►
From YouTube: Meshery Development Meeting (July 7th, 2021)
Description
Meshery Development Meeting - July 7th, 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
everybody
welcome
to
the
measuring
development.
Call
today
is
the
7th
of
july
and
before
we
get
started
with
today's
topics,
we
have
another
meeting
which
will
happen
on
saturday,
this
saturday,
10th
of
july,
and
the
timings
are
given
on
the
on
the
meeting
minutes.
So
it
will
be
to
identify
some
bugs
in
measuring
ctl
and
we
will
we
will
join
in
call
and
we
will
walk
through
we'll
test
measures
it'll
end
to
end.
A
So
if
that
that's
something
that
interests
you
please
join
this
saturday
and
if,
if
you
are
having
problems
with
the
time
just
ping
on
the
thread
in
measuring
ctl
in
the
slack
channel,
so
yeah
so
for
new
contributors,
this
might
be
a
good
good
way
to
get
started
with
using
measuring
and
testing
measuring
and
lee
has
shared
a
test
plan.
A
So
our
next
release
is
0.6
and
we
are
testing
measuring
components
throughout.
So
we
will
be
testing
measuring
server
and
machinery,
adapters
measuring
ui,
all
sorts
of
things.
It
is
not
just
limited
to
measuring
ctl,
but
we
will
also
have
separate
bug
hunting
meetings
for
for
the
other
components
as
well.
A
Yeah
yeah
with
that.
I
see
a
couple
of
fresh
faces
on
this
car
so.
A
A
C
Hello,
my
name
is
devopria.
This
is
my
first
mastery
development
meet
I've
attended
a
few
meets
this
week,
so
I'm
a
first
year
student
studying
at
with
pilani
hyderabad.
D
That's
awesome,
but
I
started
to
interrupt
who
who
introduced
just
a
moment
ago.
I
missed
that.
D
Day
well,
priya
is
old
news
in
my
book,
she's
been
so
who
just
introduced
just
before
dabo
priya.
B
E
Hello,
everyone.
This
is
harshita,
I'm
pretty
new
comment
to
measure.
I
guess
I
I
just
saw
a
couple
of
I
just
made
a
couple
of
pr's
and
I'm
looking
forward
to
contribute
to
my
shari's
idea.
That's
my
main
interest
right
now
in
mashari,
yeah.
D
Awesome,
yeah
harshida.
I
think
you
have
an
an
open
pr
that,
like
I
think,
you're
waiting
very
patiently
for
review
on.
D
Cool
good
good
by
the
way,
just
for
everybody
on
the
call.
Not
all
of
you
have
enough
bandwidth,
maybe
to
support
a
webcam.
But
if
you
do
we'd
love
to
see
your
faces
so
be
be
friendly.
A
Yeah
hello:
this
is
bharath
here,
I'm
a
first
year,
undergrad
from
jit
and
well,
I'm
quite
new
to
this,
but
I
have
introduced
myself
before
and
yeah.
I
look
forward
to
contributing
and
I
came
through
a
hackathon
which
was
held
by
jid
again.
Thank
you.
A
Thank
you,
better
yeah,
any
others.
If
you,
if
you
are
having
mic
problems,
then
you
can
introduce
yourself
in
the
chat.
A
So
first
up
we
have
lee
lee.
Would
you
like
to
talk
about
the
hd
remote
provider
call
back.
D
Oh
wow,
okay,
that's
first
up!
Do
we
should
yes,
I
can
utkarsh.
Do
you
think
that
it
is
better
to
talk
about
it'll,
be
more
understandable
if
we
first
talk
about
support
for
mastery,
ctl
authing
to
mastery
server
and
then
talk
about
remote
provider
call
back
or
is
it.
B
D
D
But
I'm
happy
to
give
I'm
happy
to
talk
anyway,
give
some
context,
while
it
would
crash
bringing
this.
B
Yeah,
I
think
screen
should
be
visible
now,
so
I
recently
created
a
prn
machine.
It
was
about
being
able
to
authenticate
message,
basically
authenticate
wire
measuring
cto
to
mystery
server
and
the
pr
is
this
if
anyone's
interested.
If
you
can
come
in
here
and
look
I'll
demo,
what
exactly
it
is
about
so,
basically,
as
it
says,
it
introduces
two
new
commands
to
measure
cto.
B
One
is
a
system
login
and
another
one
is
system
logo
with
what
basically,
the
command
is
what
it
does,
but
what
it
essentially
does
is
basically
lets
you
login
via
mystery
ctl
and,
let's
see
logos,
obviously.
D
Yeah,
what
a
lucky
guy
he
didn't
get
stuck
in
some
weird
all
right!
Well,
let
me
ask
you
so
so
I'll
do
so.
A
D
Cool,
so
crush
will
interrupt
me
when
he
comes
back.
What
have
you
so
to
give
everybody
some
context
as
to
the
topic
that
karsh
is
speaking
on
he's,
you
know,
there's
a
lot
of
things.
A
lot
of
things
that
we
do
on
this
particular
call
are
talk
about
either
upcoming.
D
You
know
designs
for
how
the
measuring
should
work
and
things
that
it
should
be
doing,
changes
and
bugs
and
active
prs,
and
that
kind
of
a
thing
will
anita
who's
on
the
call
is
working
to
ensure
that
we
every
so
often
have
like
an
intro
to
what
the
heck
is.
D
Measury
and
sort
of
also
slash
what
the
heck
is
a
service
mesh
a
little
bit
and
so
she's
been
she's
working
on
getting
those
scheduled,
and
so
a
lot
of
times
we'll
take
some
time
to
explain
a
little
bit
of
that
on
on
this
call.
So
we're
not
going
to
do
that
at
the
moment.
But
but
let
me
let
me
relate.
D
Let
me
see
if
I
can
make
an
analogy
to
what
utkarsh
is
about
to
show
he's
talking
about
authentic,
a
user
authenticating
to
mesheri
and
he's
going
to
show
this
from
mesri
ctl
from
meshary's
command
line,
client
and
so
by
show
of
hands,
I
suppose
and
zoom,
since
we
never
do
this,
I
don't
know
if
I
don't
even
know
if
I
know
how
to
do
it,
how
many
of
you
are
familiar
with
docker
login?
How
many
of
you
have
ever
used?
Docker
login.
D
A
This
is
embarrassing.
It's
there
is
a
button
called
reactions
in
the
bottom.
A
D
Yeah
that
open
mouth-
that
that's
me,
okay,
thanks,
I
just
learned
something:
oh,
oh,
is
that
what
reactions
means
is
like?
Oh,
okay,
all
right,
good,
okay,
so
there's
a
few
of
us
here
that
are
familiar
with
docker
login
is
zookash
back
on.
I
don't
want
to
steal
this
thunder
necessarily,
but
so
so
when
a
user
goes
to
use
meshri,
they
meshrey
provides
identity
by
way
of
a
provider.
