►
From YouTube: Meshery Development Meeting (March 16th, 2022)
Description
Meshery Development Meeting - March 16th 2022
Agenda -
- Update on added spacing for k8s component[#5226]
- mesheryctl command to load k8s context?
- Call out for helping with E2E tests.
- Efforts to create the end-to-end integration tests being created in PR 5171 (Lifecycle for active adapters):
- Question in PR about card expected not to be present for Consul Lifecycle Management page (“Apply Service Mesh Configuration”, why?).
- Team clarified that this specific “operation” is not global, so Consul should not have this card.
- I mentioned that in my local there seems to be a misconfiguration (probably Consul adapter is not enabled) which causes lifecycle cypress end-to-end test to fail with a similar output
- Also, is there a specific scope for this PR regarding tests to be added?
We will keep this as an ongoing discussion using meshery test plan spreadsheet as a reference as we brainstorm @abhishek lee.calcote@layer5.io
Meshery Test Plan - https://docs.google.com/spreadsheets/d/13Ir4gfaKoAX9r8qYjAFFl_U9ntke4X5ndREY1T7bnVs/edit#gid=0
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
Okay:
hey
welcome
everybody.
It's
wednesday
march
16th
2022,
the
measuring
development
meeting.
It's
getting
kicked
off
a
couple
of
housekeeping
items,
there's
a
few
folks
who
send
regrets
today
and
we'll
note
a
couple
of
them
down
for
those
who
aren't
sending
regrets,
but
are
here
please
list
your
name
in
the
meeting
minutes.
A
Aditya
chatterjee
is
one
of
those
individuals
a
lot
of
times,
he'll
host.
This
call
that
a
hosting
role
that
I'll
perform
today
is
rotational.
So
if
you
want
to
host
do
so,
just
just
let
me
know
just
let
aditya
know
that
one
of
the
community
managers
know
that
one
of
the
maintainers
now
yep
all
right.
Well,
we've
got
a
couple
of
topics
lined
up
today.
We
also,
as
we
get
the
meeting
rolling.
A
A
Has
has
everyone
that's
on
this
call
right
now
introduce
themselves
on
this
call
right
now,
and
I
say
it
like
that,
because
we
have
a
lot
of
different
meetings
on
different
topics
and
so
different,
there's
a
different
audience.
Sometimes
different
people
show
up,
so
it's
always
good
to
introduce
on
that
call,
so
just
to
help
prod
folks,
just
in
case
so
amman
and
md.
A
Nice,
okay,
very
good.
I
know
you
were
on
the
websites
call
this
week
as
well,
so
it
was
nice
good,
good
commentary
by
the
way.
So
that
was
a
rough
monday.
It
was
that
daylight
savings
just
doesn't
doesn't
settle
well
when
you
spring
forward
anyway,
nice
to
have
you
on
this
call.
Md
have
you
been?
Have
you
introduced
before.
A
Okay,
great!
Well,
let's
go
over
a
couple
of
things,
just
a
quick
review
of
the
agenda
and
then
and
ask
in
case
there's
additional
agenda
items.
So
she
she's
going
to
talk
about
a
couple
of
topics.
Some
things
that
we've
got
going
on
with
adapters
some
end-to-end
tests.
Mario,
has
a
couple
of
topics
in
there.
Gaurav
is
going
to
talk
about
things
that
are
coming
up
in
the
0.7
release.
There's
a
lot,
notably
around
what
ends
up
being
kind
of
a
re.
Well,
an
overhaul
of
mesherie's
ui.
A
So
he'll
talk
about
that.
Pranav
has
been
working
on
an
extension
point.
Refactoring.
It
she's
got
some
suggestions
about
capabilities
that
mesri
has.
If
md
comes
back,
maybe
we'll
get
on
talk
about
his
some
cypress
testing
that
he's
doing.
And,
lastly,
we'll
talk
about
a
forthcoming,
an
adapter,
that's
in
the
alpha
state.
A
A
A
B
So,
assuming
that
you
can
see
my
screen,
ask
that
lee,
basically
the
schema
that
communities
components
register
there.
We
have
the
basically
the
keys,
are
you
know
cluttered
like
this,
and
the
ask
was
to
have
some
spacing
kind
of
so
this
is
just
like
lee
suggested
this
kind
of
I
it
happens
at
the
source
so
when
the
components
register
themselves
they
register
with
the
spacing.
