it
doesn't
have
an
object.
It
doesn't
have
a
construct
for
allowing
measure
e
to
connect
to
multiple
clusters
and
that
clearly
needs
to
change
that.
Clearly,
we
need
to
allow
what
you
just
said,
which
is
connecting
to
a
couple
of
different
instances
of
gke
another
eks.
A
A
So
there's
certainly
a
design
session
to
be
had
there,
but
in
because
we
don't
have
that.
For
the
moment
you
wouldn't,
if
the
moment
you
wouldn't
need
to
consider
the
for
the
moment,
you
will
master
CTL
will
only
connect
to
one
cluster
or
you
can
limit
it
to
that.
That
makes
sense.
So
so,
if
someone
runs
your
CTL
config
generate
and
they
they
export
and
generate
their
mini
cube,
compatible
config
that
that
whatever
they've
generated
last
will
overwrite.
G
A
Getting
to
the
multi
environment
would
be
that
would
be
awesome.
That'll
be
that'll,
be
one
of
those
things
where
we
actually
need
to
finally
design
a
little
bit
of
an
object,
an
object
model
for
the
for
the
project,
because
the
the
point
of
being
bringing
forth
an
environment
will
will
be,
like
will
be
to
say,
hey.
These
are
the
different
constructs.
These
are
the
different
objects
that
are
related
to
one
another,
and
how
are
those
relationships
are
done?
A
Sometimes
it's
a
great
point
like
we
actually
really
need
that
today,
because
my
Cherie
has
you
can
load
up
any
number
of
adapters
inside
inside
of
a
running
instance
of
measure
e
you
can
and
we
haven't
really
diagrammed
that
and
measure
E
is
quite
powerful
powerful.
In
that
regard,
you
can
use
a
measuring
steel
adapter
to
connect
to
any
number
of
instances.
Events
still
it's
that
adapter
is
stateless,
so
you
can
connect
to
a
bunch
of
different
sto
deployments.
A
You
can
also
load
up
any
number
of
sto
adapters
and
like
multiple
instances
of
the
sto
adapter
in
your
mesh
redeployment
and
you
you,
you
might
want
to
do
that
because
maybe
you've
got
a
multi
region
thing
going
on
our
multi
cluster
thing
going
on,
but
we
haven't.
We
have
a
simple
document
that
is
in
the
in
the
user
guide
in
the
Messer
documentation
that
talks
about
that.
But
there
isn't
a
diagram
and
so
yeah
we're
going
to
end
up
meeting
some
of
these
object.
Diagrams
object.
A
Nice,
ok,
good
good!
Well,
in
this
case,
you
can't
you
can't
go
wrong,
I!
Think,
like
yeah,
yeah,
I
guess
what
I
was
trying
to
call
out
is
we
haven't
given
as
much
forethought
as
we
need
to
to
the
command
syntax?
But
if
you
were
to,
but
so
long
as
you're
taking
a
stab
at
it,
we'll
figure
it
out
as
we
go.
E
A
E
A
E
A
I
A
Is
a
global
flag
today,
so
these
flags
here
are
global.
Some
of
them
are
I.
Think
a
couple
of
them
are
kind
of
boilerplate.
If
you
will,
this
is
help
in
this
toggle
anon
had
worked
on
a
V
for
a
version
which
is
nice
because
it'll
just
tell
you
what
version
of
my
sri
CTL
the
client
that's
running
and
more
or
less
there's
something
of
a
placeholder
for
assessing
what
version
of
measure
e.
This
server
is
running
so
very
similar
to
cube
CTL.
A
A
Sanjay,
what
I
was
trying
to
say
is
that
there's
another
global
come
flag,
config
that
I
don't
know.
Okay,
it
says
an
error.
This
flag
needs
an
argument.
I
this
command
does
should
work.
It's
related
to
you,
you
trying
to
produce
mini
qbg,
ke,
+
e
KS
configs
at
this
command
lets.
You
tell
mastery
CTL
where
your
cube
config
is
by
default.
Mastery
CTL
will
look
to
your
current
users,
home
folder
and
look
to
a
hidden
folder
in
there
and
then
a
gamma
file,
a
docker
composed,
compatible
yam
will
file,
and
so
I
do.
A
A
A
Your
things
like
that.
So
this
this
config
is
the
measure
II
CTL
config,
the
measure,
each
measure
ECT
l
today
it
controls
the
lifecycle
of
measure
e
and
so
on.
My
machine
measure
E
is
not
running,
but
if
I
did
a
mastery
CTL
start,
it
will
take
that
config
file.
Oh,
it
will
reference
that
config
file
and
basically
do
in
this
case
it's
essentially
under
the
hood.
It's
doing
something
like
a
docker
compose
up
using
that
file
like
this.
G
A
A
You
know
the
because
when
you
look
at
mastery,
just
as
the
brief
review
speaking
of
object
diagrams
and
things,
we
do
have
at
least
this
architecture,
diagram,
posted,
I
think
I'm
starting
to
lose
my
voice,
which
is
bad,
so
I
apologize,
but
measure
itself
measure
II
as
an
application.
It's
it's
these
group.
It
is
these
green
rectangles,
it's
six
or
seven
containers.
A
A
A
A
Those
adapters,
those
containers
there's
one
container
now
running
on
my
system
for
the
measure:
II
CPX
adaptor,
that's
a
citrix
service,
mash,
there's
one
for
Miss
Tia.
