►
From YouTube: Meshery Development Meeting (May 19th, 2021)
Description
Meshery Development Meeting - May 19th, 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
Good
good
good
good,
I'm
well
on
my
second
cup
of
coffee,
so
I'm
I'm
caffeinated,
I'm
good!
All
right!
Fair
enough!
I
I
just
had
to
say
hi
before
the
meeting
got
started
so
nice
to
see
you
nice
to
see.
A
B
You
know
it
was
a
little
bit
difficult
for
me
to
to
find
some
times
for
for
the
meeting,
because
at
this
time
I'm
I'm
at
work,
and
I
I
I
I
need
to.
B
Today
there
is
no
problem,
I
I
can
assist
all
the
meeting,
but
maybe
sometime
I
I
will
have
to
miss
some
some
meeting
also.
A
Yeah
yeah
makes
sense.
Sometimes
we
can
we
a
lot
of
times
we'll
talk
about
measuring
stuff
on
the
friday
community
call
as
well,
so
it
it's
kind
of
a
nice
catch-all.
A
C
C
C
I
see
a
couple
of
new
names
here.
Anyone
on
this
call
for
the
first
time.
C
Okay,
more
we've
seen
some
of
your
work.
Would
you
mind
introducing
yourself
to
the
others.
B
C
D
Yeah,
hello,
hi,
gavdam
hi,
so
I
am
gautham.
I
am
pretty
much
into
web
technologies
right
now
and
want
to
dive
deep
into
devops
and
stuff.
So
I
found
layer
5
community
just
yesterday,
so
it's
very
warm
welcome.
I
appreciate
that
I
have
had
a
talk
with
me
about
what's
actually
the
layer
5
community
and
had
a
reading
in
the
documentation.
I
found
it
very
interesting
way
to
to
means
dive
deeper
and
learn
more
about
devops.
D
C
A
Yeah
really
yeah
wow,
actually
yeah.
It's
an
interesting
thing.
If,
if
you
all
go,
look
at
the
measuring
repo,
it's
like
it's
like
spot
on.
I
think
200,
200
individuals
have
well
committed
code
or
documentation
or
something
it's
interesting
and
appropriate
to
recognize
that
that
number
is
a
lot
larger
than
that
it's
by
by
in
two
ways.
One
is
that
there's
a
lot
of
people
who
have
contributed?
A
They
just
didn't
contribute
code,
and
that
doesn't
mean
that
they're
not
contributing
there's
a
lot
of
people,
and
maybe
some
of
you
on
the
call
who
will
join
and
just
ask
questions.
A
You'll
provide
feedback,
you'll,
just
listen
that
all
counts,
like
being
a
standard,
buy
and
there's
there's
some
folks
that
have
been
on
the
call,
and
you
know
what
I'm
not
gonna
look
to
see
if
they're
here,
because
I
wouldn't
I
wouldn't
you
know
I
try
to
embarrass
people
when
I
can,
but
I
usually
end
up
embarrassing
myself
and
I
would
say
that
there's
some
people
who've
been
on
the
call
for
like
a
year.
A
I
haven't
said
anything,
but
they
consistently
show
up
and
that's
fantastic
we've,
we've
chatted
or
other
things.
They
just
don't
talk
on
the
calls
and
that's
fine,
they're
contributing
they're
being
here
and
and
helping
so
anyway.
My
point
is
not
only
do
those
contributions
count,
so
the
account
is
a
lot
higher,
but
also
there's
like
40
something
other
repositories
that
people
have
contributed
code
to,
and
so
the
count
is
actually
a
lot
higher
than
that,
but
anyway,
really
cool.
A
A
There's
one
other
individual,
so
kartik
kartik66
did
you
get
a
chance
to
introduce
oh
kartia.
A
If
you
can,
you
want
to
say
hi
and
just
introduce
if
your
audio
isn't
good
doing
so
in
chat
will
work
as
well.
It's
just
nice
to
say,
hi.
C
B
Okay,
I
was
saying
that
the
first
time
I
assist
on
this
meeting
meeting
so
I
have
to
share
my
screen
or.
A
And
so
just
for
you
know
like
a
little
bit
of
context
as
more
goes
to
present,
so
I
think
part
of
his
focus
has
been
on
measuring
ctl
and
on
expanding
the
perf
command.
Yes,.
B
A
So
there
was
an
enhancement
to
the
performance
management
capabilities
of
mesherie
in
the
v050
release,
and
it
was
for
expanding
on
the
concept
of
performance
profile
which
allows
users
to
save
different
test
configurations
that
they
want
to
run
and
easily
recall
them.