B
A
There's
just
one
one
other
comment,
and
this
was
something
that
I
don't
know
that
I
had
expressed
as
we
were
kind
of
discussing
the
challenge.
Well,
the
challenge
around
id
identifiers
and
display
names.
So
there's
one.
This
looks
pretty
great
there's
one
aspect
to
this
that
settles
slightly
uneasily,
and
that
is
that
we
are
still
in
a
situation
where
the
display
names
are
also
used,
as
the
identifiers
granted,
your
the
pr
that
you
have
up
prettifies
them
so
that
they
can
be
displayed
to
humans
and
read
by
humans.
A
A
A
A
A
B
A
Right-
and
I
know
that
that
you'll,
that
there'll
be
more
consideration
around
these
thoughts,
but
to
advance
that
thought.
It
would
be
that
this
would
be
a
universal
attribute
for
every
component,
that
it
would
have
this
additional
field
and,
as
such,
it
would
have
kind
of
special
rules
around
it
for
not
displaying
it
as
an
input
field
and
but
just
using
it
as
a
descriptor.
A
B
A
A
Just
so
so
we've
got
this
there's
kind
of
this
perpetual
challenge
and
well
in
computers
in
computer
science
about
you
know
having
an
object
and
it
having
a
name
that
name
being
unique
or
not
like.
How
do
you
reference
the
thing?
How
do
you
reference
that
object?
What
are
you
calling
it,
and
so
often
what
you're
calling
it
is,
what
we
end
up
is
what
software
ends
up
showing
to
the
users,
and
so
often
that
gets
us
software
in
trouble
one
way
or
the
other
it's
either.
A
It
gets
the
software
in
trouble
because
the
pretty
names
aren't
unique
like
who
else
can
have.
You
know
the
beginner
86
on
github.com,
only
one
person
in
the
world
right,
that's
really
frustrating,
because
that's
the
one
that
I
wanted
and
if
they
had
only
made
it
name,
spaced
it
or
give
or
allow
you
to
have
an
identifier
that
was
separate
from
your
display
name.
That
would
be
possible,
and
so
that's
the
challenge
of
the
system
like
that,
it's
like
well,
somebody
else
got
the
f5
handle
and
so
f5.
You
know
too
bad.
A
A
Right
totally
yeah
exactly
yes,
you
basically
every
time
that
you're
dealing
with
an
object.
You
basically
always
want
to
have
like
at
least
two
attributes,
an
identifier
which
is
unique
globally
and
universally
unique.
Ideally,
and
then,
even
if
your
name
spacing
that
identifier,
some
other
way,
it's
still
nice
to
have
a
single
entity,
a
single
id
and
then
metadata
or
attributes
about
that,
one
of
which
is
a
display
name
or
a
pretty
name
or
a
description
or
a
yeah
like
the
problem,
with
the
discrete
I
should
I
take
that
back,
not
a
description.
A
A
B
Okay,
can
I
move
on
to
the
next
stop
if
you
would
yeah
okay,
so
there
is
an
open
call
for
we
have
we
had
basic
end-to-end
tests
in
all
of
our
adapters,
which,
basically,
what
they
do
is
if
you
like.
If
I
go
over
this
one
end
to
intestine,
is
there
they?
Basically
they
set
a
pattern
file.
They
run
that
pattern
file
in
mesh
with
you,
know
messy
and
adapter
running
and
they
captured
the
result
and
they
update
the
docs.
So
this
is
a
three-step
process.
B
So
currently
we
are
doing
this
three-step
process
for
basic
service
mission,
installation
and
if
I
go
down
here,
we're
doing
it
for
service
mesh
with
add-ons
add-ons
like
draft
and
prometheus
and
we're
using
these
two
pattern,
files
that
you
can
see
here,
deploy.yaml
devtoy
with
add-ons
for
tablet.
We
need
to
enhance
these
tests
and
add
some
more
things
like
book
info
app.
So,
for
example,
we
already
have
a
patent
that
does
just
that.
B
So
this
pattern
and
service
mesh
pattern,
samples,
installs
your
service,
is
your
control
plane
and
then,
after
that
is
installed,
it
installs
a
book
info
app.
