►
From YouTube: Layer5 Community Meeting (Oct 15th, 2021)
Description
Layer5 Community Meeting - October 15th, 2021
Join the community at https://layer5.io/community
Find Layer5 on:
GitHub: https://github.com/layer5io
Twitter: https://twitter.com/layer5
LinkedIn: https://www.linkedin.com/company/layer5
Docker Hub: https://hub.docker.com/u/layer5/
A
Welcome
everyone
to
the
larify
community
meeting
today
is
15th
of
october.
Today
also
happens
to
be
the
last
day
of
kipcon.
I
hope
everyone
had
a
chance
to
check
out
kubecon
and
it
was
fun.
A
We
have
a
couple
of
newcomers
on
this
call.
So
like
we
do
every
time
we
have
a
tradition
where
we
harass
newcomers
as
they
join
in
the
call
for
the
first
time
so.
A
Daniel,
I
think,
is
this
your
first
time
on
this
particular
call.
Would
you
like
to
introduce
yourself.
B
C
How's,
what
green?
Well
it's
a
cucum
weeks
are
generally
pretty
packed.
I
attended
virtually
this
go
around,
which
makes
it
a
lot
easier
to
be
less
busy.
Usually
it's
like
having
it's
having
breakfast
with
people
lunch
with
people,
dinners
with
people
we
still
give.
We
gave
about
four
talks
this
time.
I
think
we've
got
one
later
today.
C
It's
a
pretty
easy
talk,
but
anyway,
kubecons
are
like
generally
really
packed.
I
think,
by
the
way,
just
as
we
talk
about
kubecon,
the
stats
before
the
conference
I
think
were
about
3
000
attending
in
person
was
the
estimation
and
then
about.
I
don't
know
if
it
was
15
000
registered
virtually,
and
so
you
can
see
the
divide
there.
Most
of
the
folks
I'm
a
little
bit.
It's
actually
would
have
been
a
nice
kubecon
to
go
to
in
person,
at
least
for
me.
C
Having
spent
a
lot
of
time
around
kubecons,
it
would
have
been
nice
to
have
it
less
busy
and
just
catch
up
with
other
vendors
other
other
people,
who've
been
there
for
a
while,
but
that's
you
know
that's
like
for
the
most
part,
that's
a
lot
of
the
folks
that
it's
like.
C
I
would
venture
a
guess
that,
like
the
the
overwhelming
majority
of
people
that
are,
there
are
vendors,
are
people
who
or
companies
that
have
something
to
sell
or
work
in
and
around
the
tech,
which
is
great,
like
there's
a
lot
of
very
interesting
things
going
on
so
anyway,
as
we
talk
about
kubecon
for
a
moment,
it's
a
good
question
for
all
of
you
who
might
have
attended
virtually
think
about
this
question.
C
C
There's
a
couple
of
reports
going
on
right
now
about
who
all
is
involved
and
I'll
go
I'll,
go,
find
it
and
share
it
with
you
guys.
But
did
anybody
see
any
new
like
a
new
project?
They
weren't
aware
of
something
that
caught
their
eye.
C
So
we
had
just
approved
psyllium
in
for
incubation
as
a
new
project,
psyllium's
focused
on
evpf
as
a
technology
very
related
to
the
tech
that
we
focus
on
here
and
we'll
become
a
focus
of
ours
as
soon
as
we
accomplish
a
few
things
together,
there
are
probably
some
that
I
would
say
are
misinformed
that,
like
ebpf
and
service
mesh
don't
go
together,
they
do.
C
There
was
one
of
the
vendors
in
the
space
had
talked
about
presented
it
like
a
little
slightly
inaccurately
but
anyway,
they're
highly
complementary.
Ebpf
is
a
way
of
speeding
up
the
same
interactions
that
service
mesh
needs
to
do
already.
C
It
just
depends
on
how
you're
implementing
the
network,
but
that
same
proxy
based
approach
can
be
done
with
ebpf.
It's
been
there
for
a
while,
actually
for
a
few
years.
So
so
don't
let
people
fool
you
if
they,
if
you
see
that.
D
Connect
with
each
other
but
yeah
this.