D
D
It
is
called
none
or
local.
D
I
don't
know
that
I've
got.
I
also
don't
know
that
I've
got
mastery
running
so,
but
most
all
of
you
who
have
used
mashri
have
seen
that
so
I'm
going
to
start
up
measuring
system
start
and
good,
and
so,
if
you
choose
so
right
now,
the
only
way
to
authenticate
to
you
know
to
to
create
a
session
with
mesh
reserver
right
now.
The
only
only
way
to
do
that
is
through
the
web-based
client
I'm
going
to
share
my
whole
screen
so
that.
D
And
it's
like
okay!
Well,
that's
good,
but
also
the
web-based
client.
Isn't
the
only
client
there's
the
cli
as
well,
and
so
wouldn't
it
be
nice
to
be
able
to
authenticate
from
any
client,
okay,
yeah
great,
and
so
that's
basically,
what
outcast
is
bringing
forth
is
a
proposed
command
called
meshrictl
system,
because
system
refers
to
mesherie
and
its
components,
measuring
ctl
system
login,
and
then
it
will
facilitate
the
oauth
interaction
or,
if
I'm
sorry,
it'll
facilitate
the
provider.
Specific
authentication,
which
is
good.
That's
a
something
for
him
to
demo.
If
he.
B
B
You
can
do
it
quickly
before
it
crashes
again,
so
I
think
he
has
given
all
the
politics.
So
this
is
the
command.
If
I
start
it
up,
it
gives
me
prompts
that
prompt.
Regarding
that
what
provider
I
want
to
select,
is
it
going
to
be
mystery,
or
is
it
going
to
be
none?
This
is
writer
supported
by
report.
Not
by
now.
Let's
say
I
select
mystery.
B
It
opens
up
the
browser
for
me,
takes
me
to
the
to
the
provider
flow,
whatever
it
be
github,
because
that's
what
I
use
and
when
it's
done
it
will
redirect
me
to
saying
basically
that
hey,
if
I
close
this
underneath
your
successfully
logged
in
and
I
get
to
get
a
prompt,
basically
not
a
prompt.
I
get
output
in
here
which
says
that
I
have
successfully
logged
in
right
now
the
tokens
are
getting
saved
inside
by
default.
It's
getting
saved
inside
dot
missionary
slash
dot
json,
and
this
is
a
token
that
got
saved.
B
It
has
information.
This
is
basically
the
same
thing
that
you
might
have
the
downloaded
earlier
by
accessing
reports
for
a
9.81
and
when
you
would
click
on
this
I'll
get
open.
This
is
the
exact
same
thing
which
is
getting
saved
onto
the
file
system:
empty
plain
text
format
right
now
at
least
this
location
gets.
Basically,
we
chose
this
location
depending
on
what
you
have
configured
in
your
conflict
in
your
context,
so,
right
now,
if
you
want
to
see
the
context
that
I'm
using
is
dev
yeah
current
context
is
dead.
B
As
you
can
see-
and
I
have
said
that
hey
I
want
to
use
my
token
difficulty-
that's
I
was
actually
testing
something
out.
So
my
token
location
is
difficulty
that
corresponds
to
actually
this
particular
location.
That's
where
the
token
got
saved
the
second
command.
That
is,
that
is
also
in
the
pr
I
added
is
the
ability
to
log
out
that
is,
you
can
do
this
and
it
will
log
you
out,
but
it
essentially
does
behind
the
scenes,
basically
clears
out
your
token.
B
Let's
just
quickly
do
the
unprovided
login
that's
going
to
be
quicker,
because
that's
just
basically
because
there
is
no
asset
authentication
that
happens
when
you
actually
select
none
provider.
So
this
is
for
this
is
the
authentication
mechanism.
So
if
someone
may
have
noticed
coming
to
the
second
engine
that
I
can
have,
if
someone
would
have
noticed
that
in
order
to
get
all
of
this
thing
working,
a
request
was
sent
to
this
particular
endpoint,
which
is
located
in
four
to
five
six
seven.
B
This
is
not
something
that
mystery
actually
spins
up
all
the
time.
In
fact,
this
address
is
basically
generated
on
the
flag,
so
when
you
were
actually
logging
in
this
thing,
this
a
temporary
authentication
server
was
created
for
you,
so
that
we
can
capture
this
token.
This
is
this
is
going
to
the
to
one
of
the
thing
that
machine
server
does
what
machine
server
does
is
basically
it
sort
of
hardwoods.
B
The
fact
that
if
you,
if
you
want
to
do
authentication
via
a
miniature
server,
it
will
hard
code
the
fact
and
and
ask
mission
provider
to
always
always
redirect
the
authenticated
request.
That
is
a
token
token
based
request
to
message
server
and
only
mystery
server
can
read
and
process
that
thing,
but
that's
not
something
that
was
needed
in
there.
What
we
actually
needed
was
it
was
a
custom
callback
model
which
is
which
is
one
of
the
agenda.
That
is,
we
should
be
able
to
configure
custom
callback
model.
B
We
should
be
able
to
let
mystery
server.
Know
that
hey
you,
although
the
default
behavior,
is
supposed
to
be
that
you
are
going
to
be
the
one
who
is
going
to
get
the
call
I'm
going
to
get
the
going
to
get
the
redirected
call,
which
will
have
the
token,
but
if
a
user
wants
to
change
that,
they
should
be
able
to
change
that
so
that
it
also
kind
of
works
for
one
of
the
use
case
that
we
had.
That
was
basically
what
happens
if
measure
is
behind
the
proxies.
B
In
that
case,
what
you
can
do,
what
you
should
be
able
to
do,
what
you
you
would
be
able
to
do
is
basically
you
can
ask
mystery
server
to
say
that
I'm
configuring
a
custom
quality
order.
It
would
be
probably
present
in
your
content
that
I
was
just
showing
it
would
probably
present
in
your
config
under
context
or
or
maybe
somewhere
else.
I'm
not
sure
that's
something
yet
to
be
proposed,
but
you
would
be.
You
should
be
able
to
configure
your
callback.
B
You
are
in
there
and
then
this
kind
of
work
around
that
we
have
right
now.
That
is,
we
create
temporary
temporary
thing.
Actually
that-
and
this
will
still
be
there-
that
is,
we
will-
we
will
still
be
creating
the
thing,
but
now
the
after
the
the
proposed
thing,
that
is,
we
should
be
able
to
point
figure
the
callback
on
the
fly
once
that
gets
gets
going.
B
What
we
will
be
able
to
do
is
basically,
instead
of
bypassing
the
authentication
flow,
we
will
be
able
to
just
use
mystery
server
for
all
of
the
things.
That
is
right
now.
