►
From YouTube: Meshery Development Meeting (June 6th, 2020)
Description
Welcome @MikeBorovik 🥳 and @tanuj_agarwal22!! 🎈
A
Surprisingly,
had
a
few
people
do
but
yeah
I
was
having
a
crisis
two
minutes
ago
with
fresh
brewed,
coffee
I
might
I
might
fall
over
if
I,
don't
I'm
gonna
run,
go
grab
that,
but
I'm
still
see
you
on
the
on
the
call.
So
I'll
be
right.
A
Now,
just
just
just
as
we
get
we
get
started
with
the
meeting.
I
just
want
to
be
honest
and
forthright
with
everyone.
This
like
at
first
blush.
This
looks
like
a
Starbucks
cup
but
turns
out
it's
just
some
generic
plastic
piece
of
piece
of
crap
put
it,
but
it's
coffee,
so
so
good!
Well
enough
of
enough
I'm
messing
around
and
it's
yeah
boys,
it's
Wednesday,
the
17th
thanks
everybody
for
coming
just
to
make
well
Michaela
people,
usually
them
McHale,
or
do
you
go
by
Mike
or.
B
A
Oh
Mike,
if
you
don't
mind,
they
introduce
yourself
to
the
other
folks
on
the
call,
if
you
would
just
do.
B
And
some
you
can't
mess
gears
and
one
so
on,
but
my
project
recent
year
since
yeah
in
my
project,
I
used
to
create
with
search
mesh
which
is
based
on
Easter,
incriminates
and
this.
Why
I'm
interested
in
solutions
like
yes
and
I
used
to
manage
it
currently
and
probably
missed
next
month,
I'll
be
transferred
to
infrastructure
comment
out
of
coding.
B
A
A
C
Hi
hi
and
I'm
in
the
in
the
boiled
car
running
from
the
work
to
refer
my
ish
culture,
because
I
need
to
go
to
see
an
apartment
and,
in
the
meantime,
I
want
to
participate
on
this
because
I
like
to
actually
start.
As
a
summary,
can
you
share
with
with
someone
who
I
mean
you
I
would
love
to
be
10
percent
of
your
guys
that
are
already
contributing
I'm,
just
so
lazy
on
setting
up
what
I
need
to
set
my
signature
blah
blah
projects.
You
know
yeah.
A
C
I
would
like
to
start
with
those
two
issues:
the
one
that
looks
simple
for
learning,
mastery
and
and
the
one
that
is
new,
which
seems
challenging,
challenging
motivating
whatever
we're
driving
for
me
so
yeah
just
to
you
know
to
to
have
it
to
have
a
point
of
start,
I'm
bad
on
being
teached
by
others,
but
I
mean
making
improvements,
yeah
I
that
that's
what
they
wanted
to
say,
but
I
could
have
also
say
it
in
private.
That
I
took
advantage
of
the
moment.
No.
A
I
did
actually
yeah
yeah.
You
know
it's
funny,
like
almost
all
the
things
that
you
do
are
things
that
I
try
to
encourage
others
to
do,
which
is
which
is
I'm
very
glad
that
you
mentioned
just
that
about
like
really
really
getting
your
first
foothold
on
an
interesting
portion
of
the
project.
Is
it's
something
that's
of
critical
concern
for
us,
because
we
have
a
lot
of
folks
come
through
and
we're
very
pleased
when
they
do
and
a
lot
of
times.
A
You
know
we
really
try
to
make
sure
that
everyone
feels
comfortable,
but
also
that
people
find
an
area
of
interest.
It's
really
easy
for
them
to
find
an
area
of
interest,
because
there's
so
many
technologies
that
we
cover.
But
if,
if
a
little
bit
of
extra
time
isn't
given
to
them
upfront
a
lot
of
times,
they
they
don't
find
that
foothold,
and
so
there's
actually
a
program
I'm
glad
that
you
mentioned.
You
mentioned
this
Edina,
because
it's
a
good
thing
to
talk
about
well
in
context
of
Mike
who's
joined,
but
also
to
news.
A
This
might
be
your
first
time
on
the
call
as
well
so
good
for
you
to
hear
and
I
think
to
news.
You
already
heard
about
this
mesh
mates
program,
which
is,
which
is
to
say
that
there
are
a
few
individuals
in
the
community
who've
been
here
long
enough,
that
they're
they're
able
to
and
willing
to
pair
up
one
on
one
with
a
newcomer
to
help
them
get
familiar
with
the
projects,
walk
them
through
the
different
repositories.