This
experience
has
been
like
great,
like
I
connected
with
so
many
people
and
all
attended
many
sessions
and
and
the
fact
that
yesterday
I
joined
the
layer
5
community.
So
then
again,
people
were
like
welcome
people
welcome.
They
forget
for
us,
like
quite
quite
an
experience
like,
and
then
I
got
a
message
from
you
on
linkedin
like
I
really
appreciate
that.
C
Oh
yeah,
it's
well,
you've
already
got
a
nickname
too,
or
I
guess
yeah.
C
Nice
yeah,
no,
that's
great
question
for
you,
as
you
connected
with
people
virtually.
Was
that
a
lot
of
that
through
slack
or
were
there
other
venues.
D
I
guess
it
was
like
what
I
should
do.
I
had
my
session,
so
I
used
certain
sessions
and
all
so
in
the
meanwhile
like
they.
I
used
to
have
some
time
like
I
don't
have
like
another
session
like
on
10
10.
For
me
this
is
night
over
here
in
india.
D
So
so
like
let's
say,
I
have
a
session
on
10
p.m,
and
then
I
have
a
session
on
one
1
pm,
so
I
have
some
time
like,
like
suppose
I
have
time
for
one
hour,
so
I
usually
go
around
the
boats
and
all-
and
usually
I
also
have
sessions
with
different
maintainers
so
yeah.
So
I
talked
to
them
and
mostly
I've
connected
with
many
people
on
the
boats
like
the
companies.
D
They
were
quite
welcoming,
even
though
everything
about
humanities
went
above
my
head,
but
in
general
I
used
to
ask
about
the
company
like
because
that
is
quite
digestible
for
me
like
how
the
company
was
made
and
all
and
the
story
behind
the
company.
So
people
really
like
that
and
yeah.
That
is
how
I
started
connecting
with
people.
C
Make
sense
anybody
else
have
reflections
on
kubecon.
C
Nice,
okay,
hey
just
we
were
like
there's
a
lot
of
we've
got
about
like
a
third
of
the
people
that
normally
show
up
to
the
call
or
about
a
half.
We
might
be
able
to
move
through
the
call
a
little
bit
quickly.
I'm
going
to
see
if
I
can
facilitate
that
before
the
venue
takes
back
the
the
mic,
a
couple
of
things.
C
It
was
a
big
week
for
us
all
of
you
on
announcing
some
of
the
work
that
had
been
done
so
measuri
and
the
service
mesh
performance
projects
had
recently
gone
into
the
cncf,
their
sandbox
level
projects.
C
C
There's
one
of
the
was
an
analyst
on
devops.com
picked
up,
took
interest
and
wrote
up
a
few
thoughts
and
bommy
who's
on
the
call
she's
going
to
post
that
on
layer,
5.
io
under
the
news
section.
C
C
All
right,
all
right,
all
right,
cool,
okay,
push
and
push
can
relate,
I
think,
but
I'll.
Let
you
guys
chat
about
that.
So
so
navendu
then,
can
tell
us
about
the
upcoming
session
a
little
bit
later
today.
A
So
we
have
the
service
mesh
performance
project
office
hours
today
at
kubecon,
so
so
this
mesh
performance
was
a
project
that
that
that
was
built
initially
by
the
larify
community
and
then
eventually
adopted
by
the
cncf.
A
So
we
have,
we
have
a
couple
of
maintenance
joining
in
from
intel
red
hat
and
hashicorp
with
us
today,
so
the
we
will
be
having
a
project
officer
session
today
in
I
think,
a
couple
of
hours
and
join
if
you,
if
you
would
like
to
learn
more
about
the
project
and
learn
about
what
the
new
initiatives
that
we
are
taking
in
the
project
and
around
that
ecosystem
yep.
So
I
will
share
the
link
to
the
event
in
the
chat
you
can
rsvp
for
the
event
and
join
in.
A
All
right
related
to
that
we
have
bi-weekly
meetings
for
service
performance
project.