B
What
happened
was
misery,
ctl
actually
queried
message
server
for
the
information
regarding
the
provider,
and
then
it
actually
went
through
the
uphold
of
the
login
process,
and
then
it
captured
the
token,
but
once
that
thing
goes
through,
that
is
so
you
can
configure
the
callback
url
machine
server
would
be
able
to
take
care
of
that
industry,
cto
would
be
or
all
of
the
clients
would
be
less
would
have
less
responsibilities
and
he'll.
Also
fix
this
fixed
issues
that
we
have
prop
with
proxies
right
now.
B
A
No,
no,
you
didn't
ask
people,
ask
people
like
get
around
to
what
you
are
explaining
samir.
Would
you
like
to
illustrate
like
the
pattern.
B
F
So
let
me
see
if
anybody
can
see
this
okay,
so
this
is
ideally
the
architecture
diagram
which
we
had
and
what
really
happens?
We
have
a
private
endpoint
which
has
been.
F
F
So
in
this
case
it
should
be
the
default
url
of
the
istio
ingress,
which
you,
you
may
create
a
load
balancer
or
something,
but
in
this
case
what
happens
in
our
cases.
So,
ideally
the
request,
if
you
do
it
directly
to
this
particular
load
balancer,
which
is
ideally
a
internal
load
balancer
of
azure,
it
cannot
be
accessible,
so
you
cannot
be
authenticated.
F
So
what
ideally
you
have
to
do
is
whenever
you
want
to
access.
This
is
gateway,
go
and
hit
this
particular
gateway.
Okay,
so
I
hope
this
is
visible.
So
this
is
a
application
gateway
from
the
edge
or
side,
and
you
have
to
keep
this
so
that
it
can
go
and
internally
talk
to
this
particular
ingress
gateway.
F
Okay,
so
what
earthquake
just
said
instead
of
using
the
public,
so
the
ip
address
or
any
any
of
the
dns
of
this
in
istio
ingress?
You
can
configure
your
own
callback
url,
which
is
in
this
case,
should
be
your
application
gateway
url,
which
is
a
public
url.
In
fact,
does
that
make
sense.
D
By
the
way,
just
for
everyone
else
on
the
call
samir
had
explained
this
a
meeting
or
two
ago
and
he's
re-explaining
it
for
your
benefit.
Yes,
you
so.
F
Yeah
yeah,
so
how
the
request
flows
actually
is
now.
This
is
the
user,
and
this
is
your
website.
When
you
go
to
your
website,
then
you,
like,
you,
know,
want
to
access
some
of
these
systems
or
some
of
these
services,
which
is
running
here.
So
you
cannot
ideally
in
a
general
scenario,
what
happens
when
you
hit
something
in
the
website
here?
F
It
should
directly
go
and
hit
the
istio
grace
gateway
and
then
everything
you
can
access,
but
for
the
security
reasons,
what
we
had,
we
have
a
provision,
the
application
gateway,
which
is
a
azure
service,
because,
okay,
to
give
you
the
hole
in
pricing
azure,
so
this
is
azure
application
gateway.
Wherein
you
hit
your
website
or
any
anything
of
your
address
or
dns
name,
it
goes
to
the
name
server
and
finally,
it
goes
to
the
public
ip
of
your
application
gateway.
F
Okay,
so
from
there
it
executes
some
of
the
rules
to
identify
whether
it
is
from
a
authorized
like
authorized,
signet
signatory
or
not
anything,
and
then
it
will
after
it
is
authenticated
here
with
this
rules,
https
rules,
then
it
will
forward
this
to
the
end
point
of
these
ist
and
grace
url.
If
you
see
this
endpoint,
this
is
a
private
link
of
arjun,
so
only
the
azure
resource
can
identify
these
things.
F
In
this
case,
everything
whatever
you
see
all
are
private,
so
ist
one
grace
gateway
is
also,
though
it
is
a
ingress
gateway,
but
this
is
a
type
load
balancer
and
it
is
internal
load
balancing
so
which
is
again
like.
However,
you
can
connect
to
them
with
all
of
this
admin
plan
so
wherein
it
has
a
like
ssh
certificate
or
ssh
connection,
or
you
can
say
any
of
the
private
key
connection
which
will
allow
enable
your
application
to
do
that.
So
this
is
ideally
a
same
blog
post.
Regarding
that
you
can
just
go
around.
F
D
D
That's
so
cool,
that's
that's
really
great,
quick,
quick
verification,
and
no
doubt
this
is
in
the
blog
post,
but
I'll
lead
by
example,
and
ask
a
silly
question.
So
everyone
else
can
can
follow
the
private
link
between
the
application
gateway
and
okay.
Actually,
it
says
it
right
there
between
azure
dns.
I
was
going
to
ask
where,
if
that
dns
is,
you
know
also
in
azure-
and
it
looks
like
so
there's
your
user.
F
Yes,
so
whenever
you
see
a
private
link,
dot
central
india
dot,
so
this
is
ideally
the
middle
one.
Is
your
location
wherein
you
are
hosting
your
application
and
the
last
one
is
the
dns
name
which,
as
your
host
in
in
their
services?
So
ideally,
if
you
go
here,
you
can
see.
So
there
is
a
as
your
site
wherein
you
can
find
what
a
lot
of
the
dns
name
you
can
use.
F
You
can
use
to
have
this
pawn.
If
I
go
to
this
link.
G
F
D
Going
once
cool
like
for
some
of
us,
we're
quite
thankful
that
you
wrote
a
blog
post
because
it'll
take
a
long
time
to
digest.
What's
going
on,
yeah.
F
D
Can
that's
great,
it's
samir,
beautiful,
hey,
we're
gonna
have
to
go,
get
we're.
Gonna
have
to
go,
be
social
with
that
blog
post.
That's
that's
awesome.
Let
me
does
anybody
else.
Have
some
questions?
If
not,
I'm
gonna
show
another
diagram,
because
what
engineer
doesn't
love
looking
at
diagrams.
D
Cool
if
you
come
up
with
others,
while
we're
going
so
so
I'll
just
go
ahead
and
bookmark
this
blog
all
right.
So
all
right,
so
samir's
challenge
is
not
an
uncommon
one.
As
a
matter
of
fact,
this
same
challenge
where
samir
correct
me,
if
I'm
wrong,
but
you're
deploying
you're
you're
using
manage,
I
manage
kubernetes
service
from
azure
aks
you're
running
istio,
in
that
environment,
on
and
by
the
way
at
this
point
like,
is
it
a
single
cluster
or
is
it
are
there?
Multiple
is.
F
Okay-
and
it
has
a
it-
has
a
capability
to
scale
up
to
1000
instances,
so.
D
I
see
okay,
so
single
yeah,
single
kubernetes,
okay,
by
the
way
I'm
curious
about
that,
did
they
just
auto,
upgrade
the
control
plane
and
the
cubelets.
Like
the
nodes
I
mean.
Maybe
I
should
actually
go
use
one
of
these
managed
services,
but
do
they
give
you
full
warning
as
to
when
that's
going
to
happen
or
is
that
they
just
empower
you
to
control
your
kubernetes
version.