So
the
ask
like
the
ass
can
be
that
here,
just
like
we
have
these
the
three
steps
for
each
of
the
two
two
things
like
so
vanilla,
service,
mesh
installation
and
service
machine
solution
with
add-ons.
B
Basically,
you'll
go
with
this
pattern
and
and
then,
as
you
can
see
here,
there
are
expected
ports
so,
for
example,
when
you're
installing
with
add-on.
So
the
expected
order
is
theory
grace
and
ingress
program
prometheus,
so
you'll
have
to
check.
So,
for
example,
when
you'll
be
installing
your
book
and
for
app
the
expected
pods
would
be,
it
would
be
something
like
reviews
and
ratings
and
in
their
appropriate
namespaces.
So
the
template
is
there
and
the
pattern
file
is
there
which
to
apply.
If
anyone
is
interested
you
can
hit
me
on.
B
You
can
ping
me
on
sorry
ping
me
on
slack.
I
can
help
you
move
forward
with
this
thing
and
we
want
this
thing
to
propagate
through
all
of
the
adapters,
not
just
this
adapter
but
istio.
Adapter
is
always
a
good
start
so
that
everything
follows
from
here.
So
if
anyone
is
interested
in
writing
these
this,
basically
this
end
when
adding
stuff
in
this
end-to-end
test,
you
ping
me
on
slack
or
ping.
B
A
A
F
A
A
B
B
I
removed
it,
it
was
related
to
the
current
topic.
Basically,
the
small
thing
that
I
wanted
to
mention
is
one
caveat
was
in
basically,
if
you
want
to
test
out
something
like
booking
for
application,
anything
which
has
kubernetes
components
within
it,
then
we
would
need
the
kubernetes
components
to
be
registered
with
the
messy
server
inside
of
the
workflow.
B
But
the
thing
is
that
that
happens
when,
whenever
a
kubernetes
context
is
detected-
and
that
happens
whenever
a
request
goes
through
your
middleware
for
the
first
time
and
when
you
have
your
message:
server
loaded
up
machine
server
when
you
start
your
message,
server
and
no
request
comes
along
and
no
request
comes
and
goes
through
the
middleware.
Your
keys
contexts
are
not
loaded.
So
your
thing,
basically,
nothing
is
registered.
Basically,
the
community's
complaints
are
not
registered.
B
So
if
you
try
to
deploy
patent
file,
which
has
kubernetes
compensate,
that
request
would
fail,
although
that
request
would
trigger
that
component
generation.
So
the
I
just
wanted
to
ask
abuse,
but
I
guess
ph
is
not
here
that
is
there
an
existing
command
in
messaging,
which
can
which
I'll?
Just
basically,
if
I'll,
if
I'll,
that
command
can
help
me
to
make
a
first
request
that
goes
to
that
middleware,
so
that
you
know
I
can
trigger
that
component
generation.
B
The
first
request
I
can
go
through
that
or
the
other
hack
can
be-
that
I
make
two
pattern:
apply
requests
the
first
one
will
fail,
definitely,
but
it
would
trigger
that
competent
generation
and
after
a
while
I
go
with
the
second
thing,
but
that's
kind
of
a
hack.
So
I
wanted
to
ask
abuse,
but
I
think
he's
not
here
so
that
was
that
small
topic
at
the
bottom.
A
Now
is
that
for
okay?
Can
you
can
you
add
that
back
because
that
that
requires
some
discussion?
Actually,
the
the
sequence
by
which
you
were
just
describing
sounds
a
little
bit
off
like
like
the
ask
of
what
you're
asking
to
do
is
maybe
treating
the
symptom
and
and
the
root
cause
that
sequence,
probably
that
system
sequence
probably
needs
to
change.
A
C
Yep
so
there's
a
bit
more
details
in
the
in
the
pr,
but
basically
there's
this
so
there's
a
test
case
in
our
measuring
test
plan
spreadsheet
about
relating
to
that
specific
adapter
and
that
specific
card.
So
I'm
a
bit
confused
by
the
fact
that
we're
adding
a
test
to
make
sure
that
for
the
console
configuration
that
card
isn't
present,
maybe
just
regard
some
clarification.
But.
C
A
C
B
Yeah,
actually
I
just
removed.
I
remember
it
from
my
peer.
I
actually
I
wrote
that
it
should
not
have
applied
services
configuration
because
in
the
current
ui
it
is
not
having
it.