That
ability
hadn't
been
exposed
that
new
capability,
that
enhancement
to
like
performance
management,
had
not
been
exposed
in
mastery
ctl.
Just
yet,
and
so
moore
is
working
on
that.
B
C
Sounds
good
so,
like
is
push
on
so
push.
Could
you
demo
your
puff
list
while
we
wait
for
more
okay,
cool
yeah,
so
piyush
is
also
working
on
the
performance
command,
the
purplish
command
and
yeah.
What
he
has
been
working
on
is
to
enable
group
four
performance
profiles
in
measuring
ct.
So,
yes,
okay,.
D
C
C
C
C
You
can
see,
I
can't
I
cannot
scroll
the
terminal,
that's
like
a
back
a
disadvantage
of
this
user
experience
and
why
it
is
happening.
I
will
try
to
explain
so
the
library
that
I'm
using
from
for
this
purpose
to
paginate
these
results
is
term
box,
and
this
this
library
needs
to
be
initialized
at
every
like
whenever
I
want
to
output,
something
so
I'm
initializing
it
and
this
library
needs
to
be
closed
at
the
end
of
all
the
things
I'm
I
want
to
do.
C
This
library
needs
to
be
closed
and,
if
I'm
doing
that,
then
the
results
will
not
be
persisted
on
the
screen.
That
is,
the
only
disadvantage
and
terminal
will
go
back
to
its
original
functionalities.
C
E
C
So
this
opens
another
window
like
a
layer
kind
of
in
terminal,
and
I
can
paginate
here,
like
you
have
already
seen
the
gifs.
I
can
paginate
you
and
if
I
like,
this
is
the
final
page
and
if
I
press,
spacebar
or
control
c
at
this
page,
this
window
will
be
get
closed
like
this
layer
will
be
flushed
out.
C
Okay,
okay,
so
like
the
second
one
is
better.
In
my
opinion,
that
is
what
I
was
mentioning
on
the
pr
right
yeah,
but
I
realized
this.
I
guess
20
minutes
ago
that
I
am
not
able
to
scroll
my
terminal
after
this
kind
of
user
experience.
C
If
I
guess,
like
the
the
better
of
the
two,
is
the
second
user
experience
so
yeah,
like
my
suggestion,
is
that
like
we
can
go
with
it
for
now
and
if
there
aren't
any
other
better
solutions
to
this
or
like
another?
Another
way
is
to
like
only
show
a
set
amount
of
results,
as
we
like.
I
think
we
we
had
a
discussion
on
the
pr
threat
right,
only
set
amount
of
discussion
and
that
that
will
ensure
that
you
can
like
scroll
back
up
and
down
right.
G
I
had
a
query
regarding
the
commands
you're
running
so
normally
I
mean
I
would
want
to
see
these
results
in
another
file
or
something
so
I
would
normally
use
this
command
with
an
output
redirection.
So
how
does
your
current
implementation
behave
with
that.
G
G
Traditionally
I
again
people
anyone
can
correct
me
if
I'm
wrong,
but
normally
these
results
view
them.
Just
like
a
quick
glance.
Sure
the
term
is
fine,
but
normally
I
would
want
to
catalog
them
and
save
them
in
a
file.
So
I
feel
that
this
command
should
be
created
with
that
as
a
first
priority
rather
than
having
space
but
move
to
the
next
page.
Just
up
again,
if
anyone
objects
to
it,
please
do.
G
C
Yeah
that
makes
sense
for
now
so
yeah.
That
kind
of,
like
maybe
another
argument
to
like
use
the
like
to
limit
the
number
of
results
to
a
set
value
so
that,
like
we,
can
pipe
the
output
to
some
other
file
or
something.
G
The
thing
is,
even
if
you
limit
the
results
to
set
value
here,
what
I'm
seeing
is
it's
automatically
adjusting
the
number
of
results
based
on
the
size
of
the
terminal?
Again,
I'm
not
sure
if
I'm
right,
but
if
that's
the
case,
then
whatever
works
done,
I'm
I'm
so
sorry
you
might
have
to
look
at
it
once.
G
Be
just
to
print
out
a
huge
list
and
then,
like
you,
have
created
a
tabular
thing
so,
instead
of
it
going
to
the
next
page,
just
print,
it.
G
C
The
thing
is:
measuring
api
exposes
fat
max
25
results
at
one
time.
I.
D
C
A
Yeah
good
good
feedback,
a
couple
of
thoughts
on
that.
I'm
not
hearing
others
say
this,
so
the
last
piece
of
feedback
about
pagination
is
spot
on
so
like,
for
example,
my
user
right
now
it
if
it
were
to
retrieve
all
the
results.