So
in
the
next
meeting,
which
will
be
on
the
28th
of
this
month,
we
will
have
yakub,
who
is
the
nighthawk
maintainer
from
google
and
he'll,
be
handling
a
session
on
adapted
load
controllers
in
night
talk,
we
will
be
talking
about
nighthawks
custom
load
control
capabilities,
including
adapter
load
controllers,
as
well
as
its
distributed
load
generation
capabilities.
A
C
A
All
right,
so
we
have
the
v0.6
release
coming
up.
We
had
a
built-in
release
call
yesterday
and
we
looked
into
some
of
the
pressing
issues
that
we
need
to
fix
before
we
make
a
release,
so
one
of
the
issues
is
was
migrating
to
helm
in
measuring
and
updating
my
meshkit's
capability
to
handle
an
updated
helm
chart
so
yep
do.
C
E
Yeah,
actually,
one
last
comment
is
still
pending
and
it's
it's
just
about
typo
as
soon
as
as
soon
as
I
get
a
reply
from
darren
on
that
particular
thing,
it
would
be
moist
as
well
as
we
will
have
a
new
release
of
mishkit.
F
C
Okay,
cool
and
rude
rocks.
That's
an
item
that
you're
looking
into
yeah.
C
And
okay,
good
all
right
and
you're,
you're,
confident
or
or
I'm
just
trying
to
gauge
your
yeah.
The
invitation
in
your
yeah.
C
Yeah,
I
could
totally
understand
it.
Man
just
signal
if,
if
your
brain
is
fried
or
if
or
okay.
A
So
mystery
serial
purpose
is
the
performance
character.
Performance
benchmarking
with
measuring
ctel
is
an
area
that
is
being
worked
upon.
I
think
pu's
me
and
pierce
will
review
it
today
and
hopefully
we
can
get
it
done
today
itself
and.
A
Yeah
and
and
a
couple
of
other
small
items
like
documenting
where
users
can
get
seat,
content
for
the
remote
provider
and
yeah
and
the
release
block.
Those
are
the
only
items
that
are
left.
C
Quick,
quick
follow-up,
just
just
because
some
recent
work
had
been
done
in
this
release
on
the
nginx.
The
mesh
adapter
for
nginx
and
jared
had
some
great
notes
that
some
great
notes
that
I
hadn't
been
able
to
go
fully
through.
So
I
figured
jared
I'll
ask
you
do
we
are
do
you
are
we
have
we
put
to
bed
each
of
the
issue,
each
of
the
challenges
that
you've
bumped
into.
G
I
haven't
tried
it
since
the
other
day
when
I
believe
it
was
darren,
was
talking
about
fixing
some
of
the
issues.
I
don't
know
if
he's
had
a
chance
to
look
at
those,
yet
if
so
so
right
now
you
can't
deploy
nginx
via
messry,
but
if
you
deploy
nginx
manually,
then
deploy
it
manually
in
the
nginx
system
namespace.
G
I
believe
it
has
to
be
that
specific
namespace,
then
messaging
will
pick
it
up
and
you
can
detect
it
there,
but
that's
the
only
way
I
found
to
to
make
it
work,
so
it
it
kind
of
works.
I
I
don't
know
that
I
would
necessarily
say
production
ready,
because
there
are
a
few
workarounds
that
users
have
to
to
go
through.
C
G
I
I
think
nginx
system
was
hard-coded
in
into
the.
G
F
E
Actually
so
so,
when
meshing
does
discover
it
in
every
new
space
is
just
that
the
resolve
will
enhance
the
hard
coding.
It
was
just
something
that
missing
disclosed
it
in
every
name
space,
but
when
you
show
it
in
the
dashboard,
we
only
show
for
engineering
cycling
system.
That's
where
the
hard
coding
is
actually.
C
Got
you
so
just
to
clarify
chair
well
jared
last
time
that
you
tried
to
provision
an
nginx
service
mesh
through
mescheri
from
the
perspective
of
kubernetes,
like
as
you
go
and
verify
that
it
was
deployed
correctly?
G
No,
so
if
you,
whenever
you
go
through
the
measure,
ui
and
say,
create
namespace,
deploy
in
this
namespace
and
create
the
namespace,
it
will
deploy
the
it'll
create
the
the
new
namespace
that
you
specify,
but
it
will
deploy
nginx
inside
the
meshrenamespace,
regardless
of
which
namespace
you
specified.