You
know
for
each
of
these
types
and
as
a
user
in
the
user
interface,
when
you
hover
it's
kinda
nice,
because
it'll
tell
you
hey
that
adapter
is
it's
a
container
that's
running
locally
in
my
docker
environment
and
it
is
listening
on
port
10000
and
what
it's
listening
for
is
commands
from
measure
e
itself.
A
So,
as
a
user
I'm
sitting
here
up
here,
interacting
in
the
UI
I'm
gonna
be
clicking
things
in
the
UI
and
telling
telling
missery
to
tell
the
adapters
over
GRP,
see
and
in
the
case
of
consul
I
think
it
was
running
on
10,000.
It
would
send
these
gr
and
send
the
commands
over
G
RPC
over
port
10,000,
and
if
the
command
is
to
spin
up
an
instance
of
console
well
measure
e
and
the
mesh
readapt
err,
it
will
need
a
cubed
config
file.
A
A
Basically
the
certificate,
that's
within
there
to
go
over
and
speak
to
or
invoke
rest
calls
against
cube
api
and
that's
and
much
of
how,
today,
much
of
how
ma
cherie
works
is
almost
all
of
my
series
functionality,
not
all
of
it,
but
most
of
it
expects
that
that
you're
going
to
be
interfacing
with
kubernetes,
and
so
when
massery
starts
up
initially
as
a
server
as
a
container
that
container
that
I
started
on
my
local
system.
Just
now
it
will.
It
has
a
local
volume
mapped
to
my
user.
A
Lee
has
a
local
volume
map
to
my
user's
home
folder
and
in
my
users,
home
folder
is
also
the
default
location
for
my
cube,
config
and
so
a
mystery
since,
when
the
container
starts
it
has
access
to
that
home
folder,
it
looks
for
a
cube.
Config
file
under
you
know:
dot
cube,
slash,
config,
it'll,
open
it
up
automatically.
You
know
in
you
know,
basically
without
asking
the
user
it
opens
it
up
and
reads
it.
A
It
finds
the
you
know,
irrespective
of
how
many
clusters
and
how
many
contexts
you
have
in
there
it'll
load
them
all
up.
That's
great,
but
basically
kind
of
ignore
everything,
except
for
the
current
cluster.
Are
the
current
context.
I
forget
what
the
net
I
think
its
current.
You
know
current
cluster
and
then
current
context,
and
then
it
will
display
that
to
the
user
here
in
the
UI
they're
like
because
I
didn't
have
to
convicted
measure.
A
F
A
You
wanted
to
change
that,
though
you
you,
the
user
as
a
user.
You
would
go
to
settings
up
here
and
under
settings.
There
are
kind
of
three
tabs
there's
one
about
your
environment,
which
is
the
Conte
and
the
odd,
like
literally
there
isn't
an
object
inside
of
the
mesh
recode
called
environment,
but
we
need
to
make
one
at
some
point
so
that
you
can
connect
to
any
number
of
kubernetes
instances
right
now.
A
It
just
assumes
that
you're
only
gonna
get
to
one,
which
is
you
know,
that's,
okay,
it's
not
great,
and
in
this
UI
you
can
connect
to
your
there's,
an
area
for
configuring
adapters,
configuring
connections
to
CRO,
fauna
and
previous.
So,
under
this
environment
it
basically
okay,
well,
hey!
It
found
the
fact
that
I'm
running
I'm
using
an
iMac
right
now,
I'm
running
docker,
desktop,
which
is
really
convenient
for
running
a
single
instance
of
a
single
node
kubernetes
cluster,
which
is
what
I'm
running
so
it
found
it.
A
It
loaded
it
up
and
it
just
assumes
I'm
gonna
use
that
if
I
were
to
click
on
this
chip
is
what
it's
called
in
react.
The
component
is
a
chip
that
Nikhil
and
others
have
helped
us
create
functionality
to
do
a
quick
test
so
that,
as
a
user,
it's
just
whenever
I
click
this
it's
an
mesh
RIA
is
doing
a
quick
check
to
confirm
that
it
can
use
the
cube
config
file.
That's
loaded
up.
It
can
use
that
those
credentials
to
speak
to
the
kubernetes
api
and
I
think
it
invokes
like
a
get
kubernetes
version.
A
Number
api
version
number
something
something
simple
that
doesn't
really
matter.
It's
just
like
a
it's
sort
of
conception.
It's
a
quick
ping.
If
you
will
and
that's
why
we
see
ping
is
literally
not
ICMP,
but
if
it's
a
REST
API
call,
that's
a
goodness
me,
but
it
could
be
the
case
that,
like
look
I
also,
it's
also
true
on
this
same
system
that
I
am
running
mini
cube
and
in
part.
Yes,
that's
because
I
spent
the
cost
of
a
small
car
to
buy
this
iMac.
That
has
this
one
has.
A
You
know
32
gigs
of
memory
that
the
laptop
we
just
got
has
64
gigs
of
memory,
and
so
I
don't
want
to
buy
another
computer
for
a
long
time.
So
it's
so
they
loaded
up
with
memory.
That's
great!
The
point
is
this
machine
is
also
has
I,
believe
mini
cube
installed
right
now,
it's
stopped,
but
if
I
were
to
start
it
I
can
then
once
it
started,
I
can
go
back
into
measure
e,
UI
and
and
remove
the
you
remove
my
connection
to
docker
desktop
by
clicking.
A
You
know
by
clicking
the
X,
then
I
would
come
over
here
to
load
up
my
other
cube
config.