It's
a
couple
thousand
results
which
would
really
hammer
on
the
server.
A
If
I
were
to
you,
know,
export
not
only
just
the
list
of
results
which
is
a
couple
thousand,
but
also
the
full
details
of
those
results
which
is
probably
the
what's
of
real
interest,
and
so
there
is,
you
know,
there's
a
use
case
for
like
the
use
case
for
exporting
all
of
the
users.
Results
is
really
probably
a
migratory
thing,
either
or
they're
tearing
down
the
server
and
they're
kind
of
done
with
it.
So
it's
a
good
okay.
A
So
there's
a
couple
couple
of
thoughts
like
one
is
that
with
local
providers,
the
number
of
results
that
we
would
anticipate
someone
would
have
in
their
system
is,
is
relatively
less
as
compared
to
the
number
of
results
that
someone
might
collect
over
time.
A
If
they're
using
a
remote
provider
that
offers
persistence,
that
remote
provider,
one
of
those
being
the
the
mescheri
remote
provider,
users
of
the
other
client,
the
mesh
reui,
they
can
go
in
and
check
on
individual
rows
in
the
performance
table
to
download
those
results
there
as
well
on
an
individual
basis.
A
I
don't
recall
if
there's
a
bulk
option,
I
think
there's
a
bulk
option
for
downloading
many
results
at
one
time,
I'm
still
that
might
be
in
a
paginated
way
in
part,
because
it's
just
like
it's
an
infrequent
use
case.
Probably
that
people
want
to
download
that
many
results
at
a
given
time.
A
You
know
they're
able
to
export
some
things
to
potentially
to
file,
but
maybe
not
all
of
them
at
at
one
time
in
the
use
case,
like
so
by
the
way
for
everyone's
edification,
with
respect
to
there's
kind
of
two
constructs
with
respect
to
performance
management,
so
one
of
them
is
a
performance
profile
which
is
just
a
set
of
configuration
items
about
the
test
that
you're
going
to
run
what
endpoint
or
endpoints
you
want
to
send
load
to
how
intensely
you
want
to
do
it
and
for
how
long
and
what
type
of
traffic
that
kind
of
thing
that
configuration
that's
the
profile.
A
A
A
It's
a
bit,
but
so
that's
food
for
thought,
as
we
think
about
this,
like
the
user
experience
of
the
first,
the
high
level
construct
of
looking
at
first,
the
performance
profile,
so
I
might
just
be
interested
in
results
for
my
soap
test,
not
my
other
profile,
and
so
maybe
there's
just
there's
50
results
in
there
and
50
results
in
the
other
one
I
don't
know,
and
so
yeah
allowing
the
user
to
there's
it's.
A
I
think
you
know
search
would
achieve
much
of
you
know
like
there's
a
limited
amount
of
experience
you
can
provide
through
the
through
a
cli
people.
Might
we
do
anticipate
people
to
run
mesherry
in
their
ci
pipelines
and
run
performance
tests
there?
A
A
So
if
jubril
is
on,
I
don't
know
that
he
is,
but
that's
a
great
use
case
for
the
github
action.
A
So
it's
a
little
bit
of
a
challenge
because
you
can't
hammer
the
the
server
with
you
know,
an
in
endless
amount
of
results
plus
isn't
really
useful
to
the
user
to
have
them
all
exported,
at
least
like
it
is
useful,
but
it's
not
the
general
purpose
use
case.
It's
a
special
migratory
thing
in
which,
in
which
case
people
might
want
to
access
the
database
or
have
a
d.
You
know
access
to
the
raw
rest
api
itself
and
write
their
own
scripts
kind
of
around.
That.
E
A
Mean
that
we
don't
want
to
facilitate
the
creation
of
a
performance
profile
from
here,
which
you
can
already
do
or
the
lookup
of
one
or
looking
up
what
the
configuration
of
one
is
and
then
the
invocation
of
one
and
like
the
retrieval
of
the
results
and
those
things
are
all
possible.
You
can
output
the
results
to
a
file
but
of
the
use
case
that
you
mentioned,
like
bulk
output
of
all
the
results
to
a
file
yeah,
that's
a
good
use
case
for
either
the
ui
client
or
for
a
remote
provider.
G
All
right
the
bulk
output
again,
I
was
a
little
hesitant
about
that
for
this
reason,
but
my
main
thing
was,
since
we
are
what
is
a
filtering
or
looking
at
a
particular
profile,
it
would
still
be
useful
to
be
able
to
do
that.
That
was
my
only.