It
always
puts
it
in
the
measuring
name
space
and
since
it
needs
its
own
dedicated
name,
space
that
doesn't
work
very
well
and
also
I
had
to
manually
create
a
cluster
role
binding
for
the
pre-installed
service
account.
G
G
Otherwise,
those
pods
would
would
crash
the
center
or
loop.
Actually,
no,
they
wouldn't
enter
do
it
because
they
would
just
finish
their
jobs.
G
And
eventually
it
would
time
out
it
took
a
long
time
to
time
out,
though-
and
I
didn't
realize
that
it
timed
out,
because
I
waited
quite
a
while,
so
I
started
just
deleting
things
and
then
like
10
or
15
minutes
later,
it's
like,
oh
in
the
metro
ui.
I
got
the
notification
that
it
finally
timed
out,
so
I
ended
up
just
usually
nuking
the
everything
and
starting
from
scratch.
C
C
Rudraksha,
or
is
that
do
you
guys
anybody
have
a
guess
as
to
why
it
is
that
the
engine
x,
control
plane
is
being
provisioned
in
the
measuring
name
space?
Is
it
probably
due
to
that
lack?
Maybe
maybe
the
lack
of
cluster
role
binding
for
now,
like
like.
C
C
Between
probably
I
guess,
between,
ashish
and
rudraksh,
probably
ashish
did
like
did-
did
each
of
the
three
things
that
jared
just
mentioned,
or
that
I
think
the
two
that
he
mentioned
and
then
like
the
third
one
about
the
hard-coded
nginx
system.
Namespace,
do
each
of
these
three
things
make
sense
to
you.
C
So
the
cluster
role
binding
that
sound
that
seems
straightforward
to
make
sure
that
that's
included
and
included
up
front
like
nsm
needs
to
be
able
to
run
that
job
and
run
it's
you
know
with
well
or
that
job
needs
to
run
as
part
of
the
provisioning
and
if
it
doesn't
have
the
correct
permissions
to
be
able
to
do
that,
then
yeah.
G
G
Yes,
the
the
pre-install,
I
forget,
the
the
full
string,
but
it's
a
pre-installed
service
account
in
the
measuring
name
space.
It
gets
generated.
C
So
jared
has
a
meeting
in
like
45
seconds
any
last
questions
for
jared
before
he
has
to
go.
G
G
C
There's
some
hpe
folks
who
are
excited
when
we
talk
about
nginx
service
mesh
because
of
its
use
of
spiffy
inspire
so
yeah
yeah.
I
know
it's
I'm
really.
I
didn't
think
that
I
had
lost
faith
that
we
would
be
able
to
it
wasn't
until
the
helm
chart
came
around
that
that
my
faith
was
restored,
that
the
adapter
could
like
live
and
breathe.
So
it's
good.
C
Assuming
that
you're
feeling
comfortable
here
with
these
first
two,
we
might
crash
if
you're
aware
of
this
item
quick
question
is
the
salute-
is
like
the
proposed
solution
here
to
either
mandate
that
when
nginx
service
mesh
is
provisioned,
that
we're
measuring
is
only
provisioning.
It
in
like
that
we're
hard
coding
that
namespace
from
the
beginning,
which
is
inflexible,
but
is
a
solution
or
are
we
considering
using
other
mesh
sync
data
other
than
that
namespace.
E
Actually
so
one
of
the
thing
is
obviously
we
can't
do
that
to
suggesting.
Second
thing
is
because
now
every
installation
is
held
based
and
we
have
very
easy
method
to
do
overrides,
so
we
can
start
annotating
stuff
that
we
provision
that
would
be,
but
again
there
would
be
a
an
issue.
That
is
what
if
engineering
services
already
existed
so
so
as
a
fallback,
we
can
get
back
to
okay.
If
it
already
existed,
it
must
exist
in
this
namespace
and
then
we've
detected,
but
yeah.
C
C
E
I
think
it
is.
