►
From YouTube: Meshery Development Meeting (May 5th, 2021)
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
A
And
if
you
haven't
already
put
down
your
names
on
the
meeting
minutes
and
add
any
agendas
that
you
might
have
to
the
topics,
so
let's
not
waste
any
time
and
get
started
with
aditi.
B
B
Naveen,
can
you
scroll
down
to
the
implement
okay,
okay,
I'll
share?
Okay,.
B
So
the
problem,
the
issue
I'm
trying
to
fix
is
the
earlier
efforts
to
in
the
direction
of
the
operator
life
cycle
just
involved,
using
the
fixed,
fixed,
hard-coded
urls
from
the
master
branch
and
since
like
since,
like
the
other
measuring
manifests,
these
need
to
be
versioned
and
measuring
operator,
and
the
other
manifests
are
in
a
separate
repository
with
different
versions.
We
we
have
to
have.
We
have
to
have
a
separate
versioning
for
them
in
the
config.yaml.
B
So
my
my
approach
is
this
and
I
wanted
to
seek
the
community's
feedback
so
first
version
there
will
be
one
version
key
with
the
two
versions
measuring
measuring
operator.
B
So
the
first
thing
I
did
was
change
change
the
config
file
to
accept
two
versions,
and-
and
this
is
my
comment
for
the
same
so
as
you
can
see
from
a
string
and
string-
it's
now
a
map
and
string,
and
I
have
just
changed
certain
functions
to
to
to
take
it
as
measuring
instead
of
just
click
conversion.
B
The
second
thing
I
did
was
extracting
this
and
using
it
to
create
get
operator
urls.
Now
now,
as
I
said,
we
need
to
download
these
based
on
operator
version
and
this
removed
the
hard
hard
coded
urls,
and
this
this
was
like
the
most
key
commit
of
the
ones
I've
made
till
now.
B
Yes,
see
fetching
manifest,
stick
or
takes
operator
version
and.
B
And
this
song
this
was
initially
a
hardcore
video
changed
and
this
this
retrieves
of
urls,
based
on
version
now
since
since
naveendo
had
brought
up
a
good
point
that
for
now
our
measuring
has
its
own
channel
channel
key.
Also
and
the
operator
will
need
one
as
well.
For
now
that
will
be
phase
two
where
similar
to
operator
I'll,
have
a
channel
for
measuring
and
measuring
operator
separately.
B
C
C
You're,
occasionally
going
in
and
out
which
might
make
it
such
that
I'm
going
in
and
out
a
little
bit.
Are
you
able
to
hear
me.
C
Cool,
I
so
aditi,
I'm
not
sure
if
you'll
be
able
to
hear
all
of
what
I'm
about
to
say.
I
hope
so.
The
reason
that
I
interrupted-
and
I'm
sorry
for
doing
so
is
because
you
have
the
best
of
intentions
and
a
good
idea,
but
I,
but
seeing
that
you're
implementing
according
to
the
design
we
we
need
to
review
the
design
or-
and
that's
not
because
you
shouldn't
go
off
and
do
excellent
things.
C
But
that's
just
because
I
don't
know
that
we'll
have
a
separate
release
channel
for
the
operator,
the
operator
it
should
follow
along.
It's
a
component
of
the
meshery
server
like
the
adapters.
It
should
be
treated
in
the
exact
same
fashion
as
the
adapters.
The
adapters
follow
a
stable
and
edge
release
channel.
Just
like
mesh
reserver
does
you
should
be
able
to
pin
to
a
specific
version
of
measuring
operator
if
you
so
choose.
C
If
there
was
a
use
case,
or
rather
if
your
deployment
calls
for
that
for
some
reason
that
would
probably
normally
be
because
there's
a
bug
in
like
the
latest
stable
edition
or
something
like
that
or
because
you
know
you
weren't
comfortable
moving
forward
with
the
latest
release.
You
wanted
you,
as
a
user,
wanted
to
pin
to
an
older
version,
but
the
concept
of
having
a
separate
another
release
channel
is,
if
I
was
interpreting
this
correctly,
that
we
should
chat
through
that
more
or
we're
it.
C
Actually,
maybe
I'm
reflecting-
and
this
is
just
me-
is
it
so
abhishek
and
niven.
You
know
everyone
else
on
the
call
please
chime
in
adida,
maybe
I'm
putting
more
words
in
your
mouth
than
what
you
were
saying.
It
might
be
that
what
you
were
saying
is,
you
might
have
been
saying
the
same
things
like
hey
the
operator
is
a
component.
It
should
be
versioned.
D
C
Should
do
some
diligence
to
version
it?
We
should
enable
the
user
to
pin
to
a
certain
version
if
they
want
to.
Yes,
the
operator
should
als.
So
just
like
the
adapters
and
just
like
mesh
reservoir,
the
as
we
go
to
version
the
operator,
it
should
have
a
stable
release
and
a
edge
release,
and-
and
I
and
maybe
I
was
misinterpreting
what
you
were
saying,
because,
though,
that
stable
release
and
that
edge
release,
that's
one
in
the
same
as
mesh
reserver's,
stable
release
and
edge
release
like
while
it's
that's
a
confusing
statement.
C
What
I
just
said,
because
technically
the
operator
just
like
the
adapters
can
be
released
at
any
given
time
whenever
they're
ready,
because
they
they
are
their
own
component,
they
just
need
to
provide
guarantee.