A
Output
to
a
file,
they
absolutely
can,
as
a
matter
of
fact,
it's
becoming
standard
output
or
standard
flags
for
output
to
yaml
or
output,
to
json
that
that
should
be,
if
that's
not
in
measure
ctl
now
for
perf,
and
it
should
be
it's
there
for
patterns
ready,
but
so
that's
yeah.
That's
that's,
definitely
use
cases
to
retrieve
the
performance
results.
You
know
like
the
full
set
of
results
for
a
given.
A
Oh,
more,
okay,
more
just
gotta
go
but
good
yeah,
good,
good
feedback.
C
C
So
I
I
guess,
like
there
are
some
things
to
look
into
more.
Are
you
available
to
look
into
the
perfume
command
right
now?
Oh.
B
So
what
I
have
done
yet
is
I'm
able
to
to
view
the
performance
profile
as
a
road
text
and
also
as
a
json.
B
B
I
think
I'm
going
to
finish
the
work
in
today
or
tomorrow,
and
I
will
submit
a
new
message
request
with
a
profile
id
instead
of
profile
name
and
also
to
refer
to
my
my
code
to
use
to
use
extract
from
from
this
model.
C
That
looks
good.
The
the
the
output
you
have
currently
is
good,
like
I
have
a
quick
question
like
do
we
or
do
we
also
add
like
in
this
command?
Do
we
also
have
add
the
ability
to
view
the
particular
result,
or
is
that
a
separate
command
was
that
in
the
issue.
B
B
B
C
B
C
Okay,
yeah,
that
makes
sense
here,
that's
good,
any
other
feedback
for.
C
A
Keep
the
feedback
coming.
It's
it's
really!
It's
really!
It's
good
feedback
that
you
gave
on
at
that
last
and
it's
really
important
that
people
give
feedback.
G
G
So
he's
done
a
pretty
good
job.
I
think
it
covers
up
from
what
I
can
understand
covers
people
might
need,
as
of
now,
only
once
people
start
using
it
and
implementing
it.
I
guess
more
use
cases
would
be
shed
light
upon.
So.
H
H
A
Yeah
in
okay,
speaking
of
have
you
provided
abhishek
kumar,
that
feedback
on
nighthawk.
Yes,.
A
A
C
Awesome,
I
guess
then
thank
you
more.
That
was
excellent.
It
was
a
good
implementation
and
see
you
next
in
the
next
meeting.
B
C
Next,
we
have
to
talk
about
pattern
capabilities
registration.
Do
you
want
to
share
your
screen.
C
F
So
I
think
I
have
previously
demoted
patents.
What
parents
are
so
the
today.
Today's
demo
is
a
recent
edition
in
adapters
that
was
made.
It
was
regarding
that
so
earlier,
what
used
to
happen
was
if
you
would
spin
up
mystery
so
if
for,
if
you
are
deploying
it
on
humanities
or
on,
maybe
maybe
in
dr,
so
what
would
happen
is
that
for
some
reasons,
machine
server
is
starting
with
starting
up
really
slow
and
basically
adapters
are
starting
up
quite
late.
F
Then
what
would
happen
is
that
adapter
will
try
to
reach
out
to
messi
server
in
order
to
register
its
capabilities,
but
it
would
actually
fail
because
the
server
is
just
not
there.
That's
it
I
mean
so.
Basically,
the
registration
process
would
fail.
So
the
recent
addition
was
regarding
retries,
that
is
now.
If
so
now,
what
I,
what
I've
done
is
I
have
started
up
mysteries
is
the
adapter
is
the
adapter
on
my
system.
It's
not
complaining
right
now,
because
behind
the
scenes
is
actually
retrying
the
registration
process.
F
I
haven't
yet
started
mystery
server,
I'm
starting
it
up.
Had
this
been
the
previous
like
without
retries
implementation.
What
would
have
happened
is
you
would
have
seen
an
ugly
error
here
that
this
endpoint
was
just
not
reachable
so
I'll
go
ahead
and
register
capabilities.
F
So
let
this
thing
come
up
and
you
will
see
that
as
soon
as
the
server
comes
up
in
a
while
it
will
machine,
adapter
will
retry
the
registration
process
and
it
did
so
the
the
logs
that
we
are
getting
here.
These
are
basically
the
capabilities
that
that
sd
adapter
has
just
registered
with
the
machine
server,
and
this
is
this.
Retries
is
basically
based
on
exponential
back
off
that
that's
the
burden
that
we're
using
for
return,
and
the
other
thing
is
that.
F
F
I
don't
think,
I'd
be
able
to
go
into
complete
depth
what
this
entire
army
is,
but
basically,
what
we
are
doing
is
we
are
from
provisioning
seo
service
mission
here,
provisioning,
a
provisioning
draft
instance
and
provisioning
prometheus.
