►
From YouTube: Meshery Development Meeting (August 10th, 2022)
Description
Meshery Development Meeting (August 10th, 2022)
Agenda:
[Pranav] Fix bug on SMI conformance page PR#5956
[Pranav] Fix meshery-operator Helm chart PR#5972 PR#5976
[Pranav] Support for discovering ReplicaSets in K8s Cluster PR#143
[Pranav] Updated key_values table schema PR#144
[Xin/Hersh] CNCF Lab scheduled performance test failure.
[Ashish] Update on adapter status/ Call for tomorrow’s CI meeting.
[Yash] Meshery-ui-restructuring branch updates Open PRs Tracker
[Leonard] Meshery-ui-restructuring (Performance page)
[Alpha] Review on `app import` design spec
Review query on `filter import` design spec
[Aditya.C] Discuss Meshery Catalog support in Meshery UI.
Consideration of:
A git-based versioning of each catalog content.
Gating and curation catalog entrance of content.
Matt Young https://github.com/meshery/meshery.io
[Abhi] Need For Eslint Github Action #5986, #6010
[Gaurav] Bug fix on empty state PR#6009
[Gaurav/Abhi] Call for volunteer to write more E2E Cypress tests for Meshery
[Harshit] Centralising SVGs from meshery project only
Join the community at https://layer5.io/community
Find Meshery on:
GitHub: https://github.com/meshery
Twitter: https://twitter.com/mesheryio
LinkedIn: https://www.linkedin.com/showcase/meshery/
Docker Hub: https://hub.docker.com/u/layer5/
B
B
All
right
well,
the
live
stream
is
up
and
we
are
off
it's
the
mystery
development
meeting
for
wednesday
august
10th
and
hopefully,
your
those
that
are
on
the
call,
see
the
meeting
minutes
that
are
being
shared
you're
in
significant
luck
today,
if
you're
joining
I'm
hosting
today,
so
my
name's
lee,
which
means
there
will
be
you
know
some
plausible
jokes
or
some
things
that
I
say
that
might
sound
like
jokes
and
so
humor
me
for
elicit.
You
know
like
force
a
chuckle.
B
Occasionally,
if
you
would,
I
expect
that
there's
some
folks
on
the
call
who
are
going
to
drop
some
additional
agenda
items
like
now
is
that
time
before
we
get
rolling
drop
in
any
items
that
you
have
folks
like
abi
should
probably
put
his
toward
the
top
as
well
as
ashish.
Hopefully,
you'll
have
one.
If
you
don't,
you
can
excuse
yourself
from
the
call
those
are
kind
of
part
of
the
standing
rules
here,
just
for
ashish,
not
for
the
rest.
So
it's
nice
to
so
we
by
the
way.
B
B
My
point
of
that
whole
discussion
was
to
say,
if
you've
introduced
on
another
call,
like
the
websites
call
on
mondays,
feel
free
to
introduce
again
here,
because
different
faces
different
folks
show
up
to
different
calls.
So
it's
nice
to
get
to
know
you
and
to
help
break
the
ice.
Let
me
call
out
a
couple
of
folks,
so
we've
got
and
alex
or
two
on
the
call
and
so
alex.
Mr
john,
did
you
mind
saying
hi
real
quick
to
everyone?
Just
do
a
quick
introduction.
C
B
That's
awesome,
awesome
good
good
to
have
you
alex
good
to
have
had
you
on
the
websites
call
as
well
alex
we
just
noting
that
you
mentioned
ux
just
you
know,
there's
an
open
design
review
app
down
here.
That
might
be
of
interest
to
you.
I
think
there's
another
open,
ux
design
review
as
well.
It
might
not
be
listed
in
today's
notes,
but
if
you
scroll
down,
you
might
find
another
one
cool.
If
you're
on
the
call
feel
free
to
drop
your
name
into
the
meeting
minutes,
we
tally
them
up.
B
B
D
C
D
E
B
B
B
B
F
Watch
I'll
share
my
screen
and
quickly
show
because
the
pias
are
already
merged,
so
there
was
a
problem
on
our
implementation
on
a
smi
page
that
was
like,
if
you,
if,
before
this
pl,
if
you
would
have
gone
there,
you
wouldn't
have
seen
this
pagination
functionality
because
it
was
not
working
properly.
Moreover,
this
per
page
function
was
also
working,
so
I
just
fixed
that
now
it
works
fine,
so
we
could
see-
and
also
we
couldn't
see
the
total
number
of
results
which
we
actually
have
in
in
our
database
for
smi
conformance
test.
F
F
But
the
problem
was
that
since
measuring
operator
also
mentioned
operating
service,
to
have
the
port
10000
and
the
same
port
is
used
by
measuring
hd
adapter
as
well,
because
of
which
both
were
conflicting
and
we
weren't
able
to
use
measure
after
at
the
same
time
when
mentioned
operator,
is
deployed
in
the
cluster.
