►
From YouTube: Meshery Development Meeting (July 6th, 2022)
Description
Meshery Development Meeting (July 6th, 2022)
Agenda:
[Pranav] Fix formatting in Pattern configurator PR#5700 PR#188
[Ruturaj] Fix app onboard silent fail #5691
[Jared] Update on mesheryctl mesh command #5375
[Ashish] Issue of unhandled malformed config in mesheryctl #issue
[Ashish/Abhi] GraphQL subscription: Design Spec, Pull Request #5686
[Harkirat] Mesheryctl user friendly error message Pull Request #5661
[Uzair] Add undeploy in Filters, deployment modal in missing components, false positive snackbar #5765
[Ashish] Meshery Applications design spec
[Ashish] Help needed on propagating this PR across adapters.
[Alpha] Discussion on `mesheryctl app import` and `mesheryctl filter import` (Issue #5766 and #5767)
[Abhi] Fix problems with context switcher click
[Abhi] Fix All the failing tests and need for moving with cypress tests
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/
A
B
So
the
problem
was
that
actually,
when
we,
when
machine
register
the
workload,
then
it
gets
all
the
schema
for
that
and
then
it's
converted
into
our
json
form,
and
then
we
see
it
in
the
ui.
So
the
problem
was
that
we
were
like
formatting
it
twice
once
in
the
back
and
then
second
time
in
the
front
end
because
of
it.
B
A
By
the
way,
just
a
little
feedback
on
this
thing,
it
would
be
nice
to
have
like
a
dick.json
and
instead
of
like
loading
it
here,
we
can
have
a
json
file
somewhere,
which
we
can
edit
and
basically,
that
dictionary
that
map
strength
string.
It
can
be
basically
unmarshalled
from
that
json,
so
that
things
are
a
little
bit
more
configurable
and
out
of
the
as
much
as
out
of
code
as
possible.
So
if
someone
else
wants
to
add
something,
they
just
have
to
enter
a
new
key
value,
but
into
that
json.
A
C
Yeah
yeah,
by
the
way,
this
is
a
good
discussion
just
for
I
can't
help
but
comment
that
for
those
that
write
in
a
particular
programming
language,
even
though
there
are
those
that
feel
comfortable
with
the
language-
and
there
are
those
that
don't
even
for
json
or
yaml
or
markdown,
there
are
those
that
feel
comfortable
with
that
and
those
don't
and
you
would
think
even
for
mark
it's
true
of
markdown
even
is
just
some
people
anyway,
the
there's
a
certain.
So
so
anyway.
C
C
C
You
know,
I
don't
know,
and
they
add
qos
on
requests.
Is
they
have
this
new
quality
of
service
word
that
doesn't
fit
the
formatting
very
well,
and
so
you
know
a
custom
thing
needs
to
be
added,
and
so
does
that
go
into
a
central
dictionary
or
does
that?
Is
that
potentially
relegated
you
know
is
there?
C
I
don't
know
that
this
is
the
place
where
we
want
to
spend
tons
and
tons
of
time
on,
but
but
another
piece
of
food
for
thought
that
what
I
was
trying
to
say
is
that
individual
users
might
want
to
customize
the
dictionary.
Probably
the
first
use
case
is
just
adapters
as
opposed
as
opposed
to
users,
but.
C
D
C
D
C
Okay,
okay,
I
was
just
gonna
assign
myself
like
two
points
like
two
bonus
points
for
making
the
right
guess
as
to
yeah.
That
was
awesome.
How
did
you
get
okay,
okay,
good?
Okay,
I'm
gonna
keep
my
secrets
to
myself
because
those
those
those
bonus
points
are
hard
to
come
by
so
anyway.
You
just
did
it
by
the
name
anyway,
all
right.
Fine!
I
cheated
anyway
nice
to
have
you
here,
but
by
the
way,
so
jacob
is
your
last
name
and
you
see
the.
D
C
E
E
Okay,
like
I
want
to
work
on
the
back
edge
of
measuring
and
so
far
I
fixed
three
issues.