That
would
be
for
meri,
I'm,
sorry
for
mini
cube,
and
so
that
would
be
under
well
right
now.
My
default
config
is,
is
here
under
you
know:
dot
cube
config
if
I
were
to
go
over
to
mini
cube
and
do
export.
Actually,
did
you
see
that
just
now
you
see
how
just
now
these
the
arrangement
of
these
files
switched
around?
A
Well,
something
happened
and
that's
because
the
the
this
file
was
just
updated,
I,
don't
think
it
was
updated,
based
on
mini
cube
being
present,
but
anyway
we
can
load
it
up
and
actually
actually
that's
right.
It
was
changed
based
on
the
fact
that
I
just
started
making,
because
the
current
context
now
is
when
it
loaded
up
that
file.
A
It's
able
to
read
through
and
look
at
what
different
clusters
and
what
different
contexts
I
had
and
since
mini
cube,
just
started
in
the
time
that
you
and
I
that
we
were
just
talking
it
mini
cubed,
updated
my
cube.
My
config
file
I
mean
now
I've
loaded
that,
if
acquire
in
to
measure
e-
and
it's
saying
like
hey,
you
know
actually
measure
he
can't
it
is
not
able
to
reach
kubernetes
so
kubernetes
config
could
not
be
validated
we're
unable
to
ping
the
kubernetes
server,
which
is
an
okay
error.
A
It
could
be
even
more
helpful
if
it
gave
the
user
something
to
kill
like
if
we
were
able
to
click
on
this,
and
it
would
give
you
a
modal
that
said,
here's
troubleshooting
steps
for
what
to
do
to
figure
that
out
and
the
troubleshooting
step
sanjay.
In
this
case,
it's
actually
like
in
the
future,
based
on
the
work
that
you'll
do.
We
may
be
able
to
auto
remediate
that
error,
because
the
error
that's
happening
right
now
is
probably-
and
I
don't
know,
but
it's
probably
the
fact
that
I
don't
have
the
right
credentials.
A
Maybe
inside
of
my
cube
config,
to
talk
to
the
kubernetes
instance
running
inside
of
the
mini
cube
via.
If
I
go
back
to
my
turtle,
mini
cube
did
start
up
great
it's.
You
know
it
did
a
thing
cube.
Ctl
is
now
configured
to
use,
mini,
cube,
okay
and
that's
why
my
config
file
was
updated
literally
before
our
eyes
a
second
ago.
So
so,
if
I
do
a
cube
CTL
again,
no.
A
I'm,
assuming
because
the
mini
cube
the
configuration
inside
my
cube,
config
is
probably
references
like
an
old
mini
cute-cute
can
keep
CTL
config
you.
So
if
we
go
look
at
that
inside
of
my
my
credit
might
keep
my
cube.
Config
file
inside
of
there
I
do
have
a
few
different
clusters.
These
are
the
same
clutch.
You
know,
docker
desktop
I
have
one
called
kubernetes.
I
have
another
one
called
mini
cube,
and
these
were
the
same
ones
that
we
were
seeing
in
the
community.
A
In
the
messy
UI,
we
were
actually
seeing
the
context
names
so
kubernetes
admin,
we
saw
a
context,
name
called
docker
for
desktop.
We
saw
one
called
docker
desktop
and
we
saw
one
called
mini
queue.
Now
this
this
one
for
mini
cube,
is
using
this
certificate,
and
this
so
I
guess
the
point
of
the
the
thing
is:
if
I
were
to
go
export,
my
mini
cube,
config
right
now,
I,
don't
I,
don't
my,
irrespective
of
massery.
Just
my
user
can't
even
talk
to
the
kubernetes
instance
running
inside
of
mini
queue.
A
I
G
I
A
So
to
remediate
the
issue
manually-
and
this
is
something
that
I
think
as
you
do
work
here-
you
will
improve
the
whole
experience
because,
because
people
won't
have
to
go
do
manually,
what
we're
about
to
do,
but
within
the
mesh
read
documentation
users
if
they
go
look
at
the
supported
platforms.
Mini
cube
is
a
supported
platform.
There's
a
Quick
Start
Guide
for
each
platform-
and
you
know
part
of
that
is
telling
you
that
you
need
to
be
running
a
certain
version
of
you
know
requirements.
You
would
start
up
mini
cube
in
this
case.
A
A
A
Here
in
my
local
users
folder,
so
it
has
the
public
improper
as
the
private
key
for
interfacing,
with
mini
cube
I
as
a
as
a
local
test
here,
I.
Don't
if
I
were
to
do
cube,
CTL
and
I.
Forget
the
you
guys,
remember
the
syntax
to
be
able
to
use
it.
A
different
cube
config
like
you're,
able
to
point
to
a
new
file.
A
A
So
so,
if
that
did
work,
it
is
trying
to
reach
mini
cube,
and
it
doesn't
look
like
it's
able
to
the
point
being
is
that
normally
people
would
run
that
command,
they
would
export
their,
they
would
minify
and
flatten
their
cube
config
file.
They
would
then
go
into
the
mesh
or
UI
that
I'll
bring
up
in
the
measure
UI
they
would
go
ahead
and
they
would
go
ahead
and
load
up
that
config
file,
config
mini
cube,
yellow.
H
A
My
sure
you
will
read
it,
it
will
try
to
you,
know,
interface
with
cube
API,
and
it's
saying
that
it
can't,
which
is,
which
is
I'm
sure,
true,
that
it
can't
that
even
I
couldn't
do
it
from
the
CLI
just
from
the
terminal.