F
Yes,
so
whenever
you
kind
of
you
know
configure
this,
then
there
is
a
like,
you
know
configuration
wherein
you
can
tell
I
can
do
or
I
can
go
until
this.
Okay,
if
you
can
give
100,
then
until
it
is
100,
then
you
can
like
know
they
will
not
give
you
any
alert.
So
after
that
they
will
just
give
you
an
alert
that
your
quota
has
been
filled.
Would
you
really
like
to
go
beyond
that?
So
no.
D
Okay,
okay,
did
they
give
you
a
lot
of
choice
over
what.
F
Kubernetes
version
is
running
yes,
so,
ideally
that
is
a
node
pool
which
you
have
to
define
and
that
in
that
node
pool
you
have
to
give
what
is
the
version.
So,
ideally
it
will
be
already
in
the
latest
one.
So
currently
it
is
version
19,
so
you
can
go
to
18
version
and
you
can
also
go
to
preview
version
of
20
or
21.
I
guess.
F
A
freebie
for
me
yeah,
but
one
thing
is
whatever
scale
happens,
all
will
have
the
same
kubernetes
version,
because
master
node
doesn't
scale
because
that
is
a
ideal
or
you
can
say
that
is
a
primary
node
wherein
kubernetes
is
there,
but
the
same,
like
you
know,
configuration
what
you
have
to
do
for
the
node
pool
that
will
be
applied
for
all
the
scaling.
D
D
So
specifically,
mastery
ctl
as
a
client
of
mesh
reserver
to
authenticate
and
make
secure
requests
to
you
know
to
ensure
that
each
of
the
measuring
ctl
commands
that
are
being
executed
are
in
fact
authenticated
and
that
that
user
has
permission
to
invoke
to
stop
mesh
restart,
measuring,
deploy
a
mesh
on
all
those
things.
D
So
that's
really
foundational
to
some
things
to
come
with
respect
to
authorization.
So
it's
one
thing
to
identify.
Have
I
authenticated
the
user
and
say
you
know
this
token?
Here
represents
the
the
verified
proof
that
you
are
you
who
you
say
you
are,
but
whether
or
not
you
have
the
permission
to
deploy
another
instance
of
istio
or,
for
example,
that's
future
functional
authorization,
future
functionality
to
come,
I'm
tossing,
and
so
in
accordance
with
samir's
use
case.
D
He
he
you
know,
so
I
don't
want
to
repeat
things
that
crush
had
said
I,
but
in
essence
with
meshary
running
behind
the
istio
ingress
gateway
here,
basically,
with
mesh
re
running
on
the
service
mesh
as
well,
it's
the
the
traffic
and
the
interactions
that
happen
with
mesh
reservoir
need
to
flow.
You
need
to
follow
this
flow
and
as
a
user,
when
you
go
to
authenticate
to
mescheri
server,
you
can
you
know
right.
Currently
you
have
choice
of
a
local
provider
or
a
remote
provider.
D
The
remote
providers
bring
a
little
more
functionality.
One
of
the
things
that
they
bring
is
persistence
so
it
because
sort
of
like
to
remember
the
performance
test,
you've
done
your
configuration
settings,
etc,
and
so
the
the
remote
provider
that's
available
today
is
is
it's
called
meshri
in
the
drop-down
when
you
sign
in
and
the
interactions
that
and
then
when
you,
when
you
sign
in
you
have
this.
I
know
this
is
quite
the
eye
chart.
Let
me
move
some
of
the
you
know
and
you
don't
have
to
digest
this,
and
this
is.
D
But
this
is
the
architecture
and
the
interactions
that
happen
for
measury
remote
provider
from
sri
cloud.
D
You,
you
all,
are
no
doubt
familiar
with
the
fact
that
you
get
presented
a
choice
of
you
know:
different
identity
providers
like
which
one
you
want
to
sign
into.
If,
if
you
want
to
use
message
cloud,
if
you
want
to
use
mesh
remote
provider
well
in
in
the
the
way
that
those
interactions
happen,
you
know
clearly
there's
an
authentication
that
goes
on
with
the
provider
of
your
choice
and
then
you
know
a
token
a
jot
is
created
and
that
gets
handed
off
and
there's
interactions
that
happen
between
measurey
server
and
the
remote
provider.
D
This
is
this
example
is
specifically
what
happens
with
measuring
cloud.
People
can
write
other
remote
providers
to
do
different
things
if
they
want,
but
this
isn't
a
canonical
implementation.
If
you
will
that's
sort
of
the
example
well
and
the
challenge
in
samir's
environment
is
that
the
return
path
for
once
that
once
you've
authenticated
and
it's
time
to
pass
back
your
token?
Well,
the
as
utkarsh
was
saying
there
was.
D
You
know
there
was
a
lack
of
consideration
for
the
fact
that
that
callback
url,
like
where
meshary
cloud
needs
to
call
back
to
mesherie's
server,
that
that
needs
to
be
configurable,
because
that
can
be
totally
environment
dependent
and
as
samir
is
highlighting
it.
As
he's
been
kindly
explaining
to
me
about
how
azure
dns
works
and
how
azure
is
like
that
url,
you
know
whether
it's
this
or
or
whether
it's
something
else
is
well,
that's
totally
accustomed
to
samir's
deployment.
So
that
needs
to
be
a
configurable
item.
D
He
needs
to
be
able
to
tell
mesherry
hey.
This
is
where
I
want
you
to
call
back
to,
and
so
that's
what
this
is
sort
of
defining
it
and
it
turns
out
to
to
achieve
that
is
relatively
easy
in
terms
of
adding
the
ability
for
the
person
who
deploys
measury
to
just
provide
that
a
custom
callback,
url.
D
So
so
a
lot
of
time
spent
on
authentication
today
and
kind
of
how
the
flows
and
things
more
than
I
would
expect
anyone
who's
seeing
those
diagrams
for
the
first
time
to
really
digest
the
nice
thing
is
samir's
got
a
very
well-written
post
that
we
can
all
go
absorb.
So
any
other
questions
or
comments
on
this
topic
before
we
go
to
the
next.
D
H
Yeah,
that
sounds
good,
so
let
me
share
my
screen.
First,
while
things
are
coding.
H
So
basically,
it's
your
few
of
the
charts
which
were
there
in
the
new
version
of
http
we're
not
working
with.
Now
they
were
using
custom
charts
due
to
which
it
was
breaking
things
for
us,
and
now
it
is
fixed,
I'm
not
sure.
H
H
Just
use
this,
it
doesn't
be
one
option
and
yeah.
They
should
be
working
now
and
yeah.
That's
a
small
update.
I
guess
we're
going
to
get
deep
into
why
this
was
not
working
was
because
they
shifted
to
a
custom
grafana
chart
instead
of
the
default
layouts,
and
we
had
to
like
edges
to
that
particular
thing
in
the
background
to
get
those
charts.