We
do
enforce
a
certain
images,
that's
not
disregarded,
that's
not
a
limitation
from
our
side
linker,
it
doesn't
works
properly.
At
least
it
didn't
used
to
work
properly
in
other
namespaces
because
they
have
stuff
hardwood
in
their
manifest.
So
that
was
our
reason
to
do
that.
C
And
I
think
the
same
kind
of
goes
for
istio
and
istio
system
is
that,
like
in
concept,
the
project
says
that
you
could
change
that
to
whatever
you
want
to,
but
it's
sort
of
like
an
in-concept
thing
like
conceptually,
you
could
and
then
you're,
probably
going
to
run
into
bugs
and
so
yeah
I
don't
know.
No
one
has
really
pushed
back
on.
C
So
yeah,
okay,
sorry
so
hey
next
next
topic
for
bommy
and
anita
and
where's
the
list
of,
and
for
a
lot
of
the
rest
of
you
for
all
of
the
rest
of
us.
I
had
a
new
topic,
a
different
one,
something
that
it'd
be
nice
if
everybody
thought
about
for
a
minute.
So
one
of
the
things
that
we
we've
one
of
the
big
items
that
we've
been
working
on
in
the
community,
what
service
mesh
patterns
gosh?
C
Where
are
they
so
this
sort
of
functional
section
right
here,
surface
mesh
patterns
patterns
patterns,
so
there's
a
catalog
there's
a
list
of
patterns
that
are
being
defined
and
each
pattern
like
a
pattern,
is
basically
a
function
of
a
service
mesh
like
if
you
want
to
configure
a
service
mesh
to
do
something
great
described
in
a
pattern,
and
these
patterns
are
being
there's
a
list
of
them,
a
catalog
of
them
being
curated
for
like
basically
like
a
best
practice
like
a
good
starting
point.
C
A
great
template,
okay,
that
can
be
represented
as
yaml,
can
also
be
represented
visually
to
help
people
grasp
what
that
pattern
is
about
this
visual.
This
visualization
is
an
example
of
am
I
sharing
my
screen.
No
okay.
Somebody
else
says:
okay,
I'm
sorry.
I
was
sitting
here
thinking
that
I
was
sharing
my
screen,
but
okay
yeah,
so
the
screen,
that's
being
shared,
is
great
like
like
there's
a
there's,
so
so,
okay,
so
to
jump
to
the
to
the
chase,
we'll
we
will
have
a
a
catalog
of
patterns.
C
Aditya
chatterjee
or
meshmate
aditya
gave
gave
a
crack
at
what
a
catalog
might
look
like.
C
The
thing
is
is
like
all
of
us,
so
I
suspect
the
so
that
you
know
I
don't
know
what
you
all
think,
but,
like
here's
a
decent
start,
you
all
probably
have
an
opinion
because
unless
you
don't
live
on
the
planet,
earth
like
you,
have
shopped
online
before
you've
looked
through
online
catalogs.
That's
basically
what
this
is
like.
There's
a
picture
of
the
thing.
There's
a
description
of
the
thing.
There's
a
unique
identifier.
Like
that's
pattern,
number
49
and
you
know
here's
a
description.
What
it
does?
C
Here's
a
here's,
a
an
image,
a
visualization
of
what
it
does
and
eventually
there'll
be
like
an
order
button.
It's
you
know,
it
won't
literally
be
an
order
button,
but
it'll
be
like
an
import
to
my
measuring
button.
C
C
C
This
I've
seen
better
and
that's
not
a
reflection
on
aditya.
That's
just
like
he
got
us
started.
Let's
what
else
can
we
do?
Is
there
any
interaction
on
there?
Vendi
when
you
click
on
a
on
a
pattern,
does
it
bring
up
a
modal
and
then
show
you
like
further
details
or
anything.
A
A
C
C
There
needs
to
be
a
paragraph
or
or
a
bulleted
set
of
like
additional
descriptors.
So
so
each
of
these
are
going
to
have
an
image,
an
id,
a
name,
a
short
name,
a
long
description
associated
yaml,
like
literally
the
the
pattern
itself
over
time.
Maybe
they'll
have
versions.