Just
now,
and
because
of
that
because
of
that
failure
we
don't
get
a
small
chip.
You
know
a
small
little
blue
chip
that
says
you're
connected
to
kubernetes.
G
A
A
Think
I
think
we
played
through
a
scenario,
and
this
is
where
we
should
write
some
of
this
down
inside
the
design
doc
about
what
these
scenarios
are
and
why
it
makes
sense
to
have
a
certain
syntax,
but
but
one
of
those
scenarios
would
be
so
I'm
a
user
I've
got
darker
desktop
and
I've
got
mini
cube.
So
basically
I've
got
to
I'm
running
two
of
the
supported
platforms.
A
If
you
will,
if
I
can
use
the
platform
term,
which
I
think
is
appropriate
to
describe
the
fact
that
I'm
running
a
kubernetes
instance
inside
of
docker
desktop
and
I'm
running
a
kubernetes
instance
inside
of
mini
queue,
so
I've
got
two
different
clusters:
they've
got
they're
there
they're
totally
separate
they
each
have
their
own
certificate.
They
each
have
generated
their
own
certificates
and
and
have
their
own
service
accounts
and
all
that
all
that
stuff
and
so
from
a
measuring
perspective.
A
A
G
That's
my
personal
TV,
so
I
will
go
back
to
my
previous
person
if
talk
about,
if
he
only
asked
for
pet
form
name
from
user,
so
your
pet
rom
name
means
my
like
DK
DK,
meaning
right
right.
So
in
gke
formula,
if
I
have
two
clusters
right,
so
how
will
I
distinguish?
Because
if
I
only
take
the
pet
from
name
from
user
and
run
the
run,
the
jakey
jakey
come
on
in
background?
G
I
won't
be
able
to
distinguish
between
multi
clusters
if
I
am
getting
two
clusters,
if
I
have
two
cluster
of
the
cure,
or
forgive
any
case
in
particular
region
I
have
to
testify,
will
give
thee.
How
in
a
distinguished
okay,
I
want
these
data
or
this
data.
But
if
I
ask
for
context
user
coops,
it
will
context
as
part
of
the
user
request.
I
can.
If
you
refuse
okay
for
eks
I
want
this
content.
G
That
context
could
be
cluster
name
or
it
could
be
your
something
is
word,
and
that
is
what
I
am
thinking
of
in
on
Miller
nine.
My
thing
is:
I'm,
assuming
when
user
east
using
a
missus
ETL
at
the
same
time
he
already
has
the
access
of
his
poopsicle
environment.
It
could
be
on
TK
e
or
e
KS
or
donÃtö.
He
already
has
the
SS.
G
We
don't
know
he
don't
have
to
set
up
that
particular
accessed
once
more
in,
even
if
you
want
to,
he
can
use
the
standard
process
of
GK
or
ek
and
once
that
setup
is
done
instead
of
pairing.
The
cloud
at
him
query
the
config
file
and
get
there
that
particularly
you
use
a
context.
That
is
my
understanding,
but
I
can
read
more
on
documentation
so
that
I
can
see
for
tempting
make
sense
or
not
III.
A
Configured
their
own
cube,
CTL
in
their
own
environment,
to
work
as
it
should.
You
know,
as
they
would
normally
run
cube
CTL
commands
to
interface
with
GK
great,
so
then,
okay,
then
the
work
that
we
have
to
do
to
connect
measuring
to
you
know
to
who
that
cube,
config
file
in
the
right
way
would
be
is
kind
of
the
second
part
of
what
you're
saying,
which
is
okay,
great
as
I.
Look
at
that
syntax,
you
know
geez,
considering
that
in
gke
I
could
be
running.
A
You
know
ten
different
clusters
and
then
each
one
of
those
clusters
could
have
ten
different
contexts
like
wouldn't
we
want
to
include
that
cluster
name
and
context,
name
potentially
in
the
command
that
we're
creating
for
Mestre
CTL.
That
makes
a
ton
of
sense
the
reason
that,
when
I
think
that
the
answer
is
yes
that
would
that
would
be
good.
It
would
be
answers.
A
Yes,
it
would
be
good
to
have
that
in
part,
because,
while
there
are
ten
clusters
in
that
gke
environment,
maybe
I
either
don't
have
permission,
my
user
and
and
consequently
sort
of
the
measure,
user
or
I
don't
want.
My
only
want
Mary
to
interface
with
one
of
them
and
specifically
one
context,
because
I'm
just
not
ready
for
Mary
to
do
anything
else,
I,
just
I,
don't
want
it
to
touch
any
of
the
other
things.
I
want
to
play
around
with
a
ste
or
something
over
here.
A
So
I
want
to
be
very
explicit
and
just
say
this
cluster
that
context
that
that's
great
I
hope
that
my
hope
is
over
time.
People
are
comfortable
with
measure
e
having
broad
access
to
multiple
clusters
and
multiple
contacts
such
that
when
they
run
that
the
command
is
also
capable
of
just
having
them
point
to
the
provider
or
the
platform
the
permit
may
be.
Providers
may
be
a
better
whatever
provider
platform
that
they
might
run
mesh.
Your
CTL,
you
know,
config,
generate
gke
and
then
yeah.
What
we'll
come
back
is:
okay,
well,
great!
A
Here's
these
ten
clusters,
here's
the
multiple
contexts
in
each
of
them.