We
are
also
specifying
the
dependencies
in
here
that
should
be
deployed
after
st
prometheus
and
some
other
things
like
enabling
mpls
automatic
cycle
pro
automatic
cycle
injection
on
certain
name
spaces
and
these
kind
of
things.
So
when
we
go
ahead
and
apply
this
pattern,
this
this
process
usually
takes
time.
F
But
what
happens
behind
the
scenes
is?
It
creates
a
plan
of
execution.
It
figures
out
that
istio
needs
to
be
deployed
first
and
then
prometheus,
and
basically
these
are
dependencies
that
it
has
figured
out.
So
istio
would
be
prometheus
deployed
first
and
once
that
is
completed,
it
will
go
ahead
and
try
to
provision
prometheus
and
then
grafana.
F
Let
me
show,
let
me
try
to
show
yeah,
so
I
provisioned
these
things
and
they
have
just
recently
came
up
so
grafana
12
seconds
and
basically
all
of
the
things
that
are
actually
answered
to
provision.
These
are
provisioned
now
in
my
in
my
tubebuddies
cluster,
and
these
are
the
messages
that
I
got.
This
is
something
that
needs
to
figure
out,
but
yeah,
and
similarly
we
can
also
run
pattern
of
that
apparently,
and
that
will
basically
read
the
same
gamut
and
will
delete
the
resources
that
were
provisioned
using
that
particular
yaml.
F
F
And
the
message
that
you
would
get
in
here
is
that
the
the
seo
service
mesh
is
deleted,
the
trades
are
basically
removed
and
prometheus
and
grafana
are
also
been
deleted.
So
now
this
is
pretty
much
it.
H
Yeah,
so
one
one
question
so
here
when
you
are
installing
it
through
the
patterns,
okay,
so
from
where
exactly
get
the
configurations
like,
if
I
have
some
like
configuration,
which
is
mostly
user
defined
configurations,
not
the
default
for
prometheus.
F
Okay,
yeah,
actually
because
this
instance
is
so
patterns
right
now.
So
when
I
registered
the
trades,
it
had
something
called
that
it
supports
prometheus
and
when
it
said
that
it
supports
stevia
adam
and
so
when
it
said
that
it
supports
grafana
sdr
from
prometheus
jr1.
F
So
it
basically
meant
that
it's
going
to
provision
the
add-on,
basically
the
istio
that
actually
comes
by
default
from
there,
so
the
default
one
right
now
there
is
right
now
patent
doesn't
have
a
capability
of
provisioning,
a
custom
prometheus,
but
definitely
that
can
be
added,
so
it
would
also
it
would
kind
of
look
like
maybe
prometheus.
Something
and
type
would
be
prometheus
and
namespace,
as
you
say,
and
then
maybe
something
like
this.
That
is
where
you
get
in
your
basically
place
in
your
custom,
configuration
that's
that
is
completely
doable.
F
It's
just
that
we
haven't
reached
to
it.
Yet
patterns
are
flexible
in
that
manner.
That
is
these
internal
things.
They
are
completely
dependent
on
the
type
of
thing
that
you're
trying
to
provision.
F
Yeah
right
now,
right
now,
the
capability
that
that
we
have
created
is
only
for
room
istio,
add-on,
okay,
okay,
yeah.
C
A
You
know,
as
you
know,
we'll
hold
off
until
dhruv
or
soham,
or
somebody
can
demo
it
later
about
what's
next
for
patterns,
but
just
to
clarify
real
quick
samir
did
you
I
missed
it?
Did
you
say
that
you
are
I'm
assuming
that
you're
running
prometheus
as
an
operator
separately
and
independently
of
istio
or
are
you?
A
Okay,
you're
not
getting
some
istios
istio
native
metrics
from
a
disparate
prometheus,
deploy
from
a
separate
prometheus
deployment.
Yeah.
A
H
No
that
we
didn't
try,
because
what
we
have
done
is
like
you
know
we,
as
we
I
told
earlier,
we
were
most
interested
in
keali
wherein
we
get
the
isto
envoy
related
stuff,
wherein
we
use
the
prometheus
for
that.
But
what
happened?
So
we
didn't
want
to
overload
it
with
different
instance
of
prometheus.
H
A
Okay,
to
take
it
with
a
grain
of
salt,
that
was
just
a
you
know
off
the
cuff
thought,
actually
quick
question
for
you,
because
I
spent
very
little
time
with
kiali.
Do
you
is
it
the
case
that
kiali
functions
mostly
with
a
prometheus
operator
setup,
but
is
mis
or
does
it
not
function
at
all.