Personally,
I
think
I'm
new
to
the
cloud
world
yeah
before
measuring.
I
contributed
a
few
things
to
to
a
kubernetes
based
integration
platform
called
kmuke,
so
I
hope
to
learn
more
about
golang
and
service
mesh
here
and
contribute
as
much
as
I
can.
E
C
C
F
Yeah,
hello,
everyone
myself
manik
from
india
and
I'm
currently
studying
in
my
fourth
year,
which
is
a
pre-final
year,
I'm
doing
an
integrated
course
b-tech
process
in
iit
from
triplett
and
I'm
a
full
stack
developer.
I
have
experience
of
one
and
a
half
year
in
full
stack
developer
and
right
now,
I'm
studying
a
little
bit
devops
so
go
kubernetes
and
docker,
and
I've
solved
two
pr's
on
mercury
and
currently
made
up
sorry.
Two
issues
solved
two
issues
on
mystery
and
currently
created
a
pr
on
one
more
issue.
F
C
C
G
G
G
So
it
was
a
recommendation
from
a
friend
and
how
I
could
take
my
goal:
golang
skills,
as
well
as
basically
teaming
skills
to
the
next
level.
So
yeah,
that's
why
I
was
like
hey,
okay.
I
think
this
is
the
right
place
to
be,
and
it
looks
like
it
is
the
right
place
to
be
so
I'm
looking
forward
to
contributing
and
yeah
doing
my
best,
oh
and
then
I'm
just
a
little
bit
of
my
background.
G
So
I
I
I
kind
of
have
like
about
three
years
of
experience,
but
then
I'm
majorly
in
it
has
been
qa,
but
then
along
the
way,
it
has
been
react.
Redux
not
js
express,
and
since
I
was
like
in
qa
for
quite
a
while,
I
was
like
okay.
Now
I
want
to
go
back
to
programming
and
I
feel
I
want
to
go
with
golang
and
I
just
feel
like:
okay
yeah.
C
G
C
C
A
H
H
I
think
couple
a
week
ago,
so
the
problem,
what
the
problem
that
was
there
was
it
was
not
so
the
deploy
endpoint
the
api
deploy
endpoint
actually
expects
you
to
send
it
a
pattern
file
to
actually
deploy
the
application.
So
we
need
to
first
send
it
send
the
application
file
to
slash
pattern
which
will
respond
back
with
an
auto
generated
pattern
file,
and
then
we
send
that
pattern
file
to
slash
deploy
and
then
it
deploys
earlier.
H
We
were
just
sending
the
application
file
to
slash
deploy,
which
was
not
worth,
therefore,
silently
failing
so
yeah.
This
pia
just
first
fetches
the
pattern
file
and
then
sends
out
the
pattern
file
yeah.
So
I
think
she
reviewed
this
just
a
couple
hours
ago
and
yeah.
His
comment
like:
why
can't
we?
Why
can
we
not
support
urls?
So,
just
to
answer
that
this
is
an
off
board.
A
A
H
A
Okay,
so
assuming
you
can
see
my
screen,
so
this
was
the
issue
was
in
mystic,
cpn
and
so
the
issue
to
describe
the
issue.
In
short,
it
was
that
this
was
my
mesh
config
and
when
I
use
this
mesh
config
for
any
command
like
in
my
case,
I
was
using
michael's
ctl
app
on
board
something
yeah
I
was
using
mesh
deploy,
so
it
should
have
given
me
a
descriptive
error,
which
should
have
been
reflective
of
the
fact
that
the
gamma
itself
is
malformed.
A
So
if
I
tried
to
gamma
dot
unmarshall
this,
that
would
result
in
an
error,
but
I
checked
and
viper
dot
unmarshall
doesn't
result
in
an
error.
So
what
happens
as
a
result?
Is
it
actually
unmarshals
it
into
the
struct,
and
so
the
struct
is
empty
so
because
the
structure
is
empty,
current
context
is
also
empty.
So
the
result
is
current
context
not
set,
but
that
technically
is
not
that
it
shouldn't
say.