Here's
a
big
fat
cube
config
file
if
you
will
and
that
when
the
users
go
into
the
mesh
or
UI
that
this
UI
nikhyl
would
be
enhanced
to
would
be
enhanced
such
that
what
they
would
be
seeing
is
the
ability
to
that
there
would
be
an
environment
construct
and
they
would
say:
yeah,
okay,
I,
great
mysteries,
connected
to
my
eks
stuff
and
my
gke,
so
so
I'll
click
on
gke
and
then
in
there
yeah.
A
Ok,
it
sees
ten
clusters,
but
I
just
you
know
right
now,
I
just
want
to
interface
with
like
the
the
first
cluster
and
then
after
you
click
that
drop-down
or
whatever
it
is
for
the
first
cluster
yeah.
Okay,
I'd
want
it
to
work
with
this
context,
so
I
think
both
both
are
potentially
true
that
hopefully
in
the
future,
while
we
are
while
you're
looking
at
this
sanjay
from
a
mesh
Reese
ETL
perspective,
that's
it's
fantastic!
It's
the
right
first
step,
I!
Think
in
the
future
that
capability
that
you
are
gonna
be.
A
Writing
would
probably
also
be
leveraged
by
the
mesh
mesh
reserver
itself
to
parse
through
the
cube
config
file,
which
it
does
today.
It
parses
through
and
it
looks
for
different
clusters
in
different
contexts,
and
it
presents
that
to
the
user
and
it
would
be
in
here
it
would
be
nice
and
powerful
in
the
UI
as
well
to
be
able
to
identify
what
cluster
in
what
context
I
think
I
think
both
are
good
things
to
support.
A
A
So
I'm
gonna
yeah
and
that's
it
that's
a
great
yeah
yeah
I'm,
going
back
now
to
load
up
my
other,
my
original
cube,
config
file
and
when
I
do
measure
II
the
server
it
it
does,
have
the
ability
to
parse
through
the
cube,
config
file
and
pick
out
different
context,
names
and
so
in
some
respects
sort
of
implicitly
it's
able
to
parse
through
that
cube
config
file,
understand
the
fact
that
there
are
multiple
clusters.
It
doesn't
expose
the
cluster
name
here
just
yet.
Rather
it
focuses
on
the
context
name.
A
A
A
Some
people
are
going
to
want
to
deploy
measure
E
as
an
application
running
inside
the
cluster
as
well,
and
so
we'll
have
some
nice
use
cases,
or
you
know
some
nice
discussions,
I
think
about
about
that.
That
might
actually
be
the
easiest
way
over
time,
because
if
measure
is
just
another
application
inside
the
cluster,
I
think
that
our
challenge
there
will
probably
be
more
around
getting
the
right
like
service
account
created
for
measure
e
as
an
app
so
that
it
can
access
other
contexts
and
things.
A
A
C
A
A
D
Yeah
actually
dark
mocha
was
very
helpful
to
me.
Clear
discern,
provide
a
description
that
you
wanted
to
show,
so
I
will
soon
make
appear
for
that,
a
low
of
value
model
for
the
provider
and
then
I
will
look
into
that
of
showing
the
details
of
the
provider,
whether
below
or
beside
I
will
drive
or
both
of
the
options
and
share
details
which
one
will
look
whiter
than
we
can
move
for
them.
A
A
A
A
C
A
A
A
Nice,
okay,
very
good
whoa,
so
so
that
we
had
daylight
savings
time
changeover
this
week,
which
means
that
our
meeting
time
will
change
a
little
bit.
So
for
those
of
you
who
were
here
this
last
hour,
maybe
we
should
conclude
the
meeting
and
let
you
know
let
you
guys
go
and
then
for
dungeon
and
Nevena
was
just
on
I
think
he
dropped.
A
A
A
H
A
Sorry
about
that
phone
somebody
keeps
calling
me
so
yeah,
so
great
I
mean
obviously
I'm
excited
about.
This
makes
me
gives
me
a
chuckle
I
thing
that
strikes
me
one
of
the
ways
to
kind
of
solve
the
the
challenge
that
you're
facing
right
now.
The
challenge
that
you're
facing
I
think
is
the
fact
that
there's
a
small
needle
there's
a
small
needle
and
then
there's
also
kind
of
the
base
of
the
needle.
If
you
will
that
that
semicircle
yeah.
A
A
H
A
H
A
A
H
A
Now
it's
a
hard
yeah.
This
is
great,
I
mean
it's.
It's
a
big
improvement
and
the
thing
is
is
like
it's
a
it's
pretty
important,
because
every
single
person
that
comes
to
the
project,
whether
they're,
contributing
or
using
or
they're
they're
going
to
go
through
these
instructions.
They're
gonna
touch
this
area
and
actually,
when
you
look
at
it
on
layer,
five
dot,
io
/
Missouri,
it's
also
not
great.
It's
it's!
J
A
Yeah,
it's
so
easy
to
intervene
and
Fong's
yeah
I,
don't
know
if
you
guys
have
ever
seen.
Watch
someone
give
a
PowerPoint
presentation
and
when
they
do
it's
like
they
you
they
tried
to
use
every
animation
that
was
possible
and
and
what
I'm
saying
is
like
it's
easy
to
go
overboard
on
the
animations
and
and
really
distract
people,
but
no
animations
at
all
clean.
That
makes
for
a
boring
presentation.
So
it's
just
like
a
little
bit
of
them.
Some
subtle
use
is
nice.
Okay,.
G
Yes,
so
I've
I
set
up
the
development
environment
last
week.