So
I
switch
that
back
to
cluster
ip
so
that
there
isn't
any
so
that
the
ports
are
not
same.
Ports
are
not
being
used.
F
So
that's
the
second
thing
which
I
updated,
and
the
third
thing
is
that
I
added
the
functionality
like
like
you
should,
if
you
use
actually
then
like
make
sure
it
has
a
capability
of
I'm
discovering
the
object
which
you
have
in
a
cluster
and
then
showing
it
on
its
own
extension,
that's
mesh
map.
So
initially
we
don't.
We
didn't
have
the
functionality
to
discover
replica
sets
in
our
cluster.
F
So
I
quickly
added
that
also
so
now,
if
you
use
mesh,
if
you
have
access
to
mesh
mouse
and
you
go
to
your
benchmark
visualizer
and
you
could
see
that
in
the
filter
now
you
see
a
new
field,
new
kind
that
you
have
and
if
you
click
on
that
it
will
automatically
show
you
all
the
replicas
in
your
question.
So
if
you
are,
you
don't
have
access
to
this?
F
You
should
probably
go
to
layer
five,
but
I
have
a
website
and
fill
the
form
for
this,
so
you
could
get
a
chance
to
use
it
in
the
better
form,
better
state
and
the
fourth
thing.
That
was
that
I
one
more
thing
it's
similar
to
this
third
problem.
That
was
that,
especially
when
it
discovers
object,
then
it
adds
them
into
different
tables
so
that
it
can
be
referenced
by
different
components
of
measuring.
So
there
is
a
table
that
is
key
value
pair,
which
stores
the
labels
and
annotation
of
the
different
object
discovered.
F
So
the
problem
was
that
it
was
just
it
was
showing
duplicate
data,
duplicate
labels
and
annotation,
and
that
problem
was
just
because
we
were
not
using
the
correct.
I
correct
attribute
as
a
primary
key
and
this,
and
that
was
causing
this
problem
because
of
which,
on
every
reboot
or
measuring
it,
was
again
discovering
it
and
again
storing
it,
because
the
primary
key
was
not
the
ideal
one.
I've
explained
everything
over
here,
so
I
changed
that
functionality.
I
changed,
updated
to
database
schema
for
for
and
now
kind
key
and
value.
F
Four
things
are
used
primarily,
so
they
works
as
a
composite
key.
The
reason
for
this
is
because,
because
like
here
id
can
be
multiple
in
the
same
database,
this
is
because
id
differentiates
an
object
and
or
an
object
can
have
multiple
labels
or
annotation,
and
it
could
be
multiple
values
for
same
object
in
a
database,
so
hence
we
cannot
individually
use
only
one
primary
according
to
me,
so
I
made
four
of
them
so
that
for
each
object
each
or
each
of
them
should
be
unique.
Only
then
it
should
be
added.
F
Otherwise
it
shouldn't
be
added.
So
now
it
works
fine.
So
I've
also
explained
that,
with
an
example,
for
example,
for
this
particular
object
like
it
was
for
this
same
idea,
it
was
showing
30
entries
in
our
database
table,
but
actually
it
was
just
two
which
which
we
can
see
after
my
changes.
So
this
is
still
in
discussion.
B
F
B
F
B
Just
kidding!
That's
great!
Thank
you.
As
a
matter
of
fact
I
so
so
the
whoa
whoa,
I
missed
one,
I'm
sorry!
I
got
there
were
other
items
happening,
but
so
you
have
an
update
on
the
helm
charts
for
measuring
operator.
Oh
okay,
I
see
yeah
good
good,
real,
real,
quick,
okay,
very
good,
so
there's
another
pr
open
on
with
respect
to
helm,
charts
and
that's
from
serene.
G
B
G
G
So
we
had
these
two
charts
as
two
different
charts,
and
this
can
this
actually
cause
a
lot
of
issues
when
you
already
have
an
operator
deployed
and
from
let's
say
this
chart
and
some
basically,
someone
tries
to
deploy
and
measure
along
with
the
operator
using
this
chart.
So
there
are
conflicts
that
come
every
now
and
then
and
serene's
pr
was
about
consolidating
these
two
charts
into
one
yeah
here.
G
Here
is
her.
She
would
be
able
to
talk
a
little
bit
more
about
it.
I
haven't
tested
it
right
now
on
my
system
but
yeah.
So
this
is
her
pr,
and
here
she
has
deleted
few
of
the
previous
basically
she's
deleted
the
she's
consolidated
both
of
those
into
one
chart
and
after
we
get
this
tested
and
in
messy
ctl,
as
well
as
mesh
to
ui
message
server.
G
G
G
We
have
a
bunch
of
them,
so
we
these
three
columns,
recently
got
added,
so
we
are
now
arm
compatible
across
all
of
the
adapters,
as
well
as
the
version
core
components,
which
is
to
say
that
I
had
actually
talked
about
it
in
the
last
call
as
well.