You
know
some
some
level
of
compatibility
with
the
other
latest
editions
of
the
other
components
you
know
like,
and
so
I
guess,
as
a
manner
of
phrasing
like
it's
not
inaccurate
to
phrase
it
just
like
you
are,
which
is
each
of
those
components
have
their
own
stable
release,
channel
and
edge
release
channel.
That's
one
way
of
phrasing.
C
It
like,
I
think,
that's
probably
accurate.
I
think
it's
probably
also
accurate
to
say
the
the
project:
hey
as
a
user.
Are
you
running
mescheri?
Yes,
you
are
oh,
okay,
are
you
is
your
measuring
on?
Are
you
subscribed
to
the
edge
channel
or
to
the
stable
channel,
and
when
someone
asks
that,
and
when
someone
characterizes
that
there
would
only
be
one
response,
it
wouldn't
be
like?
Well,
you
know
this
adapter,
it's
on
the
stable
channel
and
the
server
is
on
this.
B
Okay,
okay,
so
so,
instead
of
having
a
separate
operator
channel
I'll
use
the
measuring
channel,
for
the
same
intents
and
purposes,
that's
it
right!
That's
what
okay!
Okay.
C
Yeah
I
wish
I
wish
I
had
biceps,
because
then
I
like
do
the
like
a
real,
a
really
emoji.
You
know
yeah.
You
know
I
have
to
go
work
on
my
biceps.
Just
for
that
yeah
perfect
cool.
I
think.
That's,
that's!
Probably
a
it's
just
a
slight
tweak.
It's
like
85
percent
in
alignment
already
and
just.
E
A
A
Okay
cool,
so
what
I
was
saying
is
that,
let's
say,
for
example,
measuring
5.2
is
supports
machinery,
operators
version
one
and
five
point
three
supports
measure
operators
version
two
and
basically,
if
someone
goes
and
modifies
the
config.yaml
so
that
they
wanted
to
run
5.2
with
machine
operators,
version
2,
then
it
shouldn't
get
supported.
So
is
that
like
we
will
need
a
validation
layer
over
this
right.
B
C
And
that's:
okay:
I
don't
think
that,
like
as
a
project
we'll
we'll
have
to
first
establish
a
framework
wherein
we're
we're.
Basically,
we
have
a
compatibility
matrix,
there's
an
open
issue
to
create
one
and
as
a
project,
maturity
level,
we're
not
to
a
point
wherein
a
we're
defining
that
you
know
and
can
keeping
that
in
hard
terms
and
b.
C
I
don't
know
that
we
ever
necessarily
will
I'm
not
sure.
I
think
we
what
we
would
do
is
we
would
make
specific
rules,
conventions
about
how
releases
are
done
and
how
versions
are
tracked
and
we.
It
would
be
something
like
something
like
this.
So
so
most
of
you
are
all
familiar
with
the
notion
that
we're
using
semantic
versioning
and
there's
a
little
bit
of
semantics
to
this
style,
this
convention
of
versioning.
But
it
goes
something
like
you
can.
C
You
can
use
different
labels
and
things,
but
it's
base
it's
basically
x
dot
x,
dot
x,
maybe
you've
got
other
things
that
you
put
behind
it
and
things,
but,
but
basically
it
goes.
You
know.
C
C
Version
numbers
that
there's
no
compatibility
guarantee
across
all
the
components
like
this
would
be
something
that
we
would
work
towards,
but
within
the
miner
and
patch
that,
if
your
server
version
it's
to
to
give
an
example,
if
you're
running
a
component,
that's.
C
1.1.5
on
that
for
the
server
and
then
the
operator
is
1.1.1
that
what
we
might
get
to
is
like.
Oh
that's,
fine,
like
so
long
is
your
you
know.
Compatibility
is
guaranteed
from
here
on
down
that
type
of
a
thing:
we're
not
there
yet
and
anyway.
So
the
point
is,
I
think
that
that's
a
bad
ask
right
now
to
to
try
to
like
the
framework
and
the
the
ci
systems
and
stuff
that
that's
not
we're
not
to
a
point
by
which
there
is
a
list
to
refer
to
and
then
to
keep
track
of.
C
C
But
is
not
a
concrete
thing
rather,
I
think
the
only
concrete
statement
is
today
there
isn't
there?
Isn't
there
isn't
a
list
to
go
to
to
say
this
works
with
this,
and
so
today,
it's
mostly
mostly
what
we're
shooting
for
with
users
and
with
all
of
ourselves
is
that
if
you're
subscribed
to
a
given
channel
of
which
there
are
two
and
channel
I
refresh
my
memory.
C
Okay,
basically,
that
if
you're,
if
you're
running
a
configuration
of
the
and
by
the
way,
we're
really
talking
about
mesh
config,
the
config.yaml,
which
is
where
you
control
the
configuration
of
your
mastery
deployments
that
you
know
if
you
like
these
are
going
to
be
the
two
most
common
configurations.
You
might
have
a
measuring
deployment,
that's
running
latest,
stable
or
latest
edge,
so
you're
living
vicariously
here
or
not
back
here,
you're
living
dangerously!
C
Here,
sorry-
and
you
know,
you're
being
you
know
safe
here-
and
that
yeah
there
may
be
okay,
so
there's
the
server
there's
individual
components
released
individually
and
some
of
them
are
released
together
today,
so
the
ui
is
released
together
with
server
there,
it's
possible.
You
can
split
them
out,
but
they're
just
released
together.
Today,
measuring
ctl
release
separately
adapters
separately
operators
separately.
C
There
is
you
can
if
you
are
an
advanced
user,
and
you
have
a
particular
use
case
to
do
so
that
within
the
config
you'll
also
see
things
like
what
what
adapters
you're
running
and
by
default.