A
A
A
D
D
A
Excited
about
that,
okay,
okay,
fair
enough
in
into
you
into
this
meeting,
to
news
you
and
please
don't
take
offense,
just
there's
just
some
background
noise,
but
but
anyway,
great
alright
good.
So,
let's,
let's
dive
in
today,
we've
got
a
few
topics
to
cover.
Well,
the
first
one
here
is
is
actually
more
process.
We
I
was
just
chatting
with
Mike
about
the
various
meetings
that
we
have
and
the
notion
that
this
one
has
run
a
little
bit
like
a
stand-up
in
the
past.
A
That
has
been
like
that
and
we're
hoping
to
get
a
little
bit
back
to
that
such
that
such
that,
if
you're,
you
know
working
on
something.
This
is
your
opportunity
to
kind
of
show
it
off
get
some
feedback
if
you're
stuck
somewhere
to
ask
about
that
or
if
we've
got
new
and
upcoming
work
to
something
like
the
measure
II
operator,
to
kind
of
cover
that
and
discuss
it
at
large,
so
that
folks
can
so
that
everyone
can
participate
so
on
that
only
I
think
so.
A
Apologies
Shubham,
but
but
because
mike
is
gonna,
all
he's
only
got
30
minutes.
I
would
have
I'm,
gonna,
pre
pen,
I,
wouldn't
put
pushes
work
and
the
demonstration
of
it
up
front
because
it
would
be
wonderful
to
get
mike
said
well,
everyone's
opinion,
but
but
might
might
be
of
interest
to
Mike.
Well
with
that
cush
you
want
to
you,
want
to
take
over
and
tell
people
about
what
you've
worked
on.
E
E
We
have
to
expose
the
law
generator
settings
in
the
UI
so
that
we
can
specify
the
performant
s
their
faults
and
when
user
will
have,
when
user
will
submit
this
defaults.
So
in
this
performance
ey,
the
defaults
will
no
load
from
the
settings.
The
concrete
requests
queries
per
second
duration
in
load
generator.
This
four
are
the
attributes
which
you
will
be
able
to
set
from
this,
and
this
is
the
UI
which
which
will
give
you
the
ability
to
set
the
defaults.
E
So
this
was
the
PR.
This
PR
involves
like
there
were
some
missing
handlers
for
the
gate
request.
There
were
some
errors
and
the
handlers
which
of
which
was
already
there
and
the
complete
new
UI
for
the
set
for
the
settings
performance
default.
So
I
will
just
quickly
show
you
a
demo
for
this
right
now.
A
E
So
I'll
just
give
you
a
demo
from
the
from
the
server.
So
here
is
the
performance
test.
Ui,
and
here
we
have
these
following
fields:
concurrent
requests,
queries
per
second
duration,
enlarged
and
load
generator.
So
what
you
need
to
do
every
time
is
whenever
you
want
to
run
a
performance
test,
you
have
to
specify
it
as
same
the
service
mesh
type.
You
are
a
tourist
and
these
following
four
fields.
E
A
A
A
E
So
I
just
show
you
a
demo
right
now,
so,
as
you
can
see
right
now,
this
settings
which
are
persisted
over
here
are
six
concurrent
requests.
Six
queries
per
second
and
duration
is
six
minutes
in
the
load
and
it.
So
these
are
the
defaults
which
I
had
already
set
it
by
testing
so
I'll,
just
change
them
for
concurrent
requests.
E
A
Well,
there
are
new
they're
new
fields,
great
for
us
to
let
someone
configure
those
might
make
sense
to
be
able
to
pull
into
that
preferences.
It's
worth
a
discussion
as
to
whether
or
not
those
do
make
sense.
I
if
you
think
that
they
my
perspective,
is
that
they
generally
do.
A
E
A
A
A
Well,
Chris:
do
you
want
to
just
kind
of
foreshadow
to.
C
Feedback
I'm
not
sure
I,
understood
that
it
is
about
it's
about.
Like
you
run
or
not
performance
test
through
an
app,
then
you
have
the
request,
header
and
working
and
I.
Don't
know
the
number
of
stairs
that
you
want
to
execute,
or
this
is
I.
Don't
know
where
to
put
this.
It
is
for
app,
it
is
for
our
app
I,
don't
know
right.