Maybe
they'll
have
excuse
me
like
contributor
names
or
maybe
they'll
have
an
official
mark
like
this
is
an
official
one,
that's
supported
by
the
project.
C
If
you
really
get
it
like
it,
we
probably
won't
take
it
that
much
further
like
because
what
you
would
what
you
also
could
describe
as
like
well
well,
when
should
I
use
this
pattern
and
like
what's
an
anti-pattern
like
when?
Should
I
not
use
this
pattern
and
like
what
what's
configurable
about
this
pattern,
so
you
have
this
duration
based
canary,
so
does
that
run
for
five
minutes?
Or
could
I
change
that?
It's
like
okay,
those
are
other
things
you
might
we
might
want
to
describe.
C
We
don't
have
to
consider
for
all
of
that.
Necessarily
now
it's
okay!
If
this
evolves,
we
there's
a.
We
wouldn't
want
to
publish
all
of
that
like
what.
What
are
the
caveats
of
the
pattern
when
all
the
scenarios
in
which
you
would
want
to
consider
to
use
it
all
the
anti-pattern
blah
blah
blah
like
otherwise
we're
going
to
publish
a
book,
and
there
is
a
book
being
written
on
surface
mesh
patterns,
and
so
this
is
this
will
so
so
people
can
go.
C
Maybe
we'll
go
sponsor
that
book
and
like
give
it
away
for
free
or
something,
but
the
point
isn't
about
selling
the
book.
The
point
is
not
to
like
plagiarize
the
book
and
go
publish
the
whole
thing
you
know,
but
anyway,
anyway,
we
need
a
pattern
catalog,
so
people
have
ideas.
C
The
monday
call
mondays
websites
call
great
call
to
talk
more
in
depth.
If
people
have
ideas-
and
you
want
to
just
flesh
them
out
in
figma,
there's
a
couple
of
figma
design
projects
you
can
that
have
mockups
of
meshri.io.
C
C
D
So
in
one
of
the
previous
calls
we
had
discussed
about
this,
I
think
strategy
was
also
present,
so
we
had
discussed
about
keeping
filter
there
based
on
pattern
types
so
that
if
someone
wants
to
come
and
search
for,
say
a
deployment
pattern,
we
filter
it
beyond
base
of
that
so
somewhere
near
the
top.
We
have
a
filter
based
on
type
of
the
pattern,
say,
deployment
or
other
types
of
patterns.
C
Yep,
I
totally
forgot
by
the
way,
let
me
send
a
link
in
chat
and
steal
the
ball
for
just
a
moment
and
we'll
get
off
of
this
subject.
I
actually
forgot
that
so
we
were
talking
about
meshri.io,
it's
a
great
place
for
the
patterns
that
there's
a
list
of
the
rest
of
like
a
60
like
the
the
the
60
patterns
that
are
being
defined.
Now
those
are
out
on
layer,
5.
io
under
the
book
called
service
smash
patterns
so
to
zane's
point
like
they
are.
C
These
predefined
patterns
are
categorized
and
this
implementation
done
in
gatsby.
You
has
a
filter,
it
doesn't
take
it,
it
doesn't
really
take
it
any
further.
You
know
you
can't
oops
looks
like
the
filter.
Has
that
bug
you
can't?
You
know
click
or
you
can't
there's
nothing
else
published
here,
but.
G
A
All
right
I'll
just
skip
this
topic
for
now,
but
utkarsh.
If
you'd
like
to
talk
about
app
mesh.
F
F
So
we
have
the
app
adapter
with
us.
Now,
it's
like
all
the
adapters,
it's
very
specific,
like
app
mesh
images,
are
stored
in
ecr's
and
private
repository,
so
to
pull
the
app
mesh
controller
and
other
container
images.
You
need
to
use
a
cluster
which
is
authenticated
with
and
is
able
to
pull
images
from
those
private
repositories.
F
So
we
can
see
here
all
the
usual
functionality
you
can
play
with
namespaces.
You
can
click
here
and
deploy
or
uninstall.
I
have
it
installed
in
my
cluster
right
now,
so
I
won't
do
it
again
again:
uninstall
it
install
it
other
than
that
we
have
all
of
the
functionality
that
you'd
expect
from
an
adapter
and
okay
nice.