So
last
week
on
my
system,
it
wasn't
working,
sto,
wasn't
working
I
guess
it
was
taking
a
lot
of
memory
along
with
some
other
problems,
because
it
was
single,
node,
installation
and
I
already
had
installed,
using
mini
cube
without
using
the
configuration
file
and
then
I
was
using
cube
ATM.
G
So
there
were
some
changes,
but
I
set
up
the
mystery
and
memory
and
sto
on
a
virtual
machine,
and
now
it's
working,
fine
I've
sent
the
PR
related
to
the
issue
of
removing
of
supporting
basically
one
point
for
the
versions
greater
than
one
point
four.
So
in
versions
greater
than
one
point,
four
demo
file
is
basically
deleted
in
issue,
so
I've
updated
that
and
like
others
are
there
still
errors
that
are
coming
due
to
lending
or
some
test
errors,
but
we'll
have
to
look
into
that.
I
G
G
A
So
so,
maybe
quickly
or
briefly,
I,
don't
I,
don't
know
if
this,
if
you
guys
use
similar,
get
aliases
but
I've
I
think
these
are
now
listed
inside
the
contributor
doc.
The
contributor
guide
for
the
contributing
markdown
rather
I
end
up
using
one
like
this.
This
is
sort
of
an
abbreviation
for
check
out.
Apr
and
I
ended
up
giving
it
the
PR
number.
A
A
A
A
A
H
D
B
A
Alright,
so
it
so,
the
the
new
version
of
mystery
sto
is
built.
It's
running
locally.
Massery
is
running
locally
I'll
go
to
localhost
90-80
in
my
browser.
It'll
ask
me
to
off:
do
it
quick,
I'm,
not
able
to
ping
kubernetes?
So
the
let
me
go
and
do
this
I
was
in
our
last
meeting
with
the
other
guys
we
were
I
was
running
both
mini
cube
and
docker
desktop.
At
the
same
time,
because
we
were
talking
about
some
work,
that
Sanjay
is
gonna,
do
in
master
CTL,
so
right
now,
I'm
stopping
mini
cube
and
I'm.
A
A
A
A
A
A
Okay,
so
as
I
go
to
test
out
the
functionality.
Now
that
your
functionality
Naveen
go
her
and
the
menus
have
been
updated
based
on
the
work
or
the
issues
around
him
TLS
and
non
MTS
and
stuff
that
you
and
I
had
talked
about
so
so
I'll
try
with
M
TLS
I'll
be
watching
for
events.
While
we're
doing
that
we'll
come
over
here
and.
A
A
G
One
thing
here
which
needs
to
be
mentioned
is
we
selected
is
to
you,
let
us
just
to
you
with
MDL
s,
but
what
we
are
actually
doing
is
just
installing
East,
your
demo
demo,
without
that
the
demo
odds,
because
the
latest
version
already
is
basically
by
default,
support,
I,
guess
MPLS.
So,
but
in
the
previous
versions,
that
was
a
separate
file.
A
So
that
would
then
then
we
can
just
go
out
and
down
that
version
based
on
based
on
their
choice
and
that
at
some
point
to
particularly
maybe,
as
you
enter
into
the
next
phase
around
you
know,
use
of
the
sto
CTL
go
client,
then
then
will
hopefully
be
able
to
it'll
make
it
easier
for
us
to
get
specific
about
with
MPLS
or
without
or
you
know,
allowing
the
users
to
specify
certain
options.
Your
configuration
deployment
options
so
to
me
hey
from
a
functional
perspective,
the
your
PR
is
working.
A
A
A
A
A
There
should
be
some
virtual
services
or
a
virtual
service.
That
was
also
deployed.
Sorry,
it
was
if
we
described
that
book
info
virtual
service.
It's
just
saying:
hey
it's
a
path
based
routing,
that's
going
on!
So
if
I
use
a
request
to
go
to
this
URL
redirect
that
route,
that
traffic
to
this
destination,
to
this
host
to
this
service
route,
that
traffic
to
this
service
called
product
page
and
the
thing
that's
called
product
page-
is
here
this
service
and
what
we
need
to
make
sure
that
we've
got
exposed
externally.
A
H
A
We
go
look
there
is
on
port
in
the
requests
received
will
be
redirected
to
3
1,
380
I'm,
sorry,
oh
yeah
and
oh
great,
so
that
that
should
be
redirecting
to
product
page.
Eventually
that
should
get
to
product
page.
Let's
see,
I
go
to
my
local
host
480
with
slash
product
page
I
do
I
do
get
to
again
frog,
and
if
I
refresh
and
refresh
we
get
the
default
kind
of
round
robin
and
load
balancing.
A
C
A
Amazing,
how
much
we
had
to
talk
just
just
to
get
to
this
I
don't
mean
there's
no
reflection
of
you
at
all.
I
just
mean
it's
so
funny
how
how
how
program
how
software
engineering
works.
You
know,
because
your
questions
were
spot
on
were
things
that
I
hadn't.
You
know
III
hadn't,
sat
down
to
think
about
and
yeah
like
ultimately
understanding
that
sto
as
a
project
is
moving
away
from
helm
and
as
moving
over
to
sto
CTL
as
the
preferred
method
of
deployment
great,
we
would
want
to
follow
suit
and
so
yeah.
A
This
is
kind
of
written
poorly
in
terms
of
the
English
that
was
used
to
describe
this,
but
it's
basically
saying
that
use
of
this
package
like
when
we
when
you
pull
in
that
package,
where
that
all
right,
maybe
it's
sucky,
it
might
be
here
that
you
there
would
be
a
comment
right
above
it.