I
don't
know
the
reason
behind
the
adapter
is
not
supporting
the
command.
It
actually
doesn't
support
at
the
current
time,
so
applying
custom
services
configuration
so
probably
yeah,
so
once
we
have
it
so
we'll
have
it
in
the
ui
also,
but
as
leader
told
that
it
is,
it
is
something
that
is
like
required.
B
B
C
B
A
C
So
yeah,
I
think
that
the
reason
this
that's
failing
is
because
somehow
the
during
the
ci
run
it's
not
configured
correctly,
because
I
experienced
the
same
error.
But
when
I
looked
at
the
I
did,
the
site
npm
runs
eye
open.
So
I
noticed
that
even
though
the
test
thinks
it's
so
it's
doing
from
both
from
a
link
and
from
a
button
navigation,
it's
actually
not
in
the
correct
page.
C
B
So
so
so,
first
of
all,
actually
actually
there
is
a
bug
that
is
stopping
this,
so
that
is
visit
console
by
link.
So
when
you
visit
console
by
link,
it
doesn't
actually
visit
there.
So
I
think
I
demoed
it
in
the
previous
call
too,
so
that
was
one
fix
yeah,
so
once
we
did
it
so
these
two,
but
these
two
things
will
be
automatically
fixed
like
the
tests
are
like
intact
and
they're
checking
the
correct
things
so
other
than
that
sometime.
It
does
work
by
the
way.
C
So
yeah,
in
my
case,
I
think
that
the
adapter
isn't
enabled
so
that
might
be
one
part
of
the
problem
on
my
side.
So
I'll
try
again
once
I
can
make
sure
that's
enabled,
and
so
so
so
then
so
then
I
want
to
know
like
let's
say
this:
it's
configured
correctly
so
so
that
card
is
not
there
right.
So
that's
not
part
of
a
bug
that
were.
Is
that
correct?
B
C
Yes,
yes,
okay,
yes
yeah
at
least
for
now
sure,
of
course,
we'll
keep
working
on
this
yeah
I'll
support.
Abhishek
as
I
can,
and
another
thing
is
so.
I
would
just
like
to
define
the
scope
of
the
vr
like
it's
just
okay,
add
some
end-to-end
tests,
but
maybe
it
has
been
talked,
and
maybe
I
wasn't
part
of
that
discussion
or
I
wasn't
paying
that
much
attention
like.
G
C
H
B
Yeah,
so
actually,
actually
great
so
I'll,
add
the
description
by
the
way
for
the
things
that
are
working
right
now
we
can
have
you
know
we
can
try
to
just
click
on
button,
deploy
things
and
that's
how
we
would
be
going
forward.
Actually
I'll,
add
more
description
here,
like
the
number
of
items,
one
two,
three
four
pointers,
and
actually
we
can
work
together
on
this.
C
Okay
sounds
sounds
good
because
you
know
if
we're
we're
talking
about
end
to
end
test
like
the
current
one
like
it's,
not
bad,
because
it's
doing
some
checks,
but
but
I
I'm
not
sure
if
it's
covering
a
full
end-to-end
scenario,
you
know
like
yep,
but
I'm
sure
there's
more
to
it.
Right
like
what
yeah.
B
Yeah,
actually
it
doesn't
so
you
are
right
there
and
it's
it's
because
a
few
bugs
there
that
has
been
listed
here
and
we
just
need
to
you,
know
clear
it
off.
So
once
these
things
are
done.
So
we
are,
you
know
out
of
friction,
so
we
can
do
anything.
We
have
to
just
locate
the
button
id
and
just
click
on
it
and
it
will
deploy
and
just
write
the
expectations.
C
B
B
C
Implement
it
by
today,
tomorrow,
sure
I
know
there's
we
need
to
think
like
what's
the
scenario
we
want
to
focus
on
like
make
sure
we
cover
something
that
the
user
really
values.
As
you
know,
it
should
work
as
expected
and
specifically
to
this,
the
life
cycle.
So
it's
all
like
an
entire
functional
area.
So
it's
not
something
to
answer
right
now,
but
I'm
sure
we
can
keep
that
discussion
ongoing.
So
we
can
make
progress
on
that
yeah
sure,
thanks
for
getting
invited
sure.
Thank
you
thanks.