So
that
was
the
reason
why
this
was
happening.
H
D
Yeah,
actually,
actually,
you
know
what
I'd
be
willing
to
bet
that
the
majority
of
the
individuals
on
this
call
have
not
seen
these
charts
like
and
that
we
we're
trying
to
work
to
change
that
we're
trying
to
work
to
make
it
where
actually,
it
would
be
the
inverse
that
most
people
when
they
deployed
mescheri,
they
would
like
they
would
see
some
of
these
metrics.
Some
of
these
charts
were
like
basically
right
when
they
bring
up
mescheri
for
the
first
time,
and
so
this
effort
that
that
dhruv
is
giving
here
is
helping
us
get
there.
D
In
the
meantime,
though,
it
would
be
really
great
for
all
of
you
to
be
able
to
connect
to
grafana
and
see
some
metrics
play
with
these
things.
Ayush
has
been
making
some
enhancements
and
the
way
that
the
charts
work
and
he's
he's
ready
to
work
on
some
different
style,
charts
and-
and
so
I'm
gonna
put
a
url
into
the
chat,
that's
to
a
set
of
labs
that
all
that
that
will
take
all
of
you
through
on
setting
up
meshri,
deploying
istio,
deploying
grafana
and
connecting
meshri
and
grafano
together
and
that
way.
D
Hopefully,
then,
all
of
you
can
get
to
the
point
where
you've
got
these
charts
and
you
can
play
around
with
them.
I'd
be
curious,
particularly
for
those
that
are
devops,
oriented
or
gosh.
That's
such
a
horrible
use
of
that
word,
but
for
any
any
of
you
who
are
operations
oriented,
you
might
have
an
opinion
over
what
charts
are
being
shown.
One
of
the
questions
we
need
to
ask
ourselves:
we
need
to
answer.
D
Ask
and
answer
is
well
if,
if
we're
trying
to
facilitate
some
quick
insights
as
to
the
health
of
your
mesh
and
the
health
of
the
services
that
are
running
on
your
mesh,
like
your
kubernetes
environment
and
things
like
hey,
what
what
stats
should
we
show
like
what
would
be
helpful
to
to
people
so
that
I'll
leave
that
question
with
all
of
you.
D
Many
of
the
service
meshes
come
with
an
embedded
instance
of
an
optional
embedded
instance
of
prometheus
and
grafana
istio
does
liquidity
does
and
then
off
the
top?
My
head,
I
couldn't
prattle
off
the
rest
of
kuma
and
open
service
mastery,
but,
and
so
like.
What
I'm
saying
is
like
deploying
the
you
know.
You
know
the
the
hacking
them,
but
the
majority
of
time
people
are
going
to
want
to
have
these
insights
and
so
like.
D
It
is
the
case
that
we
don't
necessarily
measuring,
doesn't
necessarily
need
to
have
a
grafana
instance
deployed
measuring
can
directly
speak
to
prometheus.
It
can
discover
a
prometheus
is
already
in
the
environment
and
it
can
actually
query
prometheus
for
particular
metrics.
So
that
question
that
I
asked
each
of
you
basically,
which
is
what
information?
What
chart
do
you
think
should
be
shown
or
charts
the
the
world?
D
Is
your
oyster
there
a
bit
like
you
know
it
could
be
the
default
chart
that
ships
with
linker
d's
configuration
or
hdl's
configuration
in
grafana
or
it
could
just
be
a
metric
that
we
could
retrieve
from
prometheus
so
anyway
think
think
about
it
and
I'll
put
the
link
to
the
labs
in
the
chat.
G
A
I
Recently,
we
have
added
supports
to
deploy
our
service
mesh
patterns
from
the
machine,
ui
patterns
itself
and
yeah.
Let
me
just
demo
this
pattern.
First
yeah
here
we
have
the
sql
pattern
here
and
when
we
hit
this
the
reply
button
in
the
actions
you
can
check
whether
it
is
deployed
or
not
and
going
to
the.
I
I
I
A
Who's
got
a
joke
lee.
Could
you
just
fill
in
on
what
applications
are
and.
D
I
I
am
first
showing
the
department
of
the
patents
then.
D
Okay,
you
know
yeah,
let's
talk
about
applications
for
a
moment.
If
we
can,
mr
gefeller
is
on
the
call-
and
so
I'm
saying
his
name,
so
this
year's
perk
up
just
in
case
he
was
otherwise
distracted,
which
generally
isn't
the
case
as
a
side
note.
So
let's
talk
about
yeah,
we
think
about
applications
all
right.
D
So
samir
had
asked
a
question
some
long,
some
long
time
ago,
as
he
had
entered
in
and
was
familiar
entered
into
the
community
and
was
familiarizing
with
mesheri,
and
you
know,
part
of
his
question
was
hey.
You
know
this
is
this
is
intriguing.
This
is
neat,
I'm
going
to
look
to
use
it.
D
He
didn't
say
he
was
going
to
write
such
a
beautiful
blog
post,
but
but
that's
great,
he
did
say
so,
while
mescheri
is
managing
my
service
mesh,
that's
quite
helpful,
but
also
the
reason
that
I'm
running
the
service
mesh
is
because
I
have
these
services,
these
apps,
these
workloads
running
on
the
mesh,
and
so
and
am
I
going
to
have
some
running
off
of
a
mesh
potentially
as
well.
So
so
you
know
is
measuring
cognizant
of
that.
Like
does
it?
Is
it?
D
What
can
it
do
for
me,
and
it's
not
that
it's
not
that
there
aren't
other
tools
out
there
to
help
with
application
lifecycle,
management
of
services
that
are
running
in
kubernetes
or
even
on
a
service
mesh?
It's
not
like
helm,
doesn't
help
you
in
this
regard
or
or
there's
a
there's,
a
litany
of
tools.
I
didn't
mean
to
call
it
home
specifically.
What
I
meant
to
say
is
that
you
can
still
use
most
of
those
tools
to
deploy
your
apps
onto
the
mesh.
There
are
certain
apps
that
require
onboarding.
D
There
is
like
there
are
onboarding
preparations
to
do
in
order
to
get
your
app
ready
to
deploy
on
the
mesh
and
so
to
historically
or
earlier
on
and
really
sort
of.
Even
at
this
moment,
the
stable
versions
of
mescheri
do
a
tend
to
do.
Do
bake
in
a
couple
of
sample
apps
to
help
you
explore,
learn,
try
out
things.
Compare
service
meshes
benchmark
performance
across
them.