G
A
A
It
was
like,
oh
well,
no,
but
we
have
a
workshop
that
we're
giving
in
like
next
week,
and
so
we
just
never
yeah
came
around
to
fixing
it,
and
so
the
the
thing
now
with
sto
1.5
a
lot
of
the
slides
and
things
that
we
put
together
that
described
the
control
plane
and
we
know
how
things
have
changed
around
with
is
DoD.
There's
a
lot
of
updates
to
wolf
to
those
trainings.
Also,
there's
updates
to
you
steal
up
and
running
that
book.
A
Well,
Riley
just
came
back
to
me
this
week
on
asking
if
I
would
sign
a
contract
to
update
this
first
book.
This
one
D
and
I
said
that
I
would
I,
don't
have
enough
time
to
do
these
things.
But
the
reason
I
mentioned
this
to
you
guys
is
because
Naveen
and
Fonzie
and
Gunjan,
if
you
guys
there
so
I,
have
lots
of
folks
asking
me
to
talk
at
conferences.
Write
books,
write
blog
posts
to
do
things
like
that.
A
I
can't
do
them
all
and
I
don't
need
to
be
the
one
doing
them
all
the
quite
seriously.
The
a
big
reason
why
it
is
that
you
guys
are
here
and
why
I've
encouraged
you
to
reached
out
to
you
and
asked
you
to
join
in
things
is
because
it's
for
a
whole
bunch
of
reasons.
But
one
of
those
reasons
is
that
maybe
one
of
you
wants
to
work
on
part
of
a
book.
H
In
director,
oh,
we
have
shape.
People
could
either
take
a
check,
be
good,
J,
taste,
the
chicken
doctor
and
or
some
other
additives.
At
the
first
leg
of
30,
it
was
only
three
Jake
Teague
taken
the
people
in
the
chicken
stock
carbonate
and
Bush.
So
do
I
need
to
add
all
the
other
features
in
opting
also
also
dating.
A
Let's
do
it
when
you
do,
it
is
when
you
do
what
we
will
and
fangio
I
am
so
happy
that
you
joined
the
calls
by
the
way.
So
if
I
was
gonna,
suggest
that,
if
you
ever
are,
if
you
think
that
you're
ever
ready
just
to
make
to
begin
to
do
a
few
contributions,
wonderful
that
this
next
thing
that
we're
talking
about
might
be
one
of
those
areas.
A
H
A
What
I'm
saying
is,
as
you
enable
them
if
you
can
get
them
fixed,
while
we're
going.
That
would
be
wonderful
because
it's
I
think
it
flusters
contributors
when
they
go
like
like
Naveen
you're
experiencing
this
right.
Now
that
you,
you
made
us
a
few,
relatively
small
updates
to
the
East
EOC
I'm,
sorry,
the
Mesha
stele
adapter,
and
when
you
submitted
your
PR
as
it
turns
out,
it
came
back
with
a
number
of
red
x's
on
things
that
you
didn't
even
touch
yeah
and
then
all
of
a
sudden
you're
left
wondering
a
couple
things.
A
A
H
A
H
A
Davina,
what
we're
about
to
you're
about
to
make
history
I
think
if
these
small
checks
make
sense,
if
you're
able
to
correct
those
you're
about
to
become
I,
think
a
first-time
contributor
right,
okay,
you're
gonna,
be
on
the
other
side
of
this,
so
yeah
my
net
just
got
disconnected.
Oh
I
was
just
trying
to
poke
fun
at
you
and
saying
him
hey
as
soon
as
those
other
checks
are
done.
You'll
have
you'll
have
unlocked
your
first
milestone
of
being
a
contributor.
A
Take
any
opportunity
I
can
to
to
to
celebrate.
You
know:
grandma
was
over
at
over
at
a
friend's
house
and
a
former
employee
of
mine,
yogi
Paula,
and
he
apparently
had
a
couple
of
bottles
of
tequila
that
he
never
drinks
and
so
I
think
since
I
was
over.
He
thought
it
was,
would
be
a
good
excuse
to
get
rid
of
tequila.
A
H
One
more
thing:
I
want
to
tell
you
that
said
few
weeks
ago
we
were
talking
about
all
contribute
aboard
and
then
we
at
last
we
should
make
a
plan
or
all
contributor
boat
and
at
the
end
we
planned
that
we
should
make
something
for
the
community
page
what
they
are
working
on,
all
those
piers,
so
yeah
in
India.
We
have
one
thing
that
that's
what
gas
gives
somewhat
of
good
and
this
time
I
am
I
may
interrupt
that
one
positive
customer
effort
and
I
want
to
share.
One
thing
so
may
share
my
scheme
sure.
H
A
G
Guess
we
can
do
it
with
github
API
is
because
let
me
just
paste
this
here,
because
when
we
author
any
PR,
we
can
I
already
like
search
for
it
using
SBR
and
author
in
the
github
I
guess
there
should
be
some
endpoint
to
public
API
for
it
if
they
specify
project
or
something
right,
so
we
would
not
require
database.
You
can
just
tell
it
differs
from
the
API
then
right
and
list
out
the
projects
like
avi.
We
can
just
list
out
whether
it
is
in
this
organization
or
not
I
guess
it
must
be
supported.
Yeah.
A
A
To
Fonzie's
point
a
good
point
that,
like
yeah,
hey
that
does
github
api
is
our
our
public
are
open.