C
E
So
like,
whenever
you
run
a
performance
test,
you
have
to
feel
feel
all
of
this
fields,
so
there
can
be
situation
like
I
want
to
make
this
concurrent
request
always
set
to
five
queries
per
second
always
said,
and
the
duration
of
my
load
test
always
run
out
and
I
always
want
to
use
the
40
load
generator.
So
for
me,
it
might
be
a
little
bit
irritating
to
always
set
up
this
defaults
or
I
mean
there's
a
feature
where
I
can
specify
this
default,
so
that
but
I
will
not
every
time
I
run
a
performance.
E
A
C
G
C
A
A
A
Memory
is
able
to
generate
load
against
any
end
point
whether
that
end
point
is
a
workload.
That's
an
application,
that's
running
on
top
of
a
mesh
or
whether
it's
an
application
or
a
service,
that's
running
off
of
the
mesh,
and
actually
that
makes
for
a
very
good
comparison
for
people
very
high
level.
Like
hey
here,
I'm
running
my
services
on
communities,
you
know
plainly,
let
me
generate
a
load
test
and
and
see
that
mesh
reproduces
a
bunch
of
statistical
analysis
and
then
now
let
me
deploy
my
services
onto
a
mesh
and
I'll.
A
C
C
A
Yep
totally
yeah,
while
Murray
is
born
of
service
mesh
land.
This
particular
piece
of
capability
isn't
constrained
to
certain
like
yeah.
This
might
just
be
a
better
version
of
jmeter
for
people
or
or
whatever
you
know,
instead
of
just
using
w
RK
by
itself
or
Apache
bench
by
itself
that
they
might
want
to
use
this
tool
because
it
helps
facilitate
interaction
with
kubernetes.
A
You
know
setting
up
your
tests
running
them,
persisting
the
results,
doing
statistical
analysis
on
the
results
which
are
you
know,
I,
don't
know
if
it's
makes
sense
for
us
to
kind
of
demo,
some
of
that
now,
but
but
yeah
it
may
be.
Just
for
the
you
know,
Tunisian
and
Mike
and
and
and
You
Adena-
that
it
might
make
sense
to
show
what
it
looks
like.
A
A
A
By
sending
back
my
results,
my
results
set
synonymously
or
not.
Those
are
my
user
preferences.
My
measure
II
settings
for
how
my
Cherie
works
are
here
and
these
some
of
these
essentially
all
of
the
settings
that
we
see
inside
here,
the
the
settings
URL.
These
are
going
to
end
up
being
global
settings
for
a
given
measure.
Ii
deployment
measure
II
is
a
multi-user
tool
right,
many
users
can
authenticate
to
it,
use
it
run
different
tests
at
different
times
run.
A
You
know,
orchestrate
different
or
manage
different
service
meshes
and
as
users
sign
in
and
use
the
same
measure.
Ii
instance.
That's
that's
fine!
That's
great!
Now
the
settings
that
are
that
are
here,
though,
if
one
user
changes
the
settings
across
these
four
tabs,
that's
going
to
impact
the
experience
that
any
the
same
measure
II
instance
has.
A
So
these
are
global
settings
and
right
now
we
don't
clarify
this
very
well
in
our
code
and
in
the
UI,
but
but
we
will
so
these
are
global
settings.
These
are
individual
user
preferences.
There
aren't
many
of
them
today,
just
just
two,
so
these
are
per
user.
This
is
for
all
users
of
them.
This
is
per
measure,
II
instance
for
all
users.
Now,
at
some
point,
we're
gonna
find
ourselves
in
a
situation
where
we're
going
to
be
able
to
save
these
global
defaults.
A
We're
gonna
have
we're
gonna,
be
able
to
let
people
set
their
performance
test
settings,
which
is
what
cush
is
demonstrating.
We're
gonna
be
able
to
maybe
let
people
globally
turn
on
or
off
the
anonymous
sharing
of
the
sharing
of
anonymous
performance
results
and
many
other
many
of
the
other
settings,
but
on
a
per
user
basis,
individual
users
may
have
their
own
preferences.
That
might
make
a
whole
heck
of
a
lot
of
sense
to
override
the
global
defaults,
and
so
this
right
here
will
be
a
topic
of
conversation
and
discussion
in
some
of
our
forthcoming.
G
A
Totally
then
so
this
Shubham,
you,
like
you,
set
us
up.