You'll
just
see
them
listed.
C
You
would
go
ahead
and
whoops
as
a
user,
sorry
place
a
tag
behind
it
and
then
that's
that's
how
you
would
pin
your
pinning
to
actually
wouldn't
be
just
like
that
it
would
be
you
would
you
would
pin
to
that
version.
So
the
same
needs
to
well
or
I
would
suggest
a
dd.
If
you
will
consider
this
please,
I
would
suggest
that
we're
looking
at.
C
C
Oh
sorry,
I
keep
hitting
the
wrong
key.
You
grab
stable
latest
because
that's
what
they
have
configured
unless
they
decided
to
you
know
update
this
to
be
so
when
we
go
look
at
the
operator
by
the
way,
so
I
think
somewhere
in
this
design
spec,
it
says
that
we're
not
versioning
the
operator,
but
if
that's
true,
that
would
be
horrific.
That
would
not
be
good.
B
F
Perfect
so
I
had
another
point
to
make:
is
it
fine
effect
yeah?
So
I
don't
know
if
it
applies
to
this
exactly,
but
if
you
let's
say
have
the
logic
that
is
checking
against
the
you
know,
compatible
versions.
So
if
you
say
something
like
measuring
v
0.2,
I
know
no
one
uses
it
now,
perhaps,
but
just
in
general,
with
in
regards
with
version
checks
in
case
there's
a
bug
found
in
future
that
makes
it
incompatible
with
another
piece
or
another
service.
F
So
how
exactly
do
you
handle
that
so,
let's
say
there's
a
list
of
known
bugs
and
you
hard
code.
The
fact
that
you
cannot
use
one
with
another,
but
in
case
in
future
someone
discovers
okay.
This
feature:
it's
glitchy
when
you
combine
these
two.
So
how
exactly
is
that
handy?
I
don't
know
if
I
made
a
sense.
C
In
part,
in
part,
that's
what
a
dd
is.
That's
one
of
the
use
cases
that
adida
is
addressing
for
the
operator
and
it's
to
say
something
like
well.
What's
the
latest
release
of
the
operator
it
is
v.
212.,
oh,
turns
out,
there's
a
there's,
a
defect,
that's
not
compatible
with
kind
of
something
something,
and
so
in
my
environment
I
need
to
bring
it
back,
bring
it
back
a
release,
and
so
what
adidi
is
making
true
here
is
that
if
a
user
wants
to
pin
to
a
specific
version
that
they
would
apply.
F
No,
no,
I
meant.
Is
there
any
automated
checks
that.
C
Nope,
yes,
yes
and
no
yeah,
and
that
was
the
first
part
of
what
I
was
saying
is
that
there
is
you
know
when
the
project
reads:
reaches
1.0
most
projects,
don't
even
have
what
I'm
about
to
describe
by
the
way.
So
meshri
will
have
a
massive
compatibility
list.
Mastery
interacts
with
10
different
service
meshes.
Those
service
meshes
interact
with
any
number
of
different
kubernetes
releases.
There
are
90
plus
certified
kubernetes
distributions.
C
There
are
any
number
of
prometheus
versions,
grafana
versions,
but
there's
just
there's
an
astronomically
large
permutation
of
compatibility
to
be
tracked
across
all
of
these
things.
We
won't
track
those
broadly.
We
will
track
high
level
and
we
do
what
service
mesh
versions
are
compatible
with
each
of
the
adapters.
So
does
this
adapter
support
what
version
of
link
or
d
we
track
that
in
a
manual
way
and
as
the
project
goes
up,
there
will
be
more
rigor
around
tracking
that
and
having.
C
There
is
a
suite
of
automated
tests
that
occur
for
not
all
of
the
adapters,
but
a
few
of
the
adapters
every
time
before,
release
of
them
there's
a
suite
of
tests
that
basically
end
to
end
functional
tests.
Where
the
adapter
will
will
spin
it
up
in
a
temporary
kubernetes
thing,
we'll
do
functional
tests
and
then
that's
it
like,
and
then
you
we
might
manually
update
the
compatibility
matrix.
C
A
thing
yet
right,
right,
right
and
and
most
projects
most
projects
don't
ever
get
to
in
most
projects
a
lot
of
a
lot
of
commercial
enterprise.
Software
only
do
that
level
of
compatibility
testing
to
only
so
well,
I'm
not
saying
that
we
shouldn't
I'm
just
kind
of
giving
context.
B
I
have
a
question
so
in
the
example
mesh
config
you,
given
what
advantage
does
specifying
it
this
way,
have
it
compared
to
specifying
it
as
measuring
and
measuring
operator.
C
Yep,
if
you
would
maybe
layout
layout,
go
ahead
and
change
it
to
be
what
a
different
proposal.
B
I
I
meant
so
currently
what
I
have
implemented
is
there
is
measuring
latest
I
mean
this
is
under
version.
B
C
Yeah
you
yeah,
that's
right,
you
could
they
both
achieved
the
same
thing
and
we
already
it
there's.
There
becomes
no
difference.
B
Okay,
so
my
current
implementation-
I
I
don't
know
if
I
was
clear
enough-
I
was
describing
it
or
it
does
this.
So
do
you
want
me
to
redo
it
to
make
it
like
this.
C
My
shift
key
okay.
Well,
that's
what
happens
when
you
eat
over
your
keyboard?