D
D
You
may
as
well
use
helm
or
whatever
else
you're
using
of
what
gurav
and
the
vendu,
and
you
know
what
all
of
you
are
working
on-
the
purpose
of
making
applications
which,
in
my
mind,
is
somewhat
synonymous
with
the
workload
making
it
its
own
first
class
citizen
has
to
do
with
well,
it
has
that
has
to
do
with
the
fact
that
mesherie
will
hopefully
ultimately
help
with
it'll
do
a
bunch
of
things
around
applications,
but
hopefully,
like
ultimately
help
with
onboarding
services
onto
the
mesh
off-boarding
them
off
the
mesh.
D
In
the
reason,
part
of
the
reason
I
would
say
that
is
because,
like
applications
are
king,
which
is
applications,
are
why
we're
running
the
infrastructure
in
the
first
place.
So
if
we're
not
loving
on
them,
treating
you
know,
understanding
them
assisting
them
as
their
they
go
onto
the
mesh
off
the
mesh.
Then
you
know
it's
not
like.
We
wouldn't
have
a
great
tool.
It's
just
that
we're.
D
D
So
so,
if
you're,
if
you're
a
female
on
the
call
like,
I
might
say
that
that
makes
the
developer
the
queen,
because
developers
there's
more
and
more
importance
and
kind
of
being
given
to
developers
and
the
fact
that
there's
developer
defined
infrastructure,
they
make
some
of
these
early
choices
and
set
forth
how
things
are
going
to
run
and
so
for
all
the
dudes
on
the
call
that
are
developers.
Like
I'm
sorry
yeah.
I
guess
I
guess
your
your
royalty
of,
maybe
anyway,
your
queens
as
well
in
my
book,
so.
D
The
deal
here
is
michael
was
grocking
this.
After
just
a
single
sentence,
saying
hey,
we
should
have
support
for
applications
in
measuring
and
a
great
question
he
was
asking
is
okay,
you
know
all
right,
good,
hey!
You
know
what
does
that
mean
for
different
adapters?
What
does
that
mean
for
the
existing
sample?
Apps?
D
There
are
built-in
those
baked
in
sample
apps
that
that
you
can
have
deployed
by
invoking
an
operation
by
invoking
an
adapter
operation.
Those
operations
will
continue
to
live
on
for
some
time.
They're
like
internal
to
the
mechanics
of
measuring
adapters,
there's
the
way
that
mesh
reserver
interfaces
with
measure
adapters
is
using
grpc
and
there's
a
protobuf
that
defines
that
interaction.
Part
of
that
protobuf
defines
the
fact
that
an
adapter
can
bring
any
number
of
operations
to
the
table.
D
Some
of
those
operations
are
like
sample
apps
the
ability
to
deploy
them
operations
like
deploy
the
mesh
delete,
the
mesh
etc,
and
that's
a
like
it
pained
me
no
end
a
year
and
a
half
to
two
years
ago.
Now,
when
that
was
first
created,
we
created
categories
of
operations,
and
that
was
about
as
much
force
thought
that
we
could
give
to
what
that
that
design.
D
At
the
time
we
just
couldn't
expend
more
energy
to
bring
something
a
way
of
articulating
and
capturing
those
variety
of
operations
that
you're
going
to
want
an
adapter
to
perform
on
a
mesh
or
on
an
application.
D
We
didn't
have
the
there,
wasn't
anything
else
elegant
enough
at
the
time,
as
a
matter
of
fact,
at
the
time,
part
of
the
solution
that
we're
using
around
service
mesh
patterns.
Well,
some
of
that
underlying
technology
wasn't
necessarily
available,
and
so,
as
we
shift
to
include
an
alternative
style
of
operations,
we
have.
D
I
think
that
we
should
use
that
port
man
to
mesh
ops
in
a
different
way.
Externally.
I
think
that's
a
good,
but,
and
now
there
becomes
these
these
pattern.
Ops
are
these
operations
that
are
performed
defined
and
registered
and
invoked
in
a
much
more
flexible
style
of
system
through
service
mesh
patterns,
and
eventually,
I
hope
you
know
eventually
like
the
thought
is
that
those
will
probably
displace
these
older
mesh
older
operations
and
the
more
static
kind
of
grpc
protobuf
defined.
D
Method,
but
in
the
meantime,
both
of
those
style
operations
can
be
invoked
on
an
adapter
and
adapters
capable
of
taking
care
of
both
of
those
right.
Now,
it's
really
the
istio
adapter
that
we've
been
focused
on
building
that
capability
into
the
patterns
operations
capability,
and
so,
if
you're
into
golang-
and
you
want
to
make
an
impact
on
adapters
like
go,
come
talk
to
go,
go
say
so
in
the
measuring
channel
because
that's
needed
so
gurav.
I
took
a
lot
of
your
time.
What
I
was
trying
to
say
is
that
we,
you
know.
D
D
But
we
should
much
more
in
a
first-class
way,
allow
focus
on
allowing
people
to
bring
their
own
apps
and
have
life
cycle
management
over
those
apps
auto
discover
those
apps
interface
with
source
code
management
for
those
apps,
those
apps
being
primarily
like
kubernetes,
manifests
and,
and
so
graft
you
want
to
say,
take
it
away.
I
So
currently
for
applications,
we
will
be
having
more
samples
and
applications
that
will
be
added
to
the
application
ui,
with
reference
to
the
pattern
files
and
later
on.
Our
deploy
button
deploy
part
for
the
application
will
also
be
here,
and
you
can
set
your
apply
applications
over.
It
also.
I
A
Yes,
thank
you.
We
are
still
missing
that,
like
we
wish,
we
we
need
to
add
the
apply
button
in
applications
right.
A
Thank
you,
goro.
We
are
really
looking
forward
to
having
the
ability
to
deploy
applications.
A
A
But
if,
if
you
are
busy
elsewhere,
then
feel
free
to
jump
off
the
call-
and
we
will
see
you
next
in
the
next
community
call
hopefully-
and
if
you
are
a
newcomer,
we
have
a
call
tomorrow,
the
newcomers
call
so
you
can
join,
join
the
call
and
it
should
help
you
get
oriented
with
layer,
5
projects
with
that
bye,
everyone
and
hello,
everyone
who
are
still
on
the
call
push.
Would
you
like
to
show
your
recent
peer.
A
Okay
I'll
share
like
what
have
I
done
yeah
and
if
you
could,
could
you
give
some
context
to
what
the
health
checks
are
and
yeah
sure,
okay,
sure
so?
Health
checks
are
something
to
verify
the
health
of
some
components
related
to
measuring
just
for
an
example.
A
There
are
we
want
to
verify
if
docker
is
running
or
not.
If
kubernetes
is
running
or
not,
the
deployed
adapters
are
running
correctly
or
not.
Okay.
These
are
some
just
examples
and
we
have
various
functions
to
work
with
them,
and
this
topic
is
kind
of
related
to
reworking
of
all
these
health
checks
to
be
triggered
more
like.
We
previously
had
various
functions
for
docker.
We
had
a
function
for
kubernetes,
we
had
a
function
and
therefore
adapters.