Yeah
I,
you
know
their
rate
limited,
which
I,
don't
think
will
be
an
issue
for
us
because
that
you
know
people
aren't
going
to
be
hammering
on
it
so
yeah.
So
so
long
as
we
don't
need
to
have
a
private
token
stored
somewhere
and
guns.
A
If
it
doesn't,
then
it
would
be
using.
You
know
it
have
a
fair
bit
of
JavaScript
in
there
to
be
able
to
make
all
those
REST
API
calls
and
then
sort
of
compile
the
you
know
there.
It's
simple
calculations,
but
how
many
commits
and
so
yeah,
if
they
that's
great,
like
just
like
you
were
sending
to
all
contributors.
A
Yeah,
that's
cool
gan-chan
I,
agree
that,
like
hey,
what
we
have
on
mesh
for
I/o
today
is
nice,
but
there's
not
that
much
to
a
day,
because
the
site
itself
is
pretty
small,
it
even
doesn't
like.
We
actually
really
do
need
a
lot
of
other
content.
If
you
were
things
like
hey,
how
do
you
get
started
with
mastery?
A
Does
it
I
mean
that
the
docks
are
okay
that
doesn't
really
make
it
easy
to
consume?
The
fact
that
you
can
deploy
messy
locally
measuring
inside
kubernetes
that
in
what
versions,
by
the
way
in
the
being
like
what
versions
of
sto
are
supported?
What's
the
difference
like
it,
some
of
the
work
I
think
a
lot
of
times.
We
forget
to
go
back
and
update
the
docks,
but
the
dock
site
is
pretty
easy
to
update
it's
just
a
Jekyll
site:
okay,
the
docks
site.
A
D
A
H
A
G
Lee,
actually
I
was
checking
the
recommendation
and
there
we
are
like
for
every
adapter.
We
are
having
adapter
skaters
there
right
right.
So
how?
How
is
it
like?
How
is
the
status
updated?
Is
it
like
hard-coded,
or
is
it
using
some
API
e
for
checking
status
from
github
like
if
you
are
using
Travis,
then
we
can
fetch
whether
the
all
tests
are
passing
for
the
adapter
of
not
yeah.
It's.
A
Something
that
Ashish
does
a
really
good
job
of
catching
is
the
fact
that
we'll
make
a
release
of
mastery
and
then
we'll
forget
to
go
back
and
update
like
the
release
page
in
the
Miss
redox,
and
that's
something
just
like
the
example
that
you
were
giving
about
programmatically,
assessing
whether
or
not
what
of
the
vco
that
we
support
that
there's
a
ton
of
room
for
improvement
there.
We
absolutely
that's
not
that
hard
to
do
we
as
part
of
the
build
process.
We
should
be
identifying
what
those
version
numbers
are,
and
then
we
can
reference.
A
A
F
A
That's
you
know
fairly
calm,
that's
not
Universal,
but
it's
somewhat
common
for
the
other
adapters.
By
the
way,
I
think
I
told
some
of
you
that
we
just
Lera
v
just
expanded
and
formalized
its
partnership
with
hashey
Corp,
and
because
of
that
they
should.
We
should
be
seeing
a
console
engineer
pop
up
in
our
meetings
occasionally
and
be
helping
with
that
exact
thing.
That
I
just
said
that
console
has
a
couple
of
edge
and
I'm
sorry
an
edge
and
a
stable
if
they
should
be
helping
with
that.
A
More
than
that,
though,
measure
E
as
a
project,
we
should
think
about
it
that
the
don't
today
have
much
control
over
and
users.
Consequently,
don't
have
much
control
or
easy
control
over
what
version
of
measure
e.
It
is
that
they're
running
today,
when
we,
when
naveen's
PR,
is
merged.
Ten
minutes
later.
Fifteen
minutes
later,
if
a
user
is
sitting
there
running
measure
e
measure
e
will
detect
that
there's
a
new
version
of
the
docker
image
for
sto
from
measure
e
is
do
for
that
adapter
available.
A
It
will
download
that
new
image
and
reload
the
users
running
instance
with
it,
which
other
one
hand,
is
awesome
because
it's
like
hey
it's
an
auto
updating
software
feature.
That's
great.
On
the
other
hand,
that's
not
what
everyone
is
comfortable
with
or
wants
to,
or
as
good
for
everyone,
because
sometimes
they
are
breaking
changes,
and
so
we
don't.
We
need
to
improve
our
release
process.
A
The
only
thing
that
really
really
gets
versioned
and
users
are
in
control
of
their
destiny,
so
speak
is
with
measure
e
CTL,
because
a
new
Mestre
CTL
binary
is
not
put
into
the
releases
folder
on
github.com
unless
we
create
a
new
release.
There's
a
whole
document
inside
of
the
Google
team
drive
that
talks
about
the
measure
e
released,
build
and
release
strategy
and
a
lot
of
what
I'm
saying
is
outlined
in
there.
But
a
lot
of
what
I'm
also
saying
is
not
done
so.
A
So
I
brought
this
up
gungeon
as
a
or
Ashish,
rather
as
just
an
exam
SVG.
I
can't
remember
if
I
gave
this
to
one
of
you
guys
or
not,
but
this
this
is
kind
of
a
kind
of
a
neat
one
and
on
layer,
five
dot,
IO
you'll
start
to
see.
You'll
see
this
little
guy
popping
up,
and
so,
if
I
didn't
give
you
this
one
sheesh,
this
is
kind
of
a
fun
one
to
play
with.