I
guess
it's
disgusting,
okay,
so
the
reason
to
stick
with
the
same,
like
there's:
no,
if
we
were
starting
from
scratch,
you
you
might
do
it
the
way
that
you're
doing
it
since
we're
not
starting
from
scratch
and
there's
convention,
you
have
to
overcome
the
inertia
and
prove
that
your
design
is
better
and
that's
not
specific
to
this.
That's
just
a
general
statement
and
so
indulges,
and
this
is
not
this
is.
B
C
Yep,
okay,
so
that
goes
back
yeah.
We
should
have
designed
spec
reviews,
so
I'm
not
sure
who
was,
if
hopefully
abhishek
in
the
venue
I
suspected
the
two
that
you
might
have
been
speaking
to
and
I've
been
the
the
direct.
So
we
should
absolutely
treat
it
like
an
adapter,
no
difference
same
thing:
okay,
the
only
difference
is,
it
has
a
crd
in
addition
to
a
deployment
like
there's,
there's
just
no
there's
no
yeah,
I
mean
I'll
try
to
help
you
think
through
this
and
all
of
us,
and
that
is
well
gosh.
C
What's
special
about
it
or
what's
different,
why
would
it
be
treated
different?
So
if
you
look,
I
brought
in
just
I
just
copied
my
config
file
like
an
example
of
my
config
file,
to
get
to
get
a
more
full
picture.
So
so
here's
here's
a
copy
of
my
config.
It
starts
off
with
just
high
level.
It
has
like
two.
C
I
think,
there's
two
constructs
right
now:
contexts
and
tokens
and
by
the
way
this
is
an
area
of
need.
This
is
like
an
underformed
area
that
isn't
used
well
in
measuring
ctl
today,
but
so
there's
two
things.
Well,
there's
three:
I
guess
it's
kind
of
top
level,
there's
the
current
context,
tokens
and
then
the
full
description
of
the
contexts,
and
so
the
reason
I'm
it's
not
that
you
don't
know
this
at
dd,
but
the
reason
I'm
sort
of
covering,
that
is
to
say,
okay.
Well,
let's,
let's
analyze,
load
freshly
into
our
minds.
C
C
Well,
you
know
these
aren't
going
to
be
the
only
custom
controllers
or
custom
resources
or
whatever
term
we
want
to
use
to
describe
them,
and
so,
while
we
might,
while
they
might
get
so
to
your
point
about
like
categorizing
them
here
is
like
is,
I
think,
well
aligned
with
with
my
point
of
we
should
treat
them
as
much
like
adapters.
I
didn't
say
it
like
this,
but
I'll
rephrase
it.
We
should
treat
them
as
much
like
adapters
as
possible,
which
would
be
like
a
sort
of
okay.
C
Don't
don't
mislabel
it
like
go
ahead
and
give
it
a
it's
the
appropriate
label
and
then
but
but
the
reason
that
I'm
putting
it
underneath
here
and
here
is
because
that's
doing
two
conventions
at
once,
and
it's
like
one's
per
perpendicular
to
the
next.
It's
it's
not
it's!
C
You
could
you,
even
if
you
did
what
you're
just
suggesting
which
you'd
prob
you
might
go
like.
Okay!
Well,
hey
that!
Actually
you
know
that,
for
whatever
reason,
that's
that's
just
nicer!
So
let's
just
put
it
under:
let's
just
move
them
all
together
under
it
would
kind
of
still
look
the
same.
It
would
be
like
hey.
What
are
these
all
subscribed
to?
C
Well,
they're
all
subscribed
to
edge
because
they're
under
edge
and
so
and-
and
there
might
be
use
cases
that
you
can
think
of
that
where,
if
we
wanted
to
get
super
sophisticated
where
okay
well,
how
sophisticated
we
let
users
get
today,
it's
like
well
they're,
never
going
to
run
istio,
so
they
can
delete
that.
Okay,
that's
the
sophistication!
That's
the
customization!
That's
nice!
C
Also,
there's
that
you
know,
for
whatever
reason
they
want
to
perpetually
hang
out
on
edge
latest
here,
while
they
want
to
hang
back,
you
know
on
some
they
want,
so
we
let
them
customize
and
manipulate
in
terms
of
now.
An
interesting
thing
is
they
can
cross
pollinate
their
release
channels.
I
believe
I
don't
know
that
we've
ever
tested
this,
so
this
is
kind
of
an
interesting
thing.
It's
like
well,
what's
what
you
know
bought
this?
This
deployment
called
bob's
kubernetes.
It
subscribed
to
the
edge
channel,
but
maybe
they
wanted
to
use.
C
And
I
think
that
that's
feasible,
we
should
see
if
that
gets
honored.
I
am
willing
to
bet
you
50
us
dollars
that
it
is
honored.
Oh
boy,
are
we
recording
okay?
Well
aditi?
C
Why
don't
you
go
test
it
privately
and
if
I'm
wrong,
then
you
accept
my
bet
and
then
we'll,
but
so
anyway,
we're
spending
a
long
time
on
this
I
didn't
mean
to,
but
but
I
mean
like
it's
good
questions,
I
I
feel
terrible
when
someone
asks
a
great
question
and
then
the
response
is
super
concise
and
it's
no
like
that.
That
doesn't
feel
good
and
no
one
learns
and
but
so.
B
B
So
so,
finally,
this
is
what
I've
already
implemented.
So
instead
of
this
I'll
do
what
you
said
this
con
along,
create
a
separate
category
controllers
and
in
that
I'll
put
measure
operator.