So
I
won't
go
over
it
again.
If
you
want
to,
I
can
go
over
but
yeah.
The
core
idea
is
that
the
we
have
something
called
as
a
core
component
like.
If
you
want
to
install
istio,
we
have
a
istio
mesh
component.
G
We
have
add-ons
that
we
give
out
of
the
box
and
those
are
now
versions.
So
for
each
service
mesh
version,
you
can
have
those
custom
add-ons
in
the
the
custom
steems
installation
under
that,
given
version
so
that
support
is
added
for
each
of
the
adapter.
And
previously
we
were
relying
on
a
node.js
based
binary,
which
kind
of
inflated
the
size
of
our
image.
G
G
I
guess
he
accidentally
opened
this
mic
anyways,
so
yeah,
so
about
the
end
to
end
tests.
There
are
a
few
discussions
around
the
end
to
interest
a
few
open
pr's,
but
I
will
be
discussing
indeed
in
depth
about
the
end-to-end
test
that
we
have
and
adding
more
end-to-end
tests
in
tomorrow's
ci
call,
as
I
mentioned
so
for
people
who
are
interested
in
github,
workflows,
end-to-end
tests
and
all
these
things
please
attend
tomorrow's
call
for
more
details.
B
B
There
have
been
just
any
number
of
issues
with
helm
in
general
and
then,
and
one
of
the
outstanding
items
is
the
fact
that
yeah
is
the
fact
that
there's
a
duplicative
set,
and
that
might
just
be
the
limited
knowledge
that
that
we
generally
have
toward
helm
or
lack
of
capability
of
helm
to
to
let
you
judiciously
determine
which
charts
you
want
to
deploy
anyway,
it's
causing
issues
so
hopefully,
hopefully
serene.
Has
it
figured
out
good
next
on
the
agenda
was
an
update
from
shin
and
hirsch.
B
I
don't
know
that
hersh
is
on
shin,
sent
an
update,
so
so
the
two
of
them
have
been
working
within
the
cncf
lab
they've
been
they've,
automated
the
scheduling
of
performance
tests
and
those
performance
tests
and
the
results
are
being
pushed
to
the
smp
spec.
The
service
mesh
performance
dashboard,
which
you
can
see
here.
B
B
H
So
what
have
I've
done
is
like
connected
the
pages
of
performance
for
five
pages
and
patterns
pages
from
the
back
end,
so
they
are
importing
them
from
the
server
itself.
Before
that
I
was
using
a
sample
data
for
for
testing
them
out
and
yeah,
and
I
I
also
made
like
a
drop
down
for
you
know
for
the
test.
H
F
D
So
pretty
much
I
have
been
trying
to
work
on
the
performance
page
so
far
and
instagram.
Right
now
and
I
know
I
know
he
doesn't
really
like
to
see
a
design
like
this,
so
very
ugly
at
the
point
so,
but
I'm
trying
to
see
if
I
get
the
ui
to
look
as
I
live
like
what
it
is
in
the
current
korean
master
branch
then
try
to
improve
or
extend
from
what
yash
has
done
to
make
sure
that
the
data
is
coming
from
the
back
end.
D
So
there
is
nothing
going
on
at
this
point
in
the
data
it's
literally
dominated
from
or
left
or
around
in
the
performance
page,
but
then
I'm
actually
working
on
it.
I
had
some
hiccup
given
the
fact
that,
and
you
have
to
figure
out
on
how
the
components
connected
to
each
other.
It
was
quite
too
many
to
have
to
install
installing
some
other
dependencies
that
is
not
currently
installed
in
the
structuring
branch.
D
Okay,
so
what
I
was
saying
is
that
I
said
trying
to
fix
this,
like
I've
been
like
some
installing
some
other
dependencies.
That's
like
packages
that
hasn't
been
installed
yet
as
well
as
trying
to
figure
out
how
the
components
are
connected
to
each
other,
and
so
literally,
there's
no
data
moving
in
this
page.
D
At
this
point
so
but
then
I
spoke
with
yash
and
he
said
somehow
he
has
risen
some
methods
that
I
could
also
use
in
the
in
the
performance
page
as
well,
so
that
we
could
connect
both
the
performance
page
and
the
profile
page
later.
So
that's
it!
So
that's!
That's
all
for
the
update
I
have
for
this
page.
At
this
point.
D
That's
great:
is
there
something
I
should
be
looking
forward
to
like
there's
something
I
should
concentrate
on
like
some
issues,
you
feel
I
probably
might
approach
right.
I
might
meet
on
the
way
so
maybe
from
now
I
could
start
thinking
about
it.
There's
some
challenge.
Currently,
I'm
asking
much
relaxing
you.
B
Yeah,
no,
I
I
don't
recall
there
being
a
challenge
with
that
component
in
the
past.