H
A
Totally
yeah
I'm
trying
not
to
totally
that's
precisely
the
goal
through
both
through
your
choice
is
actually
the
goal
and
when
you
do
it
through
the
ui,
the
idea
is
that
not
only
can
you
use
the
form-based
ui
that
is
there
today,
but
also
a
forthcoming
visual
topology
in
the
ui
that
hasn't
been
released
yet,
but
that
you
can,
but
we'll
have
to
have.
Maybe
we'll
have
to
line
up
a
demo
for
next
week
to
show
that
that
new
capability.
H
F
So
any
other
feedback.
If
anyone.
A
Has
yeah
lukas?
A
Do
we
have
the
emergence
of
a
so
there's
the
emergence
of
a
trait
registry
or
which
is
in
some
respects,
just
a
new,
a
registry
within
sight
of
mescheri
that,
where
each
of
these
adapters,
like
the
istio
adapter,
will
register
its
capabilities,
it
will
register
the
operations
that
it's
capable
of
so
that
meshri,
as
as
sort
of
the
central
intelligence
unit,
so
to
speak,
can
make
sure
there's
invoking
the
appropriate
operations
as
and
when
it
needs
to,
and
it
and
measuring
can
determine
whether
or
not
what
a
user
is
asking
it
to
do
like.
A
What's
in
this
pattern
file
here,
whether
or
not
the
adapters
are
capable
of
fulfilling
the
request.
So
that's
great:
do
we
have
the
emergence
of
a
trait
tracker
or
or
a
like
just
a
simple
spreadsheet
that
catalogs
the
traits
that
are
being
developed.
F
Nope
not
right
now
there
is
an
endpoint
with
that
which
actually
lists
the
current,
basically,
the
current
available
trades
and
trades
and
workloads
and
those
kind
of
things.
I
do
not
remember
the
end
point,
but
yeah
there
is
an
end
point
which
so
yeah.
Basically
there
is
no
documentation
of
the
trades
that
we
have
right
now,
but
there
is
an
endpoint
which
gives
without
returning
the
same
yeah.
Documentation
needs
to
be
done.
A
Okay,
nice
yeah,
no,
it
could
be.
I
mean
it
might
be
the
case
that
that
that
might
be
an
interesting
thing
to
for
users
to
see
inside
of
me,
I
mean,
obviously,
users
will
see
that
inside
of
meshri
when
they
go
to
manage
a
particular
service.
Mesh
they'll
see
all
of
the
operations
that
meshrie
is
capable
of
for
that
given
mesh.
It
might
be
an
interesting
view
for
the
measuring
administrator,
who
might
be
managing
multiple
meshes
to
just
see
a
consolidated
list
of
all
those
traits.
All
those
operations.
F
Yeah,
so
this
endpoint
actually
returns
all
of
the
trades
that
are
available
currently
available
with
basically
the
all
of
the
registered.
F
These
are
basically
one
of
the
registered
workloads
that
that
were
registered
but
registered
with
nursery
server
by
any
of
that
after
this
is
because
only
seo
app
is
capable
of
processing
home
things
right
now,
so
this
is
only
coming
from
istio,
but
basically
it
could
be
from
anywhere.
This
is
just
from.
It
is
basically
retrieving
the
data
right
now
from
the
registry
that
mushri
server
has
so
right.
F
F
Yeah,
it's
not
but
yeah.
There
are
two
end
points
the
trade,
so
you
can
get
all
the
trades
that
are
available.
You
can
get
all
the
workload
definitions
and
maybe
soon
scope
different
scope,
definitions.
Also,
it's
just
that
we
don't
use
scope,
definitions
right
now,
so
they're
not
available.
A
But
that
is,
but
it
is
possible
to
deploy
the
either
of
the
grafana
or
the
prometheus
operator
through
the
same
ui.
It's
not
as
convenient,
because
there
isn't
a
switch
button
for
you
know
there
isn't
a
toggle
button
for
it.
But
what
you
would
end
up
doing
is
pasting
in
the
manifest
for
the
prometheus
operator
into.
H
A
Right,
yeah,
yeah
right.
I
think
that
is
valid.
As
a
matter
of
fact,
I
think
like
like,
let's
assume
that
istio
didn't
have
a
challenge
exposing
some
of
the
some
istio-specific
metrics
to
external
prometheus
instances
like
let's
say
that
that
was
like
you
would
be
using
the
prometheus
operator
today.
A
A
It
needs
to
not
only
suggest
that
intelligently
to
the
user,
based
on
the
service
mesh
that
they're
administering
at
the
time,
but
it
also
needs
to
allow
individual
users
to
identify
the
charts
that
are
of
interest
to
them
persist
the
fact
that
yeah-
and
so
it
does
do
that
second
part.