A
They
have
a
separate
function
and
there
was
there
were
multiple
functions
and
if
some
developer
in
future
wants
to
interact
with
these
functions,
he
or
she
might
like
have
troubles
with
working
with
them.
A
Okay,
so
on
sunday,
on
the
last
sunday,
we
had
a
call
where
we
decided
to
rework
these
all
checks
to
work
in
a
simpler
way,
and
that
simpler
way
is
what
I'm
going
to
discuss
with
you
guys,
and
I
want
feedback
on
how
this
is
okay,
okay,
so
this
is
something
that's
called
health
check
options
where
we
have
some
kind
of
what
should
I
say,
some
options
to
be
marked
before
triggering
these
health
checks
so,
like,
let's
say,
print
logs,
if
we
want
to
print
logs
whenever
we
are
triggering
these
checks,
is
this
a
pre-running?
A
What
is
pre-running?
Let
me
show
you
so
if
you
have
worked
with
kogara
commands,
there
is
always
a
pre-running
with
each
of
these
commands
and
if
we
are
triggering
these
health
checks
from
the
spree
anything,
then
we
have
to
mark
this
as
true.
A
This
is
the
sub
command,
like
I
have
written
here,
the
command
under
which
the
health
checks
are
being
triggered.
If
you
want
to
run
docker
checks,
if
you
want
to
run
kubernetes
checks,
if
you
want
to
run
operator
checks
and
similarly
all
other
checks,
we
mark
them.
True,
if
you
want
to
run
them
okay,
so
this
is
a
health
checkers
truck.
What
it
has
is
the
options,
these
options
and
some
things
that
might
be
used
while
like
working
while
running
these
checks.
A
Okay,
so
this
is
a
function
to
create
a
new
health
checker
and
then
there
is
this
function
run
function
now.
This
is
something
that
is
easier
to
use
than
before.
There
are
multiple
functions
and
in
future
there
can
be
more
functions,
so
that
would
have
been
really
trickier
to
work
with.
So
I
have
created
a
simpler
run
function
and
according
to
these
options,
whatever
someone
has
marked
them
like
true
false
whatever
it
is,
these
things
will
run
like.
If
he
has
mark
docker
checks,
then
we
will
run
locker
checks.
A
A
Okay,
I
forgot
what
I
was
going
to
talk
about
sorry
yeah.
So,
let's
see
how
to
work
with
these
checks,
so
that
is
very
simple,
like
I
have
created
an
options
here
which
says
this
is
being
triggered
as
a
pre-running.
We
want
to
print
logs
and
we
are
passing
the
sub
command.
That
is
start
with
this
that
this
thing
is
being
triggered
from
start
command.
A
A
So
that's
how
things
work.
So,
let's
talk
about
these
options,
how
these
options
will
work?
So
if
someone
has
marked
print
logs
are
true,
whenever
someone
will
trigger
these
these
checks,
these
things
will
be
printed
like
docker
docker
is
not
running.
I
failed
to
start
docker.
Docker
is
running
things
like
that.
I
can
initialize
given
its
client,
so
we
are
checking.
A
So
if
someone
is
marked
logs
false
in
that
case,
we
actually
return
errors,
actual
errors,
rather
than
printing
warnings,
that
this
thing
is
not
initialized.
We
are
not
finding
docker
here,
something
like
that.
We
won't
be
printing
it.
We
will
be
returning
it
because
developer
is
actually
trying
to
grab
the
errors.
Out
of
it.
Example
can
be
here,
we
have
implemented
it
where
it
should
be.
A
It
should
be
down
here
yeah,
so
I
marked
like
prince
print
logs
false.
This
is
not
a
pre-journey.
The
sub
command
is
nothing
here
and
I
just
want
to
run
kubernetes
checks,
because
I
am
checking
if
kubernetes
is
available
or
not.
So
I
created
a
health
checker,
I
ran
it
and
I
found
that
there
are
no
errors.
A
That
means
kubernetes
is
actually
running
on
this,
this
local
machine.
Then
we
will
deploy
whatever
we
want
to
do
with
the
process.
Okay,
so
this
is
the
demo
how
it
used
for
and
yeah
that's
it.
This
is
how
it
is
simpler
to
interact
with
these
checks.
There
can
be
multiple
checks
in
future,
so
it
will
be
easier
to
talk
with
them
like
trigger
them.
I
mean,
with
these
kind
of
options,
just
marking
the
options
and
running
it
yeah.
That's
it
that's.
What
was
new
with
these.
A
A
C
J
Yeah
vj
go
ahead,
yeah,
sorry!
So
where
are
you
putting
these
checks?
Where
are
you
saying
that
I
want
this
check?
I
want
that
check.
J
A
Okay,
I
am
not
actually
on
this,
like.
I
am
not
able
to
understand
what
you
are
trying
to
say.
Maybe
I
can
answer
this
vijay.
Basically,
these
health
checks
would
so
measure
each
of
the
measure.
Residual
command
can
invoke.
A
J
I
understand
what
my
my
question
is:
how
are
you
going
to
say,
for
instance,
that
I
want
to
run
docker
check?
Do
you
have
to
go
there
and
mark
this
true
or
how
does
that
that
work,
or
are
you
going
to
have
like
a
file
that
it
reads
at
run
time
to
that
tells
it
which
checks
to
to
run
of
what
checks
are
available
and
which
checks
to
to
run?
Is
this
what
I
was
curious
about?
Okay,.
A
So
in
case,
we
want
to
run
docker
health
checks,
we
will
not
run
through
and
it
will
trigger
this.
This
command
will
keep
like
make
it
we'll
go
through
these
options
and
we'll
run
docker
checks
and
return
errors.
If
we
are
if
we
are
taking
errors
or
print
logs,
if
we
want
to
print
logs.
J
About
documents,
no,
that's
fine,
see
what
what
I
was
trying
to
say
is.
You
know
if
you
can
create
a
file
which
has
all
the
possible
checks
in
that
and
if
you
can
say
whether
it
is
true
or
false
in
that
file,
so
you
won't
have
to
go
into
the
source
code
and
change
the
source
code.
Each
time
you
want
to
run
this
yes
and
plus,
you
know,
then
every
every
application
which
wants
to
do
its
checks
can
do
its
own
checks.
You
know
one
application
can
check
off
docker.
J
A
I
get
it
okay,
so
actually,
this
is
a
command
like
there
is
a
command
for
doing
all
these
checks
with
the
messages.
Okay.
So
what
is
happening
there?
There
are
some
flags
for
this
like
there
is
a
pre-flight
flag,
okay
and
then
we
can
trigger
adapter
checks
separately.
We
can
trigger
operator
checks
separately
and
we
can
trigger
all
these
checks
combined.
G
A
Here
we
have
actually
checked
docker
and
cubenets,
because
these
are
the
three
requirements
of
running
measuring.
Okay.