Just
you
know
for
everyone's
to
refresh
your
memory
or
if
you
haven't
seen
that
particular
component
before
it's
a
it's
a
calendar
view
of
performance
results
so
for
tests
that
that
people
have
run.
I
think
I
have,
I
think
I
have
measuring
up
and
running
here
I'll.
B
B
I'm
not
sure
when
the
last
time
this
user
ran
one
okay.
Here
we
go
so
they
ran
some
on
this
day
in
concept.
Maybe
they
ran
10
total
wait.
Let's
do
that
again,
so
seven
more
yeah
yeah
that
makes
sense
so
and
what
users
are
able
to
do
is
just
kind
of
quickly
call
up
the
test
results
for
individual
test
runs
that
were
performed
at
that
day
time.
B
The
reason
that
I
mentioned
this
is
that
it's
future
there's
api,
there's
back
end
capability
for
scheduling
a
performance
test
in
the
future,
but
none
of
that
in
the
ui,
so
something
that
we
did
to
keep
in
the
back
of
your
mind,
leonard
as
you
get
to
this
area.
Is
that,
ultimately
it
we
will
ideally
want
for
people
to
be
able
to
schedule
their
performance
tests
using
this
calendar
interface.
F
C
B
H
C
H
B
B
H
H
B
Okay,
yeah,
okay
yeah,
it's
good
good
to
mention
so
you're,
saying
hey!
You
might
be
seeing
that
error,
because
the
message
server
that
you're
running
incorporates
changes
outside
of
the
master
branch
yeah,
very
good,
very
good!
That's
why
yeah!
That's
why
I
didn't
have
an
immediate
answer,
because
it
would
just
be
surprising
if
that
was
yep
makes
sense,
all
right
any
other.
So
so,
if
you
didn't,
if
this
wasn't
obvious,
there
is
kind
of
an
implicit
call
to
action
for
those
that
are
into
ux
or
those
that
are
into
product
design.
B
Has
a
particular
ui
and
it's
served
the
project.
Okay,
thus
far,
it's
time
for
an
overhaul,
both
in
terms
of
the
technical,
the
frameworks
that
are
used
to
build
the
ui,
but
also
it's
time
for
an
overhaul
like
large,
much
more
largely
of
the?
U
of
the
user
flows,
the
ux
itself
and
if
you're
here
with
a
ux
orientation
that
might
sound
compelling
to
you
and
exciting,
and
I
hope
it
does.
B
What
they're
going
after
the
user
flows,
what
they
need
to
do,
how
they
need
to
accomplish
it,
how
quickly
they
could
do
that
if
there's
a
whole
headspace
around
that,
I
totally
encourage
all
of
you
to
jump
into
that
recognize
that
there's
very
few
maintainers
compared
to
the
number
of
contributors
and
so
they'll
all
do
their
best
to
kind
of
answer.
Answer
your
questions
and
feed
you
info
as
we
go,
but
so
yeah
there's
a
real
challenging
opportunity
set
out
there
about
that
redesign.
B
B
B
C
I
hope
it's
possible
now,
okay,
so
for
password
like
for
one
week,
I
have
been
working
on
on
creating
the
design
spec
for
the
merch
recital
app
input
sub
command.
So
if
you,
if
you
guys
ask
like
what's,
the
intention
of
this
command
is
like,
if
you
have
some
of
the
application,
manifest
or
configuration
files,
and
you
need
to
import
them
onto
the
machine
but
not
deploy
them,
then
this
command
is
intended
to
sell
such
purposes.
C
It's
like
an
generic
algorithm
so
that
some
guys
can
understand
what
will
happen
and
how
to
how
the
execution
will
happen,
and
I
also
added
the
flowchart
like
data
on
like
behind
the
scenes
scenario
like
what
are
all
the
process
should
mastery
ctl
will
execute
and
if
there's
is
there
any
process
in
which
mess
measures
server
would
so
as
well
as
in
machine
cloud.
So
this
one
is
still
work
in
progress,
though,
but
somehow
after
receiving
this
suggestions,
the
folks
give.
C
B
Something
like
this
would
certainly
clarify
for
serene
why
there's
two
sets
of
charts
and
when
one
gets
used
versus
the
next,
and
so
I
think
of
I
mean
so.
I
know
that
this
flowchart
is
specific
to
this
particular
command,
but
I
would
definitely
yeah,
but
but
many
aspects
of
it
are,
you
would
tend
to
be
universal
and
and
yeah
bbs.
It's
just
a
very
helpful
reference.
I
think,
like
a
generalized
version
of
this
would
be.
You
know
very
helpful
to
have
in
the
mastery
ctl
contributor
guide.
B
B
Relatively
straightforward,
set
of
behavior
that
is
supposed
to
happen,
is
basically
a
subset
of
what
happens
today.