F
C
F
F
F
F
C
C
The
push
like
the
golden
temp
like
we
were
probably
ready
to
kind
of
move
at
this
app
mesh
adapter
into
beta,
which
would
mean
you
know
all
the
all,
the
usual
things
updates
to
measure
ctl
and
the
docker
files,
then
a
manifest
or
I'm
sorry,
a
chart.
A
sub
chart
for
this.
F
C
C
I
got
to
imagine
that
appmesh
would
run
on
aws
outposts
as
well,
although
their
docs
don't
say
it
at
the
moment.
The
app
mesh
team,
the
reason
we're
talking
about
app
mesh
adapter.
It's
been
on
the
the
backlog
for
a
long
time,
but
the
app
mesh
team
had
reached
out
asking
to
meet
up
and
learn
more
and
and
so
figure.
Maybe
we
should
go
ahead
and
get
the
app
mesh
adapter
ready
so
that
they
can
so
that
they're
included.
C
C
Rudraksha,
has
you
become
somewhat
familiar
with
app
mesh
itself?
Have
you
noticed
any
particular
namespace
naming
requirements
like
the
ones
that
we
were
just
talking
about
with
nginx.
F
No,
I
guess
there
is
nothing
as
such.
It
isn't.
You
can't
install
it
in
any
name
space
for
miss
sync.
We
have
preferred
app
mesh
system
because
simply
stating
that's
what
the
tutorial
suggests,
so
we
are
assuming
most
of
the
people
would
come
and
store
it
in
app
system,
and
that
would
sync
with
the
dashboard
thing
that
utkarsh
pointed
some
moments
ago.
F
C
E
Yeah,
so
actually,
as
everyone
probably
knows,
and
if
they
don't,
let
me
tell
you:
we
have
github
actions
for
for
for
multiple
things
that
make
sure
you
can
do
like
smi.
E
Those
things
need
up.
So
when
we
perform
those
github
options,
we
first
of
all,
we
need
the
persistence
of
those
results.
If
I'm
not
wrong.
So
for
that
you
would
need
a
long
living
token
and
not
a
token
which
dies
in
24
hours
or
basically,
if
it
dies.
E
So
recently,
we
added
that
particular
functionality
in
measure
cloud.
So
what
mission
cloud
is
now
capable
of
is?
Is
you
can
ask
mr
cloud
for
generating
tokens
within
definite
lifetime?
Not
all
tokens
are
will
be
affected,
and
neither
this
particular
interface
is
is
openly
available
so
that
you,
you
don't
generate
indefinite
lifetime
tokens,
but
definitely
this
is.
This
is
something
that
machine
cloud
will
have
access
to
that
as
it
will.
E
It
is
going
to
expose
an
interface
through
which
you
can
get
individual
tokens
once
you
do.
What
you
can
do
is,
let
me
try
to
grab
the
pr
where
this
thing
was
added.
If
you
have
access
to
in
different
tokens
use,
you
can
use
those
indefinite
tokens
in
in
github
actions,
which
is
certainly
what
we're
planning
to
do,
and
obviously
now
that
now
that
the
tokens
will
never
expire
and
this
particular
implementation
is
completely
transparent
to
mystery
service.
The
message
server
is
not
any
affected.
E
All
it
knows
is
that
there
were
of
that.
There
is
a
token
and
it
has
to
pass
it
to
a
provider,
not
a
mission
cloud
to
any
provider
you're
connected
to
right
now.
Only
massive
cloud
is
supporting
it
so
yeah.
So
basically
the
implementation
is
completely
transparent.
You
can
send
the
request.
Your
request
will
be.
E
E
Okay,
let
me
let
me
share
the
screen
in
november.
If
you
can't
share
the
screen,
because
apparently
mark
has
issues
with
everything
that
I
do
so.