You
keyed
up
a
conversation
perfectly
for
something
that
that
kanishka
and
cush
and
some
others
have
are
looking
at
and
and
have
kind
of
designed
a
little
bit,
and
that
is
a
related
concept
between
the
thing
that
we
were
talking
about
with
global
global
settings,
great
user
preferences
great,
but
now
there's
something
else,
that's
kind
of
missing
and
it's
exactly
what
you're
calling
out
Shubham,
and
that
is
that
oh
wait,
a
second!
A
So
so,
let's
go
through
a
scenario:
I'm
running
a
mesh
I
have
a
I,
have
a
nap.
It's
the
soap
pest
for
fake
Starbucks,
coffee
that
I
have
here
on
my
desk
and
I
run
that
on
a
particular
match
and
blah
blah
blah
that
I
always
you
know,
I
always
run
this
one
with
these
settings,
and
so
but
I
also
have
another
test
that
I
always
run
it's
a
capacity
test
for
such-and-such
and
I've
got
like
you
know:
15
pests
right
and
at
some
point
well,
hey
just
having
a
global
setting.
That's
not
enough!
That!
A
Isn't
a
horribly
helpful!
So
the
design
there's
a
feature
enhancement
that
there's
really
you
referred
to
as
a
performance
test
profile
or
a
performance
test
configuration
like
basically
the
ability
to
name
this
configuration
that
I've
typed
in
save
it
and
be
able
to
recall
it
so
that
I
can
have
a
bunch
of
different
types
of
tests
that
I've
defined
be
able
to
recall.
A
You
know,
pull
those
back
up
very
easily
and
just
execute
on
an
ad
hoc
basis,
any
one
of
those
profiles
next
step-
and
this
is
in
part,
why
we've
been
having
discussions
with
litmus
chaos
and
mesh
chaos.
Mesh
is
the
next
step
here
is
to
facilitate
scheduling
of
those
performance
tests.
There's
use
cases
around
scheduling
a
recurring
performance
test
to
be
able
to
manage
your
performance.
Ongoing,
there's
use
cases
that
we
have
today
for
just
ad
hoc
invocation
of
a
test
right.
A
There's
also
use
cases
for
programmatically,
invoking
a
test
which
is
which
is
why
measuring,
which
is
why
measure
eserver
has
a
REST
API
and
mastery
CTL.
The
command-line
interface
that
you're
able
to
invoke
performance
tests
from
mashra
CTL
that
REST
API
and
that
CLI
massery
CTO,
is
what
we
expect
many
teams.
Many
organizations
will
use
as
they
build
measure
e
into
their
C
ICD
pipelines.
A
B
Sees
me
I
fav
one
want
me
to
see
another
performance
top
there's
some
Direction
field.
It
is
only
in
Twitter,
but
as
for
me,
it's
not
actually
useful.
I
would
prefer
here
to
have
something
like
drop-down
list
with
predefined
playlist
like
1
second
2
minutes
and
1
hour
and
so
on.
It
will
be
I.
Think
it'll
be
more
useful.
A
A
The
user
to
sort
of
look
at
the
tooltip
and
take
a
minute
to
figure
out
okay,
well,
it
looks
like
I
could
do
like
a
1h
I
guess
that
means
one
hour,
but
so,
and
so,
instead
of
having
a
drop
down
a
pick
list
like
okay,
it's
very
clear:
you
know
that
these
means
this
now
granted.
The
pick
list
could
not
be.
It
could
be
super
long
and
maybe
not
granular
enough.
No
just
some
come
on
with
us
and
leave
it
possible
to
totally,
and
so
here's,
maybe
the
best
of
both
of
those
worlds.
A
I
was
gonna.
Ask
I
wonder
if
this
makes
sense
to
you
that
if
this
is
the
duration
drop
down
of
the
duration
and
it
has
a
drop
down
that
says
you
know
one
hour
or
what
one
minute
five
minutes.
Thirty
minute,
you
know
it
has
some
common.
You
know
choices,
but
that
it
also
lets
that
still
has
the
ability
for
someone
to
type
in
of
course,
okay,
okay,
I!
Think
that's
wonderful
feedback.
Does
anybody
I'll
ask
it
like
this?
Does
anyone
disagree
with
that
suggestion.
A
F
G
E
F
Earlier
also,
you
have
commented
that
we
don't
need
to
remove
only
slash
graph.
We
need
to
remove
anything
after
ipn
put
so
yeah
I
think
this
is
already
done,
and
this
also
has
been
updated
in
the
docs.