Today,
people
might
use
the
command
meshery
ctl
app
onboard
that
would
that
would
take
a
file,
a
helm
chart
or
a
docker
compose
file
or
kubernetes
manifest
ingest
that
definition
of
the
application
as
the
first
step
and
the
second
step.
It
would
then
also
deploy
that
application
and
what
you're
doing
is
saying.
B
B
That
probably
doesn't
need
to
be
so
long.
You
should
just
start
you
should
go.
I
I
don't
know
that
I've
ever
said
that
to
anyone
here,
he's
like
oh,
that's,
too
much
spec
but,
like
honestly,
all
you're
doing
is
putting
a
break
point
right.
So
if
there's
something
different,
then
that's
reason
to
write
it
down.
If
not.
G
The
source
type
support
in
machine
ctl,
app
on
board
and
in
quite
a
hurry,
so
that
that
command
is
also
not
very
well
tested
and
I'm
very
pretty
sure
that
there
is
a
bit
of
hard
coding
going
on
there.
So,
while
you
are
working
on
this
also
take
that
existing
command
into
consideration
and
how
that
works
with
this
one
and
because
improving
a
little
bit
of
logic,
there
would
ultimately
help
you
here
as
well,
because
you
know
at
the
end,
we
don't
want
to
duplicate
logic.
G
G
So
the
job
of
import
will
be
to
take
your
case
manifest
or
a
helm
or
anything
and
then
convert
that
into
a
because
we
in
our
central
database.
We,
the
format
in
which
we
store
data,
is
a
pattern
file.
So
at
the
second
stage
the
deployment
stage.
We
already
have
the
logic
for
that
and
we
already
have
a
lot
of
logic
for
importing
as
well
the
source
type
thing.
What
raj
had
so
yeah.
I
think
there
is.
G
G
When
you
implement,
when
you
implement
that,
just
make
sure
that
code
is
readable
in
a
way
that
it,
it
basically
speaks
that
okay,
this
is
the
this
is
where
we
are
importing,
and
this
is
the
logic
should
be
inside
of
inside
of
a
package.
It's
it
should
be
there
and
then
it
can
be
reused
across
different
commands
or
sub
commands,
so
the
logic
for
importing,
along
with
source
type
and
all
in
the
logic
for
deploying
that
we
already
have
so
just
make
sure
that
you
don't
rewrite
any
code.
That's
what
I'm
saying.
B
I
was
just
gonna
kind
of
repeat
myself
and
say
how
wonderful
it
is
to
have
that
diagram
and-
and
I
think
that
you
should
it's
fantastic
to
have
it
in
this
dock
specific
to
this
command.
It
drives
a
ton
of
clarity.
I
I
highly
encourage
you
to
pull
it.
Do
you
have
a
generalized
version
in
well
whether
it's
in
the
contributor
guide,
which
is
great,
which
would
help
people
comprehend
the
general
flow?
B
But
then
also,
I
suppose
it's
just
the
contributor
guide
it
might
make
sense
to
have
in
some
of
the
inside
of
some
of
the
commands,
specifically
maybe
a
a
more
simple
version,
because
you
can
imagine
some
users
may
not
quite
comprehend
what
what
all
transpires
when
they
do
the
onboard
command,
for
example,
if
it's
taking
multiple
steps
and-
and
so
anyway,
these
types
of
diagrams
are
really
really
great,
fair
enough
very
good.
So
I
think
mr
chatterjee
aditya
is
your.
Are
you
next
on
the
agenda.
C
B
Yeah,
in
essence,
like
I
don't
yeah,
I
am
going
to
over
simplify
again,
and
it's
really,
I
would
just
say
it's
actually
not
oversimp
oversimplifying.
It's
literally
take
a
file
from
the
take
a
file
as
input
store
that
file
okay,
understanding,
yeah
yeah
yeah,
but
that
that's
really
good.
So
the
app
command
the
filter
command
the
design
command
yep.
We
want
to
be
completely
consistent
if
we
can
with
all
types
of
content,
so
they
each
should
have
an
import.
They
each
should
have
the
equivalent
of
an
apply
apply
being
like
for
the
app
command.
C
F
E
E
A
public
design
would
mean
that
it's
a
catalog
item,
which
should
you
be,
which
you'll
be
able
to
fork
and
then
any
these
all
these
patterns
would
be
uneditable
because
they
are
bad,
they're
pre-seated,
you
don't
want
them.
We
don't
want
anyone
so
adding
inside
this
pattern.
That's
why
they'll
be
unavoidable
and
then
you
can
fork
them
and
you
can.
F
E
Forking
them
you
can
deploy
them,
but
to
make
edits
you
have
to
fold
them.
This
is
what
the
grid
view
would
look
like.
This
would
have
a
description,
the
caveats
and
considerations,
and
it
would
also
show
which
compatible
with
these
designs
and
yeah
so
to
differentiate
between
catalog
items
and
other
and
your
personal
designs.