A
A
Yes,
yeah
yeah,
nice
yeah.
Yes,
I
agree
with
abby
about
mario's
comments
being
well
placed
about
just
your
you
know,
organization,
or
actually
I
mean
it's
one
of
the
best
questions
that
someone
can
ask
not
or
in
general,
like
as
you're
working
on
anything,
it's
sort
of
like
it's,
it's
it,
which
is
just
what's
the
priority
or
like
hey,
what's
the
priority
and
and
how
are
these
organized
and
the
question
is
like
implicit
to
the
question?
A
Is
the
phrase
like
so
that
I
can
help
in
the
best
of
ways
and
that's
what
mario's
you
know
goal
is
that
that
clarification?
That
he's
asking
for
and
my
mind
immediately
go
and
so
today
that
the
tool
that
we're
using
to
try
to
capture
and
articulate
that
and
capture
it
in
granular
form?
Is
this
spreadsheet?
A
It
strikes
me
that,
like
to
mario's
example,
which
is
a
good
example,
he
was
just
suggesting
that
that
maybe
the
life
cycle
management
stuff-
maybe
that's
the
big
priority,
because
that's
what
a
lot
of
people
want
to
do
is
install
the
mesh
and-
and
we
were
just
talking
about
the
console
adapter,
and
so
my-
and
so
I
get
drawn
to
this
line,
just
as
an
example
of
like
the
lists
of
tests
that
are
within
here
and
our
our
overarching
goal
to
what
are
these
two
columns?
It's
been
a
while.
A
We
should
probably
freeze
this
row
but
covered
in
integration
test
covered
in
end-to-end
tests,
and
so
it
was
mark
it
was
row
118.
I
think,
right
now.
It's
marked
as
yes
covered
in
end-to-end
tests.
It's
like
well,
yes,
that
was
true
at
the
time
the
console
adapter
specifically
has
a
few
additional.
It's
the
only
adapter
that
uses
bats.
This
bash
automated
testing
system,
I
think,
is
the
name
of
it's
like
a
framework
around
bash
to
be
able
to
run.
A
The
batch
tests
were
being
called
from
within
github
workflow,
but
we've
kind
of
shifted
away
from
that
approach
to
eating
our
own
dog
food
or
drinking
our
own
champagne,
which
is
we
have
a
few
different
github
actions.
Those
actions
perform
end-to-end
things
it'll
perform,
you
know
a
performance
test
or
whatever
they
do,
and
so
shifting
away
from
bats
for
the
console.
Adapter
is
an
open
area
of
need
and
shifting
into
writing
more
of
these
end-to-end
tests
and
to
mario's
question
is
like
okay,
great,
like
you
know,
so
what
are
those
and
where
should
they?
A
A
C
Actually
trying
to
own
it,
can
you
go
to
line
138
yeah,
that's
actually
the,
and
where
is
it?
That's
what
caused
like
all
the
confusion,
like
probably
at
some
point
that
was
true
like
that
card
being
there
for
console,
or
maybe
it
was
just
like
some.
You
know.
Oh
we're
copying
pasting
everything
like,
but
but
maybe
that
now
that
we're
like
drilling
into
it,
yeah
cool.
So
that's,
okay,
that's
yeah!
C
D
C
Use
a
copy
and
paste,
I'm
I'm
sure
it's
no
problem!
You
know
I
want
to
own
this,
but
I
don't
want
to
remove
things
without
a
certainty.
You
know
so
this
conversation.
I
think
it's
very
helpful,
so
yeah,
basically
that's
one
okay.
So,
despite
that
for
console,
that's
like
I
think
we
can
like
refine
the
spreadsheet
as
we
go
right.
Yeah.
G
A
Between
the
two
of
you
guys
that
I
guess
what
I'm
trying
to
say
is
there
aren't
toes
in
this
right
now,
so
it's
kind
of
hard
to
step
on
anyone's
toes,
of
course,
yeah
yeah,
totally
that's
great
abby.
Well,
does
anyone
else
so
socranda
and
root,
raj
and
others?
This
is.
This
is
kind
of
more
more
on
that
ete
testing
end
to
end
testing
and
so
secronda.
I
don't
know
if
this
that's.
This
is
an
area
of
interest
for
you
or
not.
C
Sorry
to
interrupt
well,
at
my
opinion,
this
kind
of
artifact
like
it.