A
Yeah
like
but
like,
I
think
it
would
be
best
if
we
like
iterate
over
this
spec
for
a
while,
and
make
sure
that
before
we
start
implementing,
let's
make
sure
that
we
have
the
right
spec
and
then
we
can
do
it
right.
E
A
A
Okay,
okay,
so
moving
on,
we
have
samir
who
has
been
testing
out
a
measuring
to
be
deployed
to
use
to
be
used
in
a
aks
deployment.
D
Yes,
so
good
morning,
if
some
most
of
them
don't
know
me,
I'm
sami.
So
I
have
been
introduced
to
layer
5
recently
and
I'm
part
of
the
community
recently,
and
we
have
just
decided
to
start
few
of
the
kind
of
you
know.
Pocs
on
this,
there
were
few
problem
statements
which
led
us
in
fact
me
to
adapt
and
come
to
mercedes.
To
kind
of
you
know
to
a
poc
is
on
our
isto
okay.
So
just
like
you
know,
quick
thing
before
I
just
head
into
the
issues
which
I
faced.
D
C
D
So
fasting,
what
exactly
I
find
it's
something
which
is
not
conventional
for
from
a
kind
of
a
devops
guy
from
my
side
is.
I
can
see
there
are
a
lot
of
thing
here,
the
left
side,
pan,
if
you
see
I
just
use
istio,
but
I
can
see
there
is
a
whole
bunch
of
services
which
have
been
installed
here
right
here.
D
Maybe
if,
if
this
list
can
populate
depending
upon
the
service
mesh
which
I
am
using,
that
would
be
very
handy
for
me
instead
because,
let's
say
I'm
using
only
istio,
but
this
is
kind
of
no
use
for
me,
the
whole
bunch
of
other
things
which
which
it
has
okay.
So
what
I
am
telling
in
the
same
constraint,
like
you
know
in
the
same
context,
is
this
kind
of
checks
which
is
happening
service
mesh
here.
Okay,
if
this
can
this
checks
can
be
done
in
the
provider
page.
D
D
D
A
A
C
Yeah
totally,
this
is
so
samir
great
point
and
I
think
you
know
you're
not
the
first
to
I
think,
you're,
not
the
first
to
highlight
this,
the
as
a
matter
of
fact
it
used
to
work
differently.
It
used
to
work
exactly
like
you've.
You
just
said
which
is:
if
you
don't
have
an
adapter
deployed
in
your
environment,
then
you
don't
see
a
link
to
to
that
service
mesh
because
you're
not
dealing
with
that
service
mesh.
So
why
are
we
overloading
the
user
with
a
bunch
of
links
and
earlier
in
the
project?
C
It
was
because
we
wanted
to
let
everyone
know
that
there
was
lots
of
different
service
meshes
that
can
be
managed
anymore.
I
think
that
that's
well
understood
and
anymore,
the
screenshots
and
videos
and
things
that
are
done
to
help
promote
an
understanding
of
the
project
and
and
to
kind
of
market
the
project.
C
Well,
that's
a
separate
concern
from
the
real-time
user
experience
that
you,
as
a
user
who's,
choosing
to
just
to
only
interact
with
istio.
You
don't
perpetually
need
to
see
kuma
if
it's
not
going
to
be
there,
it's
just
getting
in
your
mouse's
way
and
so
that's
excellent
feedback.
We
should
execute
on
it.
There
is
already
logical,
there's
already
a
few
different
logical
conditions
within
the
menu
that
can
be
updated,
really
quick
to
to
accomplish
what
you
just
asked.
C
Moreover,
as
a
point
of
nerdy
interest,
you
can
deploy
15
istio
adapters
like
you
can
deploy
a
lot
of.
You
can
deploy
adapters
of
the
same
type
of
service
mesh,
a
lot
of
them
and
if
you
do,
the
menu
is
dynamic
and
it
will
list
out
each
adapter
independently,
but
so
so
thanks
for
that,
yeah.
D
And
another
thing
is
that
which
I
found
so
let's
say
once
I'm
going
to
profile
okay,
so
there
is
no
direct
link
to
go
to
my
overview
tab
so
which
I
discovered
to
click
on
here
so
which
is
like
you
know,
for
a
new
user
may
not
be
kind
of
a
ergonomics
way.
So
maybe
before
performance
here
can
be
some
overview
tab
and
you
can
just
like
map
it
to
this
system
dashboard
or
the
overview
somewhere.
So
yeah.
C
Yeah,
it's
a
good,
that's
fantastic!
Thank
you
for
walking
us
through
this
as
you
just
you
calling
me
so
from
my
part,
and
anyone
else
is
welcome
to
add
commentary
here
as
well.
But
for
my
part,
I'd
considered
the
same
that
like
that
yeah.
It
wasn't
necessarily
immediately
intuitive
if
you're
on
a
web
page.
C
Okay,
you
might
you
might
be
expected
to
guess
that,
but
you're
not
on
a
web
page
you're
in
you're
in
an
app
and
so
the
the
curious
if,
on
the
upper
right
hand,
corner
there's
a
few
icons
well
over
by
your
avatar
yeah
and
well,
there's
one
on
the
very
left.
C
C
No
no
see,
but
that
doesn't
mean
anything.
What
means
something
is
that
you
didn't
see
it.
Okay
and
that's
obviously
your
fault.
No,
like
that's.
Obviously
not
your
fault
is
the
thing
like
hey
yeah,
there's
a
button
there
to
do
exactly
what
you
just
said,
but
if
you
didn't
see
it
well,
that's
not
your
that's!
Not
your
fault!
I
mean
you
know
what
I
mean
like.