E
Yeah
so
as
this
particular
appeal
states
that
now
there
is
an
end
point
in
mission
cloud,
definitely
it
has
some
restrictions
built
around
it
so
that
no
one
can
abuse
this
endpoint
for
generating
random
tokens
about
after
that,
once
you
get
your
hands
on
in
different
individual
lifetime
tokens
for
a
certain
user,
then,
which
you
can
grab
the
token
and
generate
generate
a
json,
as
I
have
written
in
point
number
one
that
is
all
you
have
to
do
to
use
this
token
in
measuring
ctl
is
generate
this
particular
thing
or
dot.
E
Json
token
is
this:
if,
if
this
looks
familiar
familiar,
you're,
probably
right,
this
is
exactly
the
same
format.
What
you
get
when
you
click
on
in
the
ui
you
go
and
click
on
get
token
download
token.
This
is
the
exact
same
format,
so
all
you
have
to
do
is
provide
your
provider
and
basically
the
token
that
you
got.
You
can
place
it
in
token.
E
Now
this
now
that
you
have
this
particular
json
file,
this
can
be
placed
in
github
action,
obviously
using
secrets
or
something
you
can't
place
this
in
github
action,
and
now
you
will
have
an
action
which
does
not
have
issues
which
does
not
have
issues
with
the
with
the
token
expiry
and
all
so
I
actually
all
of
this,
like
we
did
all
of
this
thing,
to
get
the
github
action
working
so
that
it's
obviously
it's
a
fuse,
but
it's
far
more
usable,
if,
if
it,
if
you
don't
have
to
replace
the
token
every
other
day,.
A
E
Yeah
so
yeah,
definitely,
as
I
think
lee
is
pointing
more
than
start
that
we
probably
will
have
a
ui
for
creating
this
token
and
because,
obviously,
this
token
is
of
indefinite
lifetime.
So
it's
like
a
high
privilege
token.
It
doesn't
have
all
of
the
rules
assigned
to
it.
Obviously,
but
so
this
is
a
restricted
action,
so
it
won't.
The
interface
won't
be
exactly
available
to
everyone.
E
It
would
be
available
in
the
admin
panel
of
mr
clough,
probably,
and
that
is
something
that
we'd
be
probably
doing
next
so
from
there
there's
an
end
point.
We
will
create
a
interface
for
that
so
that
we
can
generate
tokens
on
the
fly
for
a
certain
user,
and
then
you
can
use
that
particular
thing
in
whatever
place
you
want
to
use.
A
Saying
would
you
like
to
mention
something.
D
Yeah
I
had
seen
this
requirement
actually
of
the
ui.
I
was
coming
to
ask
so.
C
Yeah,
please
should
be
able
to
point
you
in
the
right
direction
in
terms
of
what
type
of
ui
might
make
sense.
A
All
right
before
we
close
this
meeting
out,
I
just
wanted
to
point
out
to
this
issue
where
we
are
trying
to
improve
measuring
ctl
dogs.
So
there
is
this
tool
called
ascii
nema,
so
it
is
basically
so
here's
an
example.
So
so
this
is
to
dock.
We
can
use
this
to
document
our
measuring
ctl,
so
this
could
be
an
enhancement,
so
users
can
better
understand
what
is
happening
in
measuring
ctl
and
we
can
provide
a
guided
user
flow
so
use
this
instead
of
just
simply
the
command.
A
So
this
will
help
the
users
understand,
measuring
ctl,
better
and
understand
all
its
capabilities.
So
there
is
an
open
issue
in
measuring
that
talks
about
this.
This
could
be.
A
This
might
need
some
design
on
what
goes
into
what
goes
into
this
new
new
user
experience
so
feel
free
to
take
a
look
and
give
suggestions,
push
will
be
push,
will
be
handling
this
issue
and
please
reach
out
to
piyush.
If
you
have
any
suggestions
on
this
as
well
and
also
there
are
a
couple
of
a
lot
of
open
pull
requests
with
the
bug
bash,
as
well
as
the
new
release.
So
I
just
wanted
to
make
a
call
out
to
everyone
here
to
anyone
can
go
ahead
and
review
this
pr.
A
A
All
right,
thank
you,
everyone
for
joining.
I
hope
to
see
you
all
at
the
smp
project
office
hours
later
today.
If
not
I'll,
see
you
guys
next
week,
bye.