This
is
a
free
flight
check,
so
we
ran
the
flight
check,
which
is
checking
docker.
The
docker
is
running.
Docker
compose
is
available,
given
its
api,
we
can
create
kubernetes
client
and
we
are
also
taking
the
cuban
its
version
if
it
is
compatible
or
not,
and
there
are
multiple
flags
like
that
and
separate
checks.
Checks
can
be
triggered
like
that.
H
K
I
think
it
would
make
sense
to
have
one
of
these
options
as
an
input
to
the
system
check
commands
right,
so
I
want
to
run
only
one
check
so
right
now
we
have
p
and
p
flight
right,
so
I
think
what
dj
might
be
hinting
at
is.
I
want
to
do
only
operator
check.
How
do
I
do
that.
A
D
Yeah,
I
think
I
think
it
was
just
I
I
know
you're
tr,
so
there's
a
design
spec
on
this,
and
there
is
I'm
an
open
pr
and
push
you've,
given
consideration
to
the
fact
that
these
health
checks
need
to
be
invoked
with
high
regularity
that
and
they
need
to
be
invoked
from
different
perspectives,
some
directly
from
the
user,
some
as
a
pre-run
execution
before
another
command
is
executed
and
like
pre-run
as
part
of
cobra.
D
The
framework
that
we're
using
to
create
this
cli
you've
also
given
consideration
that
these
need
to
be
executed,
or
they
may
be
invoked
ad
hoc
by
other
commands,
as
and
when
those
other
commands
might
need
to
do
that
within
their
own
logic,
and
so
so
congrats.
Your
code
is
going
to
get
a
lot
of
use.
D
I
wonder
it
is
probably
just
fine,
like
the
logic
there
I'm
assuming
is
that
for
the
given
context,
check,
availability,
check
the
you
know,
the
availability
of
every
adapter,
that's
defined
in
the
context,
and
so
so
we
don't
need
to
further
allow
the
user
to
further
specify
just
this
adapter
and
we
could,
but
that's
probably
taking
the
use
case
a
little
too
far
like
if
they
had
10
adapters
loaded.
Maybe
they
just
want
to
check
one
okay.
D
I
do
think
that
what
becomes
it
becomes
really
important
this.
This
really
highlights
the
use
of
error
codes,
because
people
are
going
to
see
those
stinking
error
codes
like
all
the
time
can't
can't
connect
to
kubernetes
and
like
and
so
the
having
well-written
probable
causes
and
suggested
remediations
like
well-written
descriptors
of
what's
potentially
going
on
and
how
to
fix
it.
I'm
quite
quite
important
and
there's
more
than
one
there's
potentially
multiple
of
those
like
more
than
one
probable
cause
for.
D
D
We
I
think
we
have.
We
have
written
down
this
list
of
different
checks
that
are
performed,
and
so
it,
while
it's
kind
of
a
pain
in
the
rump,
it
is
advantageous
to
have
a
documented
list
of
what
those
checks
are
and
what
they
do
to
the
extent
that
it's
anything
more
than
just
like
a
ping,
so
to
speak
in
part,
because
people
are
going
to
want
to
troubleshoot
those
environments
and
understand
how
to
fix
something,
but
also
in
part,
because
that
will
help
us
as
a
set
of
contributors
and
end
users.
D
Like
reason
over
whether
or
not
a
given
check
should
be
mandatory,
invoked
like
every
single
time,
any
command
is
invoked
or
whether
a
check
should
be
invoked
only
when
a
mesh
res
only
when
a
certain
class
of
commands
are
invoked
so
like
measuring
ctl
system.
That's
like
system
is
a
category
of
sub-command.
There's
a
category
of
some
commands
under
that,
and
so
maybe
that,
like
I
guess
that's
a
question
is,
is
it
does
cobra
facilitate
for
hierarchical
implications
of
pre-running?
D
Like
can
you
say
at
the
system
level,
irrespective
of
what
the
sub
command
that's
invoked
is
yeah,
okay,
yeah
good,
it
does
all
right.
I
don't
have
anything
more
intelligible.
I
guess
that's
all
I
can
add,
but
but
like
I
do
want
to
highlight,
like
I'm,
I
mean
like
I'm
excited
to
for
my
partners,
like
I'm
excited
to
see
the
people
that
are
here
and-
and
I
think
sometimes
folks
might
join
and
say:
well,
hey,
I'm
really
in
to
operate
like
I'm
a
user.
D
I'm
really
just
here
to
like
learn
this
stuff
or
like
I'm,
into
ops
and
like
devops,
and
sometimes
I
think
that
people
say
that
sheepishly,
like
hey,
I'm
not
here
to
necessarily
write
code.
All
right!
That's
not
my
thing!
It's
like
great!
I,
like
hey,
yeah,
don't
don't
like!
Thank
you
because,
like
p,
people
need
to
come
way
in
on
this.
Like
use
these
things
and
say
some
of
what
samir
was
just
saying
like
it
would
be
helpful
if
there
was
an
auto
scale.
If
we
were
tracking,
auto
scaling.
D
Events
like
oh
yeah,
okay,
right
or
like
hey,
would
be
helpful
if
there
was
a
different
flag
that
only
let
you
I
know
that
actually
it
didn't
just
check
on
the
operator
it
like
actually
checked
on
mesh
sync
itself,
like
is
mesh
reservoir
receiving
data
from
mesh
sync.
That's
what
I
want
to
know.
I
don't
want
to
know
if
this
thinking
operator
is
like
says
it's
up.
I
want
to
know
if
data
is
actually
flowing
through
this
pipe,
nothing
and
so.
C
D
A
We
are
planning
to
propose
all
the
constants
and
all
the
config
related
stuff
into
one
single
struct.
So
we
recently
discussed
on
this
on
a
call.
So
to
so
this
I
have.
I
have
written
down
some
considerations
for
this
in
the
mystery,
cli
comments
and
documentation
document,
and
let
me
just
share
it
on
the
chat.
A
So
if
you
go
to
the
go
to
if
you,
if
you
go
to
the
link,
you
will
find
find
this
and
basically
what
we
are
trying
to
do
is
to
to
more
to
move
to
all
the
all
the
configurated
stuff
and
all
the
global
constants
into
a
central
struct,
so
that
we
can
easily
reference
this
track
to
and
as
from
this
truck,
and
this
struct
will
stay
as
the
central
source
of
truth
and
also
we.
A
So
before
we
do,
that,
like
there
might
be
better
ways
to
implement
this,
so
we
will
let
we
will
keep
this
open
for
discussions
for
for
a
while
and
then
we
will
start
working
on
it.
So,
yes,
that's
my
item
so
check
the
those
out.
I
will
also
bring
in
the
misery
ctl
channel
regarding
this
later
and
with
that.
We
are
stopping
for
today
and
thank
you
all
for
staying
long
after
the
call
finished
and
we'll
see
you
next
time.