Also
like
I
created
the
pr
in
the
doctor
and
also
updated
that
so
I,
don't
know
how
you
is
this
error
like
right,
because
if
you
follow
the
talk
you
have
to
type
only
I
can
put
not
the
slash
graph
in
the
end.
A
I
I
guess
I'll
ask
this:
I
can
I
can
verify
it
right
now
on
the
fly.
If
anyone
is
interested
in
sort
of
seeing
how
mesh
redeploys,
sto
and
mesh
we
can
connect
to
Cora
fauna
and
Prometheus,
and
we
can
kind
of
verify
the
behavior.
Your
all
right,
yeah
good
to
me,
we'll
knock
out
a
couple
birds
with
one
stone
here.
I
need
to
come
up
with
another
metaphor.
A
A
Let
me
actually
just
let
me
the
the
text
text
is
never
big
enough
like,
and
so
let
me
just
make
it
fullscreen
and
make
it
obnoxiously
huge.
So
here
I
am
man
line
jcpl
system
status,
I
am
running.
My
sri
is
currently
running
on
my
system.
It's
running
and
docker
desktop
in
the
set
of
eight
through,
however
many
cubes
seven
containers,
and
so.
A
A
My
browser's
keeps
used
as
to
who
I'm
signed
in
as
and
Twitter.
It
would
be
nice
if
I'm
signed
in
I
guess
technically
I
don't
need
to
be
signed
in
to
go
walk
through
this
demo.
So
the
all
right.
So
the
feature
that
or
the
bug
/features
/font
allottee,
that
Shubham
has
is
identifying
account
wanting
to
confirm,
is
complete
and
can
be
closed
out
is
one
around
mesh
reconnecting
to
her
fauna
and
mesh
reconnecting
to
prometheus.
So
in
my
environment,
I
am.
A
Running
it
looks
like
I'm
running
its
deal
already.
This
is
issue.
Probably
the
running
is
T
already,
which
should
be
fine
so
in
and
to
deploy,
is
to
you
I
just
go
over
to
his.
Do.
Click
on
the
plus
sign
click
on
deploy
is
deal
with
the
latest.
You
know
with
this
configuration
so
so
I've
got
sto
running
I,
don't
have
a
sample
app
I'll
do
book
info,
just
because
it's
quite
familiar
to
me
so
we'll
come
over
and
I'll
make
func
a
little
bit
smaller.
A
Let's
do
another
look
at
what
all
pod
they're
running,
so
look
like
cocaine
pose
up
as
a
14
seconds
ago.
I!
Don't
know
why
I
deployed
book
info.
It's
not
actually
that
helpful
for
this
demo,
but
rather
when,
depending
upon
when
we
when
mesh
redeploys
is
do
it
deploys
a
Steelers
control.
Plane
and
part
of
that
is
the
fact
that
a
small
instance
of
grow
fauna.
A
Prometheus
are
deployed,
so
here's
Griffin
and
Prometheus
is
also
deployed
in
the
is
do
control
plane.
A
You-
and
these
are
the
two
now
those
are
running
as
services.
Those
services
are
not
ok
and
and
they're
running
as
services
inside
the
sto
control
plane
inside
kubernetes,
now
I'm
running
measure
e
on
a
different
network,
I'm,
not
running
measure
e
inside
kubernetes,
I'm
running
measure
e
on
docker
on
a
virtual
machine
outside
of
that
network.
So
one
of
the
steps
that
I
need
to
do
is
to
expose
Prometheus
and
expose
core
fauna,
so
that
measure
e
from
the
external
Network
can
make
some
API
calls.
A
So
in
order
to
expose
Prometheus
and
graph
ona.
What
I'll
do
is
that
those
service
is
inside
the
SEO
system,
namespace
I'm,
going
to
edit
the
service
Prometheus
and
change
its
network
configuration
from
a
type
of
the
type
cluster
IP
to
type
node
port
is
that's.
What's
gonna
work
well
for
my
local
deployment
and
we'll
just
go
ahead
and
very
briefly,
do
the
same
thing
for
Agri
fauna,
a
graph
on
a
service.
A
Services
so
for
let's,
let's
do
Prometheus
first,
actually,
as
it
was
the
example
in
the
issue,
so
the
external
port
now
is
3
to
5
of
g2
and
I
should
be
able
to
on
my
local
system,
bring
up
my
local
systems.