Today,
it
does
let
a
user
identify
what
charts
are
of
interest.
A
I
think
the
the
challenge
that
it
that
or
where
measuring
needs
to
expand
is
in
in
a
couple
of
areas
and
and
I'll
just
I'll
describe
them
later,
so
that
other
people
can
demo
so
yeah.
C
Yeah,
so
the
next
item
is,
we
have
some
open
issues
on
measuring
ctl
docs.
C
So
there
are
a
lot
of
issues,
and
these
are
good
first
issues
and
I
know
there
are
some
newcomers
on
the
call
so
feel
free
to
comment
on
one
of
these
issues
and
you
can
get
assigned
to
them.
So
the
basically.
C
The
issue
is
that
we
we
have
a
doc's
website
for
measuring
and
it's
in
docs
not
measuring
attire,
and
we
have
a
command
reference
for
measuring
ctl
and
it
can
be
found
here
so
yeah,
that's
this
page,
and
so
it
basically
gives
a
documentation
for
each
of
the
measuring
ctl
commands,
and
each
of
these
commands
sub
commands
will
have
an
individual
documentation
page
and,
as
you
can
see,
some
of
the
commands
are
missing
and
the
issues
are
opened
up
to
to
fill
in
these.
C
These
missing
commands,
so
you
will
be
creating
a
page
similar
to
this,
and
so,
if
you
are
interested
in
getting
started
with
measuring
this,
this
might
be
a
good
place
to
start.
So
you
will
need
to
go
through
the
go
through
mystery,
ctl
and
test
it
out
to
write
these
recommendations.
So
it
will
be
a
good
place
to
start,
and
these
pages
are
written
in
are
written
using
jekyll.
So
if
you
know
jekyll,
then
you
you
can
easily
like
fix
couple
of
couple
of
these
issues.
C
And
yes,
I
guess
that's
it
from
my
part.
C
A
Next
week
it
sounds
like
we
might
have
a
couple
of
a
couple
of
topics
to
prepare
for
potentially
one
of
those
dhruv.
If
you
don't
mind,
maybe
taking
back
to
the
mesh
map,
call
a
request
for
demo
of
importing
a
pattern
and
visualizing
it.
I
think
that
might
be
of
interest
to
samir
and
maybe
some
others,
so
it
might
be
a
next
week
thing
another
one
that
might
be
a
next
week.
A
Thing
is
pranav
and
rude
rocks
and
a
couple
of
others
have
been
steadily
advancing
well,
garov
and
others
have
been
steadily
event
and
it's
steadily
advancing
the
filter
management
capability,
the
the
web
assembly
filter
management
capability
for
envoyboy
envoy
based
data
planes,
so
istio
being
one
of
those
and
console
and
and
the
list
and
kuma
and
the
list
of
others
that
mastery
will
ultimately
manage
the
the
dynamic
loading,
unloading
and
configuration
of
filters
in
those
data
planes.
A
And
so,
if
we
can
I'd
like
to
like
hold
ourselves
accountable
to
trying
to
make
part
of
those
conversations
a
little
more
mainstream,
because
because
pranav
kept
his
faith
and
was
working
as
a
solo,
individual
trying
to
build
up
momentum
in
that
area-
and
we
have
and
so
there's
now
kind
of,
filter
commands
that
are
coming
out
from
sri
cto.
There's
new
crud
endpoints
for
filters
in
mastery
server
and
mesh
free
cloud.
And
if.
G
We
can,
if
we
can
so,
could
I
add
something
to
what
you
were
saying
there.
Currently.
The
thing
is
that
something
I'm
noticing
and
I'm
pretty
sure
it's
my
incapability
of
also
coordinating
too
well
but,
like
all
of
a
lot
of
us,
are
sort
of
scattered
in
a
lot
of
places
and
now
that,
finally,
my
college
has
sort
of
you
know
said
that:
okay,
your
final
exams,
it's
going
to
be
postponed
indefinitely
or
something
of
that
sort
I'll
be
able
to
spend
more
time.
G
My
only
concern
is
that,
since
there
are
like
two
or
three
people
working
on
a
particular
aspect
and
since
it
sort
of
has
a
chain
reaction,
we
sort
of
need
a
system.
I
I'm
aware
that
the
design
document
is,
of
course,
really
really
important,
but
we
need
a
more,
I
guess,
a
real
time
system
to
coordinate
our
requirements
and
needs
last
meeting.