So
it's
something
for
you
know.
You
know
everyone
here
to
ponder
a
little
bit
like
what
are
we?
How
are
we
like
right
now:
hey?
Where
is
samir?
C
Well,
if
I
look
over
on
the
menus
on
the
left-hand
side,
I
don't
know
I
can't
tell
where
is
he
if
I
look
on
the
on
the
upper
right
hand
corner
hell,
I
don't
know
where
is
he
I
don't
know?
I
have
no
idea
other
than
like.
I
can
look
it
in
the
center
and
I
can
look
at
the
title
of
the
page,
but
so
why
aren't
the?
Why
don't
we
take
hyperlinks
and
make
the
active
hyperlink
light
blue.
D
C
D
D
D
Okay,
so
I
have
already
installed
my
istio
after
that
I
have
started
this
like
I
have
installed
this
particular
necessary
on
on
top
of
that,
but
it
didn't
like
you
know,
apparently
identify
that
I
had
isd
on
prime,
so
I
have
passed
on
my
comments
onto
that
and
the
logs
also.
So,
if,
like
you
know,
you
can
say
something
what
I'm
missing
right
now,
because
what
I
could
see
is
it
is
now
using
my
cluster
config
so
which
ideally
should
feed
this
system
that
I
have
still
already
installed.
A
D
Okay,
so
can
I
do
that
myself
so
in
the
hell
chart
or
somewhere
yeah.
A
It
has
to
be
changed
in
the
head
chart.
Actually,
the
template
of
the
new
chart.
D
A
A
The
crds
and
like
basically,
you
need
to
update
permissions
and
the
rules
and
that's
it.
D
Yeah
I'll
try
everything
so
maybe
I'll
try
the
airbag,
let's
see
if
it
works.
So
what
will
exactly
the
access
it
would
need?
So
I
can
help
you
with
that.
Okay
and
the
next
thing
is
yeah,
so
I
am
now
happy
that
my
mystery
has
been
installed
and
I
have
the
dashboard
and
everything
so
now
I
want
to
like
interact
from
my
command
line
so
to
interact
with
my,
like.
You
know,
message
system
from
the
command
line.
Is
it
necessary
to
have
the
docker
installed
in
my
system.
A
D
Okay,
so
those
kind
of
configuration
actually
is
it
documented
somewhere?
Actually,
I
I
didn't
find
it.
C
Yeah
who's
saying
this
is
another
great
point:
sameer
hussain,
you've
taken
a
step
or
two
a
step
or
two
to
help.
Provide
users
help
provide
user
to
help
stop
forcing
docker
on
users.
I
think.
Is
there.
D
E
Thinking
was
there
any
issue
with
the
bash
script?
Installation?
Sorry,
I
did
not
get
the
exact
issue.
D
C
C
That
is
probably
not
prominent
and
is
certainly
not
the
default,
and
so
thinking
toward
how
to
change
this
up
and
how
to
make
that
more
understood
is
pr.
I
think
I
made
a
suggestion.
I
thought
I
opened
an
issue.
Maybe
it's
in
chat,
but
the
probably
the
right
suggestion
here
is
that
yes,
people
can
users
can
supply.
C
Tags
parameters
when
they
execute
the
bash
script.
Great,
that's
perfect!
That
helps
with
automated
installation
of
messy
ctl.
You
know
a
lot
of
you
know
automated
downloading
and
installing,
but
it,
and
we
would
probably
want
to
support
a
silent
flag
on
that
as
well,
which
I
think
you
know,
we've
got
some
prior
art
for
what
we
probably
want
to
do
by
default,
like
instead
of
supporting
either
docker
by
default
or
kubernetes
by
default,
is
by
default
prompt
the
user.
It's
probably
what
we'll
want
to
do.
A
I
guess
we
have
it
first,
but
I'm
not
sure,
but
even
though
then
it
would
be
only
be
a
temporary
override,
so
we
should
be
like
I
actually
mentioned,
we
should
be
prompting
the
user
to.
You
know
like
check
their
platform
before
we
create
a
config
file
for
default.
So
I
think
that
should
work
and
for
some
is
like
to.
A
You
can
just
change
your
platform
in
the
measuring
config
file,
which
will
be
inside
the
measuring
folder,
so
just
change
that
to
kubernetes
and
you
will
be
able
to
run
mesh
residual
commands.
A
D
Yeah,
okay,
so
that's
all
from
my
side,
so
we'll
be
continuing
the
same
test.
Probably
like
you
know.
We
have
not
yet
started
so
we
are
just
configuring
so
maybe
from
tomorrow,
we'll
start
doing
all
sort
of
things
so
before
that
I'll
be
getting
in
touch
with
the
visit
to
get
the
thing
sorted
to
actually
identify
my
istio
okay
yeah,
so
yeah,
that's
from
my
side!
Thank
you.
C
That's
that's
great
samir,
that's
thank
you.
Some
of
those
are
a
little
embarrassing
and
they're
just
little
small
items
that
are
easy
to
clean
up.
So
hussein,
as
you
think
about
the
issue,
the
perspective
from
which
I
encourage
you
to
think
of
that
last
challenge
is
from
scratch
from
the
very
beginning
before
a
user
has
any
anything
about
measuring
on
their
system.
They
go
to.
You
know
they
don't
always
follow
this,
but
they
would
go
to
mashri.io.
C
They
see
that
the
bash
script
is,
you
know
one
of
the
choices
and
if
they
use
that
today
it
actually
says
it
on
the
screen.