Ip
address
and
I
know
you.
Everyone
probably
can't
see
that,
but
just
I
used
the
same
port
number
to
get
to
that
instance
of
Prometheus's.
A
So
I'm
on
that
instance
of
Prometheus,
the
URL
here,
I'm
just
gonna
go
ahead
and
grab
that
URL,
which
which
you'll
see
what
it
looks
so
so
right
now
in
my
memory
and
since
I'm,
not
you
can
see
it
like
mastery
is
yet
to
connect
to
Prometheus.
It
doesn't
know
we're
coming
Thea
says:
well,
let's
go
ahead
and
configure
that
I
come
over
to
this
screen
and
actually
there's
an
enhancement
to
be
done
here
when
I
click
on
configure
Prometheus.
A
Ideally
it
I'm
landing
here
on
this
prometheus
sub
tab,
so
I
use
I
think
you
had
enhanced
measure
II
not
too
long
ago
to
provide
deep
links
into
each
of
these
tabs.
There's
a
small
enhancement
if
we
could
from
these
two
buttons.
When
someone
clicks
on
configure
Prometheus,
ideally
on
the
sub
tab
level
on
the
sub
navigation
level,
that
they
would
go
to
the
Prometheus
time,
so
I'll
paste
in
this
URL
to
connect
to
Prometheus
and
we
reach
it
it's
configured,
that's
that's
ideal,
behavior!
A
A
A
A
So
if
we
go
back
and
look
at
this
same
thing:
okay,
yep,
nice,
Berger,
fauna.
Let's
do
it!
What
we'll
go
back
to
my
local
IP
address?
We
will
use
the
graph
on
a
port
I'll,
just
grab
I'll
just
grab
the
URL
from
here.
Like
you
know,
core
fauna
is
up
so
I'll
go
back
and
we're
gonna
we're
now
going
to
configure
you're
just
kind
of
taking
you
guys
through
here.
We're
gonna
go
ahead
and
configure
connectivity
to
crow.
Fano
are
Prometheus
connectivity.
A
Is
there
and
that's
good
and
here's
the
URL
that
that
was
given
to
me
from
the
dashboard.
There
is
no
API
key
by
default
in
the
configuration
of
Griffin,
oh
that
sto
deploys
and
so
good.
We
get
connected
and
there's
actually
a
couple
of
issues
on
this
page.
So
one
is
really
just
the
same
thing
should
little
is
just
we
want
to
trim
off
any
of
the
rest
of
this,
but
the
connectivity
works.
The
functionality
works.
That's
good
is
this
is
just
a
minor,
you
know
cosmetic
enhancement.
A
A
Will
just
so
that
I
don't
bore
the
crap
out
of
everybody.
I
will
document
the
behavior
separately,
but
yeah
to
finish
off
that
that
other
bug
is
that
unless
you
refresh
the
screen,
the
import
of
the
graph
on
boards
on
don't
doesn't
happen,
but
once
you've
refreshed
great,
you
can
actually
go
through
and
get
really
specific
about
what
boards
you
want
to
pull
in.
A
Look
at
these
statistics.
Are
these
metrics
in
context
of
the
performance
test
and
see
the
effect
on
their
workloads,
see
the
effect
on
the
service
mesh.
The
overhead
that
it
has
see
the
effect
on
their
nodes,
and
so
it's
actually
a
pretty
nice
set
of
features
that
rarely
gets
demonstrated
in
part,
because
we've
just
got
a
little
bit
of
user
experience
to
improve
about
automatically
connecting
to
Prometheus
and
core
fauna.
The
thing
is:
is
we
already.
A
We
are,
we
measure
you
already
automatically
discovers
if
there's
a
Prometheus
or
AG
or
fauna
available
a
running
rather
inside,
of
the
kubernetes
that
is
connected
to,
like
so
measure
ease
like
this
close
to
just
automatically
connecting
to
kubernetes
on
this.
It
already
discovers
a
thesis
and
graph
on
it.
We
just
need
to
like
connect
to
them
or,
at
our
present
to
the
users
say:
hey
we
found
these.
You
know
these
instances
did
you
want
to
connect
and
what
a
massively
different
experience
people
will
begin
to
have
with
measure
e
based
on
functionality.