We'd
also
add
a
filter
here
which
would
hide
the
catalog
item
so
that
you
can
only
view
your
items
and
yeah.
That's
about
it
there's.
This
is
just
a
first
iteration
that
we've
made.
E
I
see
a
lot
of
duexes
on
this
call,
if
you
guys
can
suggest.
Can
you
know
make
this
better
if
you
have
any
suggestions
on
how
to
make
this
design,
but
please
feel
free
to
do
so.
The
link
is
in
the
minutes,
but
I'll
leave
a
link
in
the
chat
as
well
yeah.
So
any
thoughts
and
suggestions
that
people
would
like
to
give.
E
E
F
E
B
B
B
They
then
want
to
share
those
designs
with
everyone
else
to
use
and
when
they
do
they'll.
Those
designs
are
published
in
that
public
facing
catalog
on
out
here
on
the
website
great.
If
these
are
highly
curated
and
well
done
and
and
made
to
be
a
little
generic,
they
can
then
be
imported
from
the
website
into
measure
ui
into
the
application
as
you're
running
it.
B
C
B
B
B
C
G
C
C
C
They're
inside
the
comic
book
also,
but
three
commuters
can
be
suppressed
by
the
contributor
and
that's
where
some
of
the
you
know
checks
get
missed
by
the
reviewers
so
adding
it
here
is
you
know
it's
not
under
control
of
anyone,
so
it
will
run
anyways,
so
it
can't
be
suppressed.
So
that's
why
it
kind
of
provides
the
actual
status
of
the
full
request.
C
A
A
This
isn't
super
well
formed
like
I
didn't.
I
didn't
prep.
You
know
questions
for
I
just
in
looking
at
the
demo
and
watching
checking
out.
I
just
got
into
figma
as
well,
which
is
really
awesome.
I
guess
I
had
two
questions,
one:
when
a
user
imports
a
design
or
when,
when
a
user
imports
a
pattern
into
their
design
or
to
that
you
know
they
start
a
design
from
a
pattern.
Is
it
sort
of
like
one
of
those
old
visual
studio
like
one-way,
wizards
right
where
it's
like?
A
It's
a
you
know
it's
basically
a
copy
and
then
that's
it
or
is
it
more
like
a
forking
and
tracking
the
delta's
kind
of
thing
so
like
if
they
wanted
to
push
changes
back
or
iterate
on
the
pattern
and
then
propose
it?
Is
that
a
workflow
that's
here
and
then
the
other
one
I'll
just
mention
just
so,
because
so
I
could
just
not
not
go
into
one
too
deep,
but
the
other
one
was
when
we're
when
you're
building
designs,
you
usually
go
in.
A
You
know
through
that
browser
you
showed
and
you
could
choose
what
kubernetes
version
that
you're
pulling
the
definitions
from
you
know.
Once
there
is
a
design,
are
there
any
workflows
either
you
know
doing
it,
sort
of
back
page
backspace
style.
You
know
where,
where
where
somewhere,
there's
like
a
github
action
running
or
there's
some
sort
of
ci
running,
because
it's
backed
by
a
git
repo
or
is
there
some
workflow
within
the
tool
that
might
test
the
various
versions
of
kubernetes?
A
B
F
A
Service,
you
know,
like
the
open
question
like,
should
the
ui
tell
me,
oh
by
the
way
you're
pulling
in
something
that
in
six
months
is
gonna
evaporate
right
or
something
like
that
like?
But
again
I
I've
been
digging
into
the
docks
and-
and
you
guys
knows
you
all
built
so
much-
it's
awesome
that
there's
so
much.
I
still
don't
know
but,
like
you
know,
one
way
to
do
it
is
kind
of
like
their.
You
know
when
they
make
a
new
design
like
if
that's
backed
by
a
git
repo
somewhere
or
something
like
that.
A
B
Yeah,
so
let
me
answer
that
last
one
first,
because
it's
probably
the
shortest
which
is
well,
which
is
so
I
just
imported
from
the
catalog,
this
istio,
this
spire
configuration
and
and
loaded
it
into
the
map.
Here,
if
I
load
oh
like
obviously,
there
might
be
a
bug
going
on
here.
The
answer
to
your
last
question
is
that,
yes,
that
you
can
take
and
merge
multiple
designs.
B
If
you
take
a
design
and
if
you,
if
you
click
it
it
loads
just
that
design,
if
you
click
and
drag
another
design
that
this
right
here
may
or
may
not
work,
because
what
I
was
doing
was
combining
the
same
design
on
top
of
itself,
and
so
so
that's
why
this
didn't
work,
but
you
could
otherwise
yeah
take
and
combine
designs
by.
You
can
do
a
merge
in
that
regard.
B
G
Yeah,
I
also
just
wanted
to
toot
my
honda
that
on
the
back
end,
actually
we
have
a
support
for
importing
patterns.