Current
context
is
not
said
because,
instead
of
yaml,
the
current
context
is
set.
A
A
A
Cool,
so
there
were
a
bunch
of
people
who
were
interested
in
this
thing
and
I
think
santosh
sent
some
screenshots
on
slack
as
well,
but
I
don't
think
I
was
able
to
convey
the
issue
properly
to
him.
So
if
someone
else
can
or
if
someone
else
can
take
up
this
issue,
then
I
will
assign
that
to
you.
If
you
show
an
interest
or
I'll
assign
to
the
people
discussing
this
thing
in
the
comments
itself,.
A
Just
one
sec
yeah,
so
here
it
is
so
this
issue.
Sorry,
this
pr
is
about
the
graphql
subscriptions
optimizations.
We
were
earlier
using
a
lot
of
channel
which
we
have
reduced
with
this
pr.
So
mostly
ashish
has
worked
on
this
pr
and
it
solves
some
issues
regarding
graphql
or
channel
inconsistencies,
while
listening
to
the
operator
as
well
as
meshing
data,
so
I
think
it
had
some
resolve.
A
D
A
No
yeah
for
from
my
side,
it's
ready
to
go,
and
although
we
have
some
different
considerations
and
everything,
but
for
now
I
think
this
will
fix
some
of
the
fix
some
of
the
buggy
things
which,
which
are
not
these
bugs,
are
not
like
very
prominent,
but
rather
intermittent
some.
Maybe
sometimes
you
would
see
them
or
not
because
of
the
channel
inconsistencies.
A
A
A
K
Yeah
so
previously
to
deploy
using
the
mesh
sub
commands,
you
had
to
know
the
url
of
the
adapter
you
were
going
to
use
which
to
me
is
not
a
very
easy
way
to
do
that.
Usually,
I
think
it's
easier
to
reference
meshes
by
their
names
rather
than
having
to
remember
where
an
adapter
is
especially
if
it's,
if
you're
new
to
measuring
ctl
or
I
mean
measuring
in
general,.
K
K
Then
it
will
deploy
so
whether
you
deploy
it
like
this,
you
have
the
dashes
in
here
or,
if
I
have
it
like
this,
with
completely
separate
words
with
mix
of
upper
and
lower
case,
or
if
I
have
it
as
a
single
string
in
quotes
with
spaces
and
all
this
that
will
parse
it
out
and
get
the
correct
name
or
you
can
select
from
a
prompt.
So
I'll
show
you
what
that
looks
like.
K
C
Yeah,
but
I'll
chime
in
chime,
in
that
this
concept
is,
is
quite
nice
in
so
much
as
the
in
so
much
as
like
the
ability
to
check
real
time
on,
like
what's
possible,
with
measuring
different
deployments,
might
have
different
adapters
or
might
have
different
applications
or
performance
profiles
or
any
number
of
things,
and
so
there's
an
aspect
of
what
of
what
you're
doing
here
jared.
That
is
not
only
facilitating
discovery
by
a
user.
That
may
not
be
aware
of,
like
all
the
options
that
they
have
and
and
what
those
options
are
called.
C
K
Right
and
ashish,
this
is,
if
you're
on
the
call.
This
is
what
you
were
asking
about
the
other
day.
I
think
so.
There
is
some
checks
to
go
on
this
validate
mesh.
It's
actually
used
for
all
three
sub
commands,
but
it's
it's
checking
that
that
the
name
is
valid.
So
if,
if
the
name
isn't
isn't
available,
I
mean
if
a
mesh
isn't
available.
K
So
there's
that
and
one
of
the
reasons
I
wanted
to
bring
this
up.
I
know
I
presented
this
a
while
back,
but
this
is
quite
a
change
lee,
like
you,
were
saying
to
the
way
a
lot
of
the
commands
work,
and
I
know
consistency
is
important,
especially
in
a
cli,
so
I
wanted
to
make
sure
that
everyone
is
understood
what
the
changes
are
and
that
it's
okay,
instead
of
just
sneaking
something
in
and
all
of
a
sudden,
the
command
is
completely
different
from
what
people
are
remember.