A
I
think
yes,
yes,
yes,
all
of
the
above,
like
yeah
feedback
on
the
fact
that
there's
a
bug
that
you
have
to
refresh
the
UI
to
get
to
feedback
on
my
ask
if
I
used
to
help
us
do
some
deep
linking
feedback
on
the
notion
that
when
we
deploy
Naveen
when
we
deploy
is,
do
we
are
I,
don't
know
if
maybe
in
still
on,
but
but
when
we
deploy
SEO
we're
not
automatically.
Oh,
he
is
there
we're
not
automatically.
A
A
But
if
we
and
that's
just
a
simple,
you
know
change
in
the
helm
chart
or
you
know,
change
in
our
sto
deployment.
If
we
were
to
do
that,
I
think
Vijay
has
already
worked
on
Naveen
I
think
you
worked
with
Vijay
on
the
work
to
do
mesh
sync
for
discovering
Prometheus
and
graph
fauna,
the
so
they're
right
there
available
to
the
UI.
A
We
just
need
in
the
UI
to
pick
up
on
the
fact
that
there
is
a
Prometheus
or
Griffin
available
and
just
Auto
connected
them
just
make
some
the
some
assumptions
that
people
would
like
to
connect
and
if
they
don't,
they
can
remove
those.
But
what
a
better
experience
it
would
be.
When
people
first
load
up
measure
e
measures
should
be
auto,
auto
connecting
to
kubernetes
auto
connecting
to
Griffin
on
Chrome.
A
You
should
be
discovering
all
that
automatically
connecting
and
people
should
just
drop
their
jaw
on
the
first
time
that
they
run
measure
ease
that
they
download
measure.
Ctl,
it
should
download
automatically
start
massery
automatically
pull
up
their
web
browser.
They
click
to
sign
in
automatically
discovers
kubernetes
automatically
discovers
Griffin
Prometheus
connects
to
them
shoot.
It
could
very
easily
show
them
the
fact
that
they
have
it
pre-existing
Griffin
on
boards
and
stuff
like
Wow.
Wouldn't
that
be
a
delightful
experience.
A
A
A
All
right
on
the
so
a
great
question
serve
I
to
let
me
read
mo
that
by
I'm,
just
gonna
remove
our
remove
our
prometheus
connection,
so
in
our
environment,
we're
connected
to
these
things.
We
are
not
connected
to
prometheus,
oh,
but
one
is
available,
so
I'd
like
to
connect
so
I'll
go
ahead
and
click
to
configure
Prometheus
when
I
do
I'm
taken
to
the
settings
screen
good,
but
the
issue
and
I'm
taken
to
the
metrics
area.
A
H
So
what
do
you
mean?
Is
that
one
Savior
in
the
environment
that
that
is
where
we
are
wanting
to
configure
Prometheus
so
right
now?
What
it
is
going
is
that
it
is
going
one
level
deep
that
is
going
to
the
matrix
part,
but
we
want
it
to
be
a
two-stage
process
so
that
it
opens
matrix,
/
dramatists,
something
like
that
right
on,
and
since
this
is
like,
we
are
using
it
on
the
same
stream.
We
cannot
even
cleans
the
API
like
the.
What
we
call
I
mean
the
URL
right
I.
H
What
I
meant
was
normally
what
we
do
is,
for
example,
say
within
if
we
have
say
in
normal
circumstances,
you
know
when
we
have
two
different
tabs.
If
you
want
to
go
to
a
specific
tab,
we
you
mentioned
see
there
out,
for
that.
So
I
was
mentioning
what
the
router
and
I'm
sorry
to
rephrase
myself,
but
since
we're
not
using
that
kind
of
a
system
over
here,
yeah
I
mean
I,
guess
it's
we
done
with
the
deep
linking
part.
H
A
H
A
A
So
yeah
one
other
one
other
note
here
is
that
so
I've
you
know
I've
zoomed
in
right
to
to
make
sure
that
everyone
can
see
what's
going
on
and
if
you'll
notice
there
are
there's
something:
that's
not
zooming
well
or
a
tokens.
Yeah
I'll
screenshot
that
but.
A
C
Yeah,
it
was
what
this
this
type
of
trials.
We
don't
really
need
to
care
if
it's
going
to
succeed
or
not
taking
advantage
of
the
moment
that
there
are
more
than
one
or
there's
more
than
one
person
more
than
two
persons,
you
kind
of
get
the
idea.
You
know
trying
right
yeah,
that's
a
good
thing
for
having
the
the
meeting
besides
them
on
being
perfect
part
of
the
development
side.
Oh
yeah,.