So
what
I'm?
You
can
recursively
have
designs
inside
of
designs
inside
of
designs,
so
that
you
can
have
a
compact
design
where
one
service
just
does
seo
installation
and
which
actually
points
to
a
whole,
another
design
which
has
all
those
things
and
then
another
service.
G
Let's
say
it
does
something
else:
it
installs
book
info
app,
so
they
actually,
if
you'll
go
inside
of
service
mesh
patterns
like
service
mesh
patterns
and
you'll
use
machine
pattern
apply.
You
can
actually
deploy
those
things
and
those
things
do
work
so
yeah
that
we
do
have
that
thing
at
the
core
that
that
was
one
point.
C
A
G
Actually,
at
the
back
end
how
we
do
it
is
that
if
you
right
now,
you
only
support
by
url.
So
if
you
refer
to
something
by
url,
it
will
fetch
all
those
services
from
that
and
then
expand
those
into
this
new
patent
into
this
new
design.
And
if
you
reference
that
in
something
else,
then
it
will
take
all
these
services,
so
it'll
kind
of
do
a
bottom-up
resolution
of
all
these
services.
G
Until
the
point
that
you
have
a
new
whole
design,
which
has
all
these,
like,
let's
say,
100
services,
but
on
on
your
screen,
you
won't
be
able
to
see
that
you
will
be
seeing
you
know
like
one
service
per
and
that
one
service
would
be
pointing
to
that
one
single
design
internally,
it
would
be
like
100,
it
may
be.
100
services.
A
B
A
Parameterizable,
like
yeah,
you
give
it
okay,
yeah
yeah
with
inputs
and
outputs.
So
so
really
you've
got
all
the
art.
You've
already
got
the
rudiments
to
make
a
general
purpose,
composability
mechanism
that
could
be
arranged
as
a
filter
graph
or
that
could
be
arranged
in
other
other
kinds
of
places.
That's
cool,
okay
and.
B
Yeah
there's
a
couple
of
ways
to
answer.
I
think
the
way
in
which
so
there's
there's
some
good
news
with
respect
to
there's
a
and
actually
it's
like
somewhat
of
a
similar
approach,
using
github
workflows
to
have
a
matrix
of
github
versions
through
which
you
know
every
every
pr.
Every
pr
goes
through
this
you
know
set
of
tests.
Part
of
the
end
to
end
tests
that
occur
are
like
provisioning
of
one
of
the
service
meshes
provisioning
of
a
sample
app.
A
G
Yeah,
actually,
if,
like
you,
have
a
conversion
of
kubernetes
and
you
load
up
a
design
that
is
not
compatible
with
that
version
of
communities,
you
will
get
validation
errors
there,
that
this
version
of
communities
actually
has
deprecated
this
thing.
So
you
can't
use
it.
So
we
will
actually
throw
those
address
at
you,
while
you're
interactively
doing
stuff
or
like
that
summary
page,
where
you
can
kind
of
like.
A
I'm
kind
of
thinking
like
what
happens
when
I
use
this
for
a
while
right
so
like
now,
I'm
I'm
running
my
stuff.
You
know
this
is
awesome.
I
can
click
drag.
I
can.
I
can
make
my
own
catalog
of
things
I
can
even
curate
them
so
that
I
can
make
patterns
that
other
people
can
use
if
I'm
moving
to
kind
of
cappy
and
cluster
as
a
service
for
my
teams
teams
being
engineering
teams
that
are
my
my
customer
in
this
sorry,
I
put
it.
A
A
At
the
same
time,
it's
kind
of
thing
when
I
saw
that
like
I
saw
that
we
don't
have
that
we're
still
doing
the
design
for,
like
the
top
level
like
it
seems
like
there's
some
sort
of
operational
view
that
might
be
like
okay.
Well,
this
pattern
needs
to
be
revisioned,
so
I'm
gonna
have
to
communicate
that
internally,
using
it.
So
you've
got
this
like.
G
Yeah,
actually
we
have.
We
have
an
end-to-end
test
which,
if
like
you,
can
share
the
the
compatibility
matrix
which
actually
runs
test
against
multiple
communities
versions.
So
we
have
right
now
few
patent
files
like
so
we
have
basic
skew
mesh
installation
booking
for
installation
and
all
those
things
and
we
actually
caught
one
thing
that
osm
had
deprecated
in
a
previous
humanities
version.
A
B
A
Yeah
like
this
is
exactly
what
I
meant
like
it's
totally
cool
to
do
it
on
the
design
side
like
we
don't
let
bad
things
in
but
like
over
time.
Things
will
rot
right
because
something
might
move
from
like
optional
to
required,
right
and
so
like
now,
you've
got
like
a
persistent
pattern
that
used
to
work,
but
now
isn't
gonna
work
right.