Let
me
let
me
show
you
this
just
because
I
want
to
clarify
so
so
that
hussain,
you
can.
You
know,
move
forward
with
this,
which
is
you
go
to
getting
started?
Okay,
we'll
choose
your
platform.
Well,
you
know
what
I'll
just
go
ahead
and
use
bash.
That's
that's
convenient
now,
if
you're,
if
I,
if
you
are
like
me,
I
don't
read
anything
more
than
I
have
to.
C
I
see
a
copy
button
there.
It
is
I'm
good
to
go.
I'm
gonna
go,
and
I
didn't
see
that.
Oh,
that's
using
docker.
Oh
okay,
well,
well,
where's
the
option
for
using
kubernetes,
okay.
Well,
there
isn't
one
for
bash.
Oh,
but
there
is
oh,
but
it's
not
documented,
oh
okay,
so
I
guess
I
have
to
go
over
here
and
use
geez
something
else.
You
know
what
I'd
rather
use
bash,
oh
wait,
but
this
only
works
with
docker.
C
Hopefully
that
helps
usana
so
who's.
It
yeah,
I
think,
is
any
of
you
agree.
It's
a
have
a
second
stanza
here.
You
know
like
make.
We
need
to
clarify
with
and
by
the
way,
not
with
words,
but
with
pictures,
docker,
logo
and
kubernetes
logo,
because
you
know,
because
I'm
I'm,
because
I'm
lazy
and
I
assume
everyone
else
is
as
well.
The
second
thing
is:
if
that
parameter
is
not
specified,
we
should
remove.
D
E
Yeah
so
basically
like
an
environment
variable
or
maybe
a
a
prompt
right,
both
of
that
should
be
okay
to
pass
to
the
measuring
script
right.
Yeah.
D
C
But
hussain,
I
think
if
you
check
in
the
mesher
ctl
channel,
this
is
probably
there's
either
an
open
issue
or
a
chat.
But
the
thing
is:
it
is
actually
critical
to
do
this
in
a
specific
way
and
to
get
it
right,
because
because
people
will
get
snagged-
and
that
is-
I
think-
there's
kind
of
three
considerations-
one-
an
update
on
mestre.io
to
have
that
second
option
to
make
it
clear
that
one
is.
C
You
know
what
one's
is
like
the
one
that
is
there
now
it
shouldn't
be
opinionated
it.
It's
the
one
for
prompting
the
user
as
secondary
considerations,
and
maybe
you
know
smaller
little
notes-
is
that
we
could
note
that
there's
a
you
know
optional
flags
that
you
can
pass
that
that,
like
when
people
want
to
eventually
automate
the
install
and
deployment
of
measury,
they
can
supply
those
environment
variables
and
it's
not
as
critical
to
have
those
as
prominent
in
that
particular
page,
because
that
page
is
meant
to
be
like
getting
started
very
quickly.
C
But
for
advanced
use,
cases
and
environment
variables
and
stuff.
There
can
be
a
link
to
we
can
you
can
figure
out
how
to
show
that
to
them
there
or
link
them
to
the
docs,
but
the
behavior
would
be
that
yeah
by
default,
they're
prompt
with
a
question.
They
can
override
it
with
a
silent
answer
using
an
environment
variable
like
you're,
saying.
C
Nice,
nice
good
right.
E
A
C
C
Yes,
it
will
be
automatic.
An
instance
of
that
will
be
automatically
deployed
when
you
go
into
measures
ui
and
you
click
to
deploy
istio
yeah
it.
It
will
subsequently
need
to
be
configured
to
allow
you
know
to
route
traffic
to
you
know
to
the
specific
you
know
to
your
app
and
stuff
yeah.
C
And
so
sameer
as
a
prelude
to
like
assuming
some
successful
next
steps
for
you,
a
prelude,
you
will
come
back
very
shortly
and
provide
feedback
saying
well,
it
would
be
wonderful
if,
in
the
measure
ui
for
managing
istio
excuse
me
in
all
the
rest
of
the
service
meshes
if
well,
if
you
exposed
a
knob
for
a
control
for
deploying
my
ingress
or
not
and
then
subsequently
and
like
in
deploying
an
egress
gateway
or
not,
and
subsequent
to
that,
if
I
choose
to
deploy
one
of
the
two
of
them
or
both
of
them,
I
would
might
also
like
to
apply
some
config
like
right
there,
because
I'm
going
to
deploy
this
app,
and
so
I'd
like
to
open
up
this
port
or
have
these
you
know
virtual
services
and
destination
rules
configured
and
that
that
level
of
detail
isn't
supported
in
the
ui
today,
mesherie
is
capable
of
that
through
pattern,
files
and
errors
getting
there,
and
so.
D
Yeah,
so
why
it
came,
to
my
mind
is
so
I
the
thought
process
was
like
so
when
I,
like,
you
know,
click
on
that
link
or
when
I
install
messaging
in
my
kubernetes
or
kubernetes
a
case
cluster.
D
It
should
ideally
directly
give
me
one
like
ib
address
to
directly
access
misery
instead
of
going
to
my
localhost.
So.
C
It's
totally
that's
another
good,
it
totally
and
I
think.
C
Adidi
and
piyush
yuka,
you
two
are
kind
of
working
on
that
very
thing,
so
push
as
you
work
through
the
system
checks.
It's
I
think
we
were
just
talking
about
this
last
night
yesterday
on
an
issue
where
we're
gonna
system
check
for
the
availability
of
yeah.