G
What
we
discovered
was
that
each
of
us
have
a
certain
responsibility
to
each
other,
so
we
need
to
work
like
this
we're
just
a
group
of
four
people,
so
we
need
to
work
with
each
other
with
our
on
our
respective
problems,
and
I
just
wanted
to
know
if
you
had
any
suggestions
to
help
coordinate
that
a
bit
better.
A
One
of
those
is
in
the
design
spec
a
small
table
that
identifies
the
various
areas.
Various
individuals
in
various
areas
that
they're
working
on,
but
also
the
various
areas
that
aren't
being
worked
on,
is
one
part
of
that.
The
other
part
is
beginning
to
outline
an
epic
and
a
set
of
child
issues
within
github
that
encapsulate
kind
of
the
full
set
of
full
scope
of
work,
and
then
the
while
github
is
far
cry
from
a
project
management
system.
It
helps
it
might
pronounce
it.
A
A
Service
mesh
now
you
know,
filter
management
like
it'll,
be
and
so
it'll
be
deep.
I
guess
is
what
I'm
saying
so
that
might
be
appropriate.
It
might
be
something
to
solicit
some
interest
on
in
the
awesome
slack
channel
and
see
if
I
think
I
think,
I'd
I'd
probably
be
up
for.
G
So
the
thing
is,
I
don't,
have
a
table
of
sorts,
but
we
have
already
made
a
list
of
responsibilities
and
what
each
of
us
are
responsible
for
and
gaurav
was
missing.
Unfortunately,
for
that
meeting
and
I'm
sort
of
in
the
dark
as
to
what
exactly
he's
been
quietly
digging
away
at.
Hopefully,
I'm
fixing
that
tomorrow,
as
I
set
up
a
meeting
with
him
and
I'll
update
the
docs
accordingly.
But
I
guess
a
weekly
meeting
or
a
bi-weekly
meeting
would
be
a
huge,
huge,
huge
help
because
it
will
just
keep
like.
A
Makes
sense
to
me,
let's
good
I'll,
let's
coordinate,
let's,
let's
do
it
yeah!
It's
we've.
We've
hit
that
juncture.
There's
this.
H
And
one
last
generic
question,
so
is
it
possible
for,
like
you
know,
voice
like
you
know,
awesome
to
manipulate
the
headers
when
it
comes
to
nba
proxy.
G
Oh
yeah,
I
can
do
whatever
you
want
with
the
thing,
so
it
again
depends
on
the
filter
so.
H
Okay,
so
is
there
any
demo
which
has
been
recorded
so
far
so.
G
A
demo
that
has
been
recorded,
okay,
I'm
I'm
sorry.
I
forgot
to
record
the
demo
so
like
I
haven't
really
okay,
so
this
is
more
sort
of
intercepting
the
current
filter
that
I
have
worked
on
this
intercepts
it
and
then
increments,
it's
from
its
own
store.
It
doesn't
actively
take
a
look
at
the
incoming
header.
Do
a
packet
inspection,
change
the
contents
and
rebuild
the
packet
and
send
it.
That
is
something
we
haven't
done
yet.
C
H
G
Right
right
right,
all
right,
so
the
thing
is
sure
we'll
definitely
explore
more
on
the
filter
aspects,
because
one
important
thing,
of
course,
is
to
build
up
the
filter,
library
and.
G
Is
it
possible
to
like
make
a
note
of
it
somewhere
or
leave
an
issue
on
image
hub
or
sorry?
There
wasn't
filter
repo
so
over
there.
If
you
leave
a
request,
we'll
take
a
look
at
it.
G
Yeah
currently
there's
a
issue
with
the
demo,
not
working,
and
I
have
a
meeting
with
abhishek
today
to
try
and
getting
try
to
get
it
working
with
the
latest
adapter,
but
do
try
with
an
older
version.
I've
been
told
that
it
does
work
5.1
if
you
use
the
adapter
version,
so
yeah.
H
H
I
G
Apologize
for
interrupting,
but
rudraksh,
perfect
timing,
I'm
going
to!
Are
you
free
at
around
10
o'clock.
B
C
Awesome,
we
are
quite
over
time
just
to
recap:
from
the
we
had
a
box,
the
buck
hunting
session
this
saturday
to
hunt
some
bugs
on
measures
area,
and
we
have
opened
up
a
lot
of
issues
for
that
and
yeah,
and
we
do.
We
also
have
a
test
plan
to
like
to
check
all
the
link
to,
and
we
use
this
as
a
guide
to
test
measure
receipt
out
and
measure
it
out
and
yeah
like
it
was
a
productive
meeting
and
yeah.
C
I
guess
that's
the
end
of
today's
call.
I
guess
we
will
meet
this
friday
on
the
community.
Call.
Thank
you.
Everyone
yeah.