A
A
I
work
with
a
lot,
so
would
that
be
nice
to
we
have
a
list
command
as
well,
so
the
same
exact
list
that
you
get
on
the
machine
mesh
deploy,
but
it's
like
a
interactable
thing
would
be
nice
to
just
have
say,
measure
ct
and
mesh
list
and
get
the
same
exact
list
so
that
in
cio
workflows.
If
you
are
dealing
with
this
thing,
we
can
use
that
command
instead
of
yeah.
A
You
know,
if
you
don't
know
beforehand,
what's
the
available
measures
and
all-
and
I
don't
think
that
would
require-
because
we
already
have
this
information,
so
that
would
be
a
little
small
command.
If
that's
not
functionally
that
much
important,
then
we
can
let
go
of
that
and
just
would
go
and
merge
this
pr,
because
the
that
was
standing.
This
was
standing
for
a
long
time
but
yeah
there
was
this
little.
I
don't
know
yeah.
A
K
K
K
H
K
K
K
K
K
H
K
K
A
K
K
K
Yeah
one
thing
I
wanted
to
point
out
with
these
commands:
so
cobra
has
some
pretty
it.
It's
pretty
advanced
has
a
lot
of
helpers
in
there
and
so
like
there's,
you
don't
need
to
pass
in
like
the
check
args.
You
have
this
this
this
function
in
here,
because
it
has
some
built-in
commands
to
actually
help
with
that.
K
C
That
type
of
a
thing
is
very
much
desired
and
is
missing
on
a
number
of
commands,
some
of
the
commands
will
be
executed.
There's
been
a
ton
of
work
on
measuring
ctl
over
the
last
two
and
a
half
years
like
to
in
in
some
areas
to
do
exactly
what
we're
saying
now,
which
is
to
avoid
things
like.
If
you
execute
a
command-
and
it
says,
error,
error
missing,
parameter,
argument:
zero,
it's
like
what
what
does
that
mean?
C
C
Granted
we
were
just
saying
it's
not
necessarily
desirable
to
give
you
full
help
every
time,
but
it
certainly
is
better
than
giving
you
some.
You
know
some
other
gobbledygook
message
that
doesn't
really
make
sense
and
and
so
for
each
one.
Hopefully
we
would
move
from
there.
One
of
the
things
that
you
just
said
about.
You
know
the
lack
of
a
need
to
do
an
argument
check
that
sounds
like
a
potential
addition
to
the
mastery
ctl
contributing
docs,
the
the
dock
that
specifically
provides
guidelines
for
mastery
ctl
itself.
K
C
K
K
K
C
K
A
J
Now
I've
added
the
error
messages
like
if
a
person
adds
something
wrong,
so
it's
like
command,
not
found
user
use,
machine,
ctl
measure
and
earlier
they
were
not
pointing
to
the
perfect
reference.
So
I
also
did
for
a
couple
of
commands,
like
I
mentioned
in
the
vr,
for
all
this
command
that
I
have
done
like
this.
If
you
do
like
this
all
this
stuff
to
my
side,
if
anyone
wanna
give
some.
A
A
A
B
So
it
is
making
the
ui
consistent
it.
The
the
grid
view
of
filters,
applications
and
patterns
were
not
having
model
and
filters
where
no
was
not
having
under
projection.
So
it
is
added
now
and
the
false
positive
of
when
we
are
deploying
the
application
or
pattern
and
then
also
it
is
showing
success
message.
It
was
because
this
data
fetch
function
was
having
the
problem,
because,
if
the
res
dot
okay
is
false,
then
this
is
executed,
but
this
doesn't
mean
the
promise
it
rejected.
B
A
Okay,
if
response.talk
is
false,
we
should
throw
the
error,
since
this
may
come
up
with
some
400
500
or
not
200
status
sequence.
That's
what
you
want
to
say
right!
Yes,
okay,
yeah!