A
To
that
point
to
allow
interactive
fixing
but-
and
then
once
that's
done
like
as
as
somebody
running
a
bunch
of
patterns,
I
would
almost
be
like
hey.
Can
I
do
a
git
log
you
know?
Can
I
see?
Can
I
see
what
changed
where
because
I'm
gonna
have
to
like
again
as
an
operator,
I'm
gonna
have
to
correlate
that
with
my
internal
systems
that
my
workflows
for,
like
cio
and
devsecops,
and
threat
modeling
and
all
that
stuff.
A
My
change
management,
rather
my
enterprise
change
management
will
have
to
you
know
somehow
have
a
linkage
or
my
cio
freaks
out,
because
I
make
a
manual
workflow
for
his
team.
So
these
are
kind
of
I'm
kind
of
thinking
about
all
of
this.
Don't
I'm
not
criticizing,
I
think
I'm
totally.
I
can't
wait
to
like
dive
in
deeper,
but
like
I'm
kind
of
coming
at
it
from
I'm
running
kubernetes
at
scale,
mindset
and
just
kind
of
like
surfacing
things
here
that
are
still
at
design
time.
B
Maybe
it's
the
circuit
breaking
one.
It's
just
used
all
the
time
that
there's
an
implicit
expectation
for
I
would
think
from
users
that
come
by
that
drive
by
the
catalog
and
want
to
use
something
that
that
they'll
be
taking
care
of
that
the
latest
kubernetes
is
supported,
or
that
like
and
that's
an
ongoing
validation
that
occurs.
So
it's
there's.
G
Yeah
and
and
also
I'd
like
to
point
out
that
in
mystery
catalog,
we
have
caveats
and
consideration
that
can
be
even
more
dynamic
that
you
think
so
we
can
have
ci
workflows,
running
for
multiple
version
of
kubernetes
and
those
caveats
and
considerations
might
point
out
that
hey
this
works
with
this
version,
and
this
works
with
that
version
and
it's
not
manually
entered
it's.
The
ci
workflow
runs
that
the
design
for
for
as
a
cron
job
for
that
version,
and
if
it
fails
it,
it
is
marked
in
the
caveats
and
consideration.
G
C
C
A
B
With
each
other
yep
very
much
so
that
with
each
other
to
get
you
know,
you
know,
I
don't
know
how
like
on
one
hand,
there's
this
thought
that
maybe
like
we
would
matt.
This
is
actually
kind
of.
This
is
a
I
think,
you'll
have
it's
good
to
plant
the
seed
in
your
mind,
because
this
might
be
something
that
you
form
some
thoughts
on
and
opinions
on
is
like
one.
B
One
thought
here
is
to
go
fork,
artifact
hub
and
like
and
build
it
into
much
more
of
a
like
broad
set
of
different
types
of
content
that
are
supported
and
who
did
it
and
ongoing
and
like-
and
this
is
more
than
the
well
for
forking-
isn't
something
we
necessarily
desire
to
do,
but
it's
also
heavier
and
more
than
the
community
can
probably
undertake
at
the
moment
part
of
the
thought
just
like
different
thoughts
that
were
out
there
was
like.
Oh
maybe
we
should
have
a
measuring
pattern
as
an
entity.
B
That's
first
class
supported
over
here.
Another
train
of
thought
had
been
yeah
that
we
might
take
is
take
something
like
gideo.
If
I'm,
which
is
an
mit
licensed
get
bait
like,
is
to
take
this
and
build
it
as
a
lightweight
gold
bait
golang
based
git
system
internal
to
mesherie
such
that
yeah,
there
is
implicit
versioning
happening
to
help
satisfy
part
of
you
know
what
what
you
were
articulating
and
then
yeah.
There's
really
that
that
collaboration,
sharing
and
ultimately
kind
of
kind
of
you
know
like
permissions
around
this
right
now.
B
A
I'll
go
make
an
issue
for
that
and
I'll
say
also
in
closing
on
a
bit
about
artifact
hub,
I
think
is
it's
worth
considering
like.
If
you
look
at
the
crosstalk
plane
project,
you
know
from
the
get-go,
they
said:
okay,
well
we're
going
to
distribute
our
cross-plane
patterns
if
you
will
or
compositions
as
oci
images,
so
now
like
anyone's
existing
infrastructure,
whether
it's
harbor
or
quay,
or
what
you
just
put
up,
you
know
they're
all
easy
to
to
version
and
distribute
as
oci
container
as
oci
images.
So
it's
like
their
artifact
format.
B
It
packages
up
ebpf
stuff
as
oci,
does
eb,
okay,
yeah
yeah
well,
fair
enough,
so
we
are
we're
over
time
and
in
the
best
of
ways,
I
suppose
good.
If,
if
you
had
items
today
that
we
weren't
able
to
cover,
please
don't
hesitate
to
just
do
a
quick,
recording
and
drop
that
into
slack
or
drop
it
onto
the
discus
forum.