It
was
a
dd
right
of
the
measuring
endpoint
and
not
just
the
fact
that
we
can
open
a
tcp
communication
with
the
the
message
endpoint,
but
also
verify
an
unauthenticated
api.
C
Oh
okay,
actually
samir,
you
might
be
talking
even
past.
So
what
I
was
saying
was
like
hey
when
you
do
mastery,
ctl
system
start
and
you
want
to,
and
it
tells
you
to
sign
in
to
measure,
and
it
gives
you
a
url
that
url
is
ideally
has
the
appropriate
port
and
also
has
verified
that
in
fact
meshri
is
there.
Is
there
reachable
and
available
yeah?
C
I
mean
it
won't
always
be
the
case
that
we
like,
because
there's
so
many
ways
to
expose
mastery's
ui
and
do
networking
and
all
these
things
and
depending
upon
where
you're
sitting
and
whether
or
not
the
network
is
available
and
et
cetera,
like
we
might
be
able
to
only
take
that
so
far,
but
of
the
subsequent
thing
that
you
just
said
were
you?
Were
you
talking
about
measuring
verifying
your
accessibility
to
your
workload,
endpoints.
D
Yeah
so
now,
like
you
know
messy,
like
you
know,
I
have
the
kubernetes
already
and
I
do
have
my
micro
services
running
and
my
like,
you
know
you
say
the
cluster
ip
are
already
there,
and
so
when
I
install
mercury,
so
my
expectation
is
so
it
should
validate.
My
like,
you
know,
cluster
ips,
which
is
basically
let's
say
with
some
level.
D
C
For
okay,
yep
good
for
accessing
mescheri's
ui
or
for
accessing
your
like
your
workload,
endpoint.
C
Yeah,
so
very
good,
let
me
ask
you
this
or
that
that
makes
a
lot
of
sense.
That
mashri
should
facilitate
a
few
things
like
one.
What's
the
health
of
my
one,
let
me
help
me
deploy
my
workloads
that
might
be
nice.
It's
not
that
people
don't
do
it
in
different
ways,
but
since
you're
right
there
it
might
be
nice
to
also
deploy
your
workload.
C
You
know
visualize
it
and
some
of
the
things
that
we
were
talking
about.
It
would
also
be
nice
if
messy
was
sort
of
telling
you
the
health
of
the
you
know.
If
the
workload
is
up
and
accessible
generate
some
load
against
it.
Let
you
test
performance
and
things
like
that,
but
also
help
you
expose
it
through
the
istio
ingress
gateway.
F
C
D
Yes,
I'm
ready
to
give
some
sort
of
instructions
like
what
should
be.
My
like
you
know,
path
routing
should
be,
and
if
it
is
required
so
then,
depending
upon,
if
there
is
a
dns
already
attached
to
my
sto
ingress
gateway,
so
it
will
just
create
some,
like
you
know,
qualified
domains
for
some
of
the
services.
C
Yeah
makes
a
ton
of
sense,
that's
good.
I
will
we'll
get
some
of
this
written
down
on
paper
so
that
we
can
so.
I
would
say
that,
like
hey,
the
tool
to
your
point
needs
to
grease
the
wheels
and
facilitate
this
and
there's
some
safe,
intelligent
decisions
and
actions
that
it
can
take.
There
are
some
decisions
and
actions
that
it
might
just
cue
up
for
you
and
suggest
that
maybe
this
is
what
you
want
to
do,
but
but
not
actually
do
it
in
case.
D
Yeah,
so
that's
my
like
you
know
thought
so
when
I
see
the
complexity
of
I
still
to
do
it
manually
with
my
yaml
file,
so
I
thought
why
not
like
you
know
if
we
have
something
which
is
managing
the
istio,
so
I
would
expect
the
iso
management
tool
would,
like
you
know,
supplement
these
kind
of
things.
So.
C
Yeah-
and
you
would
be
that's
a
great
assumption
to
make
yeah
yes,
it
should
absolutely
please
samir
great
feedback
today.
This
is
this
is
good.
C
A
Yeah
we
missed
a
couple
of
recent
works
on
boston
filters,
image
hub
and
the
like
the
new
feature
to
add
filters
in
measuring
ui.
So
we
had
a
discussion
yesterday
and
the
summary
of
it
is
available
on
the
wasam
channel
on
slack
and.
F
Yeah,
okay,
so
a
more
detailed
thing.
We
typed
it
up
in
the
specification
itself.
I
know
we're
not
supposed
to
put
in
meeting
minutes
in
the
specification,
but
this
temporary
dump
there
and
it
sort
of
helped
us
align
our
goals.
So
there
are
currently
three
people
working
on
three
different
things
on
the
project
and
a
lot
of
our
tasks.
Our
goals
are
overlapping
and
essentially
the
meeting's
purpose
was
to
sort
of
discuss
where
the
overlap
is
occurring.
F
Who
needs
to
coordinate
with
in
regards
to
what
and
also
I
guess,
gaurav
did
not
join,
and
I
am
extremely
confused
as
to
where
he
what
work
he
is
doing.
I
was
hoping
to
you
know,
get
that
clarified
now,
but,
okay,
I
guess
we'll
catch
up
I'll
catch
up
with
him
at
a
later
point
of
time.
Yeah!
That's
pretty
much!
It.
F
I
know
we
don't
have
time,
so
it's
all
right.
A
Yeah
no
no
problem
like
yeah,
so
we
are
already
done
fast.
So
I
guess
it's
time
to
wrap
up
this
meeting
and
thank
you
everyone
for
joining.
I
guess
I
will
be
your
community
caller
bye.