I
think
it
was
back
of
my
head
a
while
long
ago,
so
I
just
missed
it,
making
the
fix
and
thank
you
once
again
for
fixing
it
just
just
just
as
a
follow-up.
A
B
A
B
I
A
So
I
think
we
can
move
on
to
the
next
item
messy
applications,
design
spec.
I
would
say
I
would
say
that
if
you
have
a,
if
you
have
a
shorter
item
that
we
can
get
in
for
five
minutes
and
we
can
because
it's
like
a
more
like
a
discussion
than
an
item
itself
got
you
got
you.
So
is
it
shorter
item
or
the
next
item?
For
you?
A
L
Yeah,
it's
just
a
kind
of
a
discussion
topic
though,
rather
than
the
status
update,
so
main
lee
may
know
what
kind
of
changes
are
to
be
expected
in
those
commands
and
what
must
be
exactly
done
like?
If
should
I
import
these
settings
from
the
configuration
file
or
something
or
is
it
different
from
those
aspects.
C
I
would
use
the
pattern
import
as
an
example,
and
so
right
now
for
so
we're
talking
about
measuring
ctl,
I
think
almost
all
meeting
today.
So
this
is
good.
So
we,
if
you
take
a
look
at
the
set
of
commands
that
are
available
for
measuring
ctl
app,
an
app
being
a
helm,
chart
a
kubernetes
manifest
a
docker
compose
file.
C
The
the
commands
that
are
available
today
is
measuring
ctl,
app
onboard
off
board.
Basically,
there
are
deployment
commands
and
I
don't
believe
that
there
are
just
import
commands,
so
you
could
just
load
the
application
into
meshri
as
a
system
and
leave
it
resident
there.
Like
users,
don't
have
that
choice.
The
commands
that
are
available
today
can
do
two
steps
they
import
and
deploy
at
the
same
time
and
so
yep.
What
we
want
to
do
is
create
a
app
import,
and
then
I
thought
there
was
a
second
command
imported,
something
else.
L
A
That's
a
great
question,
so,
let's
move
on
to
my
status,
update
about,
we
have
some
problems
with
the
context
feature
earlier
due
to
the
context
reflecting
or
some
code
reflecting.
Meanwhile,
you
know
underway,
so
we've
just
missed
the
you
know.
The
click
operation
was
not
getting
fired,
it
was
not
checking
the
status
of
whether
the
docker
desktop
or
whatever
the
cluster
you
have
right
now
in
your
machine
is
connected
or
not.
So
I
think
that
is
that
is
fixed
and
you
can
see
it
in
live.
A
Second
was
fix
all
the
failing
tests
with
cyprus
tests,
so
this
is
just
to
grab
the
attention
of
everyone,
especially
the
maintainers
and
contributors,
that
all
the
e
tweeters
are
for
integ.
You
know
machine
are,
you
know,
passing
and
all
the
new
branches
or
all
the
new
tests,
all
the
new
pr's
should
have
all
the
tests
of
passing.
So
earlier
we
were
having
a
lot
of
tests
failing,
so
we
were
kind
of
ignoring
the
test
results
from
everyone's
pr
and
just
merging.
A
A
You
know
rodolfo
as
well
as
mario
2..
So
these
are
the
great
starting
point.
I
think
rohit
or
abhinav
I
had
talked
recently
is
working
on
some
functional
tests,
which
I
place
in
the
machine.
Ui
and
yeah,
so
would
love
to
get
everyone's
attention
there.
A
cyprus
is
quite
easy
and
you
won't
get
any
problem
regarding
it.
So
we
have
some
docs.
So
it's
advisable
to
follow
us.
A
And
meanwhile,
if
anyone
on
the
call
is
interested
in
writing
cyber
stress
so
they're,
you
know
far
more
than
welcome
than
being
fingers
anytime
in.
We
have
two
channels
on
the
slack
regarding
testing,
which
is
one
is
cyprus
testing
and
the
other
is
mastery
ca.
You
know
so
yeah.
We
we
need
the
attention
around
the
cypress
test
and
quality
inside
machine.