It
is
kind
of
like
a
whole
like
too
much
to
take
in,
but
I
think
if
we
use
it
as
a
reference
like
just
a
tool
as
we
go,
you
know
there's
different
things,
we're
collaborating
or
contributing
you
know
to
implement,
to
refine
and
keeping
the
quality
built
in
you
know
in
mind.
H
C
F
A
I
couldn't
find
the
desktop.
I
was
gonna,
I
don't
know
if
ashish
had
to
drop
yeah,
he
did,
and
so
let
me
do
his
job
for
him
and
that
is
to
say
for
rude,
raj
and
sekwanda
in
specific,
like
if,
if
you're
like
me
in
some
ways,
maybe
the
like,
if
I
were
to
go
to
a
minute
mark
yeah,
wait,
yeah,
wait,
say
in
the
bed
yeah
the
ten
minute
mark
no
pressure,
oh
yeah
is
very
very
quickly.
What
you
do
is
I
would
for
me.
A
A
What
is
it
you
know
to
basically
digest
what
it's
okay?
So
it's
it's
deploying
istio.
It's
expecting
pods
with
these
names
to
be
present
at
the
end.
So
this
is
this.
Is
this
is
essentially
the
assertion
statement?
It's
it
will
assert
that
these
three
pods
are
up
and
running
at
the
end,
okay,
and
so
anyway,
just
building
off
of
this
to
say:
okay,
the
next
thing
that
the
next
test
to
be
done
is
a
sample
app.
A
A
C
A
question
then,
hopefully
this
is
not
like
I'm
not
interrupting.
Oh
yeah,
I
just
occurred
to
me,
so
we
have
yaml
files
for
end-to-end
tests
for
adapters
and
it's
fine,
because
measuring
just
interacts
with
them
and
the
adapter
like
handles
working
with
a
specific
mesh
correct
right.
So
we're
kind
of
like,
like
it's
a
good
place,
to
do
end-to-end
stuff
too,
like
maybe
there's
more
to
it
in
higher
places.
But
what
about
like?
C
Would
it
make
sense
to
look
for
a
way
to
have
some
golang
or
javascript
like
something
that
can
help
us
use
an
editor
and
that
intellisense
autocomplete,
like?
I
think
the
challenge
with
those
yaml
files
is
that
we
kind
of
need
to
memorize
the
structure
of
the
yaml
and
yaml
supposed
to
be
like
easy
for
humans,
but
it
tends
to
be
unwieldy.
C
A
Having
documentation
about
what
those
yaml
files
do
about
what
those
workflows
do
is
a
good
is
a
good
starting
point.
Yeah
totally,
and
this
the
inside
the
test
plan
inside
inside
this
spreadsheet,
there's
a
another
sheet,
another
tab
sheet,
automated
tests.
So
this
isn't
the
itemization
a
high
level
itemization
a
high
level
documentation
of
the
different
workflows
that
run.
A
A
What
they're
actually
testing,
whether
or
not
they
can
start
by
themselves
or
they
get
kicked
off
from
a
different
workflow
when
they
run
like
in
part,
because
yeah
I
mean
so.
This
is
a.
This
is
one
of
those
things
that
there's
a
little
bit
of
so
this
is
related
to
what
mario
is
saying,
but
like
just
at
a
high
level.
If
you
understand
like
oh
hey,
this
particular
workflow
is,
for
this
purpose
great
and
then
there's
the
whole
problem
of
what
mario
was
saying,
which
is
like
okay
within
that
workflow,
the
yaml,
that's
written.
A
A
There's
an
overarching
perspective
that
we
need
that
we
need
to
come
together
about
what
are
the
areas
under
test?
How
are
they
being
tested
and
and
when
and
do
we
have
a
bunch
of
duplication?
Do
we
have
a
bunch
of
under
lap
kind
of
kind
of
another,
a
related,
a
tangential
thing,
but
yeah
describing
documentation
about
the
yaml
that
I
was
just
showing
like
because,
because
yeah
it
is
readable,
but
I
mean
you
can
read
a
bit.
You
don't
necessarily
know
like
what
like.
Where
does
this
link
to,
like
somebody
just
made
up
this?
A
A
Take
a
little
bit
of
this
is
why
we
should
have
a
session
on
this,
like
we
should,
if
we
don't
get
a
session
going
before
the
next
build
and
release
call
like
what
a
great
set
of
topics
for
that
build
and
release
call
to
really
yeah
yeah
yeah
and
there's
some
so
aditya
alpha
if
you're
on
part
of
what
we
were
saying
earlier
on.
The
call
is
that
the
output
of
these
end-to-end
tests,
the
results
they
get
published
into
mesh,
redox
and
aditya-
is
to.
A
Help
summarize
the
compatibility,
metrics
yeah,
because
there's
a
ton
of
so
coming
out
of
those
end-to-end
tests.
Here's
here's
an
example
of
one
that
was
run
and
actually
it
turns
out
to
be
the
istio
adapter,
that's
weird.
So
the
test
was
run,
and
this
is
coming
right
out
of
that
yaml
that
we
were
just
looking
at
actually
there's
istio
d
there's
the
ingredient.
A
This
is
exactly
what
we're
just
looking
at,
and
the
results
for
this
particular
run
were
passing
that
this
pod
came
up
and
that's
as
far
as
the
test
takes
it
and
okay,
what
mini
cube
was
spun
up
during
that
time
frame
this
one.
You
know,
so
here's
all
the
details,
this
guy.
If
you
want
to
get
more
details,
you
can
kind
of
I'm
sorry
if
you
want
to
get
so
so
alpha.
A
The
structure
of
this
page
is
such
that
there's
an
integration
test
section,
and
within
this
section
there
should
be
an
entry
for
each
measuring
component,
so
one
entry
for
istio,
kumar
and
lincoln
okay,
and
I
think
it's
one
entry
for
edge
and
for
stable.
So
I
take
it
back
two
entries
per
component,
one
for
edge.
One
for
stable
here's,
the
and,
if
you
want
to
look
at
the
historical
test,
results
for
a
component
for
that
release
channel.
A
You
can
come
here
to
see
all
the
historical
ones,
because
that,
because
you
know
they
get
built
and
run
fairly
frequently
and
so
over
time.
How
they've
been
doing-
and
this
is
great
like
this-
is
you
know
this?
They
did
this
really.
You
know,
we've
got
a
great
project
now
to
some
people
who
come
look
at
this
they're,
like
all
I
wanted
to
know,
was,
does
mesh
is
measuring
compatible
with
kubernetes
122.?
A
I
can
I
can't,
like
you
know,
and
so
the
your
goal
is
to
create
a
new
section,
not
called
above
integration
tests.
That's
a
compatibility
matrix!
That's
takes.
You
know
that
summarizes
the
data
from
here
saying
you
know
just
just
a
brief.
It's
a
much
much
briefer,
summary
saying:
here's
the
systems
that
measure
is
compatible
with
istio
113.2,
istio
113.2
through
1.9
or
whatever,
that
is
kuma
1.5
like
to
summarize
all
that
and
say
yeah,
so
because
the
integration
test
results
is
more
for
the
folks
that
are
on
this
call.
C
What
what
I'll
do
is
that,
like,
I
think,
there's
value
in
looking
at
those
failures
and
seeing
well?
Why
is
it
failing?
You
know
like
that,
helps
a
lot
understand
what's
being
tested
and
the
quality
of
the
test.
It's
not
just
like
see
it
pass
every
time
it's
about
what
what
does
it
uncover
like
is
is
a
failing,
is
a
failed
output
really
finding
a
bug
or
is
it
just
or
is
it
just
like?
C
A
Yeah
right,
yeah,
no
they're,
very
yeah,
just
the
fact
that
istio
d,
as
a
like,
using
that
example
as
a
pod,
just
the
fact
that
istio
d
is
present,
doesn't
necessarily
mean
that
it's
measure
is
going
to
be
successful.
Managing
its
life
cycle
ongoing
it's,
but
it's
the
first,
it's
a
good
step.
Of
course.
C
Yeah
and
something
else
there
is
like
a
pr,
I'm
sorry,
an
issue
yes
appear
open
for
the
for
a
ci
problem.
You
just
tagged
me
lee,
so
I
was
just
suggesting
that
if
we
could
get
some
folks
with
other
os's
to
try
to
do
an
npm
install
and
just
make
sure
that
this
doesn't
introduce
other
issues,
it's
for
the
fs